The Liscio Report

JOLTs update

2011-03-14T19:10:00-07:00

The Bureau of Labor Statistic's March 11 release of the Job Openings and Labor Turnover Survey (JOLTS) data for January showed a distinct lack of dynamism in the labor market: the level of hires and separations, expressed as a share of employment matched record lows for the series (which goes back to December 2000).

Net hiring - hires less separations - was just 0.1% of employment. Turnover - hires plus separations - was just 5.5% of employment, a record low for the series. Openings fell to 2.1% of employment (from 2.2%). That's better than the 2009 average of 1.8%, but matches the 2010 average. The numbers for the private sector were similarly tepid. January was not a sizzling month for employment.

Also, the relationship between job openings and unemployment, which had gotten out of whack, has gotten somewhat back into whack. Based on the historical relationship, unemployment in December "should" have been 7.7%, not its actual 9.4%. That gap narrowed substantially in January; the expected unemployment rate was 8.0%, vs. the actual of 9.0%. There were 5.02 unemployed persons per job opening in January, compared with 4.96 in December - and a 2002-2007 average of 2.03.

Flows out of unemployment—and out of the labor force

2011-03-05T18:23:00-08:00

Many analysts have been scratching their heads over the combination of a sharp drop in unemployment over the last few months alongside rather tepid job growth. The discrepancy is partially caused by the two different surveys from which the numbers are derived, surveys that often march to their own drummer in the short term. The BLS’s data on job flows offers additional clues.

Graphed below are some histories based on the flows numbers. The top graph shows where the formerly employed go from one month to the next. Between December 2010 and January 2011, 1.5% of the employed became unemployed, and 2.6% of the employed dropped out of the labor force. (There was a break in all the household survey numbers, because of the annual adjustments to the population controls, but the effect on these flows numbers looks to be very small.) The three-month moving averages of those two series are 1.7% and 2.7% respectively. The share becoming unemployed has been drifting lower since early 2009, but still remains rather high; the share leaving the labor force is up over the last few months, but isn’t out of line with historical averages.

The bottom graph is striking, however. The share of the unemployed finding work is very close to the recently set all-time low, and has barely improved over the last year. The share of the unemployed dropping out of the labor force, however, has been rising for more than a year, and now has exceeded those finding work for two years.

When you make a point like this, some skeptics are quick to blame excessively generous unemployment benefits and a labor force unfit for today’s demanding employers rather than a still-sick macroeconomy, suggesting that there’s less slack in the labor market than might appear at first glance.

There’s not much evidence for that. As newly minted San Francisco Fed president John Williams pointed out in a recent speech, the “natural” rate of unemployment—the one below which inflationary pressures rise—has probably risen from 5.0% before the financial crisis to 6.7% now. About half the increase is the result of extended unemployment insurance benefits (which will wane soon enough). The balance is the result of severe shocks that have hit the labor market, notably in the construction sector. Williams and his staff estimate that these increases in the natural rate are temporary, and likely to dissipate over the next few years. In any case, the unemployment rate is still considerably above 6.7%, and likely to stay there for quite a while. (Full text here: http://www.frbsf.org/economics/speeches/2011/john_williams0222.php)

SF Fed paper argues against mismatch

2010-12-05T19:02:00-08:00

In the November 8 issue of the San Francisco Fed’s Economic Letter, Rob Valletta and Katherine Kuang look at the regional and sectoral behavior of employment and find little evidence for the mismatch employment thesis currently making the rounds. If there were a serious mismatch problem, you’d expect to find major disparities in employment across geography and industry: healthy regions or sectors would show shortages of workers, and sickly ones would show surpluses. But in fact you don’t see that: regional and sectoral variation is little different now from past cycles. Valletta and Kuang conclude that the major reason for persistently high unemployment is cyclical and not structural.

