No. 12-3
What Can We Learn by Disaggregating the
Unemployment-Vacancy Relationship?
Abstract:
Rand Ghayad and William Dickens
The Beveridge curve—the empirical relationship between unemployment and job vacancies—is thought to be
an indicator of the efficiency of the functioning of the labor market. Normally, when job vacancies rise,
unemployment falls, following a curved path that typically remains stable over long periods of time. When
vacancies rise and unemployment does not fall (or falls too slowly) this may be an indication of problems of
structural mismatch in the labor market leading to an increase in the lowest unemployment rate that can be
maintained without increasing inflation (the NAIRU or nonaccelerating inflation rate of unemployment). Such a
change in the vacancy-unemployment relationship occurred once in the 1970s and persisted through the late
1980s, and we have recently observed a similar change.
This policy brief explores the nature of the recent change in the vacancy-unemployment relationship by
disaggregating the data by industry, age, education, and duration of unemployment, and by examining blueand white-collar groups separately.
The plots presented here reveal a similar pattern of increasing vacancies with little or no change in
unemployment in the recovery from the most recent recession across all categories except one: short-term
unemployment. The relationship between short-term unemployment and vacancies is unchanged. Thus, all of
the increase in vacancies relative to unemployment has taken place among the long-term unemployed.
This situation contrasts with the change in the Beveridge curve relationship in the 1970s, when an increase in
the level of vacancies without a decline in unemployment affected both the long- and short-term unemployed.
Further, in the 1970s the outward shift of the Beveridge curve was largely concentrated among blue-collar
workers, whereas the recent shift seems to have affected both blue- and white-collar workers. Finally, the shift
in the 1970s took place over several years, while the current shift seems to have taken less than six months.
Thus, there is reason to believe that the current shift may not have the persistence or the effects on the NAIRU
that the shift of the 1970s had.
Considering all the evidence together, we conclude that the current outward shift of the Beveridge curve is
likely being driven by something other than a mismatch between workers’ skills and the demands of available
jobs.
JEL Codes: D31, D63, I32
Rand Ghayed is a Ph.D. candidate at Northeastern University and a visiting graduate fellow at the Federal Reserve Bank of
Boston. William Dickens is a university distinguished professor at Northeastern University, a visiting scholar at the Federal
Reserve Bank of Boston, and a nonresident senior fellow with the Brookings Institution. Their e-mail addresses are
[email protected] and [email protected], respectively.
The views expressed in this brief are the authors’ and do not necessarily reflect the official position of the Federal Reserve
Bank of Boston or the Federal Reserve System.
This brief, which may be revised, is available on the web site of the Federal Reserve Bank of Boston at
http://www.bostonfed.org/economic/ppb/index.htm.
This version: October 2012
1
Although the U.S. economy has been recovering slowly since the last quarter of 2009, the unemployment rate
has remained stubbornly high. The persistence of high unemployment is particularly puzzling, given that the
number of job openings has been rising during the same period.
Figure 1 plots vacancy-unemployment points from 2001 on. The curved relationship between unemployment
and job vacancies depicted in Figure 1 is often called the Beveridge curve. Over the recession of the early 2000s
and the recovery from that recession, the vacancy-unemployment relationship remained remarkably constant
as it has for long periods of time in the past. However, in the recovery from the most recent recession we see
that vacancies have grown considerably without producing the normal decline in unemployment. It appears
that the Beveridge curve may have shifted outward.
There is one time in the past when the relationship shifted outward in a similar manner. During the 1970s,
vacancies rose without a normal drop in unemployment, and the Beveridge curve shifted outward for much of
the 1980s. During that period of time it was thought that the labor market was doing a worse job than usual of
matching workers and jobs, resulting in a higher NAIRU (nonaccelerating inflation rate of unemployment).
Are similar forces at work now?
This brief uncovers new facts that emerge from disaggregating the unemployment rate into various categories.
The insights from these disaggregations may reveal new information on whether the breakdown of the
vacancy-unemployment relationship today is similar to the change that took place in the 1970s and may
provide some insight as to the likely cause of the recent change.
