• Not only are financial recessions deeper and slower in recovery than in normal recessions. They also have slower recovery the greater is the credit to GDP ratio.
• ... a measure of credit based on loans outstanding, even including the role of the shadow banks, yields a conservative measure of our benchmark for where we should now be.
• We should have led the public to understand that we should measure success not by the level of the current unemployment rate, but by a benchmark that takes into account the financial vulnerability that had been set in the previous boom. We economists have not done a good job of explaining that our macro-stability policies have been effective.
• [The bailout expenditures] will probably be positive, and run to a few billion dollars. But they did literally stop a financial meltdown which was in progress.
• In sum, we economists did very badly in predicting the crisis. But the economic policies post-crisis have been close to what a good sensible economist-doctor would have ordered.

There is much more to the post, and it should be read in its entirety.

I think bullet point 3 merits additional stress. The magnitude of the calamity that unfolded in 2007 and 2008, and the difficulty in re-establishing growth, should have been placed in the context. In my view, that context includes the decade long (at least) buildup of distortions in the economy, from deregulation, non-regulation, and out-of-control fiscal policy. The challenge of educating the public remains, as attested to by the all too common refrain by some "that the recovery should have been stronger", even while demanding austerity.

If interest rates are responsive to news, most macroeconomic models agree that government “stimulus” spending crowds out private investment. In usual times, with responsive interest rates, New Keynesian models[4] typically have a strong role for monetary (Fed) policy but little or no role for fiscal policy (stimulus spending or tax rebates). In Neoclassical[5] as well as New Keynesian models, government stimulus spending diminishes private activity—especially investment—as private borrowers are crowded out of the market by government borrowing. In contrast, New Keynesian models suggest that when the interest rates relevant for investing are constrained by the zero lower bound, the crowding-out mechanism stops functioning and fiscal policy can be expansionary.

...

...in 2009, Christina Romer and Jared Bernstein published the economic bases of the Obama Administration’s $800 billion stimulus plan.[8] The cornerstones of their estimates were the multipliers reported on page 12, in a table entitled “Output effects of a permanent stimulus of 1% of GDP.”[9] These estimates came from forcing their models to constrain all interest rates at the zero lower bound regardless of the performance of the economy.[10] ...

I find this description odd, as the multipliers used are an average of the Fed's FRB/US and a private macroeconometric model (my guess is Macroeconomic Advisers, a standard in EOP and Treasury, at least back in my day). This description of FRB/US makes clear that even if the Fed funds rate was kept at zero, the long term rate would not necessarily be zero; in other words, there's a term premium, so a pure EHTS is not imposed. (And if Macro Advisers is the private model, then the same applies -- see here).

In any case, what did nominal rates actually do? I plot a slightly longer sample than Furth does (starting in 1967 rather than 1990), highlighting exactly how extraordinarily low long rates are.

Furth cites a Swanson and Williams paper that indicates a responsiveness of long term rates to announcements through 2011, as proof that there is no liquidity trap. Hence, in this interpretation, the American Recovery and Reinvestment Act (ARRA) might have crowded out investment.

In order to evaluate crowding out, one needs to examine the correlation of the stimulus package with real interest rates. This is shown in Figure 2.

A big upward shift in real rates is not apparent to me. It is true that real rates might have been even lower had there been no stimulus package. Whether that would have spurred significantly more investment is an interesting question.

The interesting thing is that according to the Swanson and Williams paper Furth cites, we are now in a liquidity trap-like situation. Now is exactly the time to undertake expansionary - not contractionary -- fiscal policy. At the very least, one would not want to cut government spending without cause. Interestingly, this implication Dr. Furth fails to note.

Now this week Bloomberg reported

Fannie Mae (FNMA), the mortgage-financier seized by U.S. regulators in 2008, will pay the Treasury Department $59.4 billion after reporting a record quarterly profit driven by rising home prices and declining delinquencies....

After its latest payment, Fannie Mae will have sent the Treasury a total of $95 billion, compared with the $117.1 billion of capital infusions that the company has received. Freddie Mac, which yesterday reported a $4.6 billion profit, will have paid $36 billion, after drawing $72 billion of aid, and Chief Executive Officer Don Layton said it may release $30.1 billion of tax-credit writedowns as soon as next quarter.

The coming large payments from the companies will probably help delay the amount of time the U.S. government has until running out of room under its debt ceiling to sometime in October, the Bipartisan Policy Center said today in a posting on its website.

But if you study Fannie Mae's 10Q report, you'll find that most of the 2013:Q1 reported profit came from Fannie's decision to recredit itself with $50.6 B in deferred-tax assets. The company's theory prior to 2008 was that, with all its previous losses, it wouldn't have to pay much future taxes as a result of carrying those losses forward. Fannie had been counting the taxes it wouldn't have to pay in the future as one of its main net assets in 2008. When Fannie was taken into conservatorship in 2008, there was a decision that maybe the enterprise would never go back to being a "private" company that owed any taxes, so those deferred-tax assets were written off as a big loss. Now the enterprise is putting them back on the books, as a result of which it claimed a huge after-tax profit for 2013:Q1. So Fannie plans to pay the U.S. Treasury (the GSE's current owner) a big dividend.

