The Dynamics of Involuntary Part-Time Employment During the Great Recession: The Ins Win

Ashesh Rambachan

Ashesh Rambachan1
1Department of Economics, Princeton University, Princeton, NJ, USA


ABSTRACT

Using micro-data from the Current Population Survey, I analyze the demographics of the involuntary part-time employed and variation in the involuntary part-time employment rate over the business cycle. During the Great Recession, the involuntary part-time employment rate more than doubled from under 2% to over 4%, an increase that dwarfs any changes in the involuntary part-time employment rate over the previous 20 years. To understand the causes of the increase in involuntary part-time employment during the Great Recession, I model the labor force as a discrete-time Markov chain and present a novel method that decomposes changes in the involuntary part-time employment rate into changes in the underlying transition probabilities of individuals across labor force states. This method allows me to construct counter-factual involuntary part-time employment rates and measure the relative contribution of labor market flows to observed changes in the involuntary part-time employment rate. I find that two-thirds of the increase in the involuntary part-time employment rate during the Great Recession is due to workers entering involuntary part-time employment from other labor force states.


INTRODUCTION

The Financial Crisis of 2007–2008 and the Great Recession from December 2007 to June 2009 led to the worst labor market conditions in the United States since the Great Depression. The official U-3 unemployment rate peaked at 10% in October 2009. Since then, it has steadily declined and is down to 4.6% as of November 2016.1 However, the U-3 unemployment rate may overstate the health of the labor market because of its strict definition of unemployment. It classifies an individual as unemployed if and only if she did not have a job in the month surveyed and actively searched for a job at some point during the previous four weeks. For this reason, policymakers increasingly turned to alternative measures of labor market slack after the Great Recession. For instance, the U-6 unemployment rate includes individuals that are working part-time for economic reasons and marginally attached workers. U-6’s more general definition of unemployment paints a different picture of the labor market than the U-3 unemployment rate. The U-6 unemployment rate also peaked in October 2009 at 17.1% and has only fallen to 9.3%, well above its pre-recession level of roughly 8%.1

Of particular interest is the increase in the number of individuals working part-time for economic reasons. An employed worker is counted as working part-time for economic reasons if they are working less than 35 hours a week, reported that they would like to work full-time, but say they cannot find full-time work because of economic conditions. These workers are also referred to as involuntary part-time employed workers. The involuntary part-time employment rate, defined as the number of involuntary part-time employed workers divided by the size of the labor force, more than doubled during the Great Recession from under 2% to over 4%.2

In this paper, I use micro-data from the Current Population Survey to analyze the involuntary part-time employed and their labor market dynamics over the last 20 years. I begin by describing trends in the involuntary part-time employment rate for the population and specific demographic groups such as males, females, and several age groups. While there are substantial differences in the levels, the involuntary part-time employment rate roughly doubled during the Great Recession for each demographic group. Moreover, unlike other measures of labor market slack, the involuntary part-time employment rate remains stuck above its pre-recession level at 1.8% in November 2007.2 I also provide basic demographic facts about the involuntary part-time employed and find that, when compared to the full-time employed, females and racial minorities make up a far larger proportion of the involuntary part-time employed.

Finally, I present a decomposition of the involuntary part-time employment rate that breaks down changes in this rate into changes in the underlying flows of individuals across labor market states. This decomposition allows me to assess the relative contribution of underlying labor market flows to the steady state involuntary part-time employment rate. Using data on individual-level transitions from the Current Population Survey, I use my decomposition to answer important questions about the dynamics of involuntary part-time employment during the Great Recession with a focus on determining the source of the increase in involuntary part-time employment during the Great Recession. Did involuntary part-time employment increase because of flows of individuals from other labor force states into involuntary part-time employment, which I call the “Ins?” Or, did involuntary part-time employment increase because individuals remained stuck in involuntary part-time employment, which I analogously call the “Outs?” This approach of using flows to analyze changes in labor market aggregates follows literature that decomposes changes in the unemployment rate throughout the business cycle into changes in underlying labor force flows between employment and unemployment.3,4

The involuntary part-time employment rate increased from 1.79% in December 2007 to a peak of 4.43% during the Great Recession. I estimate that labor force flows from individuals out of full-time employment account for 35% of the increase in the involuntary part-time employment rate during the Great Recession. Flows from individuals out of involuntary part-time employment account for 23.1% of the increase and flows from individuals out of voluntary part-time employment account for 20% of the increase. Flows out of unemployment and out of the labor force each account for roughly 6% of the increase. Therefore, the “Ins,” flows of individuals into involuntary part-time employment from other labor force states, account for roughly two-thirds of the increase in the involuntary part-time employment rate during the Great Recession with most of the “Ins” coming from full-time employment. The decomposition is repeated for males, females, and three different age groups. I find that, in each case, the “Ins” account for the majority of the increase in the involuntary part-time employment rate during the Great Recession.