Figure 1 displays an empirical relationship between data on vacancies from the Job Openings and Labor
Turnover Survey (JOLTS) of the Bureau of Labor Statistics (BLS) and the aggregate unemployment rate
obtained from the monthly household survey, also from the BLS. The data span the period January 2001
through June 2012 and are seasonally adjusted. The solid line in Figure 1 depicts a stylized Beveridge curve
that was estimated using data on unemployment and vacancy rates for the period prior to the start of the
recession. The plot reveals that by September 2009, the vacancy-unemployment points started to deviate from
the fitted curve in a counterclockwise direction, indicating a higher unemployment rate at any given level of
job openings.
One notable difference between the experience this time and what happened in the 1970s is the rapidity with
which this change took place. In the 1970s it took eight years for the outward shift to take place (see Figure 10).
The recent shift (about half the size of the shift in the 1970s) took less than a year, with most of the change
taking place in the last six months of 2009.
Another notable difference between the current period and the 1970s is the very high level of long-term
unemployment. Figure 2 shows the proportion of total unemployed who have been out of work for at least six
months.
It is interesting that the Beveridge curve starts shifting outward only after long-term unemployment reaches its
peak level. We will have more to say about this in the conclusion, but the emergence of the long-term
unemployed as a large fraction of the unemployed motivates the first break-down that we look at—the
Beveridge curve at different durations of unemployment.
2
Total Vacancies
Unemployment
and
Unemployment
Rates
by
Duration
of
We disaggregate the seasonally adjusted BLS series of monthly unemployment rates for all employees 16 years
of age and over to examine the vacancy-unemployment relationship at different durations of unemployment
in order to see whether the unemployed for different durations benefit differently from the recent increase in
the vacancy rate. We plot the aggregate vacancy rate, using data from JOLTS, against the fraction of the labor
force unemployed at different durations. Figure 3 presents this relationship for those unemployed for less than
five weeks.
The relationship between the vacancy rate and the very short-run (less than five weeks) unemployment rate is
essentially vertical—around 2 percent of the labor force is in the first five weeks of a spell of unemployment,
regardless of the level of job vacancies. There does not appear to have been any change in the relationship in
2009. Hence, evidence from the experience of individuals with unemployment spells shorter than five weeks
does not explain what we see in the aggregate plot.
Figures 4 and 5 illustrate vacancy-unemployment relationships using unemployed persons with duration 5–14
weeks and 15–26 weeks, respectively. Figure 6 combines all those unemployed for fewer than 27 weeks and
compares their fraction of the labor force with the aggregate vacancy rate.
As with the graphs for each of these durations individually, there is no evidence of an outward shift. However,
when the relationship is plotted using the fraction of the labor force that has been unemployed for more than
26 weeks, a number of interesting features are immediately apparent. First, the pattern in Figure 7 reveals a
counterclockwise outward shift that is consistent with what we see when we use the aggregate unemployment
rate. In addition to the shift, the pattern in Figure 7 shows that the vacancy and unemployment points for the
long-term unemployment group start to shift outward at the same time as the aggregate vacancyunemployment relationship breaks down.
Looking at how events have unfolded over time, we see in Figure 8 that the vacancy-unemployment points for
the short-term unemployment group cycle inward in a clockwise pattern, while those for the long-term
unemployment group move in an outward, counter-clockwise manner. At the aggregate level,
counterclockwise movements are common to all recessions because vacancies typically adjust before
unemployment to changes in labor demand. The different dynamics of the short- and long-term
unemployment-vacancy relationships suggest that the short-term unemployed benefit from the increase in the
vacancy rate far more than the long-term unemployed do.
What Happened in the ‘Seventies at Different Durations?
How does the recent change in the vacancy-unemployment relationship compare with what happened in the
1970s? As noted above, the recent increase happened much more quickly than the one in the 1970s, casting
doubt on the hypothesis that they share similar origins. Examination of changes at the disaggregated level
suggests another reason to doubt similar origins.
Figure 9 shows the vacancy-unemployment relationship using all unemployment groups for the period from
January 1960 to December 1988. The Conference Board’s help wanted index is used to construct the vacancy
3
rate, using the method suggested by Zargosky (1998). The plot clearly shows an outward shift of the Beveridge
curve.
Examining the Beveridge curves for the short- and long-term unemployment groups plotted with data
covering the 1960–1988 period, we notice that both curves shifted outward at the same time as the start of the
outward movement of the aggregate curve (Figures 10–11). This contrasts with the current period, in which the
breakdown in the vacancy-unemployment relationship is evident only for the long-term unemployed.