It's pretty amusing to see how this is getting covered by some of the press. For example, CNN's headline was Fannie Mae, Freddie Mac to help cut deficit:

U.S. taxpayers will soon reap a nearly $67 billion benefit from the recovering housing market, which will help to shrink deficits and delay the need to raise the country's debt ceiling.

To translate, you don't need to worry so much about the outstanding government debt because maybe some day Fannie is going to be a private company again that won't have to pay taxes for quite a while. Because the taxes that Fannie isn't going to pay in the future count as an asset to its current owner (the U.S. government), we can now declare ourselves to be better off financially than we were a week ago.

But I have another question about the rest of that $59 B in income earned by Fannie. The U.S. government is currently the residual claimant who receives the income from the fees that Fannie is collecting for guaranteeing mortgage-backed securities. So who is the residual guarantor of those securities?

Mortgage debt held by government-sponsored enterprises or in agency- or GSE-backed mortgage pools, 1952:Q1 - 2012:Q3. Top panel: in billions of dollars. Middle panel: as a percent of GDP. Bottom panel: as a percent of all mortgages. Data source: Flow of Funds, Federal Reserve Board, Table L217.

For the estimated impact on growth rates associated specifically with the sequester, see this post (from Macroeconomic Advisers).

For a discussion of the size of fiscal multipliers when interest rates are near zero or when slack exists, read this new survey in the New Palgrave Dictionary of Economics.

Although China’s current account surplus, which was a remarkable 10.1% of GDP in 2007, has come down to 2.6% in 2012 (Figure 1), the narrowing of imbalances appears to be rather cyclical than structural; it seems to be driven by a decline in external demand and a boost of public investment rather than by changes to the country’s perennial development model. As a response to the outbreak of the Global Financial Crisis, the Chinese government launched a huge stimulus program amounting to RMB 4 trillion (US$ 586 billion) in November 2008. Further fiscal and monetary stimulus followed in response to the anemic international demand due to the Great Recession in the U.S. and the debt crisis in Europe. As a consequence, China’s fixed investment rate increased from 41.7% of GDP in 2007 to about 48% in 2012, while the savings rate remained above 50% of GDP throughout (Figure 2), thereby shrinking the current account surplus.

How Chinese saving and investment rates will develop in the coming years is a question of crucial importance not just for China, but for the entire world economy. In a recent article (Ito and Volz, 2013), we analyzed how reforms in the financial sector, social protection, and healthcare programs affect China’s saving and investment rates and also its current account. To this end, we conducted panel data analysis using a comprehensive dataset that covers 83 countries during the last two decades. Using cross-country characteristics found in the analysis, we then attempted to unravel China’s peculiarities and shed light on its prominent role in the global imbalances debate. We also extended the cross-country analysis to provide forecasting on how China’s sectorial reforms will affect the country’s current account imbalances.

Our cross-country estimation results indicate that domestic financial reforms affect both national saving and investment, and thereby the current account balance. In particular, our estimations confirm the oft-expressed view that financial repression contributes to higher levels of national saving, or conversely that domestic financial reform would help reduce national savings rates. Furthermore, domestic financial reform is found to lead to lower levels of investment, presumably by curtailing overinvestment resulting from directed lending or other forms of distortive actions by the government.

While we find that increasing social protection expenditure increases national saving for industrialized countries, it apparently reduces national saving for developing and emerging economies. Increases in public social security spending have a negative effect on investment, which may be due to shifts of resources from traditional, more manufacturing-oriented, investment-intensive sectors to more service-oriented and less investment-intensive sectors, such as senior care and healthcare.

Our analysis also shows that reforms related to the financial sector, social safety and healthcare do matter for Chinese saving and investment, and therefore for the current account as well. We make several interesting observations pertaining to China. First, we find that the contribution of domestic financial reform to current account rebalancing has been gradually, but steadily, increasing over the years. Second, although its contribution was not as big as that of domestic financial reform, we find that capital account liberalization also contributed to a lowering of China’s current account surplus. Last but not least, we find that healthcare expenditure was another persistent negative contributor to the Chinese current account. Based on these results, we conduct scenario analysis to examine the potential impact of various hypothetical reform (and non-reform) scenarios on China’s current account, national saving, and investment. Our predictions suggest that domestic financial liberalization would indeed contribute to a rebalancing of China’s current account surplus. The impact of capital account liberalization on the current account balance, national saving and investment would be minimal, in contrast. Furthermore, our simulations suggest that China’s current account surplus would decline significantly if the country increased its public healthcare expenditure to a level comparable to that of Brazil, another major emerging economy. This effect would mainly result from a fall in the level of national savings. A strengthening of social protection systems would also contribute to a rebalancing of the Chinese economy.