The results are significant in shaping understanding of how employers respond to demand shocks during recessions and suggest that the increase in involuntary part-time employment may have been a constructive labor market response to the Great Recession. Rather than laying off more workers, employers instead cut the hours of full-time workers. Moreover, involuntary part-time employed workers are more likely to transition back to full-time employment and less likely to leave the labor force than unemployed workers. Therefore, the decision to cut hours rather than lay off workers means that many workers remained more attached to the labor market.

This paper contributes to the existing literature on involuntary part-time employment. First, unlike other papers that consider how underlying labor market flows contribute to changes in the involuntary part-time employment rate,5 this appears to be the first paper to consider how the steady state involuntary part-time employment rate is affected by labor market flows. Second, unlike other work analyzing the causes of the increase in involuntary part-time employment during the Great Recession,5,6 this paper also appears to be the first to explicitly consider differences across demographic groups in the dynamics of the involuntary part-time employment rate. The findings of this paper help continue to develop policymakers’ understanding of the involuntary part-time employed during and after the Great Recession.


DATA

I use data from the Integrated Public Use Microdata Series (IPUMS), Current Population Survey (CPS) over September 1995 to June 2014 from the Minnesota Population Center.7 The CPS is a monthly survey of 60,000 households conducted by the United States Census Bureau. It has a rotating panel structure; households are surveyed for four consecutive months, removed from the survey for the next eight months and then returned to the survey for four more months. For any given household, I observe at most eight surveys that cover a period of 16 months. I remove all respondents whose age is reported as below 18 and over 65.

In the CPS, I am able to identify individuals that are involuntary part-time employed. First, the IPUMS CPS data contains a variable that identifies whether a respondent is employed (E), unemployed (U) or out of the labor force (N). I drop all respondents that report being absent from work. Given that a respondent is E, I also observe the total number of hours she usually works each week. I drop all workers that reported their usual weekly hours vary. A respondent is labeled as part-time (PT) if the total number of hours she usually works each week is less than 35 hours. Otherwise, I label her as full-time (FT). Finally, given that a respondent is in PT, I observe her stated reason for working part-time. A respondent is labeled as involuntary part-time employed (IPT) if her stated reason for working part-time is “slack work/business conditions,” “could only find part-time work,” or “seasonal work.” Otherwise, she is labeled as voluntary part-time employed (VPT). Using this classification scheme, each respondent falls into one of five possible labor force states: FT, IPT, VPT, U, or N.

I use the CPS in two ways. First, I use it as a repeated cross-section of a sample from the U.S. population, which allows me to generate demographic information about employed workers and compute the involuntary part-time employment rate. Table 1 provides a full-set of summary statistics about the sample. Most of the respondents are prime-age individuals between the ages 25–54. Moreover, just over 50% of the sample is female and 25% composed of racial minorities. The involuntary part-time employment rate is defined as the number of workers in IPT over the number of workers in E and U. At time t, the involuntary part-time employment rate is

EQUATION 1

With this definition and the classification system described above, it is simple to compute the involuntary part-time employment rate at each month.

Next, I use the CPS as an unbalanced panel of individuals. For individuals who respond to the survey in consecutive months, I am able to track their labor force state over time and thereby observe possible labor force transitions. Since 1994, the CPS has included identification variables that allow researchers to match respondents across surveys. To construct month-to-month matches, I use a matching identifier created by IPUMS. The identifier was constructed using a matching algorithm that uses the provided identification variables from the CPS.8 There are two challenges in matching respondents across surveys. First, because of the rotating panel structure, at most 75% of those surveyed in any given month can be matched to survey responses in the next month. Second, matches based solely on identification variables may lead to incorrect matches due to migration, deaths, non-responses and coding errors.9 To adjust for these possible errors, I validate matches using the race and gender characteristics of the respondents. I am able to match, on average, roughly 71% of the sample across months, which translates into roughly 95% of all possible matches.

