2014 AretheLongTermUnemployedontheMa

• (Krueger et al., 2014) ⇒ Alan B. Krueger, Judd Cramer, and David Cho. (2014). “Are the Long-Term Unemployed on the Margins of the Labor Market?.” In: Brookings Papers on Economic Activity, Spring 2014.

Subject Headings: Long-Term Unemployment; Phillips Curve.

Cited By

• http://scholar.google.com/scholar?q=%222014%20Are%20the%20Long-Term%20Unemployed%20on%20the%20Margins%20of%20the%20Labor%20Market%22

Quotes

Abstract

The short-term unemployment rate is a much stronger predictor of inflation and real wage growth than the overall unemployment rate in the U.S. Even in good times, the long-term unemployed are on the margins of the labor market, with diminished job prospects and high labor force withdrawal rates, and as a result they exert little pressure on wage growth or inflation.

1. Introduction

A number of observers have noted that in recent years conventional Phillips Curve and Beveridge Curve models predicted greater price deflation, greater real wage declines and fewer vacancies as a result of the high rate of unemployment experienced during the Great Recession and its aftermath than actually occurred. Several economists have provided possible explanations for the missed predictions of the Phillips Curve, based on anchoring of expectations (Bernanke, 2007 and 2010) or changes in the distribution of price increases and interactions in the Phillips Curve (Ball and Mazumder, 2011). Others have shown that the price Phillips Curve relationship is stable if the short-term unemployment rate (defined as the number of job seekers unemployed for 26 weeks or less relative to the labor force) is used instead of the total unemployment rate (Gordon, 2013 and Watson, 2014), while others have shown that the Beveridge Curve relationship is stable if short-term unemployment rate is used instead of the overall unemployment rate (see Ghayad and Dickens, 2012).

This paper explores the plausibility of a unified explanation for the recent shifts in the price and real wage Phillips Curves and Beveridge Curve in the U.S.: namely, that the long-term unemployed, whose share of overall unemployment rose to an unprecedented level after the Great Recession, are on the margins of the labor force and therefore exert very little pressure on the job market and economy. The hypothesis we seek to test is that the longer workers are unemployed the less they become tied to the job market, either because, on the supply side, they grow discouraged and search for a job less intensively (e.g., Krueger and Mueller, 2011) or because, on the demand side, employers discriminate against the long-term unemployed, based on the (rational or irrational) expectation that there is a productivity-related reason that accounts for their long jobless spell (e.g., Kroft, Lange and Notowidigdo, 2013 and Ghayad, 2013). Either of these explanations would imply that the long-term unemployed are on the margins of the labor market, and have a different effect on the macroeconomy than the short-term unemployed.

Moreover, the demand-side and supply-side effects of long-term unemployment can be viewed as complementary and reinforcing of each other as opposed to competing explanations, as statistical discrimination against the long-term unemployed could lead to discouragement, and skill erosion that accompanies long-term unemployment could induce employers to discriminate against the long-term unemployed.

Motivated by the apparent stability of the Phillips and Beveridge Curves when the short - term unemployment rate is used to measure labor market slack, we assemble varied evidence to assess the hypothesis that the long-term unemployed are on the margins of the labor market. To preview our main findings, we tentatively conclude that the long-term unemployed exert relatively little pressure on the economy, although the international evidence that we have been able to assemble to this point is more mixed than the evidence for the U.S., and suggests that long-term unemployment means different things in different countries and contexts.

We first briefly review evidence from the U.S. showing that the price Phillips Curve, expected real wage Phillips Curve and Beveridge Curve are all stable if the short-term unemployment rate is used to measure labor market slack, and that the long-term unemployment rate has a comparatively modest effect when it is included in regression models. This result is consistent with Llaudes.s (2005) conclusion that the long-term unemployment rate was a much less significant determinant of price inflation and wage growth than the short-term unemployment rate in many OECD countries prior to the Great Recession. We also create two new measures of the unemployment rate, one in which the duration of unemployment is weighted by a measure of search intensity and another in which duration is weighted by the callback rate from audit studies. Both alternative measures have greater predictive power than the total unemployment rate. We also extend this analysis to estimate the Beveridge Curve in the U.K., which saw a sharp rise in long-term unemployment in the early 1980s, and Sweden, which saw a sharp rise in long-term unemployment in the early 1990s, and then a gradual decline.

Sweden and the U.K. in these periods were selected because, compared to other countries, their pattern of long-term unemployment as a share of the unemployed more closely resembles that of the U.S. over the last decade.

Next we provide a detailed profile of the long-term unemployed in 2012, and examine how the composition of the long-term unemployed has varied over time. While some notable industries (e.g, construction) and demographic groups (e.g., African Americans) are over represented among the long-term unemployed, the long-term unemployed are ubiquitous, spread throughout all corners of the economy. Fully 36 percent of the long-term unemployed last worked in the sales and service sector, suggesting that weak aggregate demand was the driver of long-term unemployment. Using the relationship between workers. characteristics and wages from 2004 to 2006 to project earnings for the long-term unemployed, we find modest changes in the composition of the long-term unemployed over the business cycle, with workers with more highly rewarded characteristics more likely to be represented among the long-term unemployed in recessions, although the differences are small.