Overall, our analysis highlights the importance of microeconomic reforms to correct macroeconomic imbalances. In particular, it provides support for the argument that a reform of China’s banking sector, preferably with interest rate liberalization at the centre of such reforms, as well as a development of China’s social security, healthcare and pension system would help the country to reduce its national saving and investment and allow domestic consumption to rise, all of which would help the country to correct its macroeconomic imbalances and get on a path of sustainable and equitable development.

References

Ito, Hiro and Ulrich Volz (2013) China and Global Imbalances from a View of Sectorial Reforms, Review of International Economics, 21(1), 57–71. [link]

Lardy, Nicholas R. 2012. Sustaining China’s Economic Growth after the Global Financial Crisis, Washington, DC: Peterson Institute for International Economics. [link]

This post written by Hiro Ito and Ulrich Volz.

S&P500. Source: Google Finance.

Rather than just look at nominal stock prices, it makes more sense to compare the price relative to earnings. Yale Professor Robert Shiller has suggested relating the current inflation-adjusted S&P500 to the previous ten-year-average of inflation-adjusted earnings. While one could reasonably argue that what happens over the next ten years is more relevant than what happened over the last, an advantage of Shiller's approach is that it is perfectly objective and one can see what kinds of values have been typical going all the way back to 1880. By Shiller's measure, the current backward-looking real P/E is up to 23.4, well above its historical average value of 16.5.

Green line: Ratio of real value of S&P composite index to the arithmetic average value of real earnings over the previous decade, January 1880 to May 2013. Red line: historical average (16.5). Data source: Robert Shiller.

If the ratio of price to historical earnings is unusually high right now, and if you expect the ratio to revert to more typical values, it suggests that you should expect a lower capital gain on stocks you buy today compared to what you would have earned if you bought at a time when the P/E was at or below its historical average. The blue line in the graph below shows the annual rate of return you would have earned by buying stocks at any indicated date and holding on to them for the next decade. That line stops in May 2003, because we don't yet know what the 10-year return of a stock purchased in June 2003 will turn out to be, and we certainly don't know what the 10-year return on a stock purchased in May 2013 is going to be. But what we do know is that in the historical record, you did indeed tend to earn a lower return on stocks if you bought them at a time like today when the P/E is relatively high.

Green line: Ratio of real value of S&P composite index to the arithmetic average value of real earnings over the previous decade, January 1880 to May 2013. Red line: historical average (16.5). Blue line: average compounded nominal rate of return on stocks purchased at the indicated date and held for ten years from that date. Data source: Robert Shiller.

So how will you do if you buy stocks today at these historically high valuations? Fernando Duarte and Carlo Rosa of the Federal Reserve Bank of New York surveyed 29 different forecasters and models for their calculation of the expected return on stocks relative to that on bonds. Obviously you want to take anybody's claim that they know where the stock market is headed with a rather large grain of salt. But it's interesting that the consensus assessment of this group is that stocks will outperform bonds by as high or higher margin as ever would have been expected over the last half century.

Expected return on stocks minus expected return on bonds from 29 different models. Source: Liberty Street Economics.

Actually there's not a contradiction between saying that stocks will do worse than average over the next decade but better than bonds, provided that bonds are going to do much worse than average over the next decade. Duarte and Rosa conclude that this is indeed the primary explanation for why the current excess return on stocks is calculated to be so high. For example, the current yield on a 10-year Treasury Inflation Protected Security is -0.62% annually. So it's not making too bold a claim that you can find something better than a guaranteed loss (in real terms) each year over the next decade.

Source: FRED.

In part the strong stock market may be signaling optimism about the real economy. But another very important factor is that stocks look good primarily in comparison to the dismal return you can expect from bonds.

The end of the semester has arrived, and as I prepared my last lecture, I checked to see how the government deficits had impacted yields. Real yields were pretty much as they were when the semester began in January.

Here’s the yield curve as of 7 May:

I was going to compare the 7 May yield curve against the 23 January yield curve (when the semester began), but they were so similar, it was pointless. Market indicators suggest negative real short term rates for the next ten years.

State and local governments avoided the massive job losses of 2008 and 2009 that affected the private sector—these governments even grew slightly during the recession. But they have been gradually downsizing ever since. The federal government, by contrast, has expanded rapidly since the recession began. Federal employment, excluding the U.S. Postal Service,[2] peaked in 2011 at 13 percent above 2008 levels. At the same time, the private sector was still mired in the slow recovery, 5.5 percent below 2008 levels. Since 2011, federal expansion has stopped, and a fifth of the recession-era expansion has been reversed.

However, most of the federal employment expansion that took place from 2008 to 2011 remains in place. Despite protestations that the additional employment associated with the stimulus would be temporary, federal employment remains as high as it was three years ago and 10 percent higher than it was before the recession. By contrast, private, state, and local employment are 2 percent to 3 percent below pre-recession levels.

The authors use the below figure to illustrate these points.

Here I provide two other ways of examining the data. Figure 1 below depicts employment levels normalized to 2007M12 levels.

