Your World by the Numbers.

Does spending money on education reduce gender equality? Or are governements just wasting their money?

2015-05-01T22:34:00.005-04:00

Governments throughout the years have allocated a significant amount of their GDP to education, in hopes that in the long run it improves upon the overall well being of the country. History has proven that investing in education prevents a vast array of social issues, such as: unemployment, poverty, environmental issues, hunger and some issues related to health; but does investing in education help reduce gender equality? This study is a cross time series analysis from 1946-2014 on all countries in the Quality of Governance Data, measuring the effect on total public expenditure on education on gender equality. I control for GDP per capita, and assistance from international agencies.

Education and Gender Equality- Importance of this study

As aforementioned, education is associated with many development outcomes and interacts with other factors to contribute to achievements related to the Millennium Development Goals set by the United Nations. According to UNICEF and UNESCO, education underpins all social progress. If the general education level worldwide is improved, global unemployment problems can be hugely tacked leading to improvements in poverty and general standards of living, including reducing the gender inequality gap (2013). Not only is it impossible to achieve gender equality without education, expanding education opportunities for all can help stimulate productivity and reduce the economic vulnerability of poor households. Gender equity in and through education examines education processes in education as well as outcomes in education to help a more gender equitable, productive and peaceful society. Education is an important tool in creating gender equity in society. Efforts to achieve gender equality through informal and formal education (2012).

The social significance of gender is that it is the device by which society controls its members, it opens and closes doors to power, property and even prestige. Like social class, it is a structural feature of society. This stratification is often referred to as the sex/gender system These types of differentiations occur not only on an interpersonal level between individuals, but also on a structural level within a given society. The shape and form of these systems can vary historically and cross culturally, as well as across populations. Thus closing the gender gap and making women equal to men has been such important part of societal development, even included in the Millennium Development Goals (2007). Additionally, it has been extensively proven through literature that improving the gender inequality gap in recent years, has increased participation of women in income-generating activities, education for girls and women becomes particularly important, especially in redressing gendered patters of discrimination.

Study Design

Variables

IV: Public expenditure on education ($)
DV: Gender equality
Z1: GDP per capita (by IMF)
Z2: Net Official Flows from UN Agencies and UNAIDS

Findings

Hausman Test p>.05

Checking for evidence of covariation (scatter plot)

Scatter Plot of Relationship

Conclusion

The data shows that historically there is not a significant relationship between public expenditure on education and gender equality. Although the relationship is not statistically significant, the charts and graphs show a positive relationship. For every dollar spend on education, there is a .00099 increase in gender equality. With this in mind the data suggests that if we want to progress upon the Millennium Development Goal of improving gender equality, perhaps it would be more valuable to allocate those specific resources to other initiatives besides educational program. For example, the data shows a positive and significant relationship between gender equality and GDP per capita, thus, perhaps investing in initiatives that incentivize financial growth result more effective when tackling the issue of gender inequality. Therefore, a more complete follow-up to this study could integrate a regression of other initiatives suggested by the literature that have the potential of improving gender equality across nations. Also, the study could be done separating countries by region, in which perhaps expenditure on education might have a bigger effect. Finally, this study poses helpful information for policymakers, specially in the departments of the government working towards gender equality to explore different alternatives besides educational initiatives.

References

Klein, Susan et al (2007). Handbook for Achieving Gender Equality through Education. Lawrence Earlbaum Associates. Retrieved from http://www.feminist.org/education/handbook.asp

Narayan, Swati (2012) Education for All: Beyond 2015. Mapping Current International Actions to Define the Post-2015 Education and Development Agenda. Paris: United Nations Educational, Cultural and Social Organization. Available from www.unesco.org/newfileadmin/MULTIMEDIA/HQ/ED/temp/Mapping_post_2015-Swati_Narayan.pdf

Nayyar, Deepak (2012). The MDGs after 2015: Some reflections on the possibilities. Background Paper. New York: UN System Task Team on the Post-2015 UN Development Agenda. Available from www.un.org/en/development/desa/policy/untaskteam_undf/d_nayyar.pdf

UNICEF and UNESCO. (2013). The World We Want— Making Education a Priority in the Post-2015 Development Agenda: Report of the Global Thematic Consultation on Education in the Post-2015 Development Agenda. Available at http://www.unicef.org/education/files/ Making_education_a_Priority_in_the_Post-2015_Development_ Agenda.pdf

Does Financial Openness Reduce Poverty?

2015-04-30T17:44:00.002-04:00

by Shona Carter

Poverty is has been a global challenge for many years as leaders around the world work together to combat its devastating effects. Although success in poverty reduction has been seen globally, there still remains persistent challenges to combating the issue. Poverty remains a big problem in Latin America. As globalization spreads throughout the world, partnerships regarding politics and economics aim to reduce poverty because it can increase shared prosperity. Financial globalization has been considered a way of improving struggling economies and thereby reducing poverty. I hypothesize that more financial integration would lead to lower levels of poverty. I will assess the impact of financial openness on poverty in Latin America because poverty remains an issue in the region and there strong data exists regarding financial integration in the region. The independent variable was obtained from a financial openness index produced by economists Menzie Chinn and Hiro Ito. The index measures the extent to which a country’s capital account is open or has unrestricted capital account transactions. Previous measures of financial openness were mostly binary variables that tell if a country is “open” or “closed”, without accounting for capital accounts that are partially open or partially closed. To overcome this, the Chinn-Ito index uses previous measures and 4 binary variables that indicate a country’s financial openness: the presence of multiple exchange rates, restrictions on current account transactions, restrictions on capital account transactions, and export proceeds surrender requirements. These variables together measure the intensity of capital controls better than a binary variable that only considers whether a capital account is open or closed.

The variable for poverty, which was obtained from the World Bank’s listing of economic indicators, is a very scarce variable that is defined by the percentage of people living on less than $1.25 per day. Due to the scarcity of the poverty variable, I created a new variable using factor analysis. Two additional variables, life expectancy and infant mortality were used to construct a new measure of poverty due to their high correlation with the original poverty variable. I will use this new variable to run a regression with financial openness variable and poverty in order to get a meaningful idea of the correlation between financial openness and poverty in Latin America. Due to the negative relationship between infant mortality and life expectancy, the factor variable has a negative relationship to the original poverty variable.

I will control for social and political indicators of globalization, as they may influence the degree to which a country is financially integrated. The social globalization variable was made using information regarding international connections, such as tourism, also the flow of information, such as internet users. It also uses cultural proximity which could be defined using the number of international businesses per capita. Social and Political globalization variables come from a Quality of Governance dataset. Political globalization measures the degree to which a country has a relationship with other countries on a government to government level. This variable includes information such as the number of embassies a country has and its memberships within international institutions. I use cross section time series analysis to create multple regressions using the STATA software. I also account for issues that arise with panel data such as autocorrelation.

Cross Section Time Series Analysis

Methods and Tables:

The statistical method that I used required me to set up the data by the country and year so that I can make the regressions. I ran a hausman test which helps determine whether I should used fixed effects or random effects in my regression. These effects would help me to obtain a valid statistical inference when there is unobserved heterogeneity ( varying unobserved components of the effect of my estimation) The hausman test supports my using fixed effects in my regressions.

Table 1 shows 5 models. The main independent variable, financial integration, and two additional independent/control variables, social and political globalization, are regressed with the original poverty variable. They are also regressed with the factor variable, as well as the two dependent variables, infant mortality and life expectancy, which are highly correlated with poverty. In the second model, only the observations for the poverty variable are regressed with the factor variable.

Table 1: Financial Integration’s Impact on Poverty in Latin America (year effects)

	(1)	(2)	(3)	(4)	(5)

VARIABLES	Poverty	Factor(for Poverty)	Factor	Infant Mort.	Life Expect.

Financial Integration	-4.633**	0.0195	-0.117***	2.654**	-0.913***
	(1.967)	(0.0744)	(0.0328)	(1.183)	(0.239)
Political Globalization	-0.00791	0.0102***	0.00347***	-0.173***	0.0122
	(0.0667)	(0.00252)	(0.00129)	(0.0471)	(0.00954)
Social Globalization	-0.304**	0.00725	0.0150***	-0.434***	0.137***
	(0.128)	(0.00485)	(0.00215)	(0.0777)	(0.0157)
Constant	17.70**	-0.188	-1.296***	89.19***	56.46***
	(7.392)	(0.358)	(0.128)	(4.632)	(0.940)

Observations	148	148	997	1,021	1,008
Number of id1	14	14	27	27	27

Standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

The first table shows the regressions with year fixed effects and the second table shows the regressions with year and country effects. Table 1 and 2 produce similar results with the coefficients aligning in the same directions.

Table 2: Financial Integration’s Impact on Poverty in Latin America (country effects)

	(1)	(2)	(3)	(4)	(5)

VARIABLES	Poverty	Factor (for poverty)	Factor	Infant Mortality	Life Expectancy

Financial Integration	-4.173**	0.655***	-0.178***	6.702***	-1.307***
	(1.819)	(0.0550)	(0.0463)	(1.623)	(0.323)
Political Globalization	-0.0681		0.0225***	-0.674***	0.143***
	(0.0566)		(0.00161)	(0.0509)	(0.0106)
Social Globalization	-0.139		0.0474***	-1.656***	0.374***
	(0.0971)		(0.00233)	(0.0749)	(0.0156)
Constant	25.68***	-0.0624**	-2.885***	138.5***	46.03***
	(3.550)	(0.0287)	(0.0787)	(3.439)	(0.814)

Observations	148	1,189	997	1,021	1,008
R-squared	0.203	0.109	0.623
Number of id1	14	32	27	27	27

Standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Table 3: Financial Integration’s impact on Poverty in Latin America (lagged independent variables)

	(1)	(2)

VARIABLES	Factor	Factor

Financial Integration	0.168	0.0432
	(0.103)	(0.0312)
Lag Social Globalization	0.0118**	0.0137***
	(0.00558)	(0.00178)
Political Globalization	0.00987***	0.00481***
	(0.00254)	(0.000848)
Lag Political Globalization	0.0135***	0.00488***
	(0.00251)	(0.000865)
Social Globalization	0.0343***	0.0131***
	(0.00543)	(0.00171)
Lag Financial Integration	-0.364***	0.0261
	(0.102)	(0.0302)
Constant	-2.854***	-1.011***
	(0.0791)	(0.178)
Standard Error/Auto. Adjustment?	No	Yes

Observations	972	972
R-squared	0.635	0.605
Number of id1	27	27

Standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

In these regressions for table 3, I decided to lag the independent variables in order to get a more accurate estimate of the true effect of financial integration. By lagging the independent variable, the models consider previous values of X for the effect on Y, and could produce the most meaningful of all the models. Tables 1, 2, and 3 show a progression of obtaining the most accurate and meaningful results.

Results:

Based on the most controlled and accountable regression in table 3, financial integration has a weak or non-significant effect on poverty.

The simpler regressions in the first 2 tables that do not include lagged variables or account for auto regression, show a significant yet inconsistent effect on the scarce poverty variable. Keeping in mind that the factor variable and poverty should have a negative relationship, these regressions do not show the expected results. For example, since financial integration is associated with a decrease in poverty in model 1 on the first 2 tables, then financial integration should lead to a positive coefficient in the factor variable in model 3. The expected direction of the factor variable is only shown in model 2 where it is regressed solely with the same units for the poverty variable. The fact that the coefficients for the factor variable and poverty go in the same direction in models 1 and 3 in the first two tables, means that those results are inconsistent.

In the third table I set out to get the most accurate regression to account for the discrepancies in tables 1 and 2 by using lagged variables with standard error adjustments. In model 1 of this table with lagged variables, I used the same process for working with panel data that I used in the previous two models. In model 2 of this table, I used a different process to calculate panel corrected standard error estimates that help with the issue of auto correlation.

The results from the first tables show that financial integration has a significant effect on the scarce variable for poverty, loosely showing that financial integration leads to a decrease in poverty, however since the poverty variable is scarce and the factor variable produces inconsistent results, model 1 is not enough to allow me to reject the null hypothesis. Therefore I would fail to reject the null hypothesis that financial integration has no relationship to poverty in Latin America. The 2 other independent variables however, show a significant relationship to life expectancy and infant mortality in the expected directions.

Conclusion:

The regressions produce inconsistent results despite high levels of significance. After making the proper adjustments to the regression over time as I have gradually done from tables 1 through 3, the effects of financial integration remain volatile. This tells me that either there is likely more to this analysis than what these particular variables contribute or financial integration is not a strong predictor of poverty levels. It does not mean that financial integration is definitely not associated with reductions in poverty, but rather my regression does not produce a stable result that controls for the issues associated with panel data.

Appendix:

Summary Statistics

Variable		Mean	Std. Dev.	Min	Max	Observations
poverty	overall	14.1763	15.71471	0.03	68.49	N = 165
	between		15.20486	1.62	61.75917	n = 14
	within		5.382453	2.527136	36.82714	T = 11.7857
Factor	overall	1.20E-10	1	-4.004508	1.487515	N = 1684
	between		0.6822429	-2.025807	0.9966946	n = 36
	within		0.741246	-2.262154	1.790473	T = 46.7778
financial integration	overall	0.4315984	0.3442466	0	1	N = 1313
	between		0.2168614	0.1247907	1	n = 34
	within		0.2723456	-0.245791	1.130454	T-bar = 38.6176
Political Globalization	overall	51.74875	19.25492	10.19148	94.72778	N = 1273
	between		17.81613	18.27359	85.22592	n = 33
	within		9.09341	25.41777	78.00189	T = 38.5758
Social Globalization	overall	39.82963	11.57027	12.46371	66.42154	N = 1273
	between		9.886109	14.93896	63.53885	n = 33
	within		6.339785	25.87805	66.74578	T = 38.5758
infant mortality	overall	45.07957	34.22633	5	202	N = 1806
	between		22.89857	13.7125	116.6556	n = 36
	within		25.74855	-20.65191	136.4481	T-bar = 50.1667
Life expectancy	overall	67.26714	7.067833	42.159	79.70502	N = 2027
	between		5.206262	52.75044	77.01164	n = 43
	within		5.132249	51.33151	79.94593	T-bar = 47.1395

***************************************************************************

Graph: Financial Integration and Poverty

A strong effect would show a clear positive relationship between financial integration and the factor variable which is the same as a negative relationship between financial integration and poverty.

Sources:

http://qog.pol.gu.se/data/datadownloads/qogstandarddata

http://data.worldbank.org/indicator/SI.POV.DDAY

http://web.pdx.edu/~ito/Chinn-Ito_website.htm

Outcomes of Economic Liberalization in Africa

2015-04-30T17:31:00.000-04:00

In the race towards development, many African countries have adopted the western conception of modernization. In the west, open market economies fast-tracked nations to modern-hood. Less barriers to market, and the less influence of government boosted the growth of enterprise in these economies. Private sector led growth has a multiplying effect, boosting growth several times higher state intervention. Private sector, especially small and medium enterprises, provide the majority of jobs, and therefore can improve the livelihood of workers. State intervention and large government spending can crowd out the private sector, limiting opportunities for entrepreneurs sell those services and solutions to the market. It also diminishes the competition to state-owned enterprise, and less competition means a less transformative economy.

Ethiopian PM at the test-run of a newly developed light rail
http://english.gov.cn/premier/news/2015/02/03/content_281475049660622.htm

Whether a country has state-led development or private led development is signaled by their economic markets. If the market is closed, this means heavy spending and intervention is occurring on the part of governments. If the market is liberalized, this means there is an openness to private investment. Many countries began with state intervention in their path to development (South Korea, Brazil, and Tanzania) but ultimately opened their markets once their domestic economy was fully developed. Countries like China, Ethiopia, and Russia continue to have state-led development although this puts limits on their private sectors.

While there is no right or wrong path to develop, one method might lead to greater human development than the other. It seems fast growth rates in many places have not reduced wealth inequality and standard of living for all. Some of the fastest growing nations, like Mozambique, have not been able to comprehensively address poverty and the standard of living. While China has rapidly industrialized it was able to lift a billion people out of poverty. Can the developmental state model ensure that growth is equal and comprehensive? To answer these question I study African countries and the effects economic liberalization has had on their human development.[1]

The two variables of interest for my study are human development and economic liberalization. I use a cross sectional time series (CSTS) with fixed effects model over 1990-2014. This model allows me to measure the effect of economic liberalization on human development over a long period of time and control for time independent effects. Changes in the explanatory variables of this model correlate to the changes in the level of human development. All the countries used in the data are in the sub-Saharan Africa region. CTST method addresses time order bias of the variables of interests. It also addresses confounding bias from things like culture, state identity, and geopolitical factors that I control for.

To operationalize economic liberalization, the dependent variable of this study, I use the Economic Freedom of the World index. This distinguished if the country has a closed market economy versus fully liberal market economy by a score on a scale of zero (closed market) to 100 (open market). To operationalize human development I use real GDP per capita as measurement.I control for the effects of a several variables, Instability, foreign aid, and economic institutional strength. Instability is operationalized by the Fragility Index provided by Fund for Peace. This measures fragility of a state on a variety of factors and gives a score on a scale of 0 to 120 (0 being the most stable, 120 being very unstable). Economic institutional strength is operationalized by the absolute economic institutional quality provided by the Heritage Foundation’s economic freedom index. This measures the strength of economic institutions by a score of 0 (worst) to 100 (best). This controls for the effects of colonial institution legacies on the continent, francophone and anglophone cultural differences in institutions, and the effectiveness of a countries central banks and finance ministers.

Foreign aid is the sum of dollar amounts received, and is distinguished between aid from international organizations and, and aid from other organizations or donors. Levels of stability are influenced by an increasing presence of non-state violent actors that might have a negative effect on a country’s modernization despite democracy, liberalization, or economic industrialization policies. Foreign Aid might influence a country’s economic growth.

Photo of MTN shop in Nigeria. MTN is a private Nigerian telecommunications firm.
http://uk.reuters.com/article/2014/03/05/us-safrica-mtn-idUKBREA240DF20140305

So what is the outcome or effect of economic liberalization on human development? My hypothesis is that economic liberalization has zero (or and infinitely small) effect in in the level of human development. In my analysis however, I found my hypothesis was wrong as there is a positive effect, which is statistically significant.

I create a regression model of of the variables I use on in this study (table one).

The regression shows that economic liberalization has a positive effect on human development. For every improved score on the economic liberalization index the Real GDP per capita increases by 1.35 units. This positive effect is also statistically significant at the 0.1% level, which means it is a substantial effect relative to the sum of GDP.

There are many reasons for this outcome. A factor I’d like to focus on is GDP measurements. I used. I used the measurement of real GDP per capita. This measurement might not have captured larger picture of inequality and human development. There are other measurements created by the UNDP that try to grasp the whole picture of wealth. Private sector can boost a country's GDP (like Mozambique, China) but if this boost is not being realized in all provinces and levels of society, it can mask larger equality and poor standards of living. In a future study, I’d like to isolate the effects further by either switching measurements, or considering a different and more robust time series statistical model.

[1] Using a Hausman test we can see that there is little difference between the fixed effects and random effects estimation. So there is little room for endogeneity. The test says it is okay to use random effects estimation. Although it is okay to use the random effects, I conclude with my model of the fixed effects would or create bias as the results would be similar.

Vital Flows: The Effect of Changes on GDP on Foreign Aid

2015-04-30T17:20:00.000-04:00

Wealthy states provide an immense amount of developmental aid to poor. In 2013, members of the Organization of Economic Cooperation and Development’s (OECD), Development Assistance Committee (DAC) provided over $135 billion in aid to poor countries around the world. [1] The OECD, a group of rich countries, promotes this kind of developmental state-to-state aid known as Official Development Aid (ODA). The OECD defines ODA as aid provided by a government that is designed to promote the “economic development and welfare of developing countries” that is concessional in nature.[2]

While changes in developmental aid from year to year likely have a number of causes, what role does economic growth in the donor states play? In this study, I set out to uncover to what extent the variance in development aid is an effect of the past (or present) variance in GDP.

While it may seem obvious that economic growth might have an effect on the aid that a country provides to developing states, on closer examination other the picture becomes more complicated. Policy makers could take a number of different factors into account when deciding how much aid to send to other states. States could commit to providing a certain amount of aid per year, regardless of domestic economic conditions. On the other hand, changes in domestic politics could effect significant policy shifts in the foreign aid budget. A quick comparison between GDP and ODA, leaving changes over time to one side for now, demonstrates a positive relationship between the two variables, as shown in Figure 1.

To attempt the estimate the effect of economic output on rich countries’ decision to release aid monies to developing states, I employed a cross-sectional time series model that looked across time and countries for evidence of an effect. Based on the availability of data for inflation-controlled variables, I focused on the period from 1990 until 2012. Although ODA is broken up into a many categories I zeroed-in on the aggregate of all of them, measured in inflation-adjusted 2013 dollars. I also logged both GDP and ODA to estimate a linear relationship.