Other Decompositions
Do other ways of disaggregating reveal any other places where the shift in the vacancy-unemployment rate is
concentrated or missing? Below we disaggregate by industry, age, education, and blue- versus white-collar
workers.
Vacancy and Unemployment Rates by Industry
The section below presents vacancy-unemployment relationships for six major industries to examine whether
the outward shift that we see in the aggregate data has been equally pronounced across all sectors of the
economy. Each plot of Figures 12–17 illustrates the relationship between the unemployment rate and the
vacancy rate in a particular industry for the period spanning January 2001 to June 2012.
Seasonally adjusted data by industry for vacancy and unemployment rates were collected from the JOLTS and
the BLS, respectively and grouped into six major categories covering construction, financial activities,
trade/transportation/utilities, professional business services, leisure/hospitality, and education/health services.
Note that these plots differ from previous plots in that sector-specific, rather than the aggregate, vacancy rates
are used. Also, to be classified as unemployed in an industry by the Bureau of Labor Statistics, a worker’s last
job must have been in that industry.
The plots show a breakdown in the vacancy-unemployment relationship across all industries at the time the
aggregate Beveridge curve was moving outward.
Vacancy and Unemployment Rates by Age
Decomposing the unemployed into different age categories and plotting the aggregate vacancy rate versus the
unemployment rate of different age groups (Figures 19–21) also reveals outward shifts across all categories.
The shifts are consistent with the pattern observed using aggregate data on unemployment and vacancies and
suggest that the outward shift is common to all age groups.
Vacancy and Unemployment Rates by Education
Looking at the relationship between the aggregate vacancy rate and the unemployment rate broken down by
education rates, the plots (Figures 22–24) show a similar pattern in which outward shifts in each educational
category occur simultaneously with the aggregate Beveridge curve shifting outward.
4
Vacancy and Unemployment Rates for Blue- and White-Collar Workers
Here, we examine the vacancy-unemployment relationship for blue-collar workers (manufacturing and
production workers, laborers and helpers, construction crafts, and transportation operatives) and white-collar
workers (all others). Similar to the other breakdowns, we see a shift outward for both blue- and white-collar
workers (Figures 25–26).
It is interesting to note that the pattern in the current recession is different from the outward shift in the
Beveridge curve in the 1973-to-1985 period. At that time there was little, if any, shift in the Beveridge curve for
white-collar workers, but a large shift for blue-collar workers (Figure 27).
Conclusion
Disaggregating the vacancy-unemployment relationship reveals some interesting new facts that may shed
light on the implications of what appears to be an outward shift of the Beveridge curve in recent years.
While the Beveridge curve for all workers appears to have shifted outward starting in 2009, data on vacancy
and unemployment rates for individuals who have been unemployed for fewer than 27 weeks reveal the usual
downward-sloping relationship with no sign of any outward shift. Interestingly, a dynamic plot of the vacancy
rate versus the short-term unemployment rate shows clockwise cycling of the vacancy-unemployment points
for the unemployed. In contrast, we see a large counterclockwise movement when the vacancy rate is plotted
versus the unemployment rate for those unemployed for more than 26 weeks. Taken together, these
observations suggest that the short-term unemployed have been benefiting more than the long-term
unemployed from increases in vacancies during the recovery.
It is widely thought that the outward shift in the Beveridge curve in the 1970s reflected a worsening of
matching efficiency—that it was hard to get suitable workers and jobs together—and that this worsened the
overall unemployment rate. We observe three differences between what happened in the 1970s and today’s
changes. First, the outward shift happened much more quickly in the recent period than in the 1970s, taking
months rather than years. Second, in the 1970s the outward shift was common to all durations of
unemployment, as we would expect from a general decline in the efficiency of matching. In contrast, the
current change has taken place only among the very-long-term unemployed, with some evidence of a slight
improvement in the tradeoff for the short-term unemployed. Finally, the outward shift in the 1970s was
concentrated among blue-collar workers. Again, this is what we might expect if shifting industry composition
were creating a greater mismatch of skills (we might expect white-collar skills to be more portable between
industries). In the current period we see that the worsening tradeoff between vacancies and unemployment is
common to both blue- and white-collar workers.
Other than the contrast between the long and short-term unemployed, the break-down in the vacancyunemployment relationship in the recent period seems to have occurred across all industries, education levels,
age groups, and among both blue- and white-collar workers.