Now, it is interesting to observe that should one not exclude postal workers (after all, many other Federal workers are distributed throughout the nation, just like postal workers), then the picture changes somewhat, as shown in Figure 2.

In other words, Federal employment is essentially back to levels of 2007M12. I think these two graphs cast in a slightly different light the authors' main contention, viz.:

...The expansion of the federal government has ultimately come to some degree at the expense of states and localities. As the federal government seeks to command more of the economy—most notably the health care sector—this troubling trend is likely to continue to move employment and power to the least transparent and accountable level of government

I guess it is an expansion of the center if the non-center shrinks. (Music side note: I almost expected the Darth Vader theme!)

Finally, I find it interesting that Sherk and Furth do not mention how little government employment grew relative to previous episodes.

As I recall, George W. Bush was President during the 2001M01-2008M12, when government employment surged. I do not recall a similar concern about centralization of power on the part of Heritage Foundation economists at the time.

In absolute terms, the change in government employment is 1.6 and 1.5 million lower in the current episode relative to the 2001M03 and 1990M08 episodes, respectively. Interestingly, this figure is very close to the 1.4 million that Chinn, Ferrara and Mignon (2013) identify as the structural change in total nonfarm payroll employment not attributable to misprediction in private employment; i.e., a good chunk of the low employment growth is directly attributable in a mechanical sense to the unique negative trend in government employment. (If you thought the previous two episodes were anomalies, see WSJ RTE.)

In my view, the NY Times assessment by Nelson Schwartz better summarizes matters:

“The drag from the government sector is quite substantial,” said Gregory Daco, senior principal economist at IHS Global Insight. “Given the fiscal headwinds, the private sector is doing O.K.”

...

“If it weren’t for the government, the economy would be stronger,” said Mr. Daco, citing the spending cuts hitting now, as well as the higher Social Security deductions for all workers and increased income taxes for top earners that began in January.

See this post for a discussion of Messrs. Sherk and Furth's previous interpretation of employment statistics.

Hard Workin' Rob works as an employee of R.C. Farms, which pays H.W. Rob $1 for every potato he grows and harvests. In a typical year, he produces 100 potatoes, and earns $100 in wages. R.C. Farms takes 20 of these potatoes and replants them in strategic locations around the island, a step that will make sure that the economy will have some new potatoes that will be produced next year. R.C. Farms sells the remaining 80 potatoes for $2 each to Crusoe's Restaurant. Hard Workin' Rob (the laborer) does not mind the mark-up because he is also the sole owner of R.C. Farms, and any profit the company makes is also part of his income. Crusoe's restaurant in turn takes the 80 potatoes it buys and turns them into 240 potato pancakes, which it sells to its customers (well, customer actually) for $1 each. The restaurant has no costs other than the potatoes-- it's a prepare-it-yourself concept in restaurants that probably works best in a beautiful island setting where everybody wants to keep the accounts very simple.

So the question is, what is the island's GDP? Clearly it's not appropriate to add the $160 in potatoes sold by R.C. Farms to the $240 in cakes sold by Crusoe's restaurant, because that would be double counting-- those potatoes that R.C. Farms sold to the restaurant were used for no purpose other than to produce cakes. The concept behind GDP is that we want to calculate the market value of all final goods and services sold (in this case, $240 in cakes), but not count sales of intermediate goods (in this case, not count the 80 potatoes used up in producing those cakes). But what about the 20 potatoes that R.C. Farms replanted rather than sold to Crusoe's Restaurant?

Suppose first that we decided to treat those 20 replanted potatoes as just another cost of doing business, thinking of them as something that got used up in the process of creating the 240 cakes. If that's our approach, then we're all done-- the island's GDP would just be the $240 in final goods (cakes) that got produced.

Again sticking with that expensing concept for the moment, Rob knows that we're also supposed to be able to calculate GDP by adding up the income earned by everyone in the economy. So to check his math, he notes that the owner of Crusoe's Restaurant earned $80 in profit, the owner of R.C. Farms earned $60 in profit, and Hard Workin' Rob (the laborer) earned $100 in wages, which indeed sums to $240, which we claimed was the economy's GDP. By the way, it was the income going to the owners of the firms that gave the restaurant's customers the funds they needed to buy all those expensive cakes.

But it's really not reasonable to say that the 20 potatoes replanted by R.C. Farms were used up in the process of making the cakes. In fact, those 20 potatoes weren't used at all in producing the cakes, and they weren't even used up, but hopefully will turn into 100 or more new potatoes for next year. The owners of R.C. Farms were enriching themselves by choosing to replant those potatoes, laying the "groundwork", if you will, for future production and profits. So it seems more appropriate to say that R.C. Farms, and not Crusoe's Restaurant, was the final user of those 20 potatoes, and that the $20 the farm paid Hard Workin' Rob for those potatoes is not an expense of getting potatoes ready to sell to the restaurant, but instead is part of the income of the owners of the farm that they decided to devote to "growing" future profits.