The OECD provides data on foreign aid as defined as ODA dating back to 1968 for the members of the Development Assistance Committee (DAC). While the OECD is a group of mostly wealthy states, the DAC states focus specifically on promoting development. I therefore decided to focus on current DAC member states 28 in total, excluding the EU, which is also a DAC member but not a state. This core group of 28 states does not include all states that provide large amounts of state-to-state developmental aid.[4] Verily, the OECD is so skewed towards Western Europe and North America, my model could not account for regional differences because too many of the states shared the same region. Still, over the time period in question, the second half of the Twentieth Century, this sample represents a majority of the worlds wealthy donor states. Although I originally intended to estimate this effect for a longer time period, inflation controlled GDP data were not available for these states before about 1990, so the time frame for this study is from 1990 until 2013. This also coincides with the fall of the Soviet Union and the addition of a number of post-Soviet states to the OECD.

To estimate the effect of variances in economic growth on the changes in foreign aid, I employed a cross-sectional time series regression model with country fixed effects, the results of which I present in the table below:

My model attempts to predict current levels of ODA from the individual interactions of countries’ GDP and a number of control variables. Importantly, although it seems implausible that ODA outlays could cause changes in GDP, I attempt to control for such reverse causality by regressing current foreign aid on the prior year’s GDP.[5] “Lagging” GAP by one year in this fashion also allows me to test the proposition that it takes policy makers some time to recognize changes in the real economy and react to them by changing ODA. My model attempts to control for variances between states in a number of ways. First, it employs country-level fixed effects, which is a process that attempts to strip out anything that is constant across time within units from the analysis. The decision to employ fixed effects instead of random effects here was also based on the result of the Hausman Test (p = 0.00). The model also attempts to control for changes in domestic politics, a potential confound, by including two variables: one representing the left, right, or center orientation of the chief executive of a state, the other representing the political orientation of the leading party in the legislature. Finally, as mentioned above, the model also includes control variables for longer term differences between states that might be confounding variables, namely, the region of the world and a country’s history as a post-soviet state.

In conclusion, based on the above results, foreign aid and gross domestic production are highly correlated correlated over time. Moreover, the analysis demonstrates that this correlation exists in a model where foreign aid could not have caused GDP, eliminating the possibility of reverse causality. A one unit increase in the previous year's GDP is correlated with a 1.871 unit increase in the amount of ODA, when political orientation of the executive an leading party in the donor state is held constant. A similar effect exists and is also statistically significant at the 0.01 level with GDP and ODA from the same year. Problematically, although the model was able to control for the political orientation of governmental leaders, [6] it was not able to control for differences between regions or post-Soviet status, and left those variables out of the regression. Although these results were statistically significant at a high level, meaning that there is a low probability that they were due to chance, it is not possible to conclude that policy makers only use GDP in making determinations about foreign aid. The results however do indicate that it may be likely that government officials do respond to changes in the economy when making decisions about foreign aid.

Appendix:

Alternative models:

[1] http://www.oecd.org/dac/stats/

[2] http://www.oecd.org/dac/stats/officialdevelopmentassistancedefinitionandcoverage.htm

[3] http://www.oecd.org/dac/stats/development-aid-stable-in-2014-but-flows-to-poorest-countries-still-falling.htm

[4] This includes large states like China and India, which were not relatively wealthy at beginning of the time frame but are now relatively much wealthier.

This also leave out states that the OECD does collect ODA data for although they are not DAC members. For these states, (Bulgaria, Estonia, Kuwait, Turkey*, and the UAE) ODA data was generally also not available for a long enough time period.

[5] Thus the GDP in 1995 is used to predict ODA in 1996, under the assumption that ODA changes in 1996 cannot have cause variance in GDP before those changes took place.

[6] In the dummy variable regression for governmental orientation one variable must be committed from the right, left, center strategy. I elected to set "Center" to the control.

Do better enrollment rates mean better education outcomes? Analysis of a Cash Conditional Transfer in Honduras

2015-04-30T16:53:00.007-04:00

When we think about experiments, we tend to picture microscopes and beakers. The advantage of natural lab experiments is that the research is in highly controlled environments where they can isolate and manipulate variables. In contrast, social experiments present more challenges because the research might use the wrong measurements, miss controlling for an important causal factor, or use an incorrect method. However, in the policy realm it is crucial to continue measuring the effects and impact of different factors in order to provide politicians with information to design better public policies.

There has been a great deal of discussion among government officials and scholars about the most effective way to break the cycle of poverty, and one of the most popular methods is using Conditional Cash Transfers (CCT) programs. In these programs, money is given to a household that meets certain criteria. These are usually related to the children going to school or regular medical check-ups. I am especially interested in testing effects in education outcomes. I want to find out if a CCT has an effect on grade progression and if the increased demand in school enrollment led to success in school, which is measured by grade progression.

Honduras is the second poorest country in Central America. In 1990, they implemented the Programa de Asignación Familiar or Family Allowance Program (PRAF) CCT. The program, PRAF II, was redesigned to improve previous weaknesses. To analyze the effect of PRAF on education performance I used the difference-in-difference technique based on panel data collected in a random experiment. To understand the validity of this experiment, we need to understand its various components.

What was the sample framework?

In 2000, the Honduran government conducted a randomized experiment in 70 of the poorest municipalities of the country. The data was collected on every individual in each program household in 2000 and 2002. The survey from 2000 was collected before the program was implemented and the survey from 2002 was collected for the current year.

Who was eligible for the program?

Education transfer was available to families with children ages 6 to 12 who were enrolled and regularly attended grades 1-4.

Health transfer was available to families with children under 3 and pregnant others who regularly attended health centers.

How are treatment and control groups structured?

In order to have a good treatment group there is a need to administer a large exogenous shock in a short period of time. The PRAF experiment had a total number of 15,350, which consists of 7,675 individuals (each individual having been surveyed in both 2000 and 2002) and 3,935 households.

Ideally, we would like to be able to compare the same treatment group and run the experiment again without giving them the treatment. Since we are not able to do that because of time restrictions, we need a counterfactual or control group that identifies the outcomes that would have happened in the absence of the treatment. We hope that random selection makes the groups as similar as possible. In this experiment, child height-for-age was used as a poverty proxy to determine which municipalities would be eligible for the CCT program. They employed a system of geographic targeting to identify the communities where the chronically poor lived. The program drew up a list of municipalities based on the findings of the 1997 census and selected those whose chronic malnutrition indices fall below a standard deviation from the mean.

Based on this preliminary selection of eligible municipalities, a random statistical sample was then made to assign those which fell into the treatment and control groups. As a result, 40 eligible municipalities were assigned to treatment and 30 into control. Any of those households from the selected treatment communities was considered eligible for the transfer if it complied with the component criteria. The quality of the estimate is mainly based on the similarity of the counterfactual to the treatments group. Otherwise, the estimate will have bias.

How to assess impact in Education performance?

The education system in Honduras faces problems of general service provision, low teacher quality, low enrollment rates and low performance. Children also tend to delay joining school, and their age on entry exceeds the normal age at that grade level. Due to the levels of poverty in Honduras, it is common that children have to work instead of going to school. Moreover, the quality of the educational service is very poor.

The cash transfers in the treatment group were available for families for each child age 6-12 who was enrolled in the first four years of primary school and attended regularly. In my test, I restricted the sample so that we are only examining children between the ages of 6 and 12. The total observations were 2,648.

I used the difference-in-difference technique because it allows me to measure the impact with the data from the pre and post implementation of the program. This method mimics an actual experiment by measuring changes in average outcome in the treatment group before and after treatment minus the difference in average outcome in the control group before and after treatment.

The program requires enrollment and attendance in school in order to receive the cash. Therefore, it will be expected that these two variables will increase because of the pre-condition and not because it is a successful program. As a consequence, only measuring these variables (enrolled or attendance) will not actually show that the program is changing behaviors in children and improving the education performance. In order to measure impact, I used two models. The first one measures the difference in enrollment and the second one that of grade progress.

It is important to include in each model some characteristics that affect enrollment and grade progress. I controlled for years of education of head of household, distance to primary school, and size of household and household characteristics such as floor composition, availability of electricity. In addition, I added the name of their municipality and if they received support from welfare program that is not PRAF.

In my first OLS regression using the variable “enrolled” I can prove that in fact there was an increase of 0 054 and it is statistically significant.

However, to find impact I applied the difference-in-difference formula to the “progressed” variable. The results show that the difference was of 0.05 and this was not statistically significant.

We can infer that even though the cash transfer motivated parents to enroll their children in school, when accounting for the original differences between the two groups and tracking those differences over time, we see that there is no significance in the results. In consequence, it appears unlikely that the PRAF program led to considerable improvements in grade progression from 2000 to 2002.

Are my results valid?

The key assumption for any difference-in-difference strategy is that the outcome in treatment and control group would follow similar trajectories in the absence of the treatment.

Common trend assumption is difficult to verify but using pre-treatment data could show that the trends are the same. Yet, even if pre-trends are the same we still have to worry about other policies changing at the same time. In this experiment, I cannot quantitatively demonstrate the common trends assumption because I do not have pretreatment data. However, the years previous to the experiment were not unusual. In other words, there was not a big economic shock or other macroeconomic factors that could have affected the municipalities included in this study. I am aware that PRAF was in place in some of the municipalities since 1990, so even if I had data from previous years the effect was already there. In addition, the number of observations and the randomization effect should help to overcome the lack of pre and post treatment data.

Conclusion:

PRAF II did not show the increase in performance results as expected. The difference-in-difference is a good estimator, but the problem is that many factors can affect the assumption that our counterfactual is almost equal as to our treatment group. Therefore, it is important to use other techniques to prove our findings. CCTs have provided good results in countries such as Mexico with Progresa and should continue to be monitored and improved; hopefully allowing people to break the poverty cycle and better their lives.

Appendix:

I used fixed effects for municipality in order to control for any differences in the municipalities such as rural vs. urban, culture, weather, agricultural conditions, etc.

Economic freedom and corruption in Latin America

2015-04-30T16:47:00.005-04:00

Economic freedom and the effects of

corruption in Latin America

After decades of economic management with solid social progress, Latin America has reached certain levels of development. This improvement of social and economic conditions over time might improve the national level of corruption perception, even though the region consistently performs poorly in the CPI (corruption perception index). Recently much attention has been devoted to testing the relationship between economic freedom and corruption, with the assumption that economic freedom leads to reduce level of corruption. Thus, this paper examines the effect of economic freedom and economic globalization on corruption perception across the Latin American region over the period 1995 to 2013, since I believe that political and trade liberalization lead to a decrease in corruption.

As part of the globalization process and as a response to the Latin American debt crisis, the region moved toward political and trade reform during the 80s and 90s. The economic effects of economic reforms have been subject of complex and holistic studies. However, the analysis of the effect on components of corruption perception is limited. Even though economists have argued that lack of competition fosters corruption and empirical findings confirm that democracy and economic freedom alone does reduce corruption; there is no cross-country study which captures the interaction effect of these two factors in Latin America over time. Therefore, this paper pretends to fill this gap of analysis due to the fact that economic freedom and corruption are factors that undermine the development growth of the region.

The theory and literature suggest that among the most corrupt nations, greater economic freedom does not reduce corruption and may even exacerbate it, because nations respond differently to levels of economic freedom, depending on their level or stage of development conditions (Billger et al, 2009). Thus, the effect of economic freedom on corruption may depend on how efficient government intervention is in countries which are more highly “regulated” or “have high levels of economic freedom”. In line with this argument, Giavazzi and Tabellini (2005), who use panel data to look at the relationship between economic and political liberalizations and their effects on growth, investment, inflation, institutional quality, and corruption find that in isolation, increased economic and political liberalization decrease corruption. Therefore, greater openness may reduce corruption, but the more corruption there is, the more rent-generating trade barriers there will be (Treisman, 2000). The degree to which a country is open to trade, then, is clearly endogenous; however, whether or not a country is open to trade, as well as whether the country decides to liberalize or not, is less likely to depend on the perceived levels of corruption.

Now focusing on Latin America, in order to examine the effect of economic freedom on corruption perception, I use data from the World Bank and the Quality of Governance Institute for the 20 countries of the Latin America region for the period 1995 to 2013. In this case, corruption perception (CPI) is the depended variable of the model and the main independent variables are economic freedom and economic globalization. Also, since I’m using the CPI I might include measurements of political participation and income size (GDP per capita) because they can cause a variation on CPI and economic freedom .(for more details see table B).

Furthermore, in order to have some idea of the probable relation with corruption, Figure A shows the Economic Freedom index and the CPI corruption index across the 20 countries of the Latin American region for the year 2000, and illustrates the clear positive association between the indices, implying that higher economic freedom is associated with lower corruption

Additionally, I examine the relationship of economic freedom and their interaction effect on corruption using Pooled Time-Series Cross-national Regressions with Fixed Effects[1]. I regress corruption (log) on prior corruption and prior scores of the Independent variables. Table A. presents the results. I found that using previous years of corruption perception, on average and controlling for Trade freedom index (column 2) 1% increase on economic freedom is associated with a reduction of 17% of corruption perception index. The results is statistical significant. However, even though the results confirm the hypothesis that Economic freedom reduces corruption perception; the results are not always significant. Column 1 shows that controlling for other variables such as income, political participation and economic globalization, economic freedom index does not have statistical significant on corruption perception. The only variable that remains significant is the Income. In this sense, it may be suggested that economic size (Income) is also associated with a decrease on levels of corruption perception with statistical significance.

Table A. Pooled Time-Series Cross-national Regressions with Fixed Effects.

(1 year Lag)

*DV: Corruption Perception (log),* *from 1995 to 2013*	(1)	(2)	(3)
Corruption Perception (Log)	0.61^***	0.62^***	0.64^***
	(13.91)	(15.40)	(15.68)
Economic Freedom Index (Log)	-0.17	-0.17^*	-0.17^*
	(-1.80)	(-2.46)	(-2.37)
Income (GDP) (Log)	-0.049^*
	(-2.01)
Political Participation (Log)	0.034
	(0.73)
Economic Globalization	-0.00039
	(-0.31)
Trade Freedom Index (Log)	0.24^**	0.18^**
	(2.96)	(3.28)
Constant	0.55	0.41	1.13^***
	(1.21)	(1.14)	(3.91)
R-squared	0.41	0.43	0.41
Observations	276	325	325

Note: Absolute value of t statistics in parentheses. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001

It is important to note that while cross-country time series regressions of economic freedom and corruption give results in line with the argument that corruption is reduced by an increase in economic freedom, this may hide heterogeneity at sectoral level, particularly in the degree and nature of government intervention (Graeff et al, 2003). In other words, I notice an ambiguous correlation among the components of economic freedom and corruption. This suggests that the weight of the components of the aggregate index influences the effect on corruption, implying that the results are sample-sensitive.

In summary, this paper assesses the effect that economic freedom has on corruption perception across Latin America. I found a linear relationship that supports the hypothesis that trade liberalization leads to a decrease on perception corruption since the results are statistical significant. Therefore, I might think that, after the trade and openness reforms in the 80’s and 90’s, the region trends to reduce the corruption perception. However, the historic trend of high corruption perception on countries of this region may cause the slowly change on corruption perception. It is likely that as long as in the region enters into the global economy, the level of corruption would decrease.

Bibliography

* Pieronia, L.; d'Agostinob, G. (2013). Corruption and the effects of economic freedom. European Journal of Political Economy. Volume 29, March 2013, Pages 54–72. Retrieved from: http://www.sciencedirect.com/science/article/pii/S0176268012000456?np=y

* Costa Tavares, Samia (2007). Do rapid political and trade liberalizations increase

corruption?. European Journal of Political Economy. Volume 23, Issue 4, December 2007, Pages 1053–1076. Retrieved from: http://www.sciencedirect.com/science/article/pii/S017626800600053X

Appendix

Table B. Descriptive Statistics

Variables	Mean	Std. Dev.	Min.	Max.	Detail
*Economic Freedom Index*	57.80	9.94	19.5	85.3	Index score range from 0 to 100, where 100 represents the maximum economic freedom
*Corruption perception index (CPI)*	3.44	1.39	1.4	7.9	10-point scale in which a score of 10 indicates very little corruption and a score of 0 indicates a very corrupt government.
*Economic Globalization*	57.80	9.94	19.5	85.3	Scale from 0 for non-flows of trade and investments and 100 for highly integration to global economy
*Trade freedom*	70.51	8.29	44.4	88.0	Index scores range from 0 for minimum degree of trade freedom to 100 for maximum degree of trade freedom.
*Income*	3928.10	2707.84	314.2	13559.1	GDP per capita in current US$
*Political participation*	0.41	6.00	-99.0	1.0	Index scores range from 0 for completely absent or disrespected political participation rights to 1 for their full presence and respect
Observations	277

Table C. Hausman Test

	(b)	(B)	(b-B)
	Fixed	Random	Difference

Economic freedom (log)	-.1652047	-.0019991	-.1632057
Income (GDP) (log)	-.0492979	.0146905	-.0639884
Political Participation (log)	.0335217	.0788459	-.0453242
Economic Globalization	-.0003927	.0004396	-.0008323
Trade freedom (log)	.23913	.0992944	.1398356

Chi2(6) =	79.23
Prob>Chi2 =	0.0000

**Luz Angela Serrano

** SIS-750- Spring 2015

**Blogpost 2

**********

* Preamble

set more off

ssc install estout

cd "G:\"

**What is the effect of trade freedom and economic globalization on corruption across LA?

** ID: Economic freedom index: eco_freedom

** ID: Economic globalization: dr_eg

** DV: Corruption perception: corrup_index

** Z1: Income: gdp_pp

** Z2: Political participation: political_p

*Test for skew:

sktest corrup_index eco_freedom dr_eg gdp_pp political_p eco_freedom

sum corrup_index eco_freedom dr_eg gdp_pp political_p, detail

* Transforms variables,

gen log_gdp = ln(gdp_pp)

gen log_trade = ln(fr_trade)

gen log_politics = ln(political_p)

gen log_corrup = ln(corrup_index)

gen log_eco_freedom = ln(eco_freedom)

**Use panel data

encode country, gen(country1)

xtset country1 year

sort country1 year

***Estimate the CSTS model w/FE estimator (with lag DV)

xtreg log_corrup l1.log_corrup l1.log_eco_freedom l1.log_gdp l1.log_politics l1.dr_eg l1.log_trade, fe

* Test for the appropriateness of RE vs FE (Hausman test)

xtreg log_corrup l1.log_corrup l1.log_eco_freedom l1.log_gdp l1.log_politics l1.dr_eg l1.log_trade, fe

estimates store fixed

xtreg log_corrup l1.log_corrup l1.log_eco_freedom l1.log_gdp l1.log_politics l1.dr_eg l1.log_trade, re

estimates store random

hausman fixed random // note that p<0.05. MUST USE FE!!! RE BAD!

**** Table (regression).

quietly xtreg log_corrup l1.log_corrup l1.log_eco_freedom l1.log_gdp l1.log_politics l1.dr_eg l1.log_trade, fe //

eststo clear

eststo: quietly xtreg log_corrup l1.log_corrup l1.log_eco_freedom l1.log_gdp l1.log_politics l1.dr_eg l1.log_trade, fe

eststo: quietly xtreg log_corrup l1.log_corrup l1.log_eco_freedom l1.log_trade, fe

eststo: quietly xtreg log_corrup l1.log_corrup l1.log_eco_freedom, fe

esttab using file_name1.rtf, replace label /// creates a table in the file "file_name1.rtf"

stats(r2_a N, fmt(2 0) labels("R-squared" "Observations")) /// adds R-sq and N

legend varlabels(_cons Constant) ///

addnote("Note: Absolute value of t statistics in parentheses.") ///

b(a2) t(2) /// "b(a2)" will allow the slope coefficient to go to 2 significant digits if necessary. **only 2 decimals for each number.

title({\b Table 1. Pooled Time-Series Cross-national Regressions, with Fixed Effects})

eststo clear

***Table: Descriptive statistics

eststo clear

estpost tabstat dr_eg eco_freedom corrup_index gdp_pp political_p, ///

listwise statistics(mean sd min max) columns(statistics)

esttab using ps4sum.rtf, replace label ///

cells("mean(fmt(2) label(Mean)) sd(fmt(2) label(Std. Dev.)) min(fmt(1) label(Min.)) max(fmt(1) label(Max.))") ///

nonumbers extracols(5) ///

title({\b Table 1. Summary Statistics})

eststo clear

*** Question: Does economic freedom impact in corruption perception (2000)?

* Inspect the variables

codebook corrup_index eco_freedom

* Is there evidence of covariation?

twoway (scatter corrup_index eco_freedom) (lowess corrup_index eco_freedom) (lfit corrup_index eco_freedom)

*Is the line a good representation of the non-parametric regression? NO!

twoway (scatter log_corrup log_eco_freedom) (lowess log_corrup log_eco_freedom) (lfit log_corrup log_eco_freedom)

* Estimate the regression of log(Y) on log(X): log(Yi) = alpha + beta(log(Xi)) +ui (reg logY logX)

* H0: beta=0; HA: beta!=0

reg log_corrup log_eco_freedom

* Display your estimates! reg logY logX; pred yhat; gen ylin=exp(yhat); (line ylin X, sort)

reg log_corrup log_eco_freedom

predict mort_hat

gen mort_lin = exp(mort_hat)

twoway (scatter corrup_index eco_freedom) (lowess corrup_index eco_freedom) (line mort_lin eco_freedom, sort)

twoway (scatter corrup_index eco_freedom, jitter(2) msize(small) mcolor(black)) ///

(lowess corrup_index eco_freedom, lcolor(blue) lpattern(dash)), ///

title("Figure A. Effect of Economic Freedom Index on Corruption perception", size(medium) color(black)) legend(off), ///

(line mort_lin eco_freedom, sort clcolor(red)), ///

ytitle("Corruption Perception Index (CPI)", size(small)) ///

xtitle("Economic Freedom Index", size(small)) ///

note("Source: World Bank World Development Indicators and Quality of Governance", size(vsmall)) ///

caption("Note: Dashed line is lowess estimate (bandwidth = 0.7). Solid line is OLS prediction", size(vsmall)) ///

graphregion(fcolor(white) lcolor(white)) ///

plotregion(fcolor(white) lcolor(black)) //

[1] I used the Hausman test to determine the appropriateness of Fix Effects over Random Effects. The model includes country fixed-effects (p<0.05) to allow any trends in corruption perception to vary from country to country. (without fixed-effects, the results are unchanged)

BP2: Oil Prices and SUV Sales in the US

2015-04-30T16:39:00.002-04:00

The Effects of Oil Price Changes on Light Truck Sales in the US

Robert Avakian

Introduction

In late fall of 2014 US sales of vehicles in the light truck category increased considerably. This increase coincided with a sharp fall in oil prices and steady economic growth. Some news reports linked the drop in oil prices to the spike in SUV sales raising the question as to whether changes in the price of oil have an impact on the sales of large vehicles such as SUVs, trucks and vans.