Any explanation of the change in the vacancy-unemployment relationship has to account for its pervasiveness
across industry groups, blue-white collar occupations, age and education groups, concentration among the
long-term unemployed, and absence from the short-term relationship.
5
One reason why the Beveridge curve relationship for the long-term unemployed has apparently shifted may
be a change in the desirability of the long-term unemployed to employers. It is possible that the long-term
unemployed increasingly comprise workers whose skills are not suited to available jobs. However, if this is the
case why do we not see some outward shift in the short-term relationship as well? Further, the mismatch
hypothesis is called into question by the fact that the vacancy-unemployment relationship has shifted in all
industries, while in the 1970s only the workers who were previously employed in blue-collar industries were
affected. Another possibility is that the long-term unemployed in this recession may be searching less
intensively—either because jobs are much harder to find or because of the availability of unprecedented
amounts and durations of unemployment benefits. This seems to be a more likely explanation, although if a
drop in search intensity is due only to the difficulty of finding jobs, it again raises the question why we would
not see that phenomenon at shorter durations as well.
6
Figure 1: Monthly Vacancy and Unemployment Rates, Jan 2001 to Jun
2012
Figure 2: Proportion of Unemployed Who Have Been Unemployed
for More Than 26 weeks
Percent Long-term unemployed
Vacancy rate
50
Sep-2009
.035
40
.03
30
.025
20
.02
10
Beveridge curve
.015
Sep 2009
0
Jul 2009
.04
.06
Unemployment rate
.08
1940
.1
1960
1980
Time
2000
2020
Source: CPS and JOLTS. Data are monthly rates, span the period 2001:m1–2012:m6, and are
seasonally adjusted.
Source: CPS and JOLTS. Data are monthly and seasonally adjusted.
Figure 3: Monthly Vacancy and Unemployment Rates, Using
Unemployed Persons with Duration Less Than 5 Weeks
Figure 4: Monthly Vacancy and Unemployment Rates Using
Unemployed Persons with Duration of 5–14 Weeks
Vacancy rate
Vacancy rate
.035
.035
.03
.03
.025
.025
.02
.02
.015
.015
.01
.014
.016
.018
.02
.022
.015
.02
.025
Unemployment rate
.03
.024
Unemployment rate
Source: CPS and JOLTS. Data are monthly rates, span the
period 2001:m1- 2012-m6, and are seasonally adjusted.
Source: CPS and JOLTS. Data are monthly rates, span the period 2001:m1– 2012:m6, and
are seasonally adjusted.
7
Figure 5: Monthly Vacancy and Unemployment Rates Using
Unemployed Persons with Duration of 15–26 Weeks
Figure 6: Monthly Vacancy and Unemployment Rates Using
Unemployed Persons with Duration Less Than 27 Weeks
Vacancy rate
Vacancy rate
.035
.035
.03
.03
.025
.025
.02
.02
.015
.015
.005
.01
.015
.02
.025
.03
.04
Unemployment rate
.05
Unemployment rate
.06
.07
Source: CPS and JOLTS. Data are monthly rates, span the period 2001:m1–
2012:m6, and are seasonally adjusted.
Source: CPS and JOLTS. Data are monthly rates, span the period 2001:m1–2012:06, and
are seasonally adjusted.
Figure 7: Monthly Vacancy and Unemployment Rates Using
Unemployed Persons with Duration Greater Than or Equal to 27 Weeks
Figure 8: Dynamic Plot of Vacancy & Unemployment Rates for
Short-Term and Long-Term Unemployment Groups
Vacancy rate
Vacancy rate
.02
.035
Feb-11 Apr-10
Apr-10
Jul-10
.03
.018
Feb-09
.025
Sep-10
Mar-09
.016
May-10
Feb-10
May-09
Jan-09
Feb-09
Mar-10
Mar-09
Dec-09
Sep-09
May-09
.02
Apr-09
Sep 2009
.015
.014
July 2009
0
.01
.02
.03
Jun-09 Aug-09
Jul-09
.02
.04
Aug-09
.03
Jul-09
.04 Unemployment rate .05
Short-term group
Unemployment rate
Source: CPS and JOLTS. Data are monthly rates, span the period 2001:m1–2012:m6, and are
seasonally adjusted.
Apr-09 Jun-09
.06
.07
Long-term group
Source: CPS and JOLTS. Data are monthly rates, span the period 2009:m1–2010:m12,
and are seasonally adjusted.