We next examine the rates at which unemployed workers find employment or exit the labor force, by duration of unemployment. Importantly, we examine transition rates both over a month and over a year or longer. Longer durations of unemployment are associated with a lower transition rate into employment, and available evidence suggests that observed duration - dependent transition rates are not primarily a result of heterogeneous job searchers (e.g., Heckman and Singer, 1984). From 2008 to 2012, only 11 percent of those who were long-term unemployed in a given month returned to steady, full-time employment a year later. This low - transition rate into steady employment is considerably lower than what would be predicted from monthly transition rates if such rates were independent over time and groups, highlighting that the long-term unemployed frequently are displaced soon after they gain reemployment. The long-term unemployed normally have a higher rate of labor force withdrawal than the short-term unemployed, although following a recession the labor force withdrawal rates for all duration groups tend to collapse to a common, relatively low level. We explore whether the process of labor force withdrawal rates gradually returning to their historical norm.with higher exit rates for the long-term unemployed. as well as a lower match rate for the long-term unemployed, can cause the Beveridge Curve to loop around a stable path following a sharp downturn. Specifically, we extend the calibration-type model of Kroft, et al. (2013) to allow for varying labor force exit rates and differential match efficiency for the long-term unemployed to project the path of the Beveridge Curve under a stable matching function. The results suggest that from 2002-07 the long-term unemployed were about 60 percent as efficient in job matching as the short-term unemployed. Using the matching function estimated for the 2002-07 period, the calibrated model does a reasonably good job capturing the rise in unemployment and shift of the Beveridge Curve in the 2008-13 period, as well as the rise in the share of unemployed workers who are long-term unemployed. Future projections predict a gradual return to the original Beveridge Curve as the share of long-term unemployment declines due to labor force exits or (less likely) transitions to employment.

An alternative explanation for the Beveridge Curve gradually returning to its original position is that a stronger labor market enables more of the long-term unemployed to transition into employment. To explore this possibility, we compare trends in long-term and short-term unemployment in different regions of the U.S. Our preliminary analysis suggests that the long - term unemployment rate has remained elevated even in low-unemployment rate states (defined as the 13 states with the lowest unemployment rates in the U.S. as of October 2013). In addition, we do not find evidence that the long-term unemployed are faring better in terms of transitioning to employment in the low-unemployment states than in the high-unemployment states. Indeed, the long-term unemployed appear to be following a similar path of transition rates both into employment and out of the labor force in both the low - and high-unemployment states. These findings suggest that the long-term unemployed will continue to encounter difficulty finding employment even if the unemployment rate continues to fall, although a stronger economy would undoubtedly raise the prospects of the long-term unemployed.

We conclude the paper by briefly considering some of the policy implications of the hypothesis that the long-term unemployed are on the margins of the labor market. Because the short-term unemployment rate has returned to its pre-recession average, one important implication.if the hypothesis that the long-term unemployed are largely on the margins of the labor market is correct.is that further declines in short-term unemployment would be expected to be associated with rising inflation and [[stronger real wage growth]]. More importantly, our findings also suggest that a concerted effort will be needed to raise the employment prospects of the long-term unemployed, especially as they are likely to withdraw from the job market at an increasing rate if they follow the same path as in the previous recovery.

The Duration of Unemployment and Inflation, Wage Growth and Vacancies This section summarizes movements in the price Phillips Curve, expected real wage growth (which we call the wage Phillips Curve), and the Beveridge Curve since the Great Recession, and provides econometric evidence suggesting that the rise in long-term unemployment has caused these historical relationships to shift. We start with evidence from the United States and then turn to two European countries that had previously experienced substantial rises in long-term unemployment: the United Kingdom and Sweden.

The Price Phillips Curve

We start by estimating a simple short-run expectations-augmented Phillips Curve for core price inflation using the following specification:

[math]\displaystyle{ π_t – π_{t-1} = α + β_{μt} + ε_t }[/math]

where. t denotes the annual average percent change in the consumption expenditures chain price index excluding food and energy items (. core PCEPI.) in year t and. t denotes the average annual unemployment rate. Expectations are captured by lagged inflation.

We restrict our sample to the period from 1976 through 2013 to avoid complications from the oil shock and stagflation of the early 1970s. Nevertheless, the results are robust to the inclusion of data back to 1960, the first year for which core PCEPI inflation can be officially estimated. Despite the parsimonious nature of this model, our specification of the price Phillips Curve illustrates the previously noted shift in the relationship between changes in price inflation and the unemployment rate since the Great Recession.

References

;

	Author	volume	Date Value	title	type	journal	titleUrl	doi	note	year
2014 AretheLongTermUnemployedontheMa	David Cho Judd Cramer Alan B. Krueger			Are the Long-Term Unemployed on the Margins of the Labor Market?						2014