Author(s): Wouter Gelade, Maud Nautet, Céline Piton
Date published: Mar 2023
SUERF Policy Brief, No 535*
By Wouter Gelade, Maud Nautet and Céline Piton, National Bank of Belgium
*This article is based on a working paper published by the National Bank of Belgium for its international biennial research conference, https://www.nbb.be/doc/ts/publications/wp/wp419en.pdf.
JEL codes: J22, D13.
Keywords: Labour supply, household decisions, risk, added worker effect.
Download: SUERF Policy Brief, No 535 (0.44 MB)
Using Labour Force Survey (LFS) microdata for 16 European countries over the period 2005-2020, we quantify the labour supply adjustments made by households when one partner is at risk of losing his/her job. When such a risk occurs, the other partner is 29% more likely to enter the labour market (extensive margin) or 52% more likely to (want to) increase working hours (intensive margin). These effects are almost as big as those of an actual job loss for the intensive margin, and a bit more than half of those for the extensive margin. Fear of job loss is thus an important additional factor influencing households’ labour supply. This is particularly true during crises, with an equally big effect of fear of job loss and actual job loss. Analysis by level of education shows that low-educated people already adjust their labour supply when fearing job loss, while the high-educated wait for this risk to materialise.
The household is a key unit for decision-making, not least on its members’ employment situation. When the household faces a shock, such as an unexpected job loss by one partner, the other partner can increase his or her labour supply. This provides a form of insurance to the household and can reduce the shock of the job loss. This phenomenon, which has been well-studied both empirically and theoretically, is often called the “added worker effect”.
Yet households might adjust their labour supply even if there is no actual job loss for any of its members, but the fear of losing their job. We show that the effect of a risk of job loss on the partner’s labour supply is substantial, both in absolute terms and when compared to the effect of an actual job loss. We therefore highlight the importance of uncertainty in household-level labour supply decisions.
Relying on labour force surveys (LFS) for 16 European countries over the period 2005-2020, we find that both at the extensive margin (entering the labour market) and at the intensive margin (raising the number of hours worked) there is a clear increase in labour supply when the partner fears a job loss. In our most demanding specification, a previously inactive partner is 29% more likely to enter the labour market if his/her partner faces a risk of job loss than couples without that risk. A working partner is 52% more likely to (be willing to) step up his/her working hours in the event of risk of job loss for a partner than without that risk.
This effect comes on top of the labour supply adjustments when a partner actually loses his or her job. Focusing on unexpected job loss (based on workers being dismissed), we estimate a 52% increase in the likelihood to enter the labour market and a 61% increase of the number of hours worked for couples experiencing an actual job loss compared to those where the partner has not lost his/her job. In other words, the effect of a risk of job loss is also substantial when compared to the impact following an actual job loss.
Figure 1: Significant labour supply adjustments occur already when the partner is at risk of losing his/her job (in percentages, compared to probabilities for couples without risk of job loss or actual job loss)
Source: LFS, authors’ computations.
This is particularly true in periods of crisis. Since a larger share of people are then (at risk of) losing their job, more partners will necessarily adjust their labour supply. Additionally, however, the effect of fearing a job loss on labour supply is also much bigger during economic downturns. In fact, it is just as big as that following an actual job loss. Risk, or the perception of risk, is thus an important factor in understanding changes in labour supply during crises.
Figure 2: During crises, the effect of fearing a job loss is almost as big as that of an actual job loss (in percentages, compared to probabilities for couples without risk of job loss or actual job loss)
Source: LFS, authors’ computations.
One might expect different reactions by gender. Nevertheless, our results show that women and men react similarly to a (risk of) job loss, in line with trends evolving away from the male-breadwinner model. The only exception is that women are more likely to (want to) raise their working hours following a job loss, which could be due to the still large gender differences in part-time work.
We also study which types of households adjust their labour supply most strongly following both fear of and actual job loss. To do so, we separately analyse couples with and without children and we also consider different levels of education for the partner adapting his/her labour supply. This helps us to understand which households are more constrained in adjusting their labour supply and/or are more sensitive to income shocks. We can also measure how well they can anticipate a job loss by already adjusting the partner’s labour supply when they start fearing loss of employment.
Results show that people with children are much more likely to increase their labour supply following a job loss. In this case, the probability to enter the labour market is almost three times bigger for households with children than for non-parent couples. One potential explanation is that the presence of children implies more responsibility and a bigger need of sufficient revenue in the household. However, these families are not more likely to increase their labour supply when only fearing a risk of job loss. Families with children face greater constraints, making labour supply adjustments more costly. This suggests that family policies, such as wider childcare availability, are not only useful to increase labour market participation, but also to help households manage labour-market-related risks.
Finally, the level of education also plays a role in labour market adjustments. Low-educated individuals step up their labour supply considerably as soon as they perceive a risk of job loss, while high-educated people wait for their partner’s job loss to materialise. The faster reaction by the low-educated could be because a job loss presents a much bigger financial risk (the income loss is proportionally higher and their level of savings is lower), and because it is more difficult for them to find a new job. This underscores the necessity for sufficient social protection for all workers, including low-income workers in non-standard forms of employment, and adequate active labour market policies to help people to rapidly find jobs when needed.
About the authors
Wouter Gelade is an economist in the Research Department of the National Bank of Belgium. He holds a Ph.D. in economics from the University of Namur and in computer science from the University of Hasselt (Belgium). His current research interest is in labour economics and social policy.
Maud Nautet is an economist in the Research Department of the National Bank of Belgium since 2010. She is specialized in labour market issues. Her current research interests are labour market developments and the situation of vulnerable groups.
Céline Piton is an economist in the Research Department of the National Bank of Belgium since 2015. She is specialized in labour market issues. In her research, she is particularly interested in understanding and improving the position of groups less represented in the labour market, such as women or people of foreign origin. She holds a Ph.D. from the ULB (Belgium).