With the matched survey responses, I compute the average labor force transition probabilities over the period of the sample. Let ltA,B denote the average probability of transitioning from labor force state A at time t to state B at time t + 1. To estimate this probability, I compute the fraction of matched respondents at time t that are in state A that transition to state B at time t + 1. I compute each transition probability for the set of labor force states {FT, VPT, IPT, U, N} and so, I estimate 25 transition probabilities in all.

Finally, because I am computing time series of the involuntary part-time employment rate and transition probabilities between labor force states, there is significant seasonality in my estimates. To adjust for any seasonal trends, I use the following procedure. Let Yt be some monthly time series. I regress Yt on a series of month indicator variables

EQUATION 2

where, for instance, Jant is a dummy variable for whether the month at time t is January. The residuals of this regression are the variation in Yt that cannot be explained by monthly seasonal effects. I add the average value of Yt  over the full sample period to the residuals, producing the seasonally adjusted series. Each time series is seasonally adjusted in this fashion.


FACTS ABOUT THE INVOLUNTARY PART-TIME EMPLOYED

Table 2: Demographic Statistics for employment groups (mean only)

Table 2 provides a series of demographic facts about the three different groups of employed workers: the full-time employed, the voluntary part-time employed and the involuntary part-time employed. There are substantial demographic differences between these groups of employed workers. Both types of part-time employed workers are more likely to be between the ages of 18–24 than full-time workers. However, compared to voluntary part-time employed workers, involuntary part-time employed workers are more likely to be prime-age workers between the ages of 25–54. Next, there are substantial racial differences. An involuntary part-time employed worker is 4 percentage points more likely to be an African American than a full-time employed worker and nearly 7% more likely than a voluntary part-time employed worker. Finally, both voluntary and involuntary part-time employed workers are more likely to be females than full-time workers.

Figure 1: Involuntary part-time employment rate: 1995–2014

Figure 1 plots the three-month moving average of the involuntary part-time employment rate from 1994 to 2014 for the full population. The effect of the financial crisis and the Great Recession are immediately apparent. From 2007 to 2010, the involuntary part-time employment rate doubled from 1.79% to a high of 4.3% in 2010. Moreover, there is substantial cyclical variation in the involuntary part-time employment rate even before the Great Recession. Beginning in the mid-1990s, the involuntary part-time employment began a sustained downward trend. It fell by 1.1 percentage points in 5 years from 2.4% in mid-1995 to 1.3% in mid-2000. This coincides with robust economic growth at the end of the 1990s. During the 2001 recession, the involuntary part-time employment rate increased by 0.8 percentage points and peaked at 2.1% in October 2003. It again declined before the Great Recession.

Figure 2: Involuntary part-time employment rate by demographic group

There are substantial differences in the involuntary part-time employment rate across demographic groups. Over the business cycle, females tend to have a much higher involuntary part-time employment rate than males (Fig. 2a). Before 2008, the involuntary part-time employment rate for females was, on average, 0.8 percentage points higher than that of males. After the Great Recession, this difference increased. After 2010, the involuntary part-time employment rate for females was, on average, 1.3 percentage points higher than that of males. There are also striking differences in the involuntary part-time employment rate for different marital groups (Fig. 2b). The involuntary part-time employment rate for single workers is higher than the involuntary part-time employment rate for the full sample over the entire sample period. Notice that for both married males and females, the involuntary part-time employment rate is consistently lower than that of the full population (Fig. 2c). Moreover, the gender gap is larger between married males and females than single males and females.

Finally, there are differences in the involuntary part-time employment rate across age groups. I consider three age groups: young workers age 18–24, prime-age workers age 25–54 and old workers age 55–65 (Fig. 2d). Young workers are substantially more likely to be involuntary part-time employed than the full population, prime-age workers, and old workers. Prior to 2008, the involuntary part-time employment rate for young workers was, on average, 3.97%, which approaches that of the overall population during the Great Recession.

In summary, the involuntary part-time employment rate, aside from differences in the average level, has followed remarkably similar patterns to those of 1995 to 2014 across each demographic group. It is strongly cyclical and for each demographic group the increase in the involuntary part-time employment rate during the Great Recession dwarfs any prior fluctuation since the mid-1990s.