Background

In the late summer of 2014 oil prices begun a drop that would take them from just over $100 per barrel in late July to a low of $43 by early January of 2015. By early November 2014 news reports began appearing that linked the drop in the price of oil to a ten year high in US sales of trucks and SUVs (The Guardian, Dec 2014). SUV brands ranging from the regular Ford Escape to the luxury Cadillac Escalade experienced sale jumps between 6 and 38% compared to the previous year. Surveys showed that fuel economy which had ranked No. 3 in 2013 on the list of reasons that people buy cars, in 2014 had dropped to No. 6 behind features like styling, brand preference, ride quality and reputation (LA Times, Nov. 2014).

At the same time however, low oil prices also coincided with a with a period of steady economic growth. Employment was back to pre-2009 levels resulting in increased disposable income, a more secure economic outlook and the possibility that oil prices were actually having a marginal or no effect at all on SUV sales.

Determining the actual effect of oil prices on vehicle sales could provide valuable insights for automakers planning marketing campaigns, climate change policymakers and regulators determining new fuel consumption and emissions standards and, behavioral economists measuring the effect of short term incentives in an environment with long-term uncertainty.
Research Design and Methodology

The study analyses the effects of oil prices on vehicle sales over a 30 year period using sales, economic and oil price data from 1985 to 2015. The time interval is chosen based on the historical availability of monthly crude oil price data. All vehicles sales data is seasonally adjusted and divided between the light truck category (SUVs, trucks and vans) and the light car category (compacts, sedans, crossovers). The study focuses specifically on the light truck category sales given their relatively low fuel efficiency and potentially higher susceptibility to high oil prices. Light cars sales are used to check whether oil prices affect the entire industry or just a specific category.

The economic variables used in the analysis include monthly data for GDP growth, GDP per capita and disposable income. The GDP growth rate and GDP per capita are used to measure the effect of economic trends. Disposable income is used as an indicator of the financial means of individuals which, can lower the susceptibility to high oil prices as well as play a major role in the choice of an SUV over a cheaper compact car. A summary statistics table is shown in the appendix.

The data is analyzed using time series Cochrane-Orcutt AR(1) regression that accounts for auto-correlation issues (see appendix for details). Because oil price (and economic indicator) changes are not believed to have an immediate effect but rather take some time to have an actual economic impact, vehicle sales are analyzed using a 1 and 3 month lead.These models assume that oil price changes in August actually impact sales in September or November rather than in August. From a behavioral economics point of view it assumes that people take 1-3 months to internalize the change in gasoline prices and determine whether to buy and what type of vehicle to buy.

Analysis

As seen in figure 1 there appears to be some correlation as oil price drops or spikes seem to be followed by changes in light truck sales. However, oil prices are not exogenous and can be influenced by economic factors, like GDP, which can also affect vehicle sales.

Figure 1

The regression analysis results in table 1 provide a more detailed analysis of the effect that oil price has on vehicle sales when taking into account economic trends such as GDP growth. Models 1-5 measure the effect of oil prices on light truck sales while model 6 measures the same effect on the light cars category.

The results in models 3 and 5 suggest that on average oil prices do have an effect on light truck sales. An approximately $1 increase in the per barrel price of crude oil leads to a 1000 unit decrease in light truck sales. GDP growth on the other hand has a significant positive effect with an average increase of 25 thousand units per 1% of GDP growth. The analysis of model 6 which measured light car sales also shows a positive effect from GDP growth but no significant effect from oil price changes.

Table 1. Truck and Car Sales by Oil Prices between 1985-2015

	(1)	(2)	(3)	(4)	(5)	(6)
DV: Seasonally Adjusted US Light Truck Sales (in thousands)	Truck Sales with 3 Month Time Lead	Truck Sales with 1 Month Time Lead	Truck Sales with 3 Month Time Lead	Truck Sales with 3 Month Time Lead	Truck Sales with 3 Month Time Lead	Car Sales with 3 Month Time Lead
Price of WTI Crude Oil	-0.189	-1.004	-1.154*	-1.793	-1.222*	-1.803
	(0.366)	(1.914)	(2.214)	(1.152)	(2.113)	(1.201)
Disposable Income per Capita		0.040*	0.015	0.014	0.015	-0.016
		(2.000)	(0.752)	(0.705)	(0.776)	(1.141)
% GDP Growth Month-to-Month		1.824	25.991**	24.344*	25.065*	20.479*
		(0.191)	(2.739)	(2.428)	(2.542)	(1.983)
GDP Per Capita		-0.006	0.005	0.015	0.013	0.032**
		(0.649)	(0.644)	(0.732)	(0.664)	(2.624)
Square of Oil Price				0.004		0.010
				(0.398)		(1.070)
Time				-0.929	-0.857	-3.403***
				(0.473)	(0.435)	(3.848)
Oil over $100					1.226
					(0.083)
Constant	605.054***	-403.970	-45.023	-169.034	-178.301	691.381**
	(9.725)	(1.245)	(0.141)	(0.388)	(0.409)	(3.174)
Observations	360	360	360	360	360	360
R-squared	0.000	0.063	0.076	0.079	0.078	0.328

Note:*Significant at 5%; **significant at 1%; ***significant at 0.1%. * p < 0.05, ** p < 0.01, *** p < 0.001
Source: Bureau of Economic Analysis (BEA) and Energy Information Agency (EIA)

Conclusions

The results of the analysis suggest that oil prices on average do have an effect on light truck sales. The fact that there is no oil price effect on light car sales but a similar GDP growth effect seems to confirm the susceptibility of the light truck category to higher oil prices. It is possible that while light cars are purchased as a necessity whenever economic conditions allow it, SUVs and trucks are seen by some buyers as more of a luxury which requires more favorable economic as well as oil price conditions in order to justify the leap in category and price.

A Bloomberg analysis suggests that "even in an environment of high gas prices, much of the growth in sales in recent years has come from...light trucks which were already up nearly 10 percent in the year through October, while cars were up a mere 1.4 percent (Bloomberg, Dec. 2014)." This interpretation can still be consistent with the above presented results as it suggests that GDP growth has a greater effect than oil prices on light truck sales but leaves room for oil prices to further influence the buyers final decision. This analysis would be consistent with the late 2014 spike of already growing light truck sales.

The above mentioned difference in pre-oil price drop sales between light trucks and cars (10% vs 1.4%) could be due to a latent demand that had been so far kept down by the effects of the financial crisis and unemployment. A further study focusing on controlling for the effects of economic crises could be conducted to further clarify the issue.

------------------------------------------------------------------------------------------------------

Appendix

Table 2. Summary Statistics from 1985-2015


	Mean	Std. Dev.	Min.	Max.
Seasonally Adjusted Light Truck Sales (in Thousands)	563.6	150.35	305.	1049.1
Seasonally Adjusted Car Vehicle Sales (in Thousands)	687.97	123.55	372.2	1254.2
Price of WTI Crude Oil	42.2	30.53	11.3	133.9
Disposable Income per Capita	30616.12	4811.37	22675.8	37832
% GDP Growth Month-to-Month	0.66	0.58	-2.1	1.9
GDP Per Capita	35566.38	11443.96	0.6	55319
Observations	364

Source: Bureau of Economic Analysis (BEA) and Energy Information Energy (EIA)

Trend Effects Concerns

Given the variance in the values of the variables shown in figures 2,3 and 4 there is no concern that steadily increasing or decreasing values over long periods of time can create trending effects that give false indications of significant effect. In regression models 4, 5 and 6 the variable "Time" is used to account for any trend effects of the growth in the date variable.

Figure 2.

Figure 3.

Figure 4.

Auto-correlation Correction

A Durbin-Watson test provided a p-value<0.001 which indicated considerable autocorrelation in the regression models. A Prais–Winsten first order auto-regressive estimation was used to correct for the effects of serial correlation.

Regression Models

Model 2 uses a 1 month lead for the dependent variable representing truck sales and is the only model where disposable income is significant. However, with a t-statistic of 2.000 the variable is barely significant and the model is dropped in further developed instead favoring the 3 month lead models.

Models 4 and 6 include a variable that has the squared values of oil prices. This variable is included to check whether rising oil prices have an parabolic or plateauing effect on car sales. The variable does not show significant effect in either model. However, its inclusion in model 4 results in the standard oil price variable becoming insignificant.

In model 5 the dummy variable "Oil over $100" is included to check whether prices over the $100 threshold have a significant effect on sales. The variable is not statistically significant.

U.S. Military Assistance and Political Stability

2015-04-30T16:32:00.001-04:00

By: Jenny Shelton

United States military assistance is an important component of U.S. relationships abroad and U.S. foreign policy. In this analysis, I examine the effects of U.S. military assistance on political stability within countries. Understanding the effects of U.S. military assistance can aid policy makers in allocating funds to best utilize the United States military and its budget. Additionally, because a fundamental aspect of international relations is forming partnerships with stable countries in key regions of the world, exploring the effects of U.S. military assistance on these countries’ stability can offer insights into optimal U.S. foreign policy.

Much of the literature on U.S. military assistance isolates specific types of security assistance, such as counterterrorism or police force training. For instance, Dube and Naidu examine how paramilitary and guerrilla groups respond to U.S. military aid in Colombia. Additionally, a RAND report observes the effects of security cooperation, but still focuses on authoritarian regimes and states with pre-existing fragility. My analysis looks at overall effects of military assistance, providing a general foundation for more specific analyses.

In this analysis, I use cross-sectional time-series model, focusing on countries receiving U.S. military assistance from 2003-2012.[i] I chose a cross-sectional time-series analysis over a quasi-experimental design because finding two similar countries, one of which received military assistance, to serve as treatment and control groups is extremely rare.[ii] In the model, I also include three control variables that might affect the relationship between military assistance and stability. Rule of law, GDP, and level of democracy might relate to both the explanatory and outcome variables; an increase in any of these variables might decrease the need for military assistance and increase stability. While I am concerned about selection bias, or whether the sample is representative of the population, with these cases, the lack of an available comparison case limits my ability to overcome this issue.

I first performed tests to determine whether the variables were normally distributed and if the model’s effects were consistent.[iii] Because the effects of military assistance are not instantaneous, I lagged political stability by two years. I regressed the lagged political stability variable on the independent variables, using a fixed effects estimator. Figure A is a visual depiction of the effect of U.S. military assistance on political stability. The graph shows a decline in political stability with increased military assistance; however, this negative relationship may be exaggerated by the extreme instability of a few cases receiving high levels of military assistance (seen in the lower right-hand side of the graph).

In contrast, the estimates of the variables depict a slightly positive relationship between U.S. military assistance and political stability (see Appendix A for a table of regression estimates). In a bivariate model (Model 1, Appendix A), a one-unit increase in U.S. military assistance is associated with a 0.055-unit increase of political stability within countries, on average. This effect is statistically significant with a p-value of less than 0.05. This value indicates that there is less than a 5% chance that the observed relationship is due to chance. The effect of U.S. military assistance is substantively similar across the models, but only retains its significance when GDP is not included in the model. In the multivariate model (Model 4, Appendix A), both level of democracy and rule of law have a positive effect on political stability within countries, with a one-unit increase corresponding with a 0.007-unit increase and 0.691-unit increase respectively, on average and holding all else constant. These effects are not statistically significant. GDP also has a positive effect; a one-unit increase in GDP is associated with a 0.215-unit increase in political stability within countries, on average and holding all else constant. This effect is statistically significant at the 0.1% level—the only statistically significant explanatory variable in the full model.

An examination of the changes in political stability in Serbia and in Montenegro from 2006-2012 supports these results. Montenegro gained its independence from Serbia in 2006, so it is reasonable to assume that these countries serve as a good comparison. After independence, Montenegro’s level of political stability increased sharply, then declined in 2009 and remained constant (Figure B). Montenegro did not receive military assistance from the United States. In contrast, Serbia, which began receiving U.S. military assistance in 2007, experienced an upward trend in political stability from 2007-2012. As the amount of military assistance to Serbia increased, political stability in Serbia rose, as well (Figures C and D).

Thus, both the cross-sectional time-series analysis and the case study of Montenegro and Serbia suggest that increased levels of U.S. military assistance are associated with an increase in political stability. However, I remain cautious about confirming causality; the military assistance variable lacked statistical significance in the multivariate model, and the level of increase was not substantively large. Despite these limitations, the analysis indicates that the use of U.S. military assistance may increase political stability, potentially contributing to U.S. relationships abroad and foreign policy.

[i] Data availability limited the observations to a ten-year span.

[ii] Discontinuity design is in appropriate because there are no clear criteria for entry. Difference-in-differences would require similar treatment and control groups, but there are few, if any, scenarios in which one of two nearly identical countries received military assistance while the other did not. Finally, I was unable to find an instrument for instrumental variable design (2SLS) that only influenced the dependent variable through the independent variable.

[iii] I found that the military assistance, rule of law, and GDP variables were positively skewed and logged these variables to correct skewness. The Hausman test yielded a p-value of 0.000, which indicated that the difference in coefficients is not systematic and the fixed effects estimator is preferred. I also tested for multicollinearity, but found no strong collinearity among the variables.

The Israel Lobby and Congressional Foreign Policy

2015-04-30T16:24:00.000-04:00

One of the most contentious issues in contemporary international relations is the US’s alliance with Israel. It is generally undisputed that Israel and the US have a close relationship – Israel is America’s biggest client for foreign aid and military assistance, receiving about $3 billion each year. However, the cause of America’s support for the small Middle Eastern nation is disputed. John J. Mearsheimer and Stephen M. Walt argue that the main cause of the American-Israeli alliance is the “Israel Lobby.” In fact, the two scholars argue that “a key pillar of the lobby’s effectiveness is its influence in the US Congress,” which largely determines the allocation of funds to Israel and general American political sentiment towards it.[1] On the contrary, others have argued that American support for Israel stems from the recognition of a mutual identity between the two countries. “The Israel Saga – of courage and will,” writes Roger Cohen, “echoes in American mythology, far beyond religious identification.”[2]

The purpose of this study is to test Mearsheimer and Walt’s hypothesis that the Israel Lobby has a strong influence on the US Congress. The study measures the effect of increased campaign donations by the pro-Israel lobby on congressional support for Israel. Mearsheimer and Walt state that, “Given the strategic importance of the Middle East and its potential impact on others, both Americans and non-Americans need to understand and address the lobby’s influence on U.S. policy.”[3] Indeed, the Middle East continues to be an international hotspot and it is important to understand the factors that govern the US’s relationship with the region.

The study measures the percentage of pro-Israel bills that each member of the House of Representatives voted in favor of across the 110^th, 111^th, and 112^th congresses. The 110^th congress, which lasted from 2007 to 2009, is used as a baseline. In 2010, in a case titled Citizens United v. Federal Elections Committee, the Supreme Court ruled that interest groups cannot be limited in the amount they spend on campaign-related contributions. In this model, then, the 111^th and 112^th congresses which follow the Citizens United decision are the post-treatment groups, in which the pro-Israel lobby is assumed to have more influence over congressional campaigns and, subsequently, is thought to have more influence on congressional matters.

Table 1. Summary Statistics

	Mean	Std. Dev.	Min.	Max.	Description
Party	N.A.	N.A	0.0	1.0	1: Republican, 0: Democrat
Congress 110	N.A.	N.A.	110	110	110^th Congress, 2007-2009
Congress 111	N.A.	N.A.	111	111	111^th Congress, 2009-2011
Congress 112	N.A.	N.A.	112	112	112^th Congress, 2011-2013
Percentage for Israel	0.80	0.25	0.0	1.0	Percent of Pro-Israel Bills House Rep Supported
Observations	804

The study observed voting patterns for 286 House Representatives, all of whom were in office between 2007 and January of 2012. The model does not measure support for Israel in the Senate, since senatorial action may be dependent on approaching election seasons. The House of Representatives, on the other hand, is constantly in election season since representatives are voted into office every two years, instead of six. The dataset excludes representatives who served for some but not all of the congressional sessions to reduce bias caused by reverse causality. In other words, one of the reasons that the representatives did not get reelected is because of their voting patterns on this issue; this would cause bias estimates in the model and, therefore, those cases are omitted. Data about voting patterns was sourced from the GovTrack.[4] The dependent variable, Percentage for Israel, measures the percent of pro-Israel bills that each House member voted in favor of in each congressional session (for a full list of variables used see Table 1). The model in the study is a cross-section time-series model with a fixed effects estimator.[5] The model uses fixed effects in order to control for time-invariant confounds that apply to representatives during their time in office. For example, a fixed effects model allows the study to control for every congressman’s specific party, state, demographic makeup of constituency, seniority within the party, and competitiveness for the representative’s seat.

TABLE 1. Support for Israel per Congress, Fixed Effects

	DV: Percentage of Pro-Israel Bills Supported
Congress 111	-0.186^***
	(-9.07)
Congress 112	-0.218^***
	(-10.61)
Party	0.000
	.
Constant	0.937^***
	(64.59)
R-Squared – Overall R-Squared – Within	.15 .20
Observations Representatives	804 286

T statistics in parenthesis. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001

The findings of the study demonstrate that an increase in funding and donation contributions by the Israel Lobby did not have a positive effect on the percentage of pro-Israel bills that House members supported (for a visual of the average decline in pro-Israel votes see Figure 1). In fact, both periods after the treatment – the 111^th and 112^th congresses – observed a decline, on average, of support for pro-Israel bills. The 111^th Congress observed an average decline of 19% in support, and the 112^th congress dropped even lower by 22% (see Table 2 for all regression estimates). In other words, although the pro-Israel lobby was allocating more funding towards congress people, this money was not translating into higher support for Israel – at least not in terms of voting for bills that are considered pro-Israel.

Although the study does not find a positive correlation between increased funding by the pro-Israel lobby and congressional support for Israel, there may be other variables that affect this relationship that are beyond the scope of this study. For example, congressional lobbying does not solely consist of donation funding – it may also consist of various meeting, briefings, and interactions between the lobbying group and congressional representatives that are difficult to measure.

[1] John J. Mearsheimer and Stephen Walt. “The Israel Lobby and US Foreign Policy.” University of Chicago, 2006.

[2] Roger Cohen. “Why Americans See Israel the Way they Do.” New York Times. August 2, 2014.

[3] John J. Mearsheimer and Stephen Walt. “The Israel Lobby and US Foreign Policy.” University of Chicago, 2006.

[4] GovTrack.us is a project of Civic Impulse, LLC which tracks Congressional operations and voting patterns. https://www.govtrack.us/about

[5] Although the Hausman test permitted a random effects model for this dataset, a fixed effects model is more accurate in reflecting the effect since it controls for all time invariant variables for each representative.

Does Community Oriented Policing have an Effect on Violent Crime?

2015-04-30T15:59:00.008-04:00

Introduction

The way a society chooses to approach issues of crime and justice is a mirror of political, social, and cultural norms and institutions. As society slowly changes and evolves, public demands and expectations of government change with it. American policing has always walked a tightrope between maintaining order through force and providing civil service. Since the 1980s the American public has increasingly expected police to “provide safety and security in a way that does not violate their privacy and freedoms.”[i] The concept of Community Oriented Policing (COP) was developed to more proactively combat crime by working with, rather than against, communities to solve problems.[ii] This study aims to assess the effect of the implementation of community oriented policing programs on crime. I find that, in the assessed communities, these programs have significant but complex impact.