The San Francisco Fed study looks at Beveridge curves. Beveridge curves plot unemployment against job vacancies. In general, the more vacancies there are, the lower the rate of unemployment. Minneapolis Fed President Narayana Kochlerlakota’s argument, outlined in a speech he made over the summer, depended heavily on the idea that the Beveridge curve suggests that the unemployment rate should be about 1.5–2.0 points below where it is, which is what leads him to the mismatch conclusion. His curve, though, was based entirely on only the most recent decade’s data.

But there’s nothing at all stable about this relationship over time. The Valletta–Kuang paper includes a graph showing how much the Beveridge curve has wiggled around over the decades. For example, Valetta and Kuang point out that the curve shifted about 4 points to the right between 1960 and the early 1980s before shifting back about 2.5 points by 1989, and that the variation in the NAIRU during that period was considerably smaller than the movement in the curve would suggest. We thought this worth a closer look.

Unfortunately, the JOLTS data only begins in 2000. Most earlier work on the unemployment–vacancy relationship used the Conference Board’s old Help Wanted (HW) index as a proxy for openings. In an earlier paper, Valletta used the period of overlap between the HW series and the JOLTS data to estimate a consistent vacancy rate series going back to 1960. He graciously shared that with us.

In the graph below you will see a set of Beveridge curves by decade—they show that a major cause of the noisiness in the long-term relation is that the vacancy–unemployment connection shifts over time. The r2’s for the decade regressions are a lot better than for the whole series, with most in the high 0.80s/low 0.90s. But there are also some noisy decades: the r2 for the 1970s is just 0.39. The 1980s aren’t so great either, with an r2 of 0.77.

The labor market is usually thought to be functioning better the further down and to the left (i.e., closer to the origin) the Beveridge curve is on the graph. The further out the curve is, the more friction, like regional or sectoral mismatches. The graph for the 2000s (which is based on data through June 2009, when the relationship prevailing earlier in the decade started breaking down) shows a much better-functioning labor market than in earlier decades, especially the 1980s.

Until, that is, we get to recent history, as shown by the dot marking the data for November 2010. It’s about two unemployment points to the right of where the curve suggests it “should” be. But it’s actually lower than the 1980s curve would predict, and not that much higher than the 1970s curve. And those were decades of major structural change in the U.S. economy—periods of major financial and real sector shocks, far more severe than those of the following two decades.

So the recent breakdown in the Beveridge curve looks to us more like a reflection of a major financial and psychological shock (one that has left employers extremely shy about hiring) than some fresh mismatch in the U.S. labor market. The fresh mismatch theory seems especially odd in light of the fact that things seemed to be functioning so well (meaning the Beveridge curve was close to the origin) for so much of the decade until the economy fell apart. The novelty is the having fallen apart, not some recent change in the labor force.

Mismatching the facts

2010-09-02T18:54:00-07:00

In a speech delivered on August 17, Minneapolis Fed president Narayana Kocherlakota claimed that there’s been a major breakdown between the relationship between the unemployment rate and the number of job vacancies reported in the BLS’s Job Openings and Labor Turnover Survey (JOLTS). According to Kocherlakota, the breakdown began in mid-2008 as the unemployment rate rose more rapidly than the JOLTS openings data suggest—a breakdown that intensified in mid-2009, as the unemployment rate continued to rise and then failed to decline significantly as the number of openings rose by nearly 20%. Kocherlakota explains this by asserting that there’s a mismatch between the skills, geography, and demography of available workers and unfilled openings. And that’s not anything that monetary policy can change: “the Fed does not have a means to transform construction workers into manufacturing workers.”

There are many things wrong with Kocherlakota’s argument. While he’s right that a regression that forecasts unemployment based on the JOLTS openings rate says that the jobless rate should have been 7.7% in June, not 9.5%, the number of unemployed fell by over a quarter-million more than the number of openings rose during the first half of 2010.

Despite that, there are still almost 5 unemployed for every opening—down from over 6 at the end of 2009, but still enormously high. Also, as the graphs below show, the major problem with the labor market is that the recession was harsh and the recovery so far has been weak. The gap between GDP growth and employment growth, though wider than it was in the recessions of the early 1990s and early 2000s, isn’t out of line with earlier downturns, like those of the early 1980s, mid-1970s, and the 1950s. Since the JOLTS data that Kocherlakota bases his mismatch thesis on only begins in December 2000, he’s missing a lot of important history.