A FLOW BASED DECOMPOSITION OF THE INVOLUNTARY PART-TIME EMPLOYMENT RATE

The population can be divided into five groups: the full-time employed (FT), the involuntary part-time employed (IPT), the voluntary part-time employed (VPT), the unemployed (U) and those that are out of the labor force (N). At each month t of the CPS, using the procedure outlined above, I observe a 5 x 5 matrix of transition probabilities across these five states. This transition probability matrix governs the flows of individuals between each of the five labor force states between month t and t + 1. Let Πt denote the transition probability matrix at time t with ith, jth entry, πti,j, denoting the probability of transitioning from state i at time t to state j at time t + 1.

 

Transition Probabilities into and out of Involuntary Part-time Employment

Table 3: Average transition probabilities for the full sample from 1995–2014

Table 3 shows the average transition probabilities in the CPS over the sample period. During this time, an IPT worker is most likely to remain IPT in the next month. An IPT worker has a 27% chance of transitioning into FT in the next month. In comparison, a U worker has a 16.5% chance and a VPT worker has a 15% chance of transitioning to FT in the next month. In addition, an IPT worker has, on average, a 5.6% chance of exiting the labor force. U workers, on average, have a 22% chance of exiting the labor force. This suggests that IPT workers have a greater attachment to the labor force than U workers. Moreover, VPT and U workers are much more likely to transition into IPT than FT workers.

Figure 3: Transition probabilities into and out of involuntary part-time employment

I next consider the time series behavior of the transition probabilities into IPT from FT, VPT, U and N (Fig. 3a). Notice that the N-IPT transition probability remained relatively unchanged from 1995 to 2014. This suggests that changes in the involuntary part-time employment rate over this period are not caused by flows of individuals from N to IPT. Moreover, prior to the Great Recession, transition probabilities into IPT do not appear to be strongly cyclical. For example, the effects of the 2001 recession are not easily discernible. In sharp contrast, the Great Recession had an enormous effect on the FT-IPT and VPT-IPT transition probabilities. The FT-IPT transition probability stood at 0.6% in December 2007 and by the fall of 2009, had doubled to over 1.3%. In December 2007, the VPT-IPT transition probability stood at 3% and by the fall of 2009, had nearly doubled to well over 5%. This suggests that a substantial portion of the increase in the involuntary part-time employment rate during the Great Recession was due to flows of individuals into IPT from FT and VPT.

Figure 3b plots the three-month moving average of the transition probabilities out of IPT. The IPT-U and IPT-N transition probabilities remained relatively unchanged from 1995 to 2014. The IPT-IPT transition probability appears to be highly cyclical, falling during recoveries and rising during recessions. For example, during the late 1990s, the IPT-IPT transition probability fell from 40% in October 1995 to 28% in June 2001. It subsequently rose to 37% after the recession in the early 2000s. Finally, there is a sharp increase in the IPT-IPT transition probability during the Great Recession. In December 2007, it stood at 35% and peaked at 49% in 2010. During the Great Recession, the IPT-FT and IPT-VPT transition probabilities declined from 29% and 19% to lows of 23% and 15% respectively.

Taken together, the substantial movements in the IPT-IPT, FT-IPT and VPT-IPT transition probabilities suggest that much of the increase in the involuntary part-time employment rate during the Great Recession is due to flows into IPT from FT and VPT, the “Ins,” and individuals remaining stuck in IPT, the “Outs.” I next assess the relative contribution of the “Ins” and “Outs” to the observed movement in the involuntary part-time employment rate during the Great Recession.