Study Design

The most effective study would focus on a single department that implemented the program in selective districts. By selecting similar districts that did not implement the program, a “difference in differences” method could be used to approximate experimental conditions by identifying deviations from a common trend in the policy group. Additionally, aggregate crime data only tells part of the story of a community policing initiative; to that end, surveys of community attitudes and perceptions would be useful to track effectiveness. A study like this was conducted by the Northwestern University’s Institute for Policy Research on community policing initiative in Chicago.[iii]

This study attempts to approximate a similar quasi-experimental design using publicly available crime data from the FBI’s Uniform Crime Reporting database to examine the impact of community policing programs on a larger scale. Several county level Sheriff’s Offices in California implemented community policing programs simultaneously in 2007[1]. By comparing rates of reported and cleared[2] crimes between counties with newly implemented programs and those that without, insight can be gained into changes in the dynamics of community-police relations on a larger scale. Data were collated from multiple sources were compared for all counties in California to determine dates of community policing initiatives.[3] Those that had implemented policies prior to 2007 were eliminated from the sample. The counties in the control group (those that had not implemented a policy by 2010) represented a diverse range of urban and rural sections of the state. The counties that implemented the policies (the treatment group) were similarly diverse. Since there were significantly more counties in the control group than the treatment group all statistics were standardized by aggregate populations of the two groups.[4]

County Sheriff’s Offices were chosen as the unit of comparison to draw attention to the wider populations served by these units outside of metropolitan areas – this helps to eliminate some of the concerns of biased results arising from specific differences between municipalities. Table 1 provides summary statistics for the variables used in the assessment.

Table 1. Summary Statistics

	Mean	Std. Dev.	Min.	Max.	Description
Violent Crime Reports	115.35	85.7	5.41	536.86	Total reports of violent crimes by county and year per 100,000 residents
Property Crime Reports	6148.16	405.62	42.36	2529.4	Total reports of property crimes by county and year per 100,000 residents
Cleared Violent Crimes	57.92	43.07	0	190.72	Total violent crimes cleared by officers by county and year per 100,000 residents
Cleared Property Crimes	77.84	72.86	0	515.51	Total property crimes cleared by officers by county and year per 100,000 residents
Violent Difference	57.43	64.95	0	526.53	Difference in violent crimes reported and cleared (per 100,000 residents)
Population	558423	623102	9686	3010232	Total population by County
Post	0.57	0.50	0	1	=0 if year < 2007, =1 if year > 2007
Treat	0.30	0.46	0	1	=0 if in control group, =1 if in treatment group
Counties: 23		Treatment:	7	Control:	16
Observations: 161
Source: CA AG, FBI UCR, BJS LEMAS

Results

In order for a difference-in-differences study to be effective, the two groups (the “control” and “treatment”) must exhibit common trends prior to policy implementation. An analysis of trends over time for reports of property crimes indicated that the control and treatment group both followed a common trend from 2004-2010. No statistically significant divergence from the trend occurred in the treatment group after the COP policy was implemented in 2007. As a result, no further analysis was conducted on property crimes. Figure 1 shows the time trends of reported violent crimes of the treatment and control groups.[5] A clear departure from the common trend is seen in 2008 when the treatment group shifts towards a greater number of crimes reported for the next three years.

Table 2 shows the results of the statistical analysis of violent crime reports. It shows that the treatment group saw a statistically significant increase of 12.82 violent crimes reported per 100,000 residents the year after the COP policy was implemented.

Table 2. Violent Crime Reports

DV: Violent Crime Reports	Pre-2007	Post-2007	Difference
COP Policy	111.14	115.84	4.7
No Policy	90.22	82.1	-8.12
Difference	20.92	33.74	12.82**

Table 3 shows the results of a similar analysis conducted on the effect of implementing a COP policy on the difference between reported violent crimes and cleared violent crimes. This shows that, while both reported and cleared violent crimes increased in the treatment groups after policy implementation, the gap between them decreased by 1.64 per 100,000 residents in 2008.

Table 3. Difference in Violent Crimes Reported vs. Cleared

DV: Reports - Clearances	Pre-2007	Post-2007	Difference
COP Policy	55.19	53.15	2.04
No COP Policy	58.75	58.35*	-0.397*
Difference	3.56	5.2	-1.64*

Discussion

The implications of the results of this study are not straightforward. At first glance, the analysis seems to suggest that community policing initiatives lead to more violent crime. The treatment group demonstrated a clear upward departure from the common trend in both reported and cleared violent crimes. Should we then simply scrap the whole notion of policing by engaging the community?

Not at all. In fact, the results here are exactly what we would expect to see given the goals of these programs. An uptick in reporting is indicative of improved relationships between police and the community. It is likely that a larger number of crimes go unreported when the community does not trust their police force. Though this study does not take into account events that could have occurred between 2007 and 2008 that could have affected crime reports, the number and variety of agencies in the treatment group suggest causality more than a simple cross-sectional analysis.

Of additional interest is the result of the analysis of the difference between cleared and reported violent crimes. After policy implementation, both reported and cleared violent crimes increased but there was a decrease in the difference between them. This suggests that not only are citizens reporting more of the crimes that are occurring but that those reports are also more substantial. There appears to be less of a divide between what citizens see as a crime and what the police see as a crime. This also suggests that there is an improving relationship between police and the community.

Conclusion

Of course, this study is severely limited in its ability to draw far-reaching conclusions about the effectiveness of these policies. It does not examine how or to what degree these programs were implemented in each agency. Were they simply throwaways on paper that led to business as usual or a concerted effort to change operations? Without in-depth individual studies we don’t have that information available. It is also unclear why a change was seen in violent crime activity but not property crime activity. One possible explanation assumes that residents were more likely to fully report property crimes than violent crimes (with lower associated investigation and arrest rates) already before the COP policy was implemented. Evolving policing norms will have a complicated impact on the community; this study showed one aspect of a significant effect of a popular policy choice.

[1] See Appendix A for a list of included counties.

[2] Police departments around the country separate crimes into two categories: reported and cleared. Crimes reported represent the total number of calls for service for a particular type of crime. Crimes cleared are those that are substantiated by a responding officer who notes a disposition for the contact. Dispositions can be official reports, an arrest, the opening of an investigation, etc.

[3] Sources include: the FBI Uniform Crime Reporting data, the State of California’s Attorney General’s Office, the Law Enforcement Management and Administration Survey.

[4] For clarity of data, per capita statistics were determined per 100,000 residents in each group.

[5] A similar trend at a different scale was found for trends of treatment and control groups regarding violent crimes cleared. These trends are shown in Figure 2.

References

[i] Bush, Michael and Dodson, Kimberly. Police Officers as Peace Officers: A Philosophical and Theoretical Examination of Policing from a Peacemaking Approach. Journal of Theoretical and Philosophical Criminology. Sept. 2014 pp.194-204

[ii] Cordner, G. W. (2010). Community Policing Elements and Effects. In R. G. Dunham, & G. P. Alpert, Critical Issues in Policing (pp. 432-449). Long Grove, IL: Waveland Press Inc.

[iii] Skogan, Wesley (1996). Evaluating Problem Solving Policing: The Chicago Experience. Institute for Policy Research, Northwestern University

Appendix A

El Dorado*	Placer*
Fresno*	Sacramento
Humboldt	San Bernardino
Imperial	San Luis Obispo
Kern*	San Mateo
Lassen	Santa Barbara*
Marin	Santa Cruz
Merced	Shasta
Modoc	Stanislaus
Monterey*	Tulare*
Orange	Ventura
	Yolo
*indicates part of treatment group

Did the Affordable Care Act Actually Destroy or Redeem Obama's Approval Ratings?

2015-04-30T15:54:00.002-04:00

By Sharon Hoeck

The passage of the Patient Protection and Affordable Care Act (also known as the ACA or Obamacare) in 2010 marked the first significant healthcare reform in the United States in decades. The ACA reforms the way health insurance is sold, aiming to “put consumers back in charge of their healthcare” by attempting to make health insurance more affordable for people with pre-existing conditions, to expand insurance coverage to young adults just graduating from college, to guarantee that individuals have the right to appeal insurance companies’ decisions not to pay their claims, and to expand Medicare. Despite its good intensions, the law has been met with opposition from individuals and politicians across the ideological spectrum who contest that the federal government is overreaching its authority and that healthcare reform should be a state-by-state decision.

Obama and his supporters argued that the electoral victory in 2012 showed that the majority of the American people approved of the law and its reforms. Republicans disagreed, ultimately shutting down the government for weeks in the fall of 2013 in an attempt to effectively repeal the ACA by defunding its programs. Reflecting this ideological split for politicians, the media has also taken sides. Depending on your preferred news source, the ACA is either the best thing or the worst thing that has ever happened to healthcare in this country and will either be the highlight or the bane of Obama’s presidential legacy. Politicians have avoided healthcare reform for years, as changes to insurance policies and Medicare are generally viewed as a dangerous issue with the potential to ruin political careers.

But does the ACA really polarize the American as much as some media outlets and politicians would have us believe? In this post, I will address this question by examining the effect of the ACA on Obama’s presidential job approval rating. To do this, I compare two states: one that has instituted a state exchange through the ACA (Minnesota) and one that has not (Wisconsin). My hypothesis is that, if the ACA is as polarizing as some lead us to believe, Minnesota’s presidential job approval ratings must have changed significantly either for the better or the worse after it passed the law implementing a state exchange.

I chose Minnesota and Wisconsin as my comparison cases because of their similarities apart from their implementation of the ACA’s state exchanges or Medicaid expansion. Both states, for example, had similar percentages of uninsured individuals prior to the ACA’s passage as well as similar median ages and similar percentages of individuals employed by small businesses. Additionally, despite Wisconsin’s more conservative state-level politics, the states have similar presidential voting records over the past twenty-three years. Most important for my purposes, however, is the similarity between the two states in terms of their approval of the job that Obama was doing as president before Minnesota passed its health exchange law in late October of 2011 (Figure 1). Both states show the decrease in approval that naturally follows a president’s election, and both states show ratings that leveled off in 2010 only to rise slightly in 2011. I can therefore assume that if Minnesota had not established a state exchange in October 2011, its presidential job approval ratings would be similar to those of Wisconsin (see Appendix A for a more detailed explanation of my rationale). Because Minnesota’s law was not signed until the end of 2011, the difference in approval rating trends after that year could be the effect of Minnesota’s full compliance with the ACA (see Appendix B for information on the approval ratings used).

While Figure 1 shows Minnesota’s presidential approval ratings decreasing at a slower rate than those of Wisconsin following Obama’s election to a second term in 2012, Table 1 shows that this difference in trend is not statistically significant and is therefore more likely due to random noise in the model than any systemic difference in opinion caused by Minnesota’s implementation of a state exchange (see Appendix C for a note on my method).

Table 1. Average Effect of State Exchange Legalization

DV: Presidential Approval Ratings	Pre-Exchange	Post-Exchange	Difference
Minnesota	47.95	48***	0.05
Wisconsin	47.6***	46.1	-1.5
Difference	0.35	1.9	1.55

Note: Differences-in-differences estimates. *, **, *** indicate statistical significance at the 10%, 5%, and
1% level respectively.

Despite the statistical insignificance of my results, comparing Wisconsin and Minnesota shows that the passage of the ACA did not result in a systematic change in public approval for the president. This result brings into question the validity of the idea that healthcare reform is a polarizing and potentially politically deadly issue for our representatives to address. The implementation of the many of the ACA’s provisions did not occur until January of 2014 and sufficient data reflecting public opinion after the law was fully enacted is not yet available. However, future data will only reflect the American public’s opinions on the ability of the law to function properly and make a difference in people’s lives. The data that I used shows that the act of drafting and passing comprehensive healthcare reform in and of itself will not destroy or resurrect a politician’s approval ratings. In this instance at least, the American public is showing themselves to be more levelheaded than some politicians and media outlets assumed. The functionality of the ACA and its programs will have much more of an effect on Obama’s legacy and long-term approval ratings than the political firestorm that surrounded the ACA’s passage. Legislators should therefore reconsider avoiding any discussion on issues that they previously believed to be off-limits or a political “third rail.”

Appendices

Appendix A: State Similarities

According to the U.S. Census Bureau, the percentage of Minnesotans without health insurance in 2009 measured at 8.8. Wisconsin’s percentage was not significantly higher at 9.5. A similar percentage of the population in both states had the potential to gain access to health insurance through the ACA, and would therefore affect approval ratings similarly.

The median age in Minnesota, at 37.4, is only slightly lower than Wisconsin’s 38.5 (U.S. Census Bureau, 2010). A similar percentage therefore stand to gain coverage through the ACA provision that allows young adults to remain under their parents’ health insurance until their 26^th birthday.

Additionally, in 2011, the Census Bureau marked the percentage of the total employed population working for small businesses at 22.9 in Minnesota and 23.5 in Wisconsin. Because the ACA mandates that businesses employing more than fifty people full-time must share health insurance responsibilities with their employees by 2016, I limited the above percentages to firms that employ between 50 and 499 people. Many have voiced the concern that small businesses, when forced to share health insurance costs with their employees, would be compelled to either fire many of their staff or significantly reduce their hours. Wisconsin and Minnesota have similar levels of small business employment and therefore relatively similar proportions of the population that would see the ACA as a threat.

Electoral results from both states are also similar, with Obama winning both states in 2008 (with 54.06% and 56.22% of the vote in Minnesota and Wisconsin respectively) and in 2012 (with 52.65% and 52.83%). Further, the popular vote (though not the electoral vote) in the two states has gone to the same candidate in every presidential election since 1992. And while Wisconsin currently has a republican governor and Minnesota has a democratic governor, the state house of both states changed party hands in 2010 elections (from Democrat Jim Doyle to Republican Scott Walker in Wisconsin and from Republican Tim Pawlenty to Democrat Mark Dayton in Minnesota). The two states are therefore prone to similar party affiliation and to a similar inherent approval or disapproval of Obama’s actions.

Despite the differences between the two states, therefore, the major systemic factors in both states that could affect their respective presidential approval ratings are similar enough to allow me to assume that if Minnesota had not established a state exchange in October 2011, its presidential job approval ratings would be similar to those seen in Wisconsin’s.

Appendix B: Data

All presidential job approval data comes from Gallup. The following is paraphrased from Gallup’s website concerning their collection method: Gallup tracks the percentage of Americans who approve of the job that Barack Obama is doing as president. Daily results are calculated from telephone interviews with approximately 1,500 individuals and are based on a three-day rolling average. Margin of error is ± 3 percentage points.

Table 2. Summary Statistics

	Mean	Std. Dev.	Min.	Max.	Description
Minnesota	50.23	5.74	45.4	61.5	Annual Averaged Presidential Job Approval Rating, 2009-2014
Wisconsin	48.53	5.10	42.4	57.7	Annual Averaged Presidential Job Approval Rating, 2009-2014
Observations	6

Source: Gallup Daily Presidential Job Approval Ratings 2009-2014

Appendix C: Method

When running the Difference in Differences regression that produced the values in Table 1, I used a lead measure of the exchange variable. For example, had my results been significant, I would have been able to use the presence of an exchange in Minnesota in 2012 to predict presidential approval ratings in Minnesota in 2013. In this way, I can justify the assumption that the ACA influences approval ratings, not vice versa.

Does Revenue Sharing Affect Competitive Balance in Baseball?

2015-04-30T15:16:00.000-04:00

From 1996-2000 the New York Yankees won four out of five World Series; in 2000 their payroll was $92.5 million, more than five times as high as the Minnesota Twins’ $16.5. In that time frame, none of the teams in the bottom half of payroll spending won a single postseason game.[1] But in 2002 and 2003 the Oakland A’s, one of the lowest-spending teams in baseball, made the playoffs both years, at one point winning twenty straight games in a row. What had changed?

In 2002, responding to the lack of competitiveness of small-market teams, Major League Baseball instituted revenue-sharing. Under that system every team contributed 31% of their revenue to a common pot, with the money redistributed to every team; as well, central funds from national broadcasts were redistributed. At the same time a luxury tax was instituted, fining teams that exceeded a certain threshold of spending. While revenue-sharing was phased in, another trend was taking baseball by storm: sabermetric analysis, more popularly known as moneyball. This approach, popularized by Billy Beane, the General Manager of the Oakland A’s, allowed small-market teams to compete with their richer big-market brothers. While the success of the A’s was certainly impressive, the question remained: had the competitive balance really been restored? Just how much had revenue sharing and moneyball changed baseball? Post-2002, did small-market teams become more competitive because of revenue sharing, and did the cost of a win go down because of the moneyball effect?

To analyze this I am using a time-series cross-sectional analysis for the years 1985-2014. The dependent variable is winning percentage (wins divided by total games). The independent variable is salary share (team payroll divided by league payroll) which controls for the league-wide increases in payroll over the thirty-year period as well as inflation. A dummy variable was created to differentiate between the pre-2002 and post-2002 periods.

Figure 1 below shows the distribution of salary share. It is relatively rare for a team to command more than 6% of the total salary of the league, and only three times has the share been greater than 8% (the Yankees from 2004-2006).

In figure 2 below the relationship is visualized by year, with 2002 indicated as the year that revenue sharing went into effect.

Somewhat counterintuitively, it would appear that revenue sharing had little effect on the distribution of salary share by year. In fact, the opposite would appear to be true. The spread becomes greater after revenue sharing, with the Yankees hitting three of their highest years of salary share after revenue sharing went into effect. This can be explained by the inauguration of the Yankees’ YES television network, which launched in 2002 and is now valued around $3.4 billion, providing the Yankees with a great deal more money to expand payroll.[2]

The effects of revenue-sharing and moneyball post-2002 are shown in Table 1 below:

Table 1. Salary and Win-Loss


Win Percentage	(1)
Share of Salary	0.02^***
	(7.26)
Post-2002	0.00 (0.33)
	0.00 (0.33)
Constant	0.43^***
	(42.85)
R-Squared	0.130
Observations	858

t statistics in parentheses. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001

The above model indicates that share of salary as a percentage of the league salary is correlated with win percentage, which should come as no surprise. A 1 unit increase in salary share corresponds to a .02 unit increase in team win percentage. To give an example, in the year 1985 the total league salary was $262 million; that means that spending an extra $2.6 million (about the difference between Baltimore and the Yankees that year) correlated to 3.24 extra wins on average. In that year three out of the four divisions in baseball were won by less than three games, meaning those 3.24 extra wins are quite valuable (whether they were worth $2.6 million is for a GM to decide). By 2014, with a total league salary of $3.19 billion, those same 3.24 wins would cost $31.9 million, or about the price of Alex Rodriguez.

The model also indicates that there is no significant difference between the periods before and after 2002. The interaction effect was negligibly small and not significant at all. This indicates that revenue sharing and moneyball did little to impact the lack of competitive balance in baseball. The effect of moneyball was short-lived, since other teams (especially big-market ones) quickly adopted sabermetrics and thus reduced the edge for small-market teams like Oakland.

One major issue with the findings is the above-mentioned phenomenon of regional sports networks like the Yankee’s YES network. This was an extra revenue stream that was instituted exactly concurrent with revenue sharing, making it difficult to assess the impact of revenue sharing. Nonetheless, the findings show clearly that salary is correlated with win percentage. Neither the effect of moneyball, nor of the revenue sharing, was enough to significantly change that relationship. These findings support the notion that salary caps (like those employed in basketball and football) create more parity and a less direct relationship between money and wins. That is bad for fans of the Yankees, but good for everyone else.

Appendix:

Table 1. Summary Statistics


	Mean	Std. Dev.	Min.	Max.
Win Percentage	0.50	0.07	0.27	0.72
Salary Share (%)	3.49	1.26	0.32	9.52
N	858

Do File:

* Preamble

set more off

ssc install estout

cd "G:\"

**merge databases

collapse (sum) salary, by (yearid teamid)

import delimited C:\Users\ad9538a\Desktop\Teams.csv, clear

keep yearid teamid franchid w l era g r

drop if year <=1984

merge 1:1 yearid teamid using G:\Stata\Blogpost_2\Salaries_collapsed.dta

** generate variables

gen winprct = w/g

gen runspergame = r/g

egen totalsalary = sum(salary), by (yearid)

gen salary_share = (salary/totalsalary) * 100

gen salarymil = (salary/1000000)

**transform

gen dummyyear =.

replace dummyyear = 0 if year < 2002

replace dummyyear = 1 if year >= 2002

** time-series reg

encode franchid, gen (team)

sort team yearid

xtset yearid team

** year v. salary graph & histogram

histogram salary_share, frequency ///

title("Figure 1. Salary Share Histogram", size(large) color(black)) //

twoway (scatter salary_share year, xline(2002)), ///

title("Figure 2. Salary Share by Year", size(large) color(black)) ///

note("Red line indicates revenue sharing") legend(off) //

xtitle("year")///

ytitle("Salary Share (%)") //

** win percentage vs. percentage of total salary of the league

xtreg winprct salary_share if year<2002

xtreg winprct salary_share if year>=2002

xtreg winprct salary_share i.dummyyear c.salary_share#dummyyear

*** Summary Statistics

estpost sum winprct salary_share, listwise

esttab using sum2.rtf, cells("mean(fmt(2)) sd(fmt(2)) min(fmt(2)) max(fmt(2))") ///

title("Table 1. Summary Statistics") nonumber replace

***** Pretty Table

eststo clear

eststo: quietly reg winprct salary_share i.dummyyear c.salary_share#dummyyear //

esttab using file_nam3.rtf, replace label ///

stats(r2 N, fmt(3 0) labels("R-Squared" "Observations")) ///

legend varlabels(_cons Constant) ///

b(2) t(2) ///

title({\b Table 2. Salary and Win-Loss})

eststo clear

[1] Levin, Richard; Mitchell, George; Volcker, Paul; Will, George. “The Report of the Independent Members of the Commissioner’s Blue Ribbon Panel on Baseball Economics”. (July 2000). http://www.mlb.com/mlb/downloads/blue_ribbon.pdf

[2] http://www.adweek.com/news/television/news-corp-acquires-49-stake-yes-network-145326

Can News protect Civilians? Media Attention and Civilian Casualties in Syria

2015-04-30T14:27:00.001-04:00

1. Introduction and Research Design

Civil wars not only include violence between fighting parties, but also against civilians who are not directly participating in the conflict (“indiscriminate violence”). The suffering of civilian populations during these wars tends to catch the attention of international media. While English-speaking news outlets arguably target an audience that does not directly participate in the conflict, many civil wars involve international actors. The parties who are fighting the war may have an interest to shape their image in a way that increases international support to their struggle.