And, the share of permanent job losers, as opposed to those on temporary layoff, hit a record high in over 40 years of data at the end of 2009, and has come down only slightly since. By contrast, the share on temporary layoff is at a record low. Clearly, that composition makes re-employment a lot harder.

In a paper prepared for a Brookings Institution panel in March, Michael Elsby, Bart Hobijn, and Aysegul Sahin review the grim pathologies of the labor market in the Great Recession. In almost every aspect, the downturn was the worst since the 1930s. Among their specific points:

While inflows into unemployment in the early part of the recession were dominated by the weaker demographics—the young, the less educated, the nonwhite—the rate of exit has been broadly similar for all subgroups.
Since the workforce is now older than it was in earlier recessions, the rise in the unemployment rate is actually sharper than it appears, since older workers are less likely to be disemployed than younger ones. Adjusting historical unemployment figures for the labor force’s changing age composition shows that this recession’s unemployment rate is a record by a significant margin.
Specifically addressing the mismatch argument, Elsby et al. find that instead of finding a divergence in outflows from unemployment between industries in structural decline and those not in decline, the rates of sectoral outflow have converged. (Outflow rates in the financial, durable goods and information sectors were all lagging the total when Elsby et al. published.) As with demographics, then, the problem is largely an aggregate one.
The dominance of long-term unemployment among the jobless in this cycle is disturbing, since the longer people are unemployed, the less likely they are to find new employment. Based on historical relations, Elsby & Co. project that the decline in unemployment could be half as rapid as it was in the mid-1980s. (Of course, if GDP growth remains weak, then that recovery would be even slower.)
Another factor likely to contribute to a slow decline in the unemployment rate: the high share of the employed working part-time for economic reasons. In July, they were 6.0% of the total employed, well over two standard deviations above the series’ 55-year average, and not far below last December’s record of 6.6%. And it’s likely that they’ll find full-time work before the currently unemployed do.
Some analysts have pointed to the extension of unemployment benefits over the last couple of years as keeping the jobless rate higher than it would otherwise be. Based on other studies, Elsby et al. estimate that emergency benefits have contributed between 0.7 and 1.8 points of the 5.5 point rise in unemployment in the recession, with the lower end far more likely than the upper. (One reason: the statistical estimates are largely based on periods when temporary layoffs rather than permanent losses were more common.)
Finally, the JOLTS data show that the quit rate was remarkably low during the recent recession. It’s picked up a bit since but remains lower than at any time before late 2008. This suggests that employed workers, who presumably have the qualifications that employers desire, remain remarkably skeptical about the possibility of finding fresh work. If it were easier than it looks for the qualified to find a job—as the mismatch theory suggests—then the quit rate would be higher.

Where we are?

2010-08-18T03:11:02-07:00

Here’s an updated guide to “where we are”—how the U.S. economy is faring relative to the average of previous financial crises around the world. Though individual details vary, we’re following the script pretty well.

In the graphs of four major indicators that below, the lines marked “average” are the averages of fifteen financial crises in thirteen rich countries since the early 1970s, as identified by the IMF. GDP isn’t shown because the experiences were so varied that the averages were meaningless. But for the indicators shown, the averages do illustrate some tendencies worth taking seriously.

Employment
Though the U.S. peaked later and bottomed earlier than the average, and also rose higher and fell harder, the trajectories of the two lines are still remarkably similar. Note that after hitting bottom, employment in the average experience grew very slowly. If we’re in for anything like that average, then we’re likely to see employment growth of only about 35,000 a month over the next year—less than a third what’s necessary just to accommodate population growth. That suggests that we could see an unemployment north of 10% in about a year.