 

Empirical Model of the Flows-Based Decomposition

As mentioned, at each month t in the CPS, I observe a 5 x 5 matrix of transition probabilities, Πt, whose entries are the probabilities of transitioning between FT, IPT, VPT, U and N. Πt is a stochastic matrix, meaning that each entry is a real number between zero and one and each row of Πt sums to one. As a result, Πt defines a discrete-time Markov chain with state space D = {FT, IPT, VPT, U, N}. For simplicity, I assume the Markov chain defined by Πt is time-homogeneous. Moreover, it is stationary because it has one recurrent class and is aperiodic (there is a non-zero probability of transitioning to each state j from any state i). This implies that there exists a steady state vector Lt* associated with Πt satisfying

EQUATION 3

Lt* is an eigenvector of Πt with eigenvalue equal to one. The entries of Lt* can be interpreted as the steady state distribution of individuals across the labor force states FT, VPT, IPT, U and N implied by the transition probabilities, Πt. The steady state involuntary part-time employment rate at time t associated with the transition probabilities, Πt, is simply the steady state number of individuals in IPT divided by the steady state number of individuals in the labor force. That is,

EQUATION 4

to construct the time series of {IPTRt*}, I compute the eigenvector associated with the largest eigenvalue of Πt for each time period t and use the previous equation above to compute steady state involuntary part-time employment rate.

Figure 4: Observed vs. steady-state involuntary part-time employment rate implied by labor force transition probabilities

The steady state involuntary part-time employment rate is highly correlated with the observed involuntary part-time employment rate with r = 0.983 (Fig. 4). Before the Great Recession, the steady state involuntary part-time employment rate tended to slightly underestimate the observed rate. Most importantly, the steady state closely tracked the observed rate during the Great Recession and the recovery. As a result, understanding changes in the steady state involuntary part-time employment rate is a useful exercise and can be used to understand what caused changes in the involuntary part-time employment rate during the Great Recession.

To compare the effects of the “Ins” and “Outs,” I construct a series of counter-factual involuntary part-time employment rates. Consider some state i. I fix all transition probabilities out of each state ji at their observed value in December 2007. Only the transition probabilities out of state i are allowed to vary as observed in the data. This generates a series of transition probability matrices, , that define a stationary, discrete-time Markov chain. Using the procedure outlined above, I compute the steady state vector  associated with  and compute the associated steady state involuntary part-time employment rate, . This shows how the steady state involuntary part-time employment rate would have evolved had only the transition probabilities out of state i varied as observed during the Great Recession. I perform this counter-factual exercise for each state: FT, IPT, VPT, U, and N.

The difference between the observed and the counter-factual involuntary part-time employment rate is interpreted as the change in the involuntary part-time employment rate that cannot be accounted for due to the transition probabilities out of state i. That is, the counter-factual involuntary part-time employment rate is interpreted to be the contribution of the transition probabilities out of state i to the observed involuntary part-time employment rate. Therefore, the counter-factual exercises estimate the contribution of the transition probabilities out of FT, VPT, IPT, U and N to the involuntary part-time employment rate. I define the contribution of transition probabilities out of FT, VPT, U, and N to be the effect of the “Ins” and the contribution of the transition probabilities out of IPT to be the effect of the “Outs.”

 

Results

The results of the counter-factual exercises for states FT, IPT, VPT, and U are presented in Figure 5. The result of the exercise for N is excluded because it is similar to that for U. Each graph plots the observed involuntary part-time employment rate and the counter-factual involuntary part-time employment rate from December 2007 to June 2014. The graph labeled “Full-time” depicts the full-time counter-factual involuntary part-time employment rate.

Figure 5: Full-Sample counter-factual exercises: Observed in blue, counter-factual in red

During the Great Recession, the observed involuntary part-time employment rate increased by 2.64 percentage points. I find that 35.4% of that increase can be accounted for by the transition probabilities (i.e. the flow of individuals) out of full-time employment. This means that had all other transition probabilities (i.e. labor market flows) been held constant, except for those out of full-time employment, the involuntary part-time employment rate would have increased by 0.91 percentage points. For the other states, I find:

  • that the transition probabilities out of voluntary part-time employment account for 19.7% of the observed increase. Alone, transition probabilities out of VPT would have led to a 0.53 percentage point increase in the involuntary part-time employment rate.
  • that the transition probabilities out of involuntary part-time employment account for 23.1% of the observed increase. Alone, transition probabilities out of IPT would have led to a 0.61% increase in the involuntary part-time employment rate.
  • that the transition probabilities out of unemployment account for 6.1% of the increase. Alone, transition probabilities out of U would have led to a 0.16 percentage point increase in the involuntary part-time employment rate.
  • that the transition probabilities out of out of the labor force account for 5.7% of the increase. Alone, transition probabilities out of N would have led to a 0.15% increase in the involuntary part-time employment rate.