In that context, one might ask whether conflict parties make an effort to refrain from indiscriminate violence when their war receives a lot of medial attention. Specifically, this short essay asks to what degree the intensity of violence against civilians in civil wars is causally influenced by media attention?

Scholarly literature in this area has previously focused on the effect of medial (human-rights related) attention on both duration of and violence in civil wars.¹ Yet these few studies use yearly country-level data and therefore are unable to track short-term variation in media attention and conflict intensity. Nor can they completely rule out that conflict intensity influences media attention more than the other way around.²

I contribute to that gap here by using daily data on news reports and violent deaths of civilians during the Syrian Civil War in 2012. Media attention is proxied with the number of news articles on the war in English-speaking newspapers.³ Daily data on civilian deaths is provided by the Center for Documentation of Violations in Syria (VDC) and the Syrian Shuhada website.⁴

I examine how the variation in news reports influences the number of civilian deaths, with the hypothesis that:

(H1) Violent deaths of civilians decrease when media attention increases.

The assumption is that decision-makers who conduct military operations are more prone to avoid civilian casualties,⁵ when they feel pressured by international attention to their struggle. The primary reason not to use a simple multilinear regression to conduct this study is that news reports on Syria are likely to be published as a result of violence in the conflict, rather than the other way around (reverse causality):

(H2) Violent deaths of civilians cause media attention to rise.

I thus use a time-series design with lagged measures of media attention to test whether the death count at one time (t) is influenced by news reports before that time (t - “lag”). The primary problem here is to decide how long it takes for (macro-level) media attention to “sink in” into (micro-level) decision-making processes to influence (macro-level) civilian casualties. I assume that this process will take between one and three days, but will test multiple lag structures to determine where the effect is strongest.⁶

I also assume that media attention is only one (out of many, possibly more important) factors taken into account by the conflict party’s decision-makers. I do control for both passed and vetoed UN Security Council resolutions on Syria, since they might spark indiscriminate violence and media attention at the same time. I further use a dummy for the period leading to the US presidential election campaign to control for a potential drop in media attention on Syria, and use seasonal dummies to account for changes in violence related to conflict intensity at temporal stages (see summary statistics in Appendix 1).

2. Analysis

Figure 1 depicts the daily number of civilian deaths and news reports over time:

The first thing to note is that the two variables move along similar lines. That media attention on Syria remains relatively low while the death count spikes after July could be a result of the political recess period during the summer in many countries, and the US presidential elections in the fall 2012. Second, since changes in media attention slightly precede changes in death counts, it appears that more attention to the conflict leads to more civilian deaths, and not less as assumed in H1.⁷

I attempt to confirm these visual impressions with a regression analysis in table 1. Models (2) and (3) use the same variables, but a lag of one and four days for media attention, respectively.⁸

Table 1. Influence of Media Attention on Civilian Deaths

DV: Civilian Deaths	(1)	(2)	(3)
Media Attention (log, t-1)	0.27^***	0.13^*
	(3.60)	(2.44)
Media Attention (log, t-4)			0.17^**
			(3.19)
Civilian Deaths (log, t-1)		0.44^***	0.43^***
		(9.17)	(9.29)
UN Resolution Passed		-0.50^*	-0.48^*
		(-2.24)	(-2.14)
UN Resolution Vetoed		1.53^***	1.52^***
		(4.34)	(4.34)
US Presidential Elections		0.13	0.14
		(1.63)	(1.76)
Season April-June		0.078	0.069
		(1.10)	(0.97)
Season July-September		0.59^***	0.58^***
		(6.77)	(6.66)
Season October-December		0.48^***	0.47^***
		(5.49)	(5.52)
Constant	3.39^***	1.67^***	1.57^***
	(11.90)	(7.10)	(6.40)
Adj. R-Squared	0.03	0.61	0.60
Observations	365	365	362

Note: Lagged OLS estimates, t-stats in parentheses. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001.

Source: Syrian Shuhada Database.

The results confirm that media attention has a positive influence on civilian deaths. On average, one additional article on Syria is associated with a 0.13 increase in civilian deaths on the next day, with a chance of 5% that this estimate is a mere coincidence. The effect is slightly stronger and more significant with a lag of four days in model (3).⁹ The problem that previous values of death counts might predict future values is not solved by including the lagged dependent variable, but a more advanced regression model shows that such autocorrelation does not significantly change the results.¹⁰

UN Resolutions appeared to result in a slight decrease of violent deaths, and vetoes in a significant increase. The seasonal dummies also confirm the significant time-variant increase in violence starting in July.¹¹ Unsurprisingly, further tests also show clear evidence that media reports increase as a result of violence, confirming the reverse causality hypothesis H2 (see Appendix 2).

3. Conclusion

These results suggest that media outlets anticipate civilian casualties and write more about it before they happen, or that there may be a causal mechanism at work that leads conflict parties to more aggressively target civilians as a result of medial attention. A macro-level study cannot unravel these causal mechanisms, nor is this evidence from a single conflict proof for a causal relationship across civil wars.

This analysis does however show that using year-data to analyze the relationship between media attention and civilian casualties (or conflict intensity in general) is highly problematic. First, changes in conflict intensity can be caused by a number of factors that can’t be controlled for with yearly data. And second, media attention is highly in flux, and is caused by civilian casualties at least as much as it predicts them.

A study that could both attend to these problems and be valid across wars would require daily, micro-level data for many more civil wars, which is unlikely to be available anytime soon.

Appendices

Appendix 1: Summary Statistics

	Mean	Std. Dev.	Min.	Max.	Description
Civilian Deaths Number	102.66	65.79	8.0	419.0	daily number of civilian casualties
News Reports on Syria	51.05	25.23	7.0	162.0	daily number of news articles on the Syrian Civil War
UN Resolution Passed	0.01	0.12	0.0	1.0	dummy variable, 1 if resolution was passed on a day
UN Resolution Vetoed	0.01	0.07	0.0	1.0	dummy variable, 1 if resolution was vetoet on a day
US Presidential Elections	0.16	0.36	0.0	1.0	dummy variable, 1 for every day after US congress recess until election day (Sep-Nov)
Season January-March	0.25	0.43	0.0	1.0	seasonal dummies
Season April-June	0.25	0.43	0.0	1.0
Season July-September	0.25	0.43	0.0	1.0
Season October-December	0.25	0.43	0.0	1.0
Observations	366

Appendix 2: Influence of Civilian Deaths on Media Attention (H2)

DV: Media Attention	(1)	(2)	(3)
Civilian Deaths (log, t-1)	0.20^***	0.18^***
	(5.76)	(5.63)
Civilian Deaths (log, t-4)			0.060
			(1.69)
Media Attention (log, t-1)		0.66^***	0.70^***
		(17.67)	(17.53)
UN Resolution Passed		-0.15	-0.18
		(-0.97)	(-1.09)
UN Resolution Vetoed		0.28	0.29
		(1.14)	(1.13)
US Presidential Elections		-0.11^*	-0.082
		(-2.10)	(-1.46)
Season April-June		-0.064	-0.046
		(-1.30)	(-0.88)
Season July-September		-0.18^**	-0.044
		(-2.95)	(-0.68)
Season October-December		-0.21^***	-0.10
		(-3.55)	(-1.61)
Constant	2.94^***	0.63^***	0.94^***
	(19.10)	(3.90)	(5.55)
Adj. R-Squared	0.03	0.61	0.57
Observations	365	365	362

Note: Lagged OLS estimates, t-stats in parentheses. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001.

Source: Syrian Shuhada Database.

_________________________________________________________________________________

1Burgoon, Brian/Ruggeri, Andrea/Schudel, Willem/Malikkalingam, Ram 2015: From Media Attention to Negotiated Peace. Human Rights Reporting and Civil War Duration, in: International Interactions 00: 1, 1–30.
Ruggeri, Andrea/Burgoon, Brian 2012: Human Rights "Naming & Shaming" and Civil War Violence, in: Peace Economics, Peace Science and Public Policy 18: 3, 1–12.

2Particularly for the effects of media attention on violence, studies struggle with endogeneity problems, and find it difficult to rule out reverse causality. See Ruggeri, Andrea/Burgoon, Brian 2012: Human Rights "Naming & Shaming" and Civil War Violence, in: Peace Economics, Peace Science and Public Policy 18: 3, 1–12.

3As per a query of the LexisNexis database (http://www.lexisnexis.com/), searching for articles related to International Relations and Security (includes tags on human rights and civil wars), and containing the words Syria, Assad, Damacus, Aleppo, Idlib, or Homs. The latter were particularly contested areas of the war. While this multiplies attention if newspapers publish when other’s do, I assume that conflict parties care more about the overall attention (“outcome”), and that a less biased approximation would actually confound the “perceived” media attention of these actors.

4The Syrian Shuhada website (http://syrianshuhada.com/) includes the VDC data. It is considered relatively authentic, considering the inherent reliability problems of casualty reports in (civil) wars. The data is publicly available, includes the names and place of the deceased, and is based on local activist information networks (“Local Coordination Committees”). See: Scharpf, Adam/Schneider, Gerald/Nöh, Anna/Clauset, Aaron 2013: Die Blutspur des Vetos. Eine Prognose zur Gefahr von extremen Massakern in Syrien, in: Zeitschrift für Friedens- und Konfliktforschung 2: 1, 6–31.
The decision to choose the Syrian case in 2012 is primarily informed by the availability of this relatively reliable micro-level data, which can rarely be found for civil wars.

5Whether consciously or unconsciously.

6I also include a lagged measure of death casualties in order to avoid autocorrelation.
Basic formal model:

7Also notable are the spikes (“outliers”) in death counts in February (military offensive on Homs) and August (bombing and massacre in Damascus), which are probably more dependent on other factors than media attention.

8I use a series of vector autoregression tests to determine which time (lag) between media attention and civilian deaths is most appropriate. Most tests suggested that a lag of one day is most appropriate, one test pointed to a four-day lag.

9The cumulative effect of media attention (LTE) equals 0.23 in model 2. The effect is stable for lag structures from one to seven days, and also holds in a Granger causality test. The latter also shows more significant results for the reverse causality hypothesis than for H1.

10Including the lagged dependent variable in models (2) and (3) still resulted in heavy autocorrelation in a Durbin’s alternative test for autocorrelation. However, a Cochrane-Orcutt AR(1) regression shows almost identical, significant results at 1% for a four-day lag structure (model 3), with a transformed Durbon-Watson test statistic of 2.23 (original: 1.15). The one-day lag structure (model 2) reports P>|t| at 0.07, and thus barely looses significance at the 5% level.

11The omitted dummy variable here is the season from January to March, significant increases in violence after July are thus relative to that period.

Does Prescribing Heroin to Heroin Addicts Reduce Heroin-related Crimes?

2015-04-30T14:15:00.000-04:00

Heroin (opiate) overdose has long been a major cause of overdose-related deaths. Many European countries conduct Opiate Substitute Treatment (OST) on heroin addicts, and have received significant success. The traditional OST prescripts methadone or buprenorphine as replacement drugs to heroin addicts. A problem with the traditional treatment is that patients heavily addicted to heroin sometimes have no reaction the replacement drugs and tend to drop out of the program.

A study released in 2009 shows that treating heroin addicts with heroin prescription is more effective than methadone treatment. Later that year, Germany legalized heroin-assisted treatment and included the new OST into their mandatory health insurance. However, despite the positive result of the study, not many countries in Europe have taken the same action. Does the legalization of prescription heroin actually have significant effect on heroin-related crimes? Or does the idea of treating heroin addicts with heroin somehow increase people’s incentives to use the drugs? To find out the answer, I compared heroin-related law offences in Germany and Italy before and after 2009, when the heroin-assisted treatment was legalized in Germany.

I collected the data from European Monitoring Centre for Drugs and Drug Addiction, which publishes annual European Drug Report and has the most comprehensive drug-related data for Europe. The data ranges from 2004 to 2012. It started 5 years before the legalization of heroin-assisted treatment. Hopefully it can provide some evidence of common trends for the two countries I chose. The data ends at 2012, 4 years after the new policy was enacted. The time range should be long enough for us to observe the effect of the policy.

The study is designed to be a difference-in-difference research. I chose Italy as a control group in comparison to Germany mainly for two reasons. First, the two countries had close rates of heroin-related crimes before 2009 as shown in Figure B below. There were only slight changes in heroin-related offences rates in both countries before 2009. For Germany, however, the percentage started to drop at a comparatively greater rate after 2009.

Source: European Monitoring Centre for drugs and Drug Addiction

Second, both countries have traditional OST programs and they use similar substitute drugs—methadone, buprenorphine, and buprenorphine-naloxone combination. Germany and Italy also have similar regulation towards prescription of these replacement drugs. Both doctors in specialized drug treatment centers and any other medical doctors can prescript replacement drugs. In most European countries that also have OST, only doctors in specialized treatment programs are allowed to do so.

Table 1. Average Effect of Legalization of Heroin Substitute Treatment

on Heroin-related Offences in Germany

DV: Heroin-related Offences	Pre-Treatment	Post-Treatment	Difference
Germany	30879.99	21098.75	-9781.24***
Italy	8893.8	10032	1138.2
Difference	21986.2***	11066.75***	-10919.45**

Note: Differences-in-differences estimates. *, **, *** indicate statistical significance at the 10%, 5%, and 1% level respectively.

Source: European Monitoring Centre for drugs and Drug Addiction

Table 1 shows the average effect of legalization of heroin substitute treatment on heroin-related offences after comparing data from Germany and Italy in 2004-2012. I estimate that the new OST program reduced, on average, 10919.45 cases of heroin-related crime, as the data indicates. The finding is statistically significant (p<0.01), meaning that there is less than 1% possibility that the observed effect happened purely by chance.

The positive effect of legalization of heroin-assisted treatment on heroin-related crimes may be explained two different views. First, the heroin substitute treatment is more effective than the traditional OST program and less patients relapse after the treatment, resulting in less demand of the heroin market in Germany. However, the observed effect could also happen because the new program created legal access to heroin, and thus the illicit business started to shrink. It is true that the legal heroin is only meant for treatment purpose, but it is hard to determine whether the prescription heroin is used elsewhere.

I would also like to address two potential problems with the research. First, the observations in the research are not large enough due to limited data. The finding should be more persuasive if there is comparative information at sub-national level, like heroin-related crime rates at county level. Second, the heroin-related offences include heroin trafficking, which is mostly cross-border criminal activity. The legalization of heroin-assisted treatment in Germany could have direct effect on the drug market in Italy, and it requires further research to determine the actual effect.

Appendix

Table 1. Summary Statistics

	Mean	Std. Dev.	Min.	Max.	Description
Germany (Cases)	26532.78	6523.35	13902	33493	Number of cases of heroin-related offences in Germany from 2004-2012
Germany (%)	11.23	2.29	6.3	12.6	Percentage of heroin-related offences in all drug law offences in Germany from 2004-2012
Italy (Cases)	9399.67	1335.55	7481	11142	Number of cases of heroin-related offences in Italy from 2004-2012
Italy (%)	12.61	1.23	10.9	14.5	Percentage of heroin-related offences in all drug law offences in Italy from 2004-2012
Year	N/A	N/A	2004	2012	Year of heroin-related offences
Pre	N/A	N/A	0	1	Dummy variable (1 before 2009)
Post	N/A	N/A	0	1	Dummy variable (1 after 2009)
Treat	N/A	N/A	0	1	Dummy variable (1 if Germany)
Control	N/A	N/A	0	1	Dummy variable (1 if Italy)
Observations	18

Sources: European Monitoring Centre for Drugs and Drug Addiction (www.emcdda.europa.eu)

The Effect of Employment Opportunities on Migration Trends Across Countries

2015-04-30T14:13:00.003-04:00

By: Ashley Holst

http://chartsbin.com

Migration is a natural part of our globalized world. People migrate for an endless number of reasons. Some migrate for short periods only to return to their home countries, while others migrate permanently in hopes that they may better their economic status and improve the livelihoods of their families at home through remittances. Improved financial status and the quest for a larger or more stable income is one of the most common pushes for migration. Labor migration is defined as the “movement of persons from one state to another… for the purpose of employment” (IOM[1]). An economic migrant may have more freedom of movement than a forced migrant or a migrant seeking refuge from persecution, but economic migrants also often feel forced to migrate in order to escape poverty and unemployment. In developing and middle developing countries unemployment, lack of opportunities, and high levels of poverty lead to increased emigration (ILO[2]). In a briefing on labor migration, the OECD[3] states “Regarding the labor market, migration is a symptom of imbalances in sending countries, such as high rates of unemployment and underemployment among low-skilled workers, low wages for skilled workers, and unmet demand for education and acquisition of skills” (2009). Despite the widespread agreement that lack of economic security and opportunities for employment lead to emigration, there is little qualitative data on the effect of lack of employment opportunities on migration.

What is the effect of lack of employment opportunities on migration patterns?

The primary dependent variable for this model is Net Migration[4], while a clear measurement of outmigration would have been ideal, net migration offers a sense of both immigration and emigration in a country. The primary independent variable is labor force participation[5]. This variable measures the percentage of the population ages 15-64 that are currently employed. By using this variable, I assume that if employment opportunities are available that people will take advantage of them.

To measure the effect of labor force participation, I wanted to look at a wide variety of countries. It would have been incomplete to answer this question only for developing or developed countries and I believe that migration trends will follow employment trends across countries, not only within countries. As examples, I observed trends in Haiti, one of the poorest countries in the world and India, one of the world’s rising economic powers. These two countries also lend themselves to a comparison of employment and migration, as Haiti is known for its poverty and unemployment issues while India is a growing hub for companies.

As seen in the graphs, migration trends seem to follow labor force participation rates. In Haiti, labor force participation rates fell between 1990 and 2000 resulting in increased outmigration in the following years, when labor force participation began to increase, outmigration was quickly reduced. The 2010 earthquake may have influenced the spike in net migration, but the trends prior to 2010 still reflect the labor force participation rates of several years previous. The observations in India show similar results, prior to 2000 labor force participation was decreasing and net migration was decreasing. After a reverse in direction of labor force participation, migration too changed directions showing a higher net migration.

Three additional variables were included in the model. Each of the control variables is a possible factor influencing migration trends: Level of Democracy, presence of conflict and GDP per capita[6]. The summary statistics for each of the variables included in the model can be seen below. I expect that level of democracy, conflict, and GDP per capita have effects on both labor force participation and net migration, thus I control for them in the regression analysis. The years included in the model are 1990 to 2012.

I use cross sectional time series panel data and a between estimator to measure the strength of the effect of labor force participation on net migration. A between estimator uses an Ordinary Least Squares regression on the mean values of the independent variables on the mean value of the dependent variable, net migration, over a given time period. This regression is useful in cases where variables are slow changing or “sluggish.” Doing a between estimator over time also reduces measurement error. Due to the lack of measurements available, between effects offers the best analysis of the strength of the effect of labor force participation on net migration.

The results in column one show that Labor Force Participation does not have a significant effect on net migration but the relationship does show the expected direction (increased labor participation resulting in a decrease in net migration which indicates increasing outmigration). The second and third columns show that GDP per capita, level of democracy, and conflict are also not significantly correlated to net migration. Increased democracy results in higher net migration (less people leaving) while an increase in the severity of conflict resulted in lower net migration (more people leaving). GDP per capita and net migration have a non-linear relationship where above a certain level of GDP per capita, net migration begins decreasing. This could be due to the fact that it is more difficult to immigrate to countries with higher GDP, thus affecting the net migration score.

The fourth column controlling for all the variables, showed significant results for GDP per capita and conflict. When controlling for the other variables, the direction of the relationship for GDP and GDP-squared became the reverse while conflict maintained direction. The primary independent variable, Labor Force Participation rate remained negative and not statistically significant.

The results show that migration is fluid and that there are many factors that have an effect on net migration rates. People chose to stay, leave, and where to migrate to based on economic, social, and physical factors. Labor force participation, or the availability of sufficient employment opportunities does not significantly affect migration patterns. This result could be due to the fact that people not participating in the formal labor markets may be earning incomes in other informal ways. It is also possible that the exclusion of undocumented migration may be playing an affect in this regression. It would be interesting to do a similar model looking at documented versus undocumented migration and these same influencing variables; the results may differ between the groups. Unfortunately accurate data on undocumented migration is difficult to obtain and would likely include significant errors. In light of these results, countries suffering from high levels of emigration should focus on reducing conflict within their country and improving GDP per capita levels.