CPI
Given the gyrations in energy prices over the last couple of years, reading the headline CPI has been very difficult. But the gyration does seem to be around the average line. And core CPI—for which we don’t have international data—is tracking the average pretty tightly. If inflation follows the script, it should continue to decline into next year. With core inflation at around 1%, it’s reasonable to expect that we could go into mild deflation sometime over the next few quarters.

Interest Rates
Rates on 10-year Treasury bonds have fallen harder in the U.S. than in other crisis-afflicted countries, but the trend is typically down for almost four years after the onset of crisis. Further declines in U.S. rates seem like a stretch, but the likelihood of an upward spike looks remote.

Stocks
Stocks fell much harder in the U.S. than they did in the wake of the average financial crisis, though they did enjoy five quarters of nice recovery. As with interest rates, further declines seem unlikely; in fact, the average increase from here would be around 7% over the next year.

So, all in all, we’re getting pretty much what we might expect out of our major economic and financial variables: a weak, choppy recovery with a deflationary undertow.

The Richmond Fed’s Take on Unemployment and Participation Rates

2010-04-08T18:07:13-07:00

In their economic brief, "Comparing Labor Markets across Recessions: A Focus on the Age Composition of the Population," Richmond Fed researchers Marianna Kudlyak, Devin Reilly, and Steven Slivinski find that controlling for the recent decline in teenagers' participation rate produces an unemployment rate of 11.3%, a new post-war high, and that much of the decline in the overall participation rate has been driven by the increasing percentage of workers 55 and older. Things are what they are, and you have to be careful about making this kind of adjustment, but this is Fed research and we are taking it seriously.

Setting the stage, they find that although the 2008 contraction in output was comparable to those of the 1957 and 1973 recessions (the prior record-setters), the 7% decline in employment was more severe than any other post-WWII recession. The 2% decline in weekly hours in the 2008 recession was not as severe as 1969's 3%, or 1973's 2.1%, but aggregating hours worked with employment produces a 9% decline, far worse than 1948's 5.7%, the previous record.

For their next comparison, they assume that the recovery began when Nonfarm Business Sector output turned positive, Q309 for the current recovery, and find that employment growth is lagging prior recoveries. Of the ten prior recoveries considered in the paper, employment continued to decline during the first two quarters of the recoveries following the 1957, 1960, 1991 and 2001 recessions, but percent declines were larger than the current decline (1.35%) only following the 2001 recession (1.8%) Positive note: Unlike the 2001 recession/recovery, weekly hours showed modest growth during the second quarter of the current recovery.

Noting that looking at labor indicators without adjusting for demographics, "may not be the best way to compare recessions," they adjust for teenagers' diminishing and the over-55 set's growing share of the work-force. In 1982, when unemployment hit 10.7%, its post-war high, teenagers constituted 7.6% of the workforce; in 2009, their share had dropped to 4%. This may be a good thing in the long-run as there is anecdotal evidence suggesting that teens are staying in school longer, and it is definitely a good thing for the unemployment rate. Teens have a volatile and high rate of unemployment, so Kudlyak et al. adjust the current rate by holding the teenagers' participation rate constant since 1982, resulting in the postwar high of 11.3% mentioned above. They go on to say that the larger share of older participants in the labor force "means the 'natural' unemployment rate is lower than it as in 1982," and that Q309's 10% is "likely further" from the natural rate than was 1982's 10.7%.

They apply the same technique to the labor force participation rate to determine how much of the current decline is cyclical and how much structural. Total participation was 58.6% in 1948, rose to 67.3% in 2000, and has declined since, hitting 64.7% in January 2010, with half of the decline occurring since December 2007. In March 2000, workers 55 and older constituted 26.8% of the working-age population, compared to 30.3% currently. Reconstructing the series using the current age composition replaces the long downward trend shown in the official series with a more modest decline that does not begin until mid-2009 and so far has only reached the level of mid-2005. Reconstructing another series that keeps 1999 participation rates constant across age groups, they show that in 2005 the official rate actually rose above what would have been predicted by demographic changes, and has only recently fallen below, suggesting that much of decline since 2000 is structural.