These results indicate that the transitions from full-time employment and voluntary part-time employment into involuntary part-time employment were the most significant determinants of the observed increase in the involuntary part-time employment rate during the Great Recession. The “Ins” account for nearly 66% of the increase in the involuntary part-time employment rate, whereas the “Outs” only account for 23.1% of the increase. Therefore, most of the increase in the involuntary part-time employment rate during the Great Recession was due to new entrants into involuntary part-time employment with the majority of these entrants coming from full-time and voluntary part-time employment.


DEMOGRAPHIC DIFFERENCES IN THE DYNAMICS OF INVOLUNTARY PART-TIME EMPLOYMENT

The analysis so far has assumed that there are no significant differences across demographic groups. However, there are key differences in the involuntary part-time employment rate between genders and age groups. I now repeat the counter-factual exercises for each subgroup to understand the demographic differences in the dynamics of the involuntary part-time employment rate during the Great Recession.

 

Gender Differences

The patterns in the transition probabilities into IPT during the Great Recession are strikingly similar for males and females (Fig. 6a & 6b). For both groups, there is a large increase in the VPT-IPT and FT-IPT transition probabilities. In addition, for both groups, there is a substantial increase in the IPT-IPT transition probability (Fig. 7a and Fig. 7b). However, there are noticeable differences in the transition probabilities out of IPT into FT and U between males and females. For example, there is an observable increase in the IPT-U transition probability for males that is not visible for females and the IPT-FT transition probability declined by a larger magnitude for males than females.

Figure 6: Transition probabilities into involuntary part-time employment by gender
Figure 7: Transition probabilities out of involuntary part-time employment by gender

Next, I estimate the contribution of transition probabilities out of FT, IPT, VPT, U and N for males and females. For males, the involuntary part-time employment rate increased by 2.39 percentage points during the Great Recession. For females, it increased by 2.97 percentage points. I find that:

  • the transition probabilities out of full-time employment account for 42.3% of the observed increase for males and only 28.3% of the observed increase for females. Transition probabilities out of FT would have led to a 1.01 percentage point and a 0.84 percentage point increase in the involuntary part-time employment rate for males and females respectively.
  • the transition probabilities out of voluntary part-time employment account for 13.8% and 26.9% of the observed increase for males and females. Transition probabilities out of VPT would have led to a 0.3 percentage point and a 0.8 percentage point increase in the involuntary part-time employment rate for males and females respectively.
  • the transition probabilities out of involuntary part-time employment account for 22.2% and 24.6% of the observed increased for males and females. Transition probabilities out of IPT would have led to a 0.53 percentage point and 0.73 percentage point increase in the involuntary part-time employment rate for males and females respectively.
  • the transition probabilities out of unemployment account for 8% and 5.7% of the observed increase for males and females. Transition probabilities out of U would have led to a 0.19 percentage point and 0.17 percentage point increase in the involuntary part-time employment rate for males and females respectively.
  • the transition probabilities out of out of the labor force account for 6.3% and 6.4% of observed increase for males and females. Transition probabilities out of N would have lead to 0.15 percentage point and a 0.19 percentage point increase in the involuntary part-time employment rate for males and females respectively.

Notice that there is a substantial difference in the contribution of transition probabilities out of FT for males and females. Transitions out of FT had a much larger effect on the involuntary part-time employment rate for males than females. This is likely because, prior to the Great Recession, males were more likely to be full-time employed than females. As a result, small changes in the FT-IPT transition probability led to larger changes in the involuntary part-time employment rate for males than females. Transitions out of part-time employment (VPT and IPT) have a larger effect on the involuntary part-time employment rate for females than males for analogous reasons. For males and females, the “Ins” into involuntary part-time employment account for most of the increase in the involuntary part-time employment rate during the Great Recession (70.4% and 67.3% respectively).

 

Age Differences

Figure 8: Transition probabilities into involuntary part-time employment by age group
Figure 9: Transition probabilities out of involuntary parttime employment by age group

I consider differences in the dynamics of the involuntary part-time employment rate between young workers ages 18–24, prime age workers age 25–54 and old workers age 55–65. Across these groups, the trends in the transition probabilities into IPT are roughly similar (Fig. 8a, 8b, & 8c). Most importantly, for each age group the VPT-IPT and FT-IPT transition probabilities increased substantially during the Great Recession. Similarly, the trends in the transition probabilities out of IPT for the tree age groups are broadly the same (Fig. 9a, 9b & 9c). During the Great Recession, the IPT-IPT transition probability increased significantly and the IPT-FT and IPT-VPT transition probabilities declined slightly.