[1] http://www.iom.int/key-migration-terms

[2] http://www.ilo.org/wcmsp5/groups/public/---ed_protect/---protrav/---migrant/documents/publication/wcms_178672.pdf

[3] http://www.oecd.org/development/povertyreduction/43280513.pdf

[4] Net Migration: Measured as the number of entering immigrants minus emigrants. Positive numbers reflect more immigrants than emigrants, while negative values reflect more emigrants leaving than immigrants entering the country. Net migration is logged to correct for positive skew.

All data collected from the Quality of Governance Standard Cross-Sectional Time Series Data Set (2015) http://qog.pol.gu.se/data/datadownloads/qogstandarddata

[5] Labor Force Participation Rate: the proportion of the population ages 15-64 that is economically active: all people who supply labor for the production of goods and services during a specified period.

[6] Freedom House Polity measures level of democracy on a 0 to 10 scale. Internationalized internal armed conflict measures conflict within a country between governments or opposition to government that requires outside intervention. Conflict is scored on a 1 to 10 scale where 1 indicates no violent incidents and 10 indicates civil war or widespread conflict. GDP per capita PPP is in $US and based on a 2005 based year. GDP per capita is logged and squared to control for skew.

Does public investment in education contribute to economic growth?

2015-04-30T13:58:00.000-04:00

Background and Design

The perceived impact that public investment in education has upon economic growth cannot be overstated. Governments worldwide invest billions of dollars and dedicate hours of human resources to the cause to promote education. Such efforts aim to improve education, which – many believe – is a key to driving greater economic growth and development. Given the immense amounts of resources – particularly money – spent towards these ends, it is important to assess whether investment in education does in fact improve economic growth, as measured by the proxy gross domestic product (GDP).

It is commonly believed that increased investment in education breeds greater economic growth. Such proponents argue that a more highly-educated labor force is easier to train, is more productive, and is even inventive, culminating in higher wages, which fuels overall growth. Empirical evidence that countries with higher levels of economic growth tend to also have higher levels of formal schooling supports this view. Therefore – proponents of this logic conclude – an increase in the population’s education levels leads to higher earnings, which consequentially increases economic output and growth rates (World Bank, 2007). However, there are no robust results that definitively prove there exists a relationship between public investment in education and economic growth. Some studies suggest that there is no relationship between the two, while others claim that public investment in education and economic growth are in fact correlated.

I seek to explore whether public investment in education is positively correlated with economic growth or not. To do so I will use a difference in difference model and will concentrate on two countries, Peru and Bolivia, chosen for their similarities across various characteristics: geography, political history, ethnicity, large indigenous populations, language, and economic growth rates. The period studied is 1998 to 2012. While the similarities between the two countries are critical, their difference are likewise acknowledged. For example, Bolivia is landlocked, a factor known to significantly limit economic growth.

For the treatment event I will use the rise of Evo Morales to power in Bolivia in 2006. Before 2006, both Peru and Bolivia had followed neoliberal guidelines for their social investment policies, which tended to restrict public expenses. Further, before 2006, both countries experienced hyperinflation in the 1980s and were mostly ruled by a white minority aligned with the Washington Consensus, which similarly restricted public expenditure in education. Given this background, I explore the effect of the rise of Morales, followed by the change in public expenditure in education, to assess whether this had a positive effect on the economic growth of Bolivia.

The model uses Bolivia as the treatment group, Peru as the control group, and two periods of time: before and after 2006. The dependent variable is GDP growth per capita and the independent variable is government expenditure in education (% of GDP). The data source is the World Bank. A difference in difference model was selected so as to create a design that resembled a randomized experiment. To achieve this, two samples that were as homogeneous as possible had to be used. However, despite the similarities mentioned above, Bolivia and Peru are not similar in gross indicators, such as their populations and GDPs; therefore, the model uses comparable variables.

Analysis

It is clear that Peru’s economic growth is higher than that of Bolivia. This fact is not surprising given Peru’s much larger economy. Bolivia’s baseline in public expenditure in education is notably higher than that of Peru. Peru’s public expenditure in education remains roughly the same, with a slight decrease, while Bolivia’s education expenditure exhibits a slight upward slope with a significant increase in 2006, the year Morales took office. After 2009, Bolivia’s education expenditure clearly declines; however, it is important to consider a possible time lag for the potential impact that Morales’ socialist policies may have. That is, given more time to take effect, Morales’ policies may actually reverse this declining trend in education expenditure – a point to consider as future data becomes available beyond 2012. (See appendix for descriptive statistics).

The difference in difference analysis shows that the difference between Peru and Bolivia’s GDP growth before and after Morales is not significant. Therefore, according to this study, the rise of Morales to power in 2006 – and increase in expenditure on education – did not have a significant effect on the GDP growth/capita. The study, however, does show that the difference of Peru’s GDP growth per capita before and after Morales was significant, as well as the difference between Bolivia and Peru’s GDP before 2006.

Finally, a regression analysis shows that there is a negative, significant relationship between public investment in education and economic growth in Bolivia during the years 1998-2012. It follows that for every 1% increase in public expenditure in education, GDP decreases in 1.5%. (Please see appendix for regression table).

Conclusion

The results empirically show that public expenditure in education does not contribute to economic growth. The rise of Evo Morales to power in 2006 was thought to have increased public expenditure in education, and, therefore, increase the GDP growth rate of Bolivia. However, the results show that, although Bolivia’s public expenditure in education did increase with Morales as president, this situation did not have an effect on the economic growth of Bolivia. Moreover, the regression analysis shows that in the case of Bolivia, public expenditure in education and GDP growth are in fact negatively correlated. Despite these results, it is important to remember that education is a long-term investment whose effects need to be studied years after initial investments. It may well be that Bolivia’s increased expenditure in education after 2006 may not yield significant results in stimulating the country’s GDP growth until many years into the future.

Appendix

Bibliography

Aghion , L. Boustan , C. Hoxby, J. Vandenbussche. “The Causal Impact of Education on Economic Growth: Evidence from U.S.”. March 2009
Ke-young Chu, Sanjeev Gupta, Benedict Clements, Daniel Hewitt, Sergio Lugaresi, Jerald Schiff, Ludger Schuknecht, and Gerd Schwartz. “Unproductive Public Expenditures A Pragmatic Approach To Policy Analysis”. Fiscal Affairs Department. International Monetary Fund., 1995.
World Bank. “The Road Not Traveled Education Reform in the Middle East and North Africa”. World Bank Report, 20007.

The Future of the International Monetary System and Optimum CurrencyArea Theory, Case Study: EUROZONE

2015-04-30T13:42:00.005-04:00

The Future of the International Monetary System and Optimum Currency Area Theory

Case Study: Eurozone

By: Emily Kaput

Optimum Currency Area theory- Major contributors, Criteria

Optimum Currency Area Theory was popularized by the work of economist Robert Mundell in 1961, and conceptualized by theorists such as Paul de Grauwe and Barry Eichengreen. The theory holds that nations who share borders (general geographical area) and adopt a fixed exchange rate regime or a single currency within their boundaries stand to benefit from increased trade and share the stress from exogenous financial shocks (like to a currency, trade deficit or an attack on foreign exchange reserve holdings abroad). However, the benefits of joining an Optimum Currency Area (OCA) must outweigh the costs- most notable of which is the sacrifice of monetary autonomy. Giving up monetary autonomy is to give up the ability to sterilize the domestic monetary base against exogenous shocks due to trade, foreign exchange speculation, and attacks on foreign reserves. This also means that nations in the OCA would only be able to rely on sound fiscal policy to maintain economic stability.

Literature on OCAs was not “quantifiable” at its conception and it experienced a “dead time” where no one was paying any attention to it. However, in 1991, theorists picked it back up and re-conceptualized the theory and operationalized data based on fixed exchange rates and pegged exchange rates (linking one’s currency to a dominant currency- most commonly the US dollar, which has historical dominance). The graph below and to the right is a table Eichengreen and Bayoumi created to show the resurgence of OCA theory papers in production over time in five year periods[i]:

There are three criteria for Optimum Currency Areas:

1. wage flexibility

2. labor mobility

3. insurance mechanisms[ii]

Popular vote categorizes the United States and the Eurozone as the only major OCAs in existence, using the US Dollar and the Euro, respectively. In this post I’ll use a Difference-in-Differences analysis to test the theory on its basis that it provides financial stability to its member countries.

Studying the theory of Optimum Currency Areas is important because the future of the International Monetary System rests on the success or failure of these Optimum Currency Areas.

Optimum Currency Areas- Case Study, Eurozone

I use the Eurozone as my treatment group (OCA) and the remaining members of the European Union as my control group. Because the members in each group in the years 2008-2010 are not equal in number, I use an average of all data. *I include in this analysis the countries who experienced severe internal shocks as a result of the crisis (Greece, Italy, Portugal, Spain, Cyprus, Ireland) because they are members of the Eurozone and as such are necessary to include to test the OCA framework.

In mid-2009, the European Union as a region experienced what is now known as the sovereign debt crisis. The unveiling of Greece's sovereign debt (along with Italy, Cyprus, Spain, Ireland and Portugal) caused a devastating lack of confidence in the European monetary system, exhibited in the mass asset dumping of the regions most indebted countries. This caused an exponential increase in the interest rates of sovereign bonds in the general Euro Area, most concentrated on the highly indebted countries. Thus, capital flows were stunted. The lack of confidence displayed towards the members of the Euro Area was devastating to the financial markets in this segment of the International Monetary System and the economies contributing to it. Media outlets across the world nay-said the OCA enthusiasts, believing that this massive, devastating shock would prove that the Euro was a terrible idea and that nations should maintain their own currencies and exchange rates.

Before these nay-sayers can be proven correct, they should look at the welfare of the nations involved. For my analysis, I use the Debt to GDP percentages and Inflation rates in each of the member countries in the control and treatment groups in the years 2008, 2009 and 2010 to determine the welfare of the nations- were members of the OCA (Eurozone) better off than non-members?

If OCA theory proves correct, then Eurozone members would experience less increase in debt to GDP percentages post-crisis (mid-2009), and lower rates of inflation (either positive or negative). [iii]

The Eurozone members experienced higher Debt/GDP percentages post-crisis (Table 2). This is unsurprising considering that several of the most indebted countries belong in the Eurozone. For a better look at economic health, we will look at inflation (Table 3).

Inflation rates in both the Eurozone and the remaining European Union members decreased. This is interesting. A negative movement in inflation rate means they both experienced Deflation, or, a general fall in prices across the economy (too little money chasing the same amount of goods in an economy). Deflation is as much an indication of economic health as inflation- and the extent to which deflation occurs is an important indicator.

Non-OCA members experienced higher rates of deflation than OCA members. They experienced 3.739% more deflation than the Eurozone. Deflation is the falling of prices in the domestic economy. This is important, because the monetary base was severely affected by the shock even though these countries could practice sterilizing monetary policies to stabilize their economies. Essentially, the shock affected the domestic economies in its most basic terms- the availability of money. Less money was chasing the same amount of goods.

Considering the deflation rates exhibited in the OCA members and non-OCA members, exchange rates in these groups are a necessary supplemental piece to this puzzle.

To the right are the average exchange rates of the OCA members and non-OCA members in the same months/years that the Debt/GDP and Inflation data were taken. The average total depreciation of exchange rates in the non-OCA members increased by 3.43 vis-a-vis the dollar, whereas OCA members experienced an average total depreciation of 0.076 from 2008-2010. This means that the currencies in the non-OCA member states devalued against the US dollar much more than the value of the Euro; rather, their currencies became weaker.

Conclusion:

The economic and financial integration in the Eurozone under the OCA framework did contribute to its response to the debt crisis in 2009. The lower rates of deflation and lower depreciation of the Euro vis-a-vis the dollar compared to non-OCA members is proof of this.

However, revisiting the OCA criteria (e.g. real wage flexibility, labor mobility, and fiscal transfers), I am certain the Eurozone does not fulfill all criteria in the framework. If the Eurozone were better integrated in all the criteria specifications their stability, when compared to nations in non-member countries, would have been more guaranteed. In the United States there is labor mobility, fiscal transfers, and real wage flexibility. In the Eurozone, there is labor mobility, theoretically. However, welfare systems and non-transferable social welfare programs act as a disincentive for citizens of nations within the Eurozone to become a truly mobile labor force. In addition to this, members of the Eurozone do not share political ideologies and fiscal government spending programs, and there is no overarching authority dictating fiscal spending across the nations, thus the fiscal transfer criteria is not fulfilled.

The path of the Euro, while it is not steady, is a financial instrument to watch. The Eurozone has two paths to choose from; to become more integrated or to disintegrate. A move towards disintegration is the possibility of the Greek exit from the Eurozone- colloquially known as “Grexit.” [vi]

But! Leading voices in this debate (Ann Krueger, Christine Lagarde, Greek FM) put the probability of Greece exiting the Eurozone very low (AEI Panel Discussion, April 2015).

The Eurozone (a partially fulfilled OCA) has exhibited more stability than non-OCA nations in the European Union. The main take away from this exercise is that Optimum Currency Area framework (if followed more closely and by more regions) could lead to a more stable and transparent International Monetary System, the stability of which would benefit the entire global economy.

[i] Eichengreen, Barry, and Tamim Bayoumi. "Exchange Rate Volatility and Intervention: Implications of the Theory of Optimum Currency Areas." Journal of International Economics 45 (1998): 191-209. Http://people.ucsc.edu/. University of California Santa Cruz. Web. <http://people.ucsc.edu/~hutch/241B/2003/Journal%20Articles/Bayoumi-Eichengreen.pdf>.

[ii] De Grauwe, Paul, and Francesco Mongelli. "The Enlargement of the Euro Area and Optimum Currency Areas." (2004): n. pag. European Central Bank. Web. <http://www.mnb.hu/Root/Dokumentumtar/MNB/Kutatas/mnbhu_konf_fomenu/mnbhu_conference/degrauwe.pdf>.

[iii] My model is simple: Y_DCGMY=α +β(treatment) + ϒ(post) + δ_dd(treat*post)+Z_DCGMYπ+e_DCGMY

Where D= debt ratio in the

C= country within the

G= group (Eurozone, EU member) in the

M= month in the

Y= year.

*Treat { 1 for treated state, 0 for control state.

*Post_y { 1 for post treatment, 0 for pre-treatment.

*My second model for Inflation Rates substitutes I for Inflation rate for Debt/GDP Percentages.

*Common trends assumption necessary to perform a DD analysis are established- countries in the European Union/Euro Area share trade, borders, and most peg to the Euro or the US Dollar (inherently connected vis-a-vis the value of the dollar because the US dollar and the Euro are the most commonly held liquid reserve currencies in the world).

[iv] Data collected April 13, 2015 http://www.x-rates.com/historical/?from=USD&amount=1&date=2009-05-20.

[v] Data collected April 13, 2015 http://www.x-rates.com/historical/?from=USD&amount=1&date=2009-05-20.

[vi] Chambers, Clem. "What Will Happen To Greece If There Is A Grexit?" Forbes. Forbes Magazine, n.d. Web. 30 Apr. 2015. <http://www.forbes.com/sites/investor/2015/02/02/what-will-happen-to-greece-if-there-is-a-grexit/>.

Gender quotas: A key to equality? A look at South Korea and Japan

2015-04-30T13:09:00.001-04:00

South Korea and Japan both face a huge, imminent economic and social problem: rapidly shrinking birthrates and slowing economies. To boot, the two countries are often at the bottom of the OECD list for gender equality, each ranking in the low 100's on the World Economic Forum list.[1] Measures have been put into place by both federal governments to create more equality for women in the hopes of eventually boosting the economy. One method used by states worldwide is gender quotas. South Korea has federally legislated gender quotas, and Japan does not. This post seeks to find the effect of gender quotas in the social sphere by looking at the effect of gender quotas on percent female enrollment in tertiary education.

Gender quotas are a controversial topic in places like the U.S. Around the world however, more countries have some type of gender quota than do not. The most important question is whether these top-down policies actually have an effect on the ground. In South Korea, this seems possible. Once more women were in South Korea’s legislature, they increased the number of women-friendly policies[2] Paid family leave increased from 8 weeks in 1998 to 65 weeks by 2012.[3] These policies, and social policies in general, go hand in hand with economic reform meant to stimulate an economy.[4] If gender quotas do increase gender equality in a state with women-friendly policies, and more women can help speed a slowing economy, gender quotas have potentially huge implications for the future by way of increasing female presence in the workforce and offering plenty of benefits to encourage them to also have children.[5]

H0: Gender quotas have no effect on female enrollment in tertiary education in South Korea.

HA: Gender quotas will have a positive effect on the enrollment of women in tertiary education in South Korea.

A difference-in-difference test will be used to test the above hypothesis. This method is especially good for measuring the effect of a policy change by selecting two similar cases and applying a treatment. Difference-in-difference also allows for a temporal element which helps lessen the risk of reverse causality. In this study, Japan and South Korea will be compared and the treatment will be gender quotas. Japan and South Korea are apt comparisons because both are in the same region, have shared history and are wealthy OECD countries. Since the legislation for quotas in South Korea was put into law in 2000, this study focuses on the years 1998-2012. Difference-in-difference is better than a cross-sectional time series analysis because it allows us to see what South Korea may have looked like without gender quotas. While not perfect, this still provides very nice answers to “what if” questions.

Table 1. Summary Statistics


	Mean	Std. Dev.	Min.	Max.	Description
% Women in Japanese Parliament	8.25	2.1	4.6	11.3	Percent women in Japanese national legislature
% Women in Korean Parliament	10.58	4.67	3.7	15.7	Percent women in Korean national legislature
Japanese Female Tertiary Enrollment	50.47	5.22	41.19	58.23	Percent gross women enrolled in Japanese tertiary education
Korean Female Tertiary Enrollment	71.96	11.76	50.07	85.32	Percent gross women enrolled in Korean tertiary education
Observations	15

The main variables will be gender quotas as the dependent variable and the female percentage enrollment in tertiary education as the independent variable.[6] The percentage of tertiary enrollment will be regressed while controlling for labor force participation. Enrollment in higher education was chosen because it can show whether these top-down policies trickle down to the societal level by breaking down cultural norms that might have otherwise prevented women from pursuing higher education. Tertiary education is also a key element women need to compete in the workforce.[7]
It is hard to quantify gender inequity in a state, but looking at social institutions is a good place to start.

The common trend assumption can be justified by looking at Figure 1. Both countries had only about a 10% difference in the percentage of women enrolled in higher education, and both were steadily increasing. However, in 2000 when gender quotas were legislated into law, South Korea begins to see a sharp spike in female enrollment, eventually far eclipsing Japan, which still continues to increase, as well. There appears to also be a clear correlation.[8]

Table 2. Average Effect of Gender Quotas on Female Enrollment in Tertiary Education

DV: % Enrollment	Pre-Quota	Post-Quota	Difference
South Korea	65.1	75.14***	13.5***
Japan	49.5***	47.34	1.77
Difference	15.6***	27.8***	11.6***

Note: Differences-in-differences estimates. *, **, *** indicate statistical significance at the 10%, 5%, and 1% level respectively.

As seen above, the results are very significant for tertiary enrollment.[9] The difference-in-difference is a positive 11.6. This can be translated as gender quotas led to an 11.6% increase in women enrolled in tertiary education. In Japan, there was only a 1.77% increase, while in South Korea there was a 13.5% increase-- a marked difference. While it is always difficult to prove causality, there does appear to be a clear correlation. With all this information, I can reject the null hypothesis. Unfortunately, gender quotas were insignificant when regressed with female labor participation rate as the outcome variable.[10] This is interesting, that even though Korean women appear to be excelling in higher education, it appears their fortunes are not improving upon entering the labor force.

This has exciting possibilities for countries like Japan, which is hoping to level the playing field for women. While there were perhaps other policies in South Korea that led to an increase in women in higher education, there is a strong enough correlation to presume that gender quotas did indeed have an effect. However, gender quotas do not appear to be enough. Even when women are excelling in higher education and gaining the credentials to theoretically compete equally with men in the labor market, and with all the legislation in place designed to encourage women in the workforce, it appears that top-down policies are helpful, but not the final answer.

[1] World Economic Forum. "The Global Gender Gap 2014." http://www3.weforum.org/docs/GGGR14/GGGR_CompleteReport_2014.pdf.

[2] Peng, Ito. "Social Investment Policies in Canada, Australia, Japan, and South Korea." International Journal of Child Care and Education Policy 5, no. 1 (02, 2015): 41-53, http://search.proquest.com/docview/1656262545?accountid=8285.

[3] Ibid, pp 48.

[4] Ibid.

[5] Subacchi, Paolo. "The G20 Are Finally Taking Gender Equality Seriously." (2015). [accessed April 23, 2015].