If we value these investment goods at cost, and regard R.C. Farms as the final user of those investment goods, we would calculate the market value of all final goods and services to be $240 in consumption goods (the cakes) plus $20 in investment goods (the replanted potatoes) for a total GDP of $260. Treating capital goods as an investment rather than a current expense thus makes reported GDP higher. From the income side, since the $20 investment spending is no longer regarded as an expense, we calculate the profits of R.C. Farms as $80 (the $160 revenue from sales to the restaurant minus the $80 cost of those goods). The restaurant still earned $80 and Hard Workin' Rob still earned $100, so GDP from the income side is 80 + 80 + 100 = 260, equal to what we calculated on the product side. Some of the income is now in the form of an imputed benefit the owner of R.C. Farms associates with the firm's decision to replant potatoes. Under this accounting scheme, the island's income now exceeds its total consumption spending, with the excess (saving) going to finance the economy's investment.

After several years of this, Rob realizes he's not getting anywhere (literally). So he decides to try to experiment with some new ideas to see if he can come up with a better way to do things. Can you get two viable plants if you take a potato and cut it in two before planting? Can you grow them in vessels, or other exotic locations? So, instead of just replanting all 20 potatoes in the usual way, R.C. Farms decides to just replant 15 potatoes, and use the others for experimentation-- its R&D budget. Maybe these experiments will produce more potatoes, maybe they won't-- mostly they'll just help Rob to learn if there's a better way to do things. The question now is, how should we treat R.C. Farm's 5-potato R&D budget?

One idea would be to say that R&D is just another expense. From that perspective, the economy's final goods are $240 in cakes plus $15 in investment, for a GDP of $255. On the income side, since we're expensing the R&D, the profits of R.C. Farms are only $75 ($160 revenue from the restaurant minus $85 in costs), which combine with $80 income from the restaurant and $100 income from the worker for $255 GDP on the income side.

But the same arguments against expensing investment could be used to say we shouldn't be expensing R&D, either. The purpose of the R&D was to produce a final good-- namely knowledge-- and producing this knowledge should directly benefit the owners of the firm. So if we decide instead to treat the R&D as a similar kind of activity as usual investment, we're led to a view that GDP consists of $240 worth of cakes, $15 worth of investment, and $5 worth of knowledge creation. That new knowledge in turn is going to now be counted as part of the income of the owners of the firm that conducted the R&D.

In other words, changing from expensing R&D (the current U.S. system) to treating R&D as a form of investment (the system that will be in place starting next quarter) would increase the reported GDP from $255 as calculated under the old system to $260 as calculated under the new. So on July 1, Rob plans to begin calculating GDP on the new basis, and report himself to be earning $5 more each year than he used to.

But the truth is Rob won't actually be any richer when he wakes up on July 1 than when he goes to bed on June 30.

In a new paper coauthored by Laurent Ferrara and Valérie Mignon, we estimate a log-levels version of Okun’s law, so that we can specify a decomposition of employment between trend and cyclical factors. Following up on intuition laid out here and here, we implement an error correction model with nonlinear short run dynamics. While it is possible to interpret the trend factors as structural in nature, it is also possible to view our trend component in a purely statistical context.

Relying on a non-linear error-correction specification over the 1950-2012 period, we find:

• If one does not account for the long-term relationship between GDP and employment, then we overestimate employment by a substantial amount.
• A standard error-correction model is able to reproduce the general evolution of employment, but underpredicts the decline in employment during the recession, and therefore over-predicts employment during the recovery.
• Using an innovative non-linear smooth transition error-correction model (STECM), we better reproduce stylized facts associated with the business cycle.
• The nonlinear model estimated over the 1950-2007 period produces ex post historical simulation results that indicate employment is still on average 1.05% below its potential level after the recession. Some share of this mis-prediction might be attributable to structural factors.

Predicting Employment using and Error Correction and Differences Model

Assume a long run cointegrating relationship between log private nonfarm payroll employment and log real GDP:

emp_t = β₀ +β₁y_t

One can estimate this relationship using a first differences specification, a first differences specification with lags, and a (linear) error correction model, over the 1950-2007 period, and then forecast out assuming knowledge of the right hand side variables. The error correction model incorporates one lag of the first differences.

Notice that neither first differences specifications fit the data very well. While the static first differences specification broadly matches the contours of actual employment, it overpredicts by a wide margin. By contrast, the dynamic first differences specification hits actual by 2012Q3, but missing completely the dip to trough in 2010Q1. The error correction model fits better, but still overpredicts on the order of 3% in log terms.

A Nonlinear in Dynamics Error Correction Model

Given the shortcomings of these linear models, we considered a smooth transition error correction model, wherein the short run dynamics vary depending upon the state of the business cycle (in this case, lagged GDP growth).