Their conclusions are a bit contradictory: the high unemployment rate of this cycle is "much higher in relative terms," than those of prior recessions, meaning there is a great deal of slack in the labor force but, since there has been less of a cyclical decline in the participation rate, there may not be as many workers outside the labor force ready to step in as the recovery continues as in prior cycles.

Economic Brief with a number of graphs here: http://www.richmondfed.org/publications/research/economic_brief/2010/pdf/eb_10-04.pdf

Philippa Dunne & Doug Henwood

Impulse Control

2010-03-13T18:09:51-08:00

Decomposing the deficit: structural, cyclical, and the fiscal impulse

Against a still struggling recovery, and continuing weakness in state revenues, talk in Washington is turning to cutting back. That may sound strange, but hear us out.

With the Obama administration released its proposed 2011 budget last month, a lot of attention has rightly been paid to the ginormous deficit—its possible corrosive economic effects and the possible difficulties in funding it. But what’s lost in those worries is what’s going to happen as the stimulus program expires.

Specifically, after the most fiscally stimulative budgets in modern history, barring major changes, policy is about to turn much tighter.

Structure of the Deficit

The federal deficit can be thought of as having two major components, structural and cyclical. The structural deficit is what the balance would be were the economy at full employment; the cyclical component is the contribution that comes from departures from full employment, i.e., the state of the business cycle.

The graph below shows changes in the structure of the deficit from 2007 through 2018. Sometimes the cyclical component is positive, lessening the deficit (as was the case in 2007); sometimes it’s negative, as it is most other years:

(These estimates come from the Office of Management and Budget, as part of the President’s budget for 2011. Despite that authorship, the estimates look cautious to us. They define full employment as an unemployment rate of 5.5%, and we don’t see the economy returning to that level until 2017. And even if they’re optimistic on magnitudes and the longer-term trend, they’re probably not wrong about the direction of things over the next couple of years.)

The best way to isolate the policy effects of changes in government spending is to look at the cyclically adjusted balance. If the deficit passively swells because of recession, that’s just the so-called automatic stabilizers switching on. But if policy adds to it, through spending increases or tax cuts, then it becomes actively stimulative.

The table below shows OMB’s estimate of the total budget deficit through 2018, and we are concentrating on the next few years. In fiscal 2010, the total budget deficit is 10.5% of GDP, 3.2% of it the result of the recession and 7.3% of it structural, or from policy. That’s up from a deficit of 9.9% of GDP in 2009, 2.4% of it cyclical and 7.6% of it from policy. In other words, while the 2010 deficit is larger than 2009’s, the main reason for that is cyclical, not structural, since there’s less stimulus and bailout spending coming this year than last. Going forward, the structural deficit is projected to continue shrinking—to 5.1% of GDP in 2012, and 3.0% the following year.

The Fiscal Impulse

Changes in the structural budget balance are sometimes called the fiscal impulse, shown in the right column. The decline in the structural deficit from 7.3% of GDP in 2010 to 5.1% in 2011 results in a contractionary fiscal impulse of 2.2% of GDP. Both the OMB and the CBO try to separate the cyclical and policy contributions to the deficit. The CBO’s number start in 1962 and end in 2011, and the OMB’s run from 2007 to 2017. If we splice the OMB and CBO series together, which is only slightly reckless since they’re pretty similar when they overlap, this is one of the starkest episodes of fiscal tightening in the last five decades, exceeded only by the 3.1% shift in 1969. A recession began in December 1969 and continued through November 1970. The second-biggest tightening was in 1987, the year of the stock market crash, which was followed a little later by a long period of sluggishness.