8
Figure 9: Evolution of the Beveridge Curve During 1960–1988
Figure 10: 1960–1988 Beveridge Curve Using Persons with
Unemployment Duration Less Than 27 Weeks
Vacancy rate
3
Vacancy rate
1969
3
1979
1978
2.5
1968
1966
1973
1969
1988
1967
1987
1986 1985
1980
1984
1974
1972
2
2.5
1968
1966
1967
1977
1965 1970
1979
1978
1973
1988
1987
1981
1976
1971
1965
2
19721974
1970
1977
1986 1980
1984
1981
1964
1.5
1983
1975
1962
1963
1964
1.5
1961
1962
1963
1960
1
4
6
Unemployment rate
8
1976
1971
1982
1960
10
1983
1975
1961
1982
1
.03
Source: CPS, Conference Board, and authors’ calculations. Data are annual averages and
span the period 1960–1988.
.04
.05
.06
.07
.08
Unemployment rate
Source: CPS, Conference, and authors’ calculations. Data are annual rates and
span the period 1960-1988.
Figure 11: 1960–1988 Beveridge Curve Using Persons with
Unemployment Duration Greater Than or Equal to 27 Weeks
Figure 12: Vacancy Rate versus Unemployment Rate in the
Construction Industry
Vacancy rate
Vacancy rate
2.5
3
1969
2.5
1979
1978
2006
2007
2005
1968
1973
1966
2
1988 1987
1967
19721980
1974
1985
1986
1970 1965
2
2004
1984
1981
2002
1.5
1976
1971
2012
2003 2008
2011
1964
1.5
1963
1960
2010
1983
1975
1962
1982
1961
1
2009
1
0
.005
.01
.015
.02
5
.025
Unemployment rate
Source: CPS, Conference Board, and authors’ calculations. Data are annual rates and span the
period 1960-1988.
9
10
15
Unemployment rate
20
Source: CPS and JOLTS. Vacancy and unemployment rates are annual averages that are
specific to the construction industry. They span the period Jan 2002- July 2012.
Figure 13: Vacancy Rate versus Unemployment Rate in the Financial
Services Industry
Figure 14: Vacancy rate versus Unemployment Rate in the Trade,
Transportation & Utilities industry
Vacancy rate
Vacancy rate
4
3
2006
3.5
2007
2006
2005
2005
2.5
2007
2004
2004
3
2012
2012
2008
2002
2002
2010
2008
2.5
2
2011
2003
2011
2003
2010
2
1.5
2009
3
4
5
Unemployment rate
6
2009
8
7
10
12
14
Unemployment rate
16
18
Source: CPS and JOLTS. Vacancy and unemployment rates are annual averages that are
specific to the trade, transportation & utilities industry. They span the period Jan 2002–July
2012.
Source: CPS and JOLTS. Vacancy and unemployment rates are annual averages that are
specific to the financial services industry. They span the period Jan 2002–July 2012.
Figure 15: Vacancy Rate versus Unemployment Rate in the Professional
& Business Services Industry
Figure 16: Vacancy Rate versus Unemployment Rate in the Leisure &
Hospitality Industry
Vacancy
Vacancy rate
4.5
4.5
2005
2007
2006
4
2007
2006
4
2005
2003
2004
3.5
2012
2008
3.5
2004
2002
2008
2002
2003
2011
3
2010
3
2012
2011
2.5
2010
2
2.5
2009
2009
4
6
8
Unemployment rate
10
7
12
8
9
10
Unemployment rate
11
12
Source: CPS and JOLTS. Vacancy and unemployment rates are annual averages that are
specific to the Leisure & Hospitality industry. They span the period Jan 2002– July 2012.
Source: CPS and JOLTS. Vacancy and unemployment rates are annual averages that are
specific to the professional & business services industry. They span the period Jan 2002–July
2012.
10
Figure 17: Vacancy Rate versus Unemployment Rate in the Education &
Health Services Industry
Figure 18: Monthly Vacancy and Unemployment Rates Plot Using
Age Group 20–24 Years
Vacancy rate
Vacancy rate
3.5
2002
4
3
2007
2006
2.5
2005
2008
3.5
2004 2003
2
2012
1.5
3
2011
4
2009
3
4
5
8
Unemployment rate
10
12
Source: CPS and JOLTS. Data are monthly rates, span the period 2001:m1–2012:06,
and are seasonally adjusted.