Where LEMP is log private employment, LGDP log real GDP, G denotes the transition function, γ is the slope parameter that determines the smoothness of the transition from one regime to the other, c is the threshold parameter, and Z denotes the transition variable. The smooth transition function, bounded between 0 and 1, takes a logistic form given by:

G(c, γ, Z_t) = [1 + exp( - γ(Z _t - c))]^-1

Where:

Z_t = 0.5 × (Δ LGDP_t-2 + Δ LGDP_t-3)

Details and references for the procedure are provided in this post. The results of estimating this over the 1950-2007 period are given y:

Most of the time, the economy is in regime 1, with regime 2 in effect usually right at the end of a recession. Estimating this model out of sample yields the following forecasts.

The extent of mis-prediction is now much reduced. Employment is over-predicted by only about 1%.

One is tempted to conclude that the gap represents the extent of employment that is reduced by structural factors. One could do that, although the first point to recall is that the mis-prediction incorporates both structural factors (say demographic trends, skills mismatch, other policies that affect the benefit of work versus leisure) as well as model uncertainty.

In order to try to determine the source of the 1% overprediction, we estimated the nonlinear model over the entire 1950-2012Q3 sample, and then predicted employment. In this in-sample assessment, the over-prediction declines to 0.6%, suggesting that at least some noticeable share (perhaps around half?) of the 1% out-of-sample over-prediction is due to factors other than a rise in structural factors. However, this conjecture awaits further investigation.

So my answer to the question posed at the beginning: “not very much”. That conclusion is in line with most analyses, e.g., Lazear and Spletzer (2012) (see also this post).

Note that we have allowed only nonlinearities in short run dynamics. We have retained a constant employment-output elasticity over the long run. Had we allowed for time variation in the long run relationship, we might very well have obtained different results. However, the theory to validate the application of the smooth transition methods to integrated variables doesn’t yet exist, to our knowledge.

Update, May 13: Data [XLS] and RATS program now online.

• The Congressional Budget Office has prepared informative slides summarizing the U.S. fiscal situation (via Business Insider)
• Betsey Stevenson and Justin Wolfers referee the Reinhart-Rogoff debate (via Free Exchange)
• Herndon, Ash, and Pollin appear to have been using faulty data for New Zealand
• Bill McBride reviews the U.S. decline in vehicle miles driven
• China's oil imports are down relative to this point last year (via Steven Kopits)

I think that there is now accumulating evidence that quantitative easing/credit easing moves by central banks can have an impact on exchange rates, even if short term rates do not move. That point has been most recently documented in an important paper by Reuven Glick and Sylvain Leduc, entitled: “The Effects of Unconventional and Conventional U.S. Monetary Policy on the Dollar”:

We document that the U.S. dollar depreciated significantly following both conventional and unconventional monetary policy surprises. Looking first at the effects of unconventional monetary policy since the end of 2008, we find that a one standard deviation surprise easing in unconventional policy leads to a roughly 40 basis point decline in the value of the dollar within 60 minutes. In turn, we find that in the conventional policy period the dollar depreciated in response to federal funds rate easing surprises, with a one standard deviation surprise easing leading to about a 6 basis point decline in the value of the dollar in the hour after announcements.

...

To enable comparison of unconventional and conventional monetary policy effects, we analyze the comovements in federal funds and long-term rates during the period when the funds rate was above its lower bound. This yields an adjustment parameter that we use to rescale the measure of unconventional policy surprises into equivalent fund rate surprises. The resulting adjusted coefficients indicate that a one standard deviation surprise unconventional policy easing leads to a 5 to 6 basis point depreciation in the dollar, magnitudes that are comparable to those for federal fund rate surprises in the pre-crisis period. This suggests that monetary policy has the same “bang-per-unit-of-surprise” as previously and that the exchange channel of monetary policy is still working as effectively as in the past.

The impact of unconventional policy surprises is shown in the figures below:

See also Rosa (2012) for another approach to measuring surprises.

Theory

Conventional macroeconomic models of the nominal exchange rate typically rely upon some sort of asset based approach. The monetary approach would say that intercountry differentials in money stocks, incomes and interest and inflation rates matter (and if purchasing power parity held instantaneously, interest and inflation rate differentials would be the same).

Given that quantitative easing (large increases in the money base) have not been manifested in correspondingly large changes in the money supply, it is hard to interpret the exchange rate changes as induced by changes in relative money stocks. In the (Dornbusch-Frankel) sticky price monetary model, if the short term interest rate is bounded at zero, then for constant expected inflation rate, the quantitative easing can’t have an effect.

Three other possibilities present. The first is that quantitative easing drives higher expected future inflation directly; the second is that large scale asset purchases (LSAPs) signal commitment to higher inflation; and the third is a monetary model is not appropriate.

An alternative approach is to interpret the results in a portfolio balance model. FRBSF President John Williams has observed that in order tounderstand LSAP effects:

one has to go back to older, more eclectic theories of asset pricing, such as Tobin’s “portfolio balance” model or Modigliani and Sutch’s “preferred habitat” theory. These assume that a range of heterogeneous investors have different preferences for different asset classes and that arbitrage across these asset classes is limited. This approach has been integrated into a modern, no-arbitrage, asset-pricing framework and has been used to guide empirical analysis of LSAP effects. (See Vayanos and Vila 2009 for a recent theoretical model, and Greenwood and Vayanos 2008 and Hamilton and Wu 2011 for empirical applications.)