Of course, the fiscal impulse can be stimulative. By the CBO’s measure, the 2009 impulse was expansive to the tune of 5.9% of GDP; by the OMB’s, it was 4.5%. These numbers are three or four times earlier maximums, like those of 1983 or 2002. And it’s not unreasonable to conclude that what economic recovery we’ve seen since mid-2009 is the result of this massive stimulus. This is apparent not only at the abstract macro level of budget numbers—it’s visible in specific cases, like the effect of cash-for-clunkers on car sales (which ebbed when the program expired) and the effect of federal support on the housing market (there would simply be no mortgage market without Washington’s support). So what is going to happen as the Fed’s purchase of mortgage securities ends in March? As the stimulus spending declines to a trickle

According to a simple regression, the fiscal impulse has an almost one-for-one influence on GDP growth. (See graph below.) So next year’s drag of 2.3% of GDP implies a brake on growth almost that large. Of course, fiscal policy, while important, is far from the only influence on growth; the regression’s r2 is .10, suggesting that it explains about 10% of the yearly variation in GDP growth (though from eyeballing the graph, we see that the extreme cases seem to matter most). That’s not beanbag, but it still leaves a lot to be explained.

A stronger influence on growth is the change in nonfederal credit; as shown in the graph below, it explains about 18% of the yearly variation in GDP growth. Together, the two explain close to 30% of the yearly variation in GDP growth. Putting the two together suggests that we’d need a very strong growth rate of 7-10% in nonfederal credit this year and next to offset the fiscal drag. We’ve only come close to that in two years out of the last 50. So we’re back to a dilemma we’ve identified many times: it looks like GDP growth needs a major revival of private credit expansion—but is that possible or even desirable?

Given that 70% of GDP growth remains unexplained by these two variables, we don’t mean them as literal projections. But it does offer support for fears that since the recovery was driven by fiscal stimulus, once that’s gone, we could be in trouble again.

Gold’s Cloudy Crystal Ball

2009-10-28T03:47:59-07:00

Given all the fiscal and monetary stimulus of the last year or two, it’s no wonder that many people are on inflation watch. And whenever that happens, many eyes turn to gold, which is revered in some circles for its power to foretell the future. Is this reputation for prescience deserved?

Not for the last 25 years, it isn’t. Long-time readers may recognize this as a theme of ours: markets generally reflect the present, and rarely divine the future. We’ve made this point concerning oil prices (which reflect, don’t lead supply and demand trends), fed funds futures (which predict next to nothing), and inflation-protected bonds (which reflect current inflation, and don’t anticipate it). And now we can add gold to the list of markets with undeserved reputations as seers.

As the graph below shows, gold did a pretty good job of anticipating moves in the CPI in the late 1970s and early 1980s, but since then the relationship has broken down. For example, the correlation coefficients for the yearly change in the gold price and the CPI during the 1974–84 period improve the further ahead you go—from 0.61 for the same month to 0.68 for a three-month lead on the CPI. They stay in the 0.67–0.68 range for leads of up to about a year (that is, gold anticipates moves in inflation). But if gold were really good at anticipating inflation, you’d expect a steeper gradient of improving correlations than just 0.07 point over the course of three months, followed by a flattening, wouldn’t you?

And after 1984, gold’s powers of prognostication collapse completely. Since 1985, the correlation coefficient for the yearly change in the gold price and the CPI is 0.01 for the same month, 0.00 with gold given a lead of three months, and –0.04 six months out. And the correlations remain slightly negative up to fifteen months out.

In fact, you could do far better in predicting future inflation just by drawing a straight line from current inflation. The correlation coefficient between the yearly inflation rate and the CPI three months out for the graph’s entire history is 0.96. As with gold, the 1974–84 experience pulls up the average; since then, the coefficient is 0.79 (lower, but still a lot better than gold). Six months out, it’s 0.58—not great, but still miles ahead of gold.

Clearly the 1974–84 period is special: a great acceleration in inflation followed by a great disinflation. Maybe we’re in for a rerun—though there are plenty of reasons why the present is very different from that past. One difference is the behavior of productivity and unit labor costs. Another is that the 1970s were the climax of a long wave of rising labor militancy. We’ve got nothing like that now.