During the Great Recession, the involuntary part-time employment rate increased by 6.21 percentage points, 2.44 percentage points and 2 percentage points for young, prime-age, and old workers respectively. I find that:

  • the transition probabilities out of full-time employment account for 23.5%, 40.1% and 41.5% of the observed increase for young, prime-age and old workers. Transition probabilities out of FT alone would have led to a 1.46 percentage point, 0.98 percentage point and 0.83 percentage point increase in the involuntary part-time employment rate for young, prime-age and old workers.
  • the transition probabilities out of voluntary part-time employment for 23.5%, 18.4% and 31% of the observed increase for young, prime-age and old workers. Transition probabilities out of VPT alone would have led to a 1.46 percentage point, 0.45 percentage point and 0.62 percentage point increase in the involuntary part-time employment rate for young, prime-age and old workers.
  • the transition probabilities out of involuntary part-time employment account for 29.0%, 19.3%, 32% of the observed increase for young, prime-age and old workers. Transition probabilities out of IPT would have led to a 1.80 percentage point, 0.47 percentage point and 0.64 percentage point increase in the involuntary part-time employment rate for young, prime-age and old workers.
  • the transition probabilities out of unemployment account for 5.5%, 6.9%, and 11.5% of the observed increase for young, prime-age and old workers. Transition probabilities out of U would have led to a 0.34 percentage point, 0.17 percentage point and a 0.23 percentage point increase in the involuntary part-time employment rate for young, prime-age and old workers.
  • the transition probabilities out of out of the labor force account for 9.3%, 4.6%, and 10.5% of the observed increase for young, prime-age and old workers. Transition probabilities out of N would have led to a 0.58 percentage point, 0.11 percentage point, and a 0.21 percentage point increase in the involuntary part-time employment rate for young, prime-age and old workers.

Again, I find that for all age groups, the “Ins” into involuntary part-time employment account for most of the observed increase during the Great Recession.


DISCUSSION

I found that roughly two-thirds of the increase in the involuntary part-time employment rate during the Great Recession can be explained by the “Ins,” flows of individuals into involuntary part-time employment. Moreover, most of the “Ins” originated from full-time employment, which may be due to full-time workers having their hours cut rather than being laid off.

This finding suggests that, while policymakers have increasingly focused on the involuntary part-time employed, the increase in the involuntary part-time employment rate may have been a constructive response by employers to the Great Recession. It is possible that many of the full-time employed workers that transitioned to involuntary part-time employment would have otherwise transitioned to unemployment if their employers were unable to reduce their hours worked. A simple, back-of-the-envelope calculation can quantify the size of this effect. During the Great Recession, the FT-IPT transition probability increased by 0.7 percentage points from 0.6% to 1.3%. In December 2007, there were 121,609,000 full-time workers in the United States.10 Therefore, 851,263 workers transitioned during the recession into involuntary part-time employment from full-time employment. For simplicity, suppose each full-time worker worked 40 hours each week and each involuntary part-time employed worker worked 20 hours each week. That is, two full-time workers must have transitioned to involuntary part-time employment to save one full-time job. This translates into roughly 425,361 workers that would have otherwise been laid off. Given that in December 2007, there were only 7,700,000 unemployed workers in the United States,11 it translates into an increase in the unemployment rate by over a quarter of a percentage point.

This effect is similar to what was observed in the German labor market during the Great Recession. During the Great Recession, Germany experienced an even larger real GDP decline than the United States but unemployment did not increase substantially.12 The structure of most labor contracts in Germany provides strong incentives for employers to avoid laying off workers, leading German employers to cut hours. Consequently, workers transitioned to involuntary part-time employment.12 These results suggest that a similar effect may have occurred in the U.S. labor market but on a much smaller scale.