[6] There are three main types of gender quotas: reserved seats, legal candidate quotas and political party quotas. The first two are legislated and enforced by federal law. The first regulates the number of women elected, and the other two forms set a minimum number of women candidates on the lists. The last is voluntary on behalf of the political party. South Korea has legal candidate quotas. (See: Gender Quota Database)

[7] David, Miriam E. "Women and Gender Equality in Higher Education?" Education Sciences 5, no. 1 (2015): 10-25, http://search.proquest.com/docview/1672897504?accountid=8285.

[8] 0.983 is the correlation between the two, to be exact.

[9] All p-levels are well below the .05 level, almost all below the .01 threshold.

[10] I also ran a difference-in-difference test for just labor force as the outcome variable, which interestingly, but perhaps logically, yielded insignificant results.

Cultural and Economic Globalization: Baseball and Foreign Direct Investment

2015-04-30T10:55:00.001-04:00

In 2014, 26.3% of Major League Baseball’s active players were from places outside the United States.[i] In the 1970 that number was only 12.3%. Baseball has gone global. So has money. In that same time frame, the total amount of foreign direct investment in the world has by over $1.2 trillion.[ii] In Figures 1-4, we can see that for several countries these increases appear to be correlated. But is one affecting the other? It is hard to separate cultural globalization from economic globalization and even harder to determine which comes first. This analysis uses the exportation of baseball players from other countries to the United States’ Major League baseball as a measure of cultural globalization and investigates its impact on the amount of foreign direct investment coming into a country, a measure of economic globalization. Specifically, the analysis explores the possibility that as a country opens itself up culturally, its increasing openness and new connections to the outside world result in economic globalization.

Figure 1

Figure 2

Figure 3

Figure 4

The globalization of sport and its impacts on nations has been evaluated extensively in the sports community. Baseball has been argued as a means to further US political interests by serving as a non-threatening, benevolent cultural institution of the United States and softens responses to US involvement in a country.[iii] There is also obvious economic benefits to transnational corporations as foreign born players create foreign born fans looking to check in on native sons in the big leagues and transnational corporations benefit immensely from the opening of new baseball markets.[iv] The explanation for the increase in foreign direct investment has often been attributed to the liberalization of trade and globalization[v] and scholars have attempted to quantify the effect of “informal institutions” on a country’s involvement in and openness to foreign direct investment, finding significant results[vi]. Cultural globalization can be considered as one those informal institutions however, scholarship has yet to explore potential measurements for cultural globalization that can be separated from economic globalization. Sports literature clearly illustrates baseball as a form of cultural globalization and even discusses its role in opening up a country’s inhabitants to outside influence. Using baseball as measurement of cultural globalization then, provides a way independent enough from economic globalization to hold a real analysis on the relationship between economic and cultural globalization.

The analysis includes the impact of other variables that could confound the results if not accounted for: previous FDI, average income, amount of exports, and the amount of natural resources.[vii] In order to ensure there was no possibility of finding results where it might be foreign direct investment impacting baseball players exported to the United States instead of the reverse, a statistical model that matches the number of baseball players in one year to the amount FDI[viii] three years later is used. The other variables in the model have been similarly lagged two years behind FDI for continuity. In short, the model is designed to account for the time it takes for impacts to occur and control for variables that differ between countries but not over time.[ix] For example, the model can control for the long standing embargo the United States has against Cuba and how that might impact both FDI and players exported to the United States.

Table 1. Regression Models for MLB Players on FDI

DV: Foreign Direct Investment	Model 1	Model 2	Model 3
Number of Players in Major League Baseball	0.04^*** (8.03)
# of Players in MLB (Lagged 3 years)		0.04^*** (7.71)	0.005^* (2.14)
Foreign Direct Investment for the Year Before			0.584^*** (22.31)
GNI per capita (constant 2005 US$)(Lagged 2 Yrs)			0.000^* (2.48)
Total natural resources rents (% of GDP) (Lagged 2 Yrs)			-0.028^*** (-3.57)
Logged Exports (Lagged 2 Yrs)			0.429^*** (9.55)
R^2 Observations	0.05 1256	0.05 1200	0.79 885

Note: OLS estimates with t-stats in parentheses.* p < 0.05, ** p < 0.01, *** p < 0.001

Sources: World Bank Development Indicators, BaseballReference.com

Table 1 above shows the complete results of the analysis. The first model shows a simple regression without considering potential bias while models 2 and 3 provide a more accurate analysis. Model 3 reflects the complete model and demonstrates that an increase in the number of baseball players does correspond to later increase FDI for a country. The coefficients of the impact for each variable are at statistically significant meaning they we can reject the concept that the correlation is the result of random chance and assert that there is a real correlation between baseballers in MLB and FDI into a country. While the size of the coefficients for the variables appear small, the fact that the dependent variable is logged means that these are percentage changes not changes in raw numbers which makes them more impactful.

This analysis shows that cultural globalization as demonstrated by the exportation of baseball players to MLB does have an impact on later economic globalization as measured by foreign direct investment. While we should better evaluate how good of a measurement of cultural globalization that players in MLB actually is before we can draw further conclusions from this model, the evidence lends itself to the argument that baseball players can affect FDI. We should note the argument that some initial foreign direct investment may have been necessary in order to groom future professional baseball players but there is still a clear uptick in FDI that follows an increase in the exportation of baseball players.

[i] http://m.mlb.com/news/article/70623418/2014-opening-day-rosters-feature-224-players-born-outside-the-us

[ii] https://www.globalpolicy.org/component/content/article/213/45740.html

[iii] Klein, Alan. 1989. “Baseball as Underdevelopment: The Political Economy of Sport in the Dominican Republic”. Sociology of Sport Journal 6, no. 2: 95-112.

[iv] Marcano, A.J. & Fidler D.P. 2000. “The Globalization of Baseball: Major League Baseball and the mistreatment of Latin American talent”. Indiana Journal of Global Legal Studies 6, no. 2: 511-577.

Gems, Gerald & Pfister, Gertrud. 2014. “Sport & Globalization: power games in a New World order.” Movement and Sport Science 86: 51-62.

[v] Volos, Ch. K, I.M. Kyrianidis, and I.N. Stouboulos. 2015. “The Effect of Foreign Direct Investment in Economic Growth from the Perspective of Nonlinear Dynamics”. Journal of Engineering Science and Technology Review 8, no. 1:1-7.

Landefeld, J. Steven & Whichard, Obie G. 2006. “The importance of, and pitfalls in, measuring globalization”. Statistical Journal of the UN Economic Commission for Europe 23, no. 2/3: 127-142.

[vi] Seyoum, Baley. 2011. “Informal Institutions and Foreign Direct Investment”. Journal of Economic Studies 45, no. 4: 917-40.

[vii] A full table of summary statistics with explanations for the variables in the model is available in the appendix.

[viii] The amount of FDI and exports were logged before inclusion in order to better account for the positive skew of the raw numbers.

[ix] A fixed effects model was employed and a Hausman Test showing the utility of using a fixed effects model over a random effects model is available in the appendix.

Appendix

Summary Statistics Table for Blog Post 2

	Mean	Std. Dev.	Min.	Max.	Description
Net Foreign Direct Investment (Logged)	21.13	2.37	11.00	26.26	The log of the net amount of Foreign Direct Investment received (Source: WBDI)
Players in MLB	7.03	20.70	0.00	160.00	Number of Players in MLB by Country (Source: BaseballReference.com)
Natural Resource Rents	5.03	7.51	0.00	58.18	Total amount received for exportation of Natural Resources (US$) (Source: WBDI)
GNI per capita	15401.80	14478.53	252.86	67582.69	Average Income for a Country’s Population (US$) (Source: WBDI)
Merchandise Exports (Logged)	23.93	2.26	18.45	28.02	The log of the amount of merchandise exported (Source: WBDI)

Hausman Test Results Table

Variable	Fixed Effects Coefficient	Random Effects Coefficient	Difference	Standard Error of the Differences
Players in MLB (3-Yr. Lag)	.0048731	.0006802	.0041929	.0018432
FDI (1-Yr. Lag)	.5836586	.7931641	-.2095055	.0177884
GNI per Capita (2-Yr Lag)	.0000174	2.18e-07	.0000172	6.60e-06
Natural Resource Rents (2-Yr Lag)	-.0278745	-.0031953	-.0246792	.0070134
Exports (Logged, 2-Yr Lag)	.4292305	.1635982	.2656323	.0394832
chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 143.55 *Prob>chi2 = 0.0000*

The Hausman test reveals that error term for the random effects model is actually correlated with the regressors. As a result, the random effects model cannot be considered accurate and a fixed effects model should be employed.

OECD Membership and Access to Education

2015-04-30T09:54:00.002-04:00

The goal of the Organization for Economic Cooperation and Development (OECD) is to promote policies that will improve the economic and social well-being of people around the world. This analysis aims to assess the impact of OECD membership on a country's ability to promote access to education. One of the key measures the OECD uses to gauge success is how well school systems in member countries can prepare young people. The OECD collects its own metrics to measure how well member states are living up to the goals of the organization. This creates the potential intrinsic bias; the organization has an obvious interest in using measures of success that paint it in a favorable light.

Using independent metrics to assess the impact of OECD’s policies is important because it offers a chance to assess whether international organizations actually have positive impacts on quality of life. Some critics claim the OECD and similar organizations are merely echo chambers for liberal thought. Many claim that the effect of international organizations is negligible because states can only act in their own self-interest interests and only cooperate only when it suits their own goals of security maximization. A strong correlation between adoption of OECD-endorsed best practices for improving access to education could also suggest ways for other struggling nations to boost their own education systems.

This research could also be used by other countries struggling to improve access to education within their own borders. In 2006, widespread protests erupted across Chile, as people grew increasingly exasperated with the privatized education system. Chile’s education had suffered for years under the Pinochet dictatorship, and while it showed some signs of recovery in the 1990s, it was still sharply criticized by for being strongly disposed in favor of the wealthy. In 2006, Chile had the most expensive education system in the world, but its primary education system ranked 119 out of 144 countries. The privatization of education meant access to education, particularly high quality schools, was strongly skewed in favor of the wealthy.

In this analysis, I use data collected from the World Bank to analyze whether joining the OECD had a measurable impact on access to education for children in Chile. I use difference-in-difference analysis to compare primary school enrollment between Chile and Argentina in the years before and after Chile joined the OECD. Access to education is measured using primary school enrollment because the data is only available through 2012, meaning there would not be significant data to measure the impact for older students. Common measures of educational quality, like literacy for example, would also be likely to show less realistic results.

While this does not offer evidence supporting the claim that all international organizations have measurable impact, or even that the OECD specifically is an effective organization, a statistically significant relationship between OECD membership and access to education in this case would offer some empirical evidence that membership in such organizations can create positive conditions for ordinary people.

I chose Chile and Argentina as my cases because while Chile joined the OECD in 2010, neighboring Argentina did not. Selecting neighboring states with shared colonial histories controls helps control for factors like region and culture. Both countries have populations that are nominally Roman Catholic, both are slowly transitioning to experience declining birth rates (which could potentially affect primary school enrollment), and both see more immigrants than émigrés.

The graph suggests that prior to 2010, when Chile joined the OECD, trends in primary school enrollment in Argentina and Chile were fairly similar. A slight deviation from Chile’s gradually increasing primary school enrollment in 2007 does not appear to detract substantially from the common trends assumption. In 2010, Chile experiences a sharp rise in primary school enrollment, while Argentina’s enrolment seems to hold steady.

The regression analysis (Appendix – Table 1) supports the assertion of a correlation between Chile’s decision to join the OECD and a sharp jump in primary school enrollment in that country. The robustness of the model is highly statistically significant, meaning that these relationships are not likely due to chance.

This is certainly not sufficient evidence to support a causal claim that joining the OECD improves access to education, but it does show that in this case, the decision to join the institution corresponded with a jump in primary school enrollments, and seems to support the OECD’s claims that it uses policy research to advise member states on ways to tangibly improve quality of life for their citizens. Joining the OECD was one of many steps Chile took towards improving the quality and equality of their education system over the past decade, but it seems to correspond with a dramatic rise in primary school enrollment.

Appendix:

Table 1: Difference-in-difference Regression

DV: Primary School Enrollment	Pre-2010	Post-2010	Difference
Chile	79.74***	101.33***	30.54***
Argentina	65.77***	79.74***	9.17
Difference	13.96***	21.59	21.42***

Table 2: Summary Statistics

Variable	Mean	Std. Dev.	Min	Max
Chile	87.39196	14.18264	74.80404	113.6486
Argentina	67.89266	5.886969	60.06174	75.55846

Does Saving Increase Micro-Finance Customers’ Ability to Repay Their Loans?

2015-04-30T00:53:00.000-04:00

By Marie-Eve G. Augier

Micro-Finance Institutions (MFIs) provide financial services to the poor, typically persons who are undesirable candidates within the formal banking sector. The industry is marked by three major challenges, increasing its scale, depth (to poorer, more remote populations), and lowering costs all whilst meeting its characteristic “double bottom line” of generating profit and improving social welfare within communities (CGAP, 2006). The research question explored here is in line with these challenges. Investigating whether a MFI model including deposit services reduces loan defaults, simultaneously considers increasing MFI liquidity and thus its ability to scale up, reach more clients and lower costs. If true, it also addresses social welfare, arguing for MFIs to provide more services to its clients in a way that does not increase their risk of default, which can lead to further impoverishment, community exile and suicide.

Deposit services could either put a strain on clients’ ability to repay loans (many MFI’s require deposits with lending) or it could provide them with a much needed buffer to aid repayment. To investigate which effect is more salient, I use data from MixMarket.org, the largest, openly available collection of MFI data. Data was available from 1999 – 2014 although not every MFI reported for every year. This left me with 2584 observations over 899 MFIs.

In addition to data on loan loss rates (dependent variable, DV) and average deposit balances per depositor as a share of host country Gross National Income (GNI) per capita (independent variable, IV) of each MFI, I included data on percentages of female borrowers, average loan balances per borrower as a share of host country GNI and MFI age as controls. I expect each control to have an effect on both the DV and IV hence the need to control for them. Female borrowers are less likely to default on loans and more likely to save than men; the higher the loan balance the riskier it is and the lower the ability of the client to save due to higher repayment installations. Lastly, older MFIs are more trusted in their communities and a certain level of trust is needed before clients feel safe depositing their money. This also allows the MFI more time to perfect its loan product and recovery methods.

In my model, I include a lag of the IV to reduce concerns of reverse causality i.e. that the DV may cause the IV rather than the converse, allowing me to say confidently that the IV changed first and thus caused the DV to change. Secondly, I graphed both DV and IV to gauge the relationships across MFIs. I noted that most observations were grouped in the lower left corner, which indicated the need to take the log of IV. After this, there is still significant skew but less than before.

Thirdly, because there are factors specific to each MFI and the country in which they are based that could affect the variables (e.g. MFI business model or a countrywide financial shock); I used a regression technique called Fixed Effects, which controls for such unobservable and incomparable MFI effects and also did one regression with dummy variables for each country. Lastly, because there was very little variation across the DV, I also did a Between Effects estimate that considers time-invariance; however, this estimation typically requires many more controls than available in this model so its results should be taken sparingly.

Table 2 shows the results. The results of the country effects model are shown in the first column. Although the sign of the estimate for the IV is negative as expected, it is not statistically significant. Only the control Young was significant at the 5% level, meaning that there is only 5% chance that this number would have been estimated if there were no true relationship. Column 2 displays the MFI fixed effects model. Unfortunately, this model returned neither the expected sign nor any significance for the IV. In the final column is the between effects model. In this model, the IV is not negative but also significant at the 5% level. The estimate suggests that if MFI1 has an average deposit balances that is 1% higher than MFI2, then its loan loss rate is 0.000064 percentage points lower than MFI2’s loan loss rate. This may seem small, but depending on how large the loan portfolio of the MFI is, this may be a sizable loss. If the loan portfolio is $1 billion dollars, the loss could mean a loss of $64,000. However, in this model, none of the controls are significant.

While I would love to seize on the last model and claim that maintaining a deposit account has a negative effect on default rates, as mentioned previously, between effects models require many more controls which is difficult to achieve due to lack of data thus the estimates might very well be biased. Thus, in conclusion, considering the results of models 1 and 2 both showed no statistically significant effect, I would say that maintaining deposits do not have a statistically significant effect on default rates. Perhaps, the effects theorized in an earlier paragraph cancel each other out across such a broad population or are more salient in say one MFI, community or country given the specific context of the sample population.

The Economy and Presidential Approval in Russia

2015-04-29T20:35:00.000-04:00

By Alisha Armas

In the United States, politicians always have one weary eye on the economy. If it plummets, they fear losing the next election, because the public will hold them accountable for the economic downturn. Perhaps most famously, Bill Clinton’s quip “it's the economy stupid” captures this sentiment, pointing to the ongoing recession in his presidential campaign against George H. Bush. In this blog, I examine the effect of the economy on presidential approval ratings in Russia to see if this same relationship holds true. Evaluating the relationship between the economy and presidential approval in Russia is of great importance, especially now that the West is using sanctions in an effort to influence Russian actions in Ukraine. I argue that despite the recent surge in Putin’s approval rating, which coincides with a decline in economic growth, poor economic performance in Russia generally causes low presidential approval ratings.

As Figure 1 shows, presidential approval appears to be connected to economic growth. For example, while economic growth remains negative, Yeltsin’s approval generally declines. Meanwhile, shortly after economic growth jumps from about -6% to about 4% in the late 1990s, presidential approval also jumps. Economic growth again declines to negative numbers just before 2009, and presidential approval falls shortly afterwards. Although economic growth was negative prior to the 1996 elections, Yeltsin was still reelected, seeming to violate the accountability mechanism. However, there was also a fear of a return to communism, as the Communist Party candidate, Gennady Zyuganov, was Yeltsin’s main challenger, and Yeltsin’s approval ratings still suffered. The election, then, may have been a choice between the lesser of two evils rather than a violation of the relationship between the economy and presidential approval.

To evaluate the effect of the economy on presidential approval in Russia, I use time series analysis with fixed effects for Russia's different presidents: Boris Yeltsin (July 1991—December 31, 1999), Vladimir Putin (December 31, 1999—May 7, 2008 and May 7, 2012—present), and Dmitri Medvedev (May 7, 2008—May 7, 2012). I measure economic performance as GDP growth, calculated as the percent change from the previous period. Although inflation also measures economic performance, inflation is associated with economic expansion and presidential approval may increase as inflation increases.

For presidential approval, I use data collected by the Levada Center, a respected survey organization in Russia, that is available from RussiaVotes.org. Respondents rate presidential performance on a scale of 1 (lowest) to 10 (highest). Because there is strong trending in presidential approval ratings, I use the change in approval rating between periods (first differences). [1] While economic data is reported quarterly, the Levada Center sometimes surveyed the population on presidential approval quarterly, and sometimes more frequently. In cases where there is more than one survey in a quarter, I calculate the average (mean) score for that quarter.

In addition, I include several policies that can affect both presidential approval and economic performance. These policies are:

Shock Therapy: A policy intended to rapidly privatize the Russian economy. It was initially very popular and Yeltsin promised that the accompanying economic hardship would only last for six months. The policy began in early 1992.
Loans for Shares: A policy implemented in September 1995 that was designed to help the government meet its financial obligations and fund Yeltsin's presidential campaign. [2] Russia's oligarchs loaned money to Yeltsin in exchange for shares in some of Russia's largest industries, which had remained under government control.
1st Chechen War: A 20-month war from December 1994 through May 1996 (peace agreement) to prevent the secession of Chechnya from Russia. [3] The war was very unpopular, in part due to widespread media coverage. [4]
Annexation of Crimea: In March 2014, Russia annexed Crimea following the EuroMaidan protests in Ukraine. The annexation was very popular in Russia, but resulted in economic sanctions imposed by the West that have hurt the Russian economy.

Each of these policies is a binary variable, which is 1 when the policy may have an effect and 0 otherwise. For each policy, except the 1st Chechen War, I measure the long-term effect, which lasts until the end of the incumbent’s presidency. I exclude the long-term effect of the 1st Chechen War because of it is too closely related to another variable and so its effect cannot be estimated. In addition, I measure the short-term effect of Loans for Shares and the 1st Chechen War, which lasts for a few months (1-2 quarters). I do not measure the short-term effect for the other policies because of a lack of data. To show that economic performance affects presidential approval and not the other way around, I use data from the previous period for economic performance and the policies identified above. While it is possible that these factors take longer than one quarter to affect presidential approval, I cannot test this because of the limited number of observations.

Regression analysis shows that economic growth has a positive effect on presidential approval: for a 1-percentage point increase in GDP growth, on average, there is a 0.05-point increase in presidential approval ratings, all else constant. This effect is statistically significant. However, the small effect of economic growth on presidential approval means that there must be large contractions in the Russian economy for presidential approval to drop in a meaningful way. Orchestrating such drastic drops in the Russian economy is not easy, as the recent sanctions against Russia for its annexation of Crimea show. Furthermore, the effect of the Crimean annexation cannot be calculated. Thus, the annexation of Crimea may affect both presidential approval and economic growth, and may affect each in different ways. Thus, although economic performance has an effect on presidential approval, its small effect makes it difficult to extrapolate policy advice from the results.

Appendix 1.