That last point is underscored by the graph below, which shows the annual number of major strikes since 1947. Strikes actually do a better job of predicting inflation than gold does. The correlation coefficient between the yearly change in strikes and the next year’s inflation rate since 1985 is 0.18. If the experience of the first nine months of 2009 holds for the rest of the year, we’ll have four—count ‘em, four—work stoppages involving 1,000 workers or more this year. That will be an all-time record low since the BLS started tracking these figures in 1947. From 1947–79, we averaged 303 a year. For 2000–2008, the average was 22 a year. So far this year, we’re down by 90% from the decade average, and more than 99% from the 1947–79 average—and it’s hard to see strike activity picking up anytime soon.

So what is the current strength in gold telling us? That a lot of money has been racing into the metal. While much of that money probably belongs to investors who think we’re going to see a major pickup in inflation—and some of it to investors who want to get ahead of the spread of that perception, regardless of their own convictions—gold’s record as an inflation forecaster isn’t something to bank on.

Bad News from the SF Fed

2009-08-06T15:18:17-07:00

The last thing we want right now is a study from a regional federal reserve bank with the headline, "Jobless Recovery Redux?" and a subhead, "Reasons for pessimism," but that's what we have from the SF Fed, and it's hard to argue with their logic.

Researchers Mary Daly, Bart Hobjin, and Joyce Kwok consider three labor market indicators--flows in and out of unemployment, involuntary part-time employment, and temporary layoffs--to project how the unemployment rate may behave when the anticipated recovery begins.

They point out that before the 1991 recession, increases in the inflow rate, the speed at which workers move into unemployment, and decreases in the outflow rate, the pace at which they find jobs, were nearly equivalent in relative terms during recessions. The resultant sharp increases in the unemployment rate were followed by rapid recoveries as firms hired back workers in improved conditions.

Increases in the unemployment rate in the 1991 and 2001 recessions, however, were driven by disproportionate declines in the outflow rate, making lack of hiring the primary culprit. And recoveries diverged from the established pattern as well: outflow rates recovered much more slowly than they had in pre-1990 recoveries. In fact, the authors cite several studies showing that current business-cycle fluctuations in the unemployment rate are driven primarily by the outflow rate.

The current recession is particularly nasty because the outflow rate is at an historic low and the inflow rate is rising in line with the recessions of the 1970s and 1980s-actually, it's shooting straight up at present-both contributing to an extremely weak labor market.

The authors evaluate several possibilities for the future. Their benchmark, no change in the outflow/inflow rates, would bring us to 10% unemployment by 2010. The "Blue Chip" consensus included in the paper posits an employment recovery slightly weaker than that of 1983 and a bit stronger than that of 1992, bringing the unemployment rate to 10% in early 2010, from where it would begin to decline. But if inflow/outflow rates behave as they did in 1992, unemployment would peak around 11% by summer of 2010, and remain above 9% through 2011.

And that's where temporary layoffs and involuntary part-time work come in. In the current recession those working part-time for economic reasons has risen dramatically and to an historical high, from 3.0% in December 2007 to 5.8% in April 2009, with significant reductions in hours across broad sectors. At the same time the share of workers on temporary-as opposed to permanent-layoff is very low. In fact it actually fell from 12.9% in Dec-07 to 11.9% in April-09. (Generally the share of workers on temporary layoffs rises during recessions: it increased from 16.1% to 20.7% between July 1981 and November 1982.) So with few workers on temporary lay-off waiting in the wings, and a large number working part-time against their wills, it seems logical that employers will expand the hours of partially-idled workers instead of taking on new employees. The authors formulate their forecast for the unemployment rate by adding the linear relationship between part-time employment and the unemployment rate between Dec-07 and April-09 to the Blue Chip consensus. This suggests labor market slack will be higher by the end of the year than at any time since WWII, the outflow rate will be historically low, and the unemployment rate will be sticky. In other words, Jobless Recovery III.