In addition to reducing the unemployment rate, transitions from full-time employment to involuntary part-time employment kept more workers more closely attached to full-time work. An involuntary part-time employed worker is nearly twice as likely to return to full-time work and one-fourth as likely to exit the labor force as an unemployed worker. In 2014, the United States economy suffered a 7.7% loss in potential output due to the Great Recession.13 This may have been worse without the large increases in involuntary part-time employment; recessions have long-term effects on growth and output through unemployment.14,15 As individuals become unemployed and fail to transition back to employment, they become more likely to leave the labor force. Because many full-time workers transitioned to involuntary part-time employment rather than unemployment, this effect may have been dampened.

To summarize, this paper provides a flows-based decomposition of the involuntary part-time employment rate during the Great Recession. It finds that much of the increase in the involuntary part-time employment rate is due to the “Ins,” transitions of workers into involuntary part-time employment from other labor force states.

 

Received: January 4, 2017
Revised: February 18, 2017
Accepted: March 22, 2017
Published: April 10, 2017


REFERENCES

  1. Bureau of Labor Statistics. The Employment Situation—November 2016. (2016).
  2. Based on author’s calculations from the Current Population Survey.
  3. Darby, M., Haltiwanger, J., & Plant, M. The Ins and Outs of Unemployment: The Ins Win. National Bureau of Economic Research Working Paper (1986).
  4. Shimer, R. Reassessing the Ins and Outs of Unemployment. Journal of Economic Dynamics 15, 127–148 (2012).
  5. Canon, M., Kudylak, M., Luo, G. & M. Reed. Flows to and from Working Part-time for Economic Reasons and the Labor Market Aggregates during and after the 2007–09 Recession. Economic Quarterly 100, 87–111 (2014).
  6. Valleta, R., Bengali, L. & van der List, C. Cyclical and Market Determinants of Involuntary Part-time Employment. Federal Reserve Bank of San Francisco Working Paper (2015).
  7. King, M., Ruggles, S., Alexander, J., Flood, S., Genadek, K., Schroeder M., Trampe, B. & Vick, R. Integrated Public Use Microdata Series, Current Population Survey, Version 3.0.
  8. Drew, J., Flood, S., & Warren, J. Making Full Use of the Longitudinal Design of the Current Population Survey: Methods for Linking Records across 16 Months. Journal of Economic and Social Measurement 39, 121–144 (2014).
  9. Madrian, B. & Lefgren, L. A Note on Longitudinally Matching Current Population Survey Respondents. National Bureau of Economic Research Working Paper (1999).
  10. Bureau of Labor Statistics. Employed, Usually Working Full-time. Retrieved from FRED, Federal Reserve Bank of St. Louis (2016).
  11. Bureau of Labor Statistics. The Employment Situation—December 2007 (2007).
  12. Burda, M. & Hunt, J. What Explains the German Labor Market Miracle in the Great Recession. Brookings Papers on Economic Activity Spring (2011).
  13. Ball, L. Long-term Damage from the Great Recession in OECD Countries. National Bureau of Economic Research Working Paper (2014).
  14. Summers, L. & Blanchard, O. Hysteresis and the European Unemployment Problem. National Bureau of Economic Research Working Paper (1986).
  15. Plotnikov, D. Hysteresis in Unemployment and Jobless Recoveries. IMF Working Paper (2014).
  • Show Comments

Your email address will not be published. Required fields are marked *

comment *

  • name *

  • email *

  • website *

You May Also Like

Sex-Related Reporting in Randomized Controlled Trials in Medical Journals

Emily Avery,1 Jocalyn Clark2 1Princeton Neuroscience Institute, Princeton University, Princeton, NJ, USA 2The Lancet, ...

A Natural Case for Taxation

Ruby Shao1 1Department of Philosophy, Princeton University, Princeton, NJ, USA ABSTRACT Libertarianism holds that ...

Reading the Nibelungen Women: Signs of their Times in the Thirteenth Century

Helena Ord1 1Department of Comparative Literature, Princeton University, Princeton, NJ, USA ABSTRACT This paper ...

Contaminated Cosmopolis: The Destabilized Nation-State in Igiaba Scego’s Beyond Babylon

Aaron Robertson1 1Department of French and Italian, Princeton University, Princeton, NJ, USA ABSTRACT This ...

Kinematic Substructure of Dark Matter and Stars in the Eris Simulation and Implications for Direct Detection Experiments

Jonah Herzog-Arbeitman,1 Lina Necib,2 Mariangela Lisanti1 1Department of Physics, Princeton University, Princeton, NJ, USA ...