In the 2012 revision of his 2010 paper, Neely lays out explicitly a portfolio balance model, within which to interpret the results that have been found thus far.

The Longer Term

The announcement effects relate to high frequency impacts. Do we know the impact over longer term? There, the potential effects are more ambiguous. As Steven Englander observes (“Some balance sheets are more equal than others…”, 4/5/2013) : “Once you allow the possibility that the balance sheet expansion can influence real interest rates, activity and asset markets, the outcome of the balance sheet expansion is ambiguous. It may raise inflation risk and even inflation expectations, but if it simultaneously raises growth, even temporarily, it is not necessarily the case that it is currency negative.”

This ambiguity is highlighted by the lack of tight correlation over time between the ratio of money bases and the respective exchange rate (US, Euro Area and dollar/euro).

For previous discussion, see [1] [2].

I’m late to the game [1], but it bears repeating: according to the BEA, government spending (either on goods and services, or more broadly including transfers) is declining as a share of GDP.

The drop is more pronounced if one accounts for the fact that GDP is at depressed levels relative to potential GDP. This is shown in Figure 2.

Government consumption and investment is also declining in real terms, relative to GDP (keeping in mind Jim's observation that total spending is relatively high, as shown in Figure 2). This decline is shown in Figure 3.

Finally, it is important to recall that what matters for the pace of debt accumulation is the deficit. The actual and cyclically adjusted budget balance to potential GDP ratio is rapidly increasing, as shown in Figure 4.

In other words, the fiscal impulse is rapidly dissipating. And it's doing so against a backdrop of slowing growth in Europe -- even Germany[2] -- and in China [3]. Whether US growth can be sustained in such a global environment is an open question in my mind, given Macroeconomic Advisers' estimate of 1.4 ppts sequester-related deduction from (annualized) growth in 2013Q2.

There is a strong historical "snap back" relationship between the strength of economic recovery and the severity of the preceding recession. Thus, recessions and their recoveries have a tendency to trace out a "V" shape. Because the current recession that began at the end of 2007 started out shallow and only recently became deep, a simple time series model of the "snap-back" dynamic implies that the current recession may have a lower-case "v" shape, although the fact that the economy is likely to continue to contract further before a recovery takes hold argues for the possibility of an upper-case "V".

However (and as the above article also discussed), when one broadens the historical reference set, as Reinhart and Rogoff suggested we should, from 60 years of U.S. data to many hundreds of years of experience of dozens of countries, the outlook for the United States in early 2009 was much less sanguine. Here is what Reinhart and Rogoff wrote in an article that appeared in the Wall Street Journal on February 3, 2009 entitled Expect a Prolonged Slump:

when one compares the U.S. crisis to serious financial crises in developed countries (e.g., Spain 1977, Norway 1987, Finland 1991, Sweden 1991, and Japan 1992), or even to banking crises in major emerging-market economies, the parallels are nothing short of stunning....

Over past crises, the duration of the period of rising unemployment averaged nearly five years, with a mean increase in the unemployment rate of seven percentage points, which would bring the U.S. to double digits.

We now all know how events turned out. The U.S. unemployment rate, which had risen from 4.7% in November 2007 to 7.3% by December 2008, would go on to climb to 10% in October 2009. Even today-- 5 years later-- the unemployment rate remains at 7.6%, far above its historical average, and above the most recent value reported at the time Reinhart and Rogoff issued their warning.

Here is what Paul Krugman wrote in July 2010, in an article in which he expressed his differences of opinion but began by acknowledging some of the areas in which Reinhart and Rogoff's research had proved crucial for understanding what had happened to the United States:

Regular readers will know that I'm a huge admirer of Ken's work, both theoretical and empirical. Obstfeld and Rogoff is the definitive work on New Keynesian open-economy macro; Reinhart and Rogoff the definitive empirical history of financial crises and their aftermath. It was largely thanks to my study of Obstfeld-Rogoff that I realized, from the get-go, that many of the arguments we were hearing about how modern macro had proved Keynesianism wrong were just ignorant; it was largely thanks to my reading of Reinhart-Rogoff that I realized, early in the game, that this was going to be a prolonged slump rather than a V-shaped recovery.

But on other issues, Krugman and others have been slower to be persuaded. Even after publication of their book in 2009, many still seemed susceptible to the latest version of the "this-time-is-different" syndrome, insisting that modern governments would be immune to the kinds of fiscal crises that Reinhart and Rogoff documented had been repeated again and again throughout history. For example, on November 22, 2009 Paul Krugman wrote:

Belgium is politically weak because of the linguistic divide; Italy is politically weak because it's Italy. If these countries can run up debts of more than 100 percent of GDP without being destroyed by bond vigilantes, so can we.