2010
2.5
6
6
Unemployment rate
Source: CPS and JOLTS. Vacancy and unemployment rates are annual averages that are
specific to the Education & Health Services industry. They span the period Jan 2002- July 2012.
Figure 19: Monthly Vacancy and Unemployment Rates Plot Using Age
Group 25–34 Years
Figure 20: Monthly Vacancy and Unemployment Rates Plot Using
Age Group 25–54 Years
Vacancy rate
Vacancy rate
.035
3.5
.03
3
.025
2.5
.02
2
.015
1.5
2
2
4
6
Unemployment rate
8
4
6
8
10
Unemployment rate
10
Source: CPS and JOLTS. Data are monthly rates, span the period 2001:m1–2012:m6,
and are seasonally adjusted.
Source: CPS and JOLTS. Data are monthly rates, span the period 2001:m1-2012:m6,
and are seasonally adjusted.
11
Figure 21: Monthly Vacancy and Unemployment Rates Plot Using Age
Group 55 +
Figure 22: Monthly Vacancy and Unemployment Rates Plot for
Unemployed Persons with Less than a High School Degree
Vacancy rate
Vacancy rate
3.5
3.5
3
3
2.5
2.5
2
2
1.5
1.5
3
4
5
Unemployment rate
6
6
7
8
10
12
Unemployment rate
14
16
Source: CPS and JOLTS. Data are seasonally adjusted and span the period 2001:m1–
2012:m6. Note: JOLTS does not report data on job openings by education.
Source: CPS and JOLTS. Data are monthly rates, span the period 2001:m12012:m6, and are seasonally adjusted.
Figure 23: Monthly Vacancy and Unemployment Rates Plot for
Unemployed Persons with a High School Degree
Figure 24: Monthly Vacancy and Unemployment Rates Plot for
Unemployed Persons with a Bachelor’s Degree
Vacancy
Vacancy rate
3.5
3.5
3
3
2.5
2.5
2
2
1.5
1.5
4
6
8
Unemployment rate
10
12
1
Source: CPS and JOLTS. Data are seasonally adjusted and span the period 2001:m1–
2012:m6. Note: JOLTS does not report data on job openings by education.
2
3
Unemployment rate
4
5
Source: CPS and JOLTS. Data are seasonally adjusted and span the period 2001:m1–
2012:m6. Note: JOLTS does not report data on job openings by education.
12
.03
Vacancy rate
.025
.02
.015
.015
.02
Vacancy rate
.025
.03
.035
Figure 26: Monthly Vacancy and Unemployment Rates Plot Using
Proportion of White-Collar Unemployed Workers
.035
Figure 25: Monthly Vacancy and Unemployment Rates Plot Using
Proportion of Blue-Collar Unemployed Workers
.06
.08
.1
.12
blue collar unemployment rate
.14
.16
.03
Figure 27: Annual Vacancy and Unemployment Rates for Blue- and
White-Collar Workers During 1973–1985
Vacancy rate
3
2.5
1979
1973
1973
1980
1974
1985
1985
1974
1977
1977
1980
1984
2
1981
1981
1976
1976
1.5
1983
1975
1983
1975
1982
0
5
10
White Collar
Blue Collar
.07
Source: Data on unemployment rate by job category (for white collar occupations) are
obtained from the public data series of the BLS’s Current Population Survey. Aggregate
vacancy rate data are from the BLS’s Jobs Opening and Labor Turnover Survey.
Data are seasonally adjusted and span the period 2001:m1–2012:m9.
Source: Data on unemployment rate by job category (for blue collar occupations) are
obtained from the public data series of the BLS’s Current Population Survey. Aggregate
vacancy rate data are from the BLS’s Jobs Opening and Labor Turnover Survey.
Data are seasonally adjusted and span the period 2001:m1–2012:m9.
1979
.04
.05
.06
White collar unemployment rate
1982
15
Source: Data on unemployment rate by job category (for blue and white collar occupations)
are obtained from the public data series of the BLS’s Current Population Survey.
Aggregate vacancy rate data are from the BLS’s Jobs Opening and Labor Turnover
Survey. Note: We use annual rate to smooth high frequency variation in the data. Data are
seasonally adjusted and span the period 1973:m1–1985:m12.
13