Full report here:

http://www.frbsf.org/publications/economics/letter/2009/el2009-18.html

Financial Crises: A Close-up

2009-06-09T19:56:32-07:00

In our issue of May 7, we wrote about the IMF’s study of 122 recessions in 21 “advanced” countries since 1960. They found that downturns associated with financial crises are far longer and deeper, and recoveries slower and weaker, than those that aren’t financially driven. Here’s a close-up of the behavior of interest rates, stock prices, inflation, and employment in the four years after the onset of the crisis in 15 major financially driven recessions identified by the IMF. (See notes below)

In a nutshell, we found that on average, interest rates and inflation generally decline for the full four years; stock prices rise steadily, though not explosively, after a two-quarter decline; and employment falls for about three years before beginning a weak recovery.

GDP experiences are quite varied, with some countries experiencing sharp, sustained declines; others, brief, sharp declines followed by a strong recovery; and still others, extended periods of stagnation. Those disparate experiences produced a very misleading “average” line, so we are not providing a GDP graph. But despite those contrasting GDP patterns, almost all countries suffered from a very weak job market for several years after the financial system hit a wall.

	GDP	Employment
Australia	-1.7%	-3.8%
Denmark	-8.7
Finland	-20.5	-19.2
France	-1.0	-1.5
Germany	-2.8	-3.1
Greece	-1.6
Italy	-2.2	-8.2
Japan 1	-3.9	-0.4
Japan 2	-3.4	-2.3
New Zealand	-7.6	-6.9
Norway	-7.9	-6.6
Spain	-0.4	-9.0
Sweden	-18.5	-13.1
UK 1	-3.4	-1.2
UK 2	-2.5	-6.2
average	-5.8	-6.3
median	-3.4	-6.2
Total decline in real GDP and total employment in recessions in the IMF sample, from pre-crisis maximum to cycle low

Comparing recent U.S. experience to the average yields the following conclusions. Interest rates, inflation, and stock prices are all down much harder than the “typical” experience. This suggests that interest rates are probably not at risk of rising sharply, but we’re almost certainly past their bottom; stock prices, even with the recent rally, are not acting with outlandish exuberance; and inflation isn’t much of a risk for some time to come. Employment, however, is tracking the average very closely, suggesting that while the worst of the job losses are probably behind us, we’ve got another year and a couple of million to go.

Interest rates: down

Although financial crises typically come with a huge fiscal cost, as revenues sag and bailout expenditures rise, interest rates on long government bonds tend to decline steadily throughout the four year period we studied. Rates fell in 12 of the 15 examples over the first 8 quarters following the crisis, and 13 of 15 over 16 quarters.

Stocks: mostly up

Stock prices do surprisingly well in financial crises, after an initial spill. Actually, prices tend to decline over the year leading up to the crisis (about 4–5% on average); the post-crisis leg down is sharper but shorter than what preceded it (about 12%). But after that decline, stock prices tend to rise, though at far from a barn-burning pace—only around 5% a year.

Inflation: down

Though a lot of market participants expect that the financial bailout will ultimately prove inflationary, the history of financial crises suggests otherwise.

Employment: down hard

Deep, sustained declines in employment are almost universal following financial crises. If these averages hold, we’re about two-thirds of the way through the contraction in numbers, and not quite that far in time.

Notes:

In the four graphs, the line labeled “average” is the mean of the 15 financially driven recessions identified by the IMF, with the quarter of the crisis’ onset set to 100. The line labeled “U.S.” is the recent experience of the United States, with the onset of crisis set to the third quarter of 2007. Horizontal axis represents quarters before and after crisis onset (“C”).

These are the crises the IMF reviewed:

Australia	1990Q2–1991Q2
Denmark	1987Q1–1988Q2
Finland	1990Q2–1993Q2
France	1992Q2–1993Q3
Germany	1980Q2–1980Q4
Greece	1992Q2–1993Q1
Italy	1992Q2–1993Q3
Japan	1993Q2–1993Q4
Japan	1997Q2–1999Q1
New Zealand	1986Q4–1987Q4
Norway	1988Q2–1988Q4
Spain	1978Q3–1979Q1
Sweden	1990Q2–1993Q1
United Kingdom	1973Q3–1974Q1
United Kingdom	1990Q3–1991Q3