Again, we now know what happened next. Krugman's "bond vigilantes" have so far left Belgium and the United States alone, but the yield on 10-year sovereign debt of Italy, Spain, and Portugal (which had all been below 4% at the time Krugman dismissed concerns about Italy), along with Ireland and Greece (below 5% at the time), subsequently spiked up to much higher levels, a development that brought significant hardships for the people in those countries, and whose final ramifications are yet to play out.

Reinhart and Rogoff (2009) summarized their core message on page 292:

All too often, periods of heavy borrowing can take place in a bubble and last for a surprisingly long time. But highly leveraged economies, particularly those in which continual rollover of short-term debt is sustained only by confidence in relatively illiquid underlying assets, seldom survive forever, particularly if leverage continues to grow unchecked. This time may seem different, but all too often a deeper look shows it is not.

Is this a lesson that any of the statements zapping across the internet over the last two weeks do anything to undermine or dismiss? You could only imagine that the answer might be yes if you've just been following the controversy at the level of headlines and soundbites and ignoring analyses of the substance of the controversy. If that's the case, please get yourself informed by reading this and this. To briefly summarize the facts that you'll find developed more fully there, some researchers at the University of Massachusetts have challenged one tiny detail of one follow-up study by Reinhart and Rogoff having to do with an issue that I have yet to even mention so far in today's discussion. That dispute concerns a measure of the correlation between sovereign debt levels and GDP growth. I say it is a tiny detail, because the dispute is completely restricted to just one of six different summaries of three different data sets in that one particular paper. The Massachusetts researchers correctly noted that there is an error in the spreadsheet which, if corrected, would have changed the number Reinhart and Rogoff should have reported for that one statistic by a few tenths of a percent. Bigger changes in that one statistic could be obtained if one adds some additional data and makes what I regard as a questionable change in methodology, but even with all the changes they want to make, the results preferred by the University of Massachusetts researchers are in fact very similar to the other 5 dataset summaries that will be found in Reinhart and Rogoff's original paper, as well as a subsequent analysis of expanded data that Reinhart and Rogoff published in 2012 (which again the Massachusetts scholars did not mention or discuss).

So how does any of that cause us to discount or dismiss the contributions of This Time Is Different? You tell me.

Quarterly growth of real GDP at an annual rate, 1947:Q2-2013:Q1. Horizontal line depicts historical average 3.1% growth.

The ongoing slow growth has brought our Econbrowser Recession Indicator Index to 9.2%, up slightly from our previous reading of 8.2%. For purposes of calculating this number, we allow one quarter for data revision and trend recognition, so the latest value, although it uses today's released GDP numbers, is actually an assessment of where the economy was as of the end of the fourth quarter of 2012, for which the growth rate is now reported to be a decidedly worse-than-mediocre 0.4%. The index would have to rise above 67% before our algorithm would determine that the U.S. has entered a new recession.

GDP-based recession indicator index. The plotted value for each date is based solely on information as it would have been publicly available and reported as of one quarter after the indicated date, with 2012:Q4 the last date shown on the graph. Shaded regions represent the NBER's dates for recessions, which dates were not used in any way in constructing the index, and which were sometimes not reported until two years after the date.

One percentage point of that 2.5% growth came from restocking inventories. In other words, real final sales grew at an even more mediocre 1.5%. Strong growth in consumption spending was offset by declines from the public sector, with lower defense spending subtracting 0.6% from the growth rate by itself, and declines in other categories of federal, state, and local spending subtracting an additional 0.2%.

Let me call attention to one often-overlooked detail of the fiscal drag. The numbers that go into the GDP calculations just reported are direct government purchases of goods and services, which fell by $19 B between 2012:Q4 and 2013:Q1 (measured in nominal dollars, reported at a seasonally adjusted annual rate). But what many people think of as "government spending" includes not just government purchases of goods and services, but also government transfer payments. These rose by $35 B between 2012:Q4 and 2013:Q1. Although the latter don't figure directly in GDP, they influence it indirectly. Insofar as many of the recipients of transfer payments have a high marginal propensity to spend out of that income, the transfer payments helped support some of the growth in consumption spending in the first quarter.

The graph below displays the longer term trends for these two components of government spending. Government purchases of goods and services (the blue line) fell significantly as a percentage of GDP from 1991-1997. They rose in 2001-2002, and again in 2008-2009. These have been falling as a fraction of GDP since 2010, and are now back to 2007 levels, or about where they were in 1994. By contrast, government transfer payments rose from 5.2% of GDP in 1965 to 12% in 1995 and 16% at the start of 2010. They've since declined to 15.4%, still above the 14.8% level of 2009:Q1.

Blue line: government consumption expenditures and gross investment as a percent of GDP, from FRED. Red line: government current transfer payments as a percent of GDP, from FRED.

Although many analysts have been describing the decline in government purchases of goods and services during the last two quarters as reflecting austerity policies at the federal, state, and local levels, another relevant factor is that spending on goods and services is being crowded out by the budget demands of making growing transfer payments. This is part of a broader pattern that has characterized the last four decades, and is something that seems likely to continue over the next decades as well.