Notes:

[1] Economic growth is already a difference between periods, and does not have the same trending problem as presidential approval. Furthermore, taking the difference in GDP growth reduces the number of observations, which is already low. Therefore, I do not use differences in economic growth.
[2] David E. Hoffman, The Oligarchs: Wealth and Power in the New Russia (PublicAffairs: New York, 2011), 313.
[3] BBC News Timeline: Chechnya
[4] While I include the 1st Chechen War, I exclude the 2nd Chechen War (early 2000s) because of collinearity and unlike the first war, there was little media coverage of the event. Therefore, the second war may not have had a large effect on public opinion.

Terrorist Attacks and International Travel

2015-04-29T12:44:00.002-04:00

by Michael Holmes

Terrorism strikes fear into the hearts of individuals. It dominates news coverage and often influences policy following an attack. When attacks occur in developing nations, they can harm international tourism to those nations, presumably striking fear into potential travelers. Terrorists also attack developed nations. If international tourism slows in developing nations, how do tourists respond to attacks in the developed world? Specifically, how did the World Trade Center attacks on September 11, 2001 impact international tourism to the United States?

Studies on terrorism generally claim that a primary goal of terrorism is to create fear. Could fear manifest itself by reducing international travel to the attacked country? International tourism demonstrates a general faith in the relative safety and security of the country a tourist visits. If people believe they risk death or injury by visiting a country, most will travel elsewhere. The September 11^th attacks are unique - using planes as a weapon - in that they could have hurt international travel as a whole, in addition to travel to the United States. I want to quantify the effect the September 11^th attacks had on international tourism to the United States and the rest of the developed world. I hypothesize that terrorist attacks against the United States caused both a small decrease in international travel overall and a substantial decrease in travel to the United States.

I use the differences in differences method, which compares international tourist travel to the United States with other countries before and after September 11^th_. I compare data on tourism (International Arrivals) to the United States with four similar countries as controls: Australia, Japan, Italy, and France. [1] I chose Australia and Japan because they are similar to the United States in that most international tourists must travel by plane. If September 11^th impacted air travel generally, it should impact all three countries. These countries in particular will allow us to observe the effect the attacks had specific to the targeted country.

Italy and France are slightly different since many international tourists arrive by other methods of transportation, but they are still good comparisons. Like the United States, France and Italy have major museums, well-visited tourist attractions, and a comparable amount of tourist traffic each year. Most importantly, the trends in tourism for these four nations were similar to the United States, even if the overall totals varied (See Figure 1). [2]

In order to make up for some of the differences in tourism totals, we must control for other factors (See Table 1). I control for the previous year’s Real GDP Per Capita because international travel generally requires advanced planning and payment.[3] A poor financial year could negatively influence travel abroad and is particularly relevant since the dot-com bubble burst during this period. I also include variables that measure the effect specific countries have on international tourism.[4] These help control for geographic, historical, and country-specific features that enhance or hinder tourism. The country controls and the Real GDP Per Capita control help account for differences in overall tourism and increase the validity of our measurements.[5]

To perform the differences in differences measure, I create a variable (Post) to mark the date of the attack.[6] I use the year 2002 to start the Post-9/11 period and justify it for a couple of reasons. First, my data is yearly and it is wiser to start in 2002 to measure September 11^th effects since the event occurred more than halfway through 2001. Choosing 2001 as the treatment year risks biasing the estimate, as there are over eight months of other explanatory variables unaccounted for. Second, many schedule international trips in advance, meaning that the true impact of the attacks might not occur until four or five months after the event. In addition to the Post variable, I include a Pre variable to mark the years prior to 2002, a Treat variable (The United States), and a Control, which represents the remaining countries.

I run the regression through 2006, which provides five years of data after the attack. I use 2006 as a cutoff because I believe the detrimental effect on the tourism industry was temporal and I want to avoid the inclusion of too many additional historical events that could influence international tourism such as the banking crisis or the weakening of the dollar.

International tourism to the United States decreased by 9.8% after 2001, but international tourism to the control countries increased by 10.1%.[7] The United States had more international tourism overall than the control nations both before and after 9/11, but while other nations experienced growth, tourism declined in the United States. Compared to international tourism in the other countries, international tourism to the United States fell 18.1%.

These results partially confirm my hypothesis. There was not a broad negative impact on international tourism, but there was a negative impact on tourism to the United States. While the attacks appear the likely cause, alternative theories are still possible. Increased security and anti-terror procedures in the United States after 2001 made flying more of a hassle, possibly discouraging potential travelers. While that theory does not take public fear into account, the increased security is still a reaction to the terrorist attacks. Other explanations could involve economics. However, these also tend to favor the terrorist attacks as the most likely cause for the tourist decline. The Dollar fell against the Euro from late 2001 through 2003. This means that vacations to the United States were cheaper for Europeans during this period, but tourism still declined.[8] While there are some other explanations, the September 11, 2001 terrorist attack was at least partially responsible for the decline in international tourism in the United States from 2002-2006. Terrorist attacks can cause damage to the tourism industry in even the strongest of nations. Since an attack in one country in an effective means of causing fear on an international scale, terrorist organizations can be expected to continue to target both developed and developing nations.

Appendix:

The percentage values are created by interpreting the log values of the tourist arrivals. The 18.1% decrease in international tourism to the US compared to the differences in other countries is the percentage change on the actual number of tourists, not the log values.

You may notice that the first value under the Difference column in Table 3 is negative, where as the corresponding coefficient in Table 2 is positive. It is negative in Table because it is measuring the difference going from the United States Pre 9/11 to Post 9/11 (a 0 to 1 shift). As a coefficient in the regression table, it measures the increase in Y going from 0 to 1, and therefore has a positive value.

[1] All countries had a high rating on the Human Development Index around September 11, 2001. I chose a Human Development Index score of .80 as my floor for selecting control countries.

[2] In order to evaluate the common trends assumption for these countries, I performed a correlation test in addition to the graphs shown above. The correlation coefficients between the United States and the following countries before 2002 are as follows: Australia=.8228; Japan=.8129; Italy=.8101; France=.7748. Correlations coefficients from 1995 to 2006 are still positively correlated, but lose strength. This is possibly a result of the hypothesized relationship.

[3] Performed a varsoc test and received results recommending a one- or four-year lag. The one-year lag made the most sense theoretically, as a previous year’s income is probably more indicative of the next year’s spending capabilities.

[4] These are dummy variables based on Country Code.

[5] Performed a VIF test for check for multicolinearity, all variables are below a 10.

[6] Post is a 1 if the Year is greater than or equal to 2002. It is a 0 if it is 2001 or earlier.

[7] -.1033 significant at 10% level and .0956 significant at 1% level. See Appendix for full Regression tables.

[8] Unfortunately the data does not say where tourists came from, so it is possible that Europe increased their US tourism and the rest of the world declined. However, this scenario also means that it was now more expensive for non-Europeans to travel to Europe, yet the European countries in the control experienced an increase.

The contribution of Official Development Assistance in the elimination of poverty

2015-03-20T18:58:00.001-04:00

by Mili Lechleiter

Official Development Assistance (ODA) is financial aid given by countries who are members of the Organization for Economic Co-operation and Development (OECD) with the purpose of promoting economic development and welfare in developing countries. In 2013, net Official Development Assistance (ODA) was 132.4 billion dollars, representing 0.30% of the donors’ combined gross national income.

The current discussion is if ODA is accomplishing its purposes of promoting economic development and welfare in developing countries. On one hand, experts believe that ODA has a negative effect on the economic and social development of developing countries. For example, Tamer (Tamer, 2013) points out that in African Countries, ODA has had a negative effect on the Human Development Index (HDI). Authors supporting this arguments argue that ODA is a form of neocolonialism of the global north over the global south and that it is doing more harm than good. On the other hand, authors such as Sachs (Sachs, 2005) and Easterly (Easterly, 2006) suggest that ODA is highly successful in improving education (measured as school enrollment) and health indicators (measured as reduction of child mortality). Overall, these arguments suggests the ODA has a positive impact in addressing development problems. Moreover, Sachs is an advocate of increasing ODA. His argument concentrates in the way ODA is broken down in per capita amounts. For example, the amount of ODA that Sub-Saharan countries received in 2002 ended up being $30 per person a year. From that total, $12 went to cover those countries’ basic needs, the other $18 were used to pay consultants, emergency relief, debt payment and debt relief operations, living only $12 per person to invest in direct development initiatives. Therefore, according to Sachs, the reason why we haven’t seen larger results of ODA investment in Africa is because ODA has not been large enough to produce results.

Study Design

Given the before mentioned context, this post seeks to explore if there is a positive significant relation in between ODA and poverty reduction that justifies the large amounts of financial aid allocated to this purpose. In order to do so, we decided to use variables from the Quality of Government Data set. The dependable variable used was “Population living below the poverty line” and the Independent variable was “Net Official Development Assistance received per capita (In constant US$)”. The original variable, Net Official Development Assistance was divided between the population numbers in order to get a per capita measurement. In order to control for co founds other variables were introduced. For example: Gross Domestic Product per capita was introduced to control for the effects of the country’s own economic development on the reduction of poverty. The variable was logged due to its negative skew given there are more low income countries than high income countries in the world.

Additionally, we introduced social variables to control for the effects of social development in the reduction of poverty. For example: average years of education of the country’s population, and Life expectancy at birth.

Finally, we wanted to control for the effects that democratic and efficient governments have on poverty reduction. First, we argue that democratic countries tend to have a better redistribution of national income and, therefore, are more likely to reduce poverty levels on their own. We used the dummy variable “democracy” with two values: 0 if the country has a dictatorial regime and 1 if it has a democratic one. The variable democracy is part of the Quality of Governance Data Set’s typology of authoritarian regimes in which using the mean of the “Freedom House and Polity Scales”, the line between democracies and autocracies is drawn at 7.5. This threshold value was chosen by estimating the mean cutoff point separating democracy from autocracy in five well-known categorical measures of democracy: those of Przeworski et al. (2000), 34 Mainwaring et al. (2001), and Reich (2002), together with Freedom House’s and Polity’s own categorical thresholds for democracy. From the 192 countries taken into account for this measure, 74 countries were considered a dictatorship and 118 were considered a democracy. For a list of countries according to their regime please see the appendix.

Second, when a government’s bureaucracy works more efficiently, its resources tend to be better allocated and, therefore, such governments are more efficient at poverty reduction tasks. We used the “Government Effectiveness – Estimate” which combines responses on the quality of public service provision, the quality of the bureaucracy, the competence of civil servants, the independence of the civil service from political pressures, and the credibility of the government’s commitment to policies. The main focus of this index is on “inputs” required for the government to be able to produce and implement good policies and deliver public goods. (This information was taken from the Quality of Governance Codebook)

Analysis

The variable ODA per capita shows us that most countries who receive ODA, receive in between $0 and $150 per person. Only 5 countries in the world receive more than $500 per person. These countries were deleted from the following graph (the countries deleted are listed under the graph).

Also, we can see in the graph below that most countries in the world have in between 10 to 50% of their population living below the poverty line.

The relationship between ODA per capita and poverty reduction is linear, but the slope of the relationship is minimal. Therefore, we decided to log the variable ODA per capita for the multivariable regression analysis.

By looking at the regression table, we can see that ODA has no significant relationship with poverty reduction either alone or when included in the multivariable model. Although, the independent variable of the model does not seem to have an effect on the dependable variable, the model does give us hints on other variables that could be effective in the fight against poverty. From the bi variate regression estimates, we can see that when regressed alone, the dependent variable and the control variables are all statically significant. Therefore, the relationship between them (Ln GDP per capita, Years of schooling, life expectancy at birth, government effectiveness) and the dependent variable is at least 90% not due to chance with the exception of ODA per capita and living in a democratic regime. We can see that in the case of education, for every year of education, the percentage of population living below the poverty level decreases in 2.72 units on average. Likewise, for every additional year of life expectancy, the percentage of the population living below the poverty line decreases in 1.04 units in average. Government effectiveness seems to have a great impact because for every unit increase in government effectiveness, poverty decreases in 17.15 units on average. GDP per capita has a significant contribution with the reduction of poverty. For every unit increase in GDP per capita, there is a 7.75 unit decrease in population living in poverty. (See Appendix for a summary of the relationships between the dependent variable and the independent variables).

When all the variables are included in the model, we see that the variables Ln GDP per capita and average years of education are not significant in explaining the percentage of population living below the poverty line.The model, however, gives us significant input regarding variables that do have a significant effect on poverty reduction. For example, we can see that for every unit increase in life expectancy at birth, the percentage of a country’s population living below the poverty line experiences a 0.87 unit decrease while controlling for the effects of government effectiveness, education, GDP per capita, and democracy. Likewise, when government effectiveness increases in one unit, the population living below the poverty line experiences a 9.96 unit decrease while controlling for the effects of the independent and control variables. Finally, it appears that democratic regimes have a higher population below the poverty line. According to the model, when a country lives in a democratic regime, its population is 10.14 times more likely to live below the poverty line than the population of countries living in a dictatorship. While controlling for all variables, the variables of life expectancy at birth, government effectiveness, and democracy had more predictive value than Average schooling years, GDP per capita, and ODA per capita over poverty reduction.

Conclusion

After analyzing the model, we conclude that there is no significant relationship between ODA and percentage of the population living below the poverty line. At a first glance, these results could encourage academics and policy makers that are looking to lower the amount of ODA given to developing countries. We need to consider, however, that the model presented in this analysis is a cross sectional analysis. The sample used was taken in a specific period of time which was the same for both variables. If ODA has any effect on poverty reduction, this effect will be seen a considerable amount of years after ODA resources have been invested. Therefore, we recommend to use a time series analysis for future models.

Although neocolonialism theorists would argue that ODA is another mechanism for developed countries to have control over developing countries’ economies, many developing countries depend on ODA to implement social programs related to health and education. It would be important to see the impact of ODA in average schooling of years, or in life expectancy. Following Sachs’ argument, ODA might not be effective in addressing the issue of economic development but it might be effective in addressing social development related to school enrollment and life expectancy. In any case, further analysis needs to be implemented in order to have a better understanding of the effectiveness of ODA in poverty reduction.

Appendix

Bibliography

· Sachs, J. (2005). The End of Poverty: Economic possibilities for our time. New York: Penguin.

· Easterly, W. (2006). The White Man's Burden: Why the West's efforts to aid the rest have done so much ill and so little good. New York: Penguin Press.

· Tamer, Christina (2013). The effects of foreign direct investment and official development assistance on the human development index. In Africa. University of Massachusetts

Why is there Still Slavery? Investigating the Causes of HumanTrafficking

2015-03-19T01:08:00.000-04:00

by Idris E.

Human Trafficking is a tremendously vexing crime. While abuses are often hidden in plain sight, the psychological effects of this crime against humanity are all too apparent. As a form of modern slavery where traffickers use force, fraud, or coercion to compel people to work or engage in sex acts, the signs are often difficult to identify and the victims are hard to find. This is especially true for child sex trafficking, where, according to international law, abusers do not need to use coercion to be considered traffickers. Yet, even though trafficking affects both sexes, it affects them differently.

Human trafficking disproportionately impacts women. In 2014, over 90% of sex trafficking cases reported to the the National Human Trafficking Resource Center in 2014 involved women, while men were only involved about 5% of the time. Yet although it is evident that human trafficking affects men and women differently, the reason for this is not as clear. Some anti-trafficking groups focus on ending the demand for sex, and imply that men's demand for sexual services is stronger than women's. Perhaps traffickers more often perceive women to be vulnerable and therefore target them for forced labor. While these factors are likely at play within individual trafficking cases or trafficking networks, I wondered how structural inequities between men and women might impact the prevalence of human trafficking and modern slavery.

National Human Trafficking Resource Center

The key question that I initially sought out to answer was "What is the effect of differing levels of opportunity for women on the prevalence of human trafficking?" I hoped to shed some light on the complexity of this seemingly simple question. I theorized that states where women had more economic and political rights would have less slavery. In theory, women with the ability to chose a job without the consent of a male relative, obtain equal pay for equal work, and relocate freely will be at a lower risk for trafficking than women who do not have these rights because traffickers would be less able to keep victims in their situations [Note 1]. Similarly, women with more political rights such as the right to vote, run for office, and petition the government might be able construct policy environments where fewer women are enslaved. This question is vitally important because answering it could help anti-trafficking advocates develop interventions to increase communities' resilience trafficking.

I constructed a statistical model using data provided by the Walk Free Foundation's 2014 Global Slavery Index (GSI). The GSI attempts to estimate the global prevalence modern slavery by extrapolating country-level estimates of the amount of trafficking to similar countries. The GSI is based on a number of sources, including random surveys that coded qualitative data about the subject's family's slavery experiences in the past five years [Note 2]. Although this extrapolation approach is problematic, and has been criticized, the GSI provides the only relatively complete estimate of the prevalence of trafficking around the world, even it it is only a very rough estimate, at best. The researchers who developed the GSI helpfully explain their methodological choices and also provide recent data for dozens of other potential predictors of trafficking. The data on women's rights come from the CIRI Human Rights Data Project and is included in the GSI data set [Note 3].

A box plot with some of the most extreme values removed (see also Figure 1b) provides a good first visualization of the relationship between women's economic rights and the prevalence of slavery. States with higher economic rights have less slavery than states where women have low economic rights or no economic rights (Level 1).

This second visualization also shows the relationship between slavery and women's economic rights. It attempts to fit a regression line to the data which have been "jittered" to differentiate the points from each other. This particular regression model (the dashed line) matches the non-parametric lowess fit line reasonably well, but still has trouble describing everything all the points in the graph.

The full statistical model shown below, estimates the impact that women's economic rights (Column 1) and women's political rights (Column 2) have on the prevalence of human trafficking holding a number of other variables (economic equality, income, development, corruption, and governance) constant [Note 4, 5].

The model explains that while increases in economic rights for women are associated with a lower percentage of enslaved people, the same effect was not borne out for increases in women's political rights [Note 6]. The predicted effect of Women's Economic Rights was relatively small: a one-unit increase in the 1-4 scale of Women's Economic Rights was correlated with a .201 unit decrease in the percent of trafficking in a country in the model that only included Women's Economic Rights.

Although the model was able to show a statistically significant relationship for the variable of Women's Economic Rights alone (Column 1), when holding other variables constant (Column 5), that could also affect women's rights and the prevalence of trafficking, the model was not able to show a relationship that could not have occurred by chance. It is also concerning that the adjusted R-squared value for human development is higher than that for women's economic rights. This means that human development explains more of the variance in the prevalence of slavery than women's economic (or political) rights does.

While there is a correlation between Women's economic rights and the prevalence of trafficking, that alone does not constitute a causal relationship. There still might be confounding variables that affect both the prevalence of slavery and Women's Economic Rights (such as human development, see appendix) that this model cannot account for. Reverse causality is also a concern here - respondents to the Women's Economic Rights survey might have felt like women have more economic rights in part because of their freedom from slavery.

More investigation would help to clarify these relationships. Although the statistical model above could not conclude that a relationship between human trafficking and women's rights existed, controlling for other relevant variables, it was able to suggest that Women's Economic Rights may be correlated with the prevalence of trafficking. This may be an important piece that can help anti-trafficking advocates develop approaches that seek to prevent trafficking, perhaps as part of a holistic development effort..

Appendix:

[Note 1]
Increased economic rights could also have the reverse effect - women would have more freedom to seek employment far away, where they lave less social support from their families and communities, potentially making them more vulnerable to traffickers.

[Note 2]
For example, if a respondent in Kenya said that his sister had been forced into domestic servitude in Canada last year, that victim would count towards the estimate for Canada.

[Note 3]
The Women's Political Rights and Women's Economic Rights measurement originally ran from 0 to 3 and were each resealed to a 1-4 scale with 1 representing no rights for women.

[Note 4]
The variable "Gini Coefficient (Flipped)" is a measure of income equality. It is the Gini coefficient, a measure of income in flipped into a measurement of economic equality. This measure ranges from a value of 0, which indicates perfect income inequality to 100 which represents perfect equality.

[Note 5]
The figures employ a quadratically transformed linear regression prediction instead of a standard OLS prediction because the outlier values (~4.0%) skew the relationships into a parabolic ones. This is done by regressing the Y variable (the prevalence of slavery) on the X variable (women's economic rights or HDI) and the X variable squared. The final model however employs standard OLS regression because the quadratic transformation did not yield a statistically significant model (see Table 3 below).

[Note 6]
This effect can also be seen in Figure 1c, where it is apparent that the states where women have some of the most political rights are the same states where the prevalence of slavery is the highest, a trend that run contrary to the trend seen with women's economic rights.

[Note 7]
Although logging GDP Per Capita did produce a statistically significant result in a potential model (shown below,) the coefficient was still 0.000 it did not make a strong contribution to the explanatory power of the rest of the model, judging by the relative differences in the R squared statistics between models 5 and 6 in Table 3.

Table 2. Summary Statistics

An interesting model, shown above, that might prove to be an effective control variable for another model is the influence of Human Development on Trafficking. In the final model (Column 5 of Table 1) Human Development was statically significant and when regressed against the prevalence of slavery by itself, a one unit increase in Human Development was associated with a 1.522 unit drop in the prevalence of human trafficking.