Temporary Employment, Informal Work and Subjective Well-Being Across Europe: Does Labor Legislation Matter?

Following OECD guidelines, we define subjective well-being as the combination of happiness level and the level of overall life satisfaction (OECD 2006).

The two indices reflect two different aspects of the labor legislation system and do not correlate with each other. The correlation is 0.2 between EPL_dismissals and EPL_temps.

When we speak about informal employment, EPL is not the only one factor influencing its proportions, the tax system and law enforcement contribute a lot to the growth of this type of jobs. But in this paper, we concentrate not on pure informal work but on precarious VS permanent employment and labor market regulations that are responsible for the labor market system in a country. That is why we speak here only about EPL as a main country-level predictor.

The ESS is a household survey conducted every 2 years in almost all European countries. The questionnaire consists of a collection of questions that can be classified into two main parts—a core section (repeated regularly) and a rotating section. Sampling on the ESS is guided by the following key principals: samples must be representative of all persons aged 15 and over (no upper age limit); individuals must be residents within private households in each country, regardless of their nationality, citizenship or language; individuals are selected by strict random probability methods at every stage; sampling frames of individuals, households and addresses may be used; all countries provide a minimum 'effective achieved sample size' of 1500 or 800 for countries with populations of less than 2 million after discounting for design effects; quota sampling is not permitted at any stage; substitution of non-responding households or individuals (whether 'refusals', 'non-contacts' or 'ineligibles') is not permitted at any stage. More information can be obtained from http://​www.​europeansocialsu​rvey.​org/​methodology/​.

For example, European Value Study or World Value Survey data do not contain any questions on type of contract but only allow to identify part-timers and full-timers.

ESS Round 2: European Social Survey Round 5 Data (2004). ESS Round 5: European Social Survey Round 5 Data (2010). Norwegian Social Science Data Services, Norway – Data Archive and Distributor of ESS data. Only these two rounds out of the six contain a special module of questions on job characteristics that we are using to describe temporary and informal workers.

We estimated the same models for 2 years separately at the first stage of the analysis and results proved to be the same and consistent with the merged sample. Having two rounds together for empirical analysis allowed us to perform a robustness check, on the one hand, and to control for changes in EPL between 2004 and 2010, on the other hand.

However only four countries out of 27 experienced deregulation of the EPL on temporary employment. So there is almost no variation between 2004 EPL and 2010 EPL in the sample. In order to control for crisis effects we also did test for the dynamics of informal and temporary work for all available 2002–2012 rounds of ESS, the data is closely aligned with official statistics: there is no sharp increase in temporary/informal work during the 2008–2010 period, meaning that the financial crisis did not stimulate the growth of informal/temporary work. The crisis effect was much higher for involuntary part-time work (Gash and Inanc 2013).

The diversification of the EPL indices allows to speak in detail on countries difference in terms of institutional regulations. Collapsing the two indices to EPL gap is not of our focus as well as the EPL on collective dismissals in not in focus of our research.

The bi-variate correlation analysis by countries showed that the correlation coefficient between happiness and life satisfaction varies from 0.58 in Portugal to 0.76 in Austria, with 0.71 for all countries sample.

The exact question on contract type is formulated like this: “Do you have a working contract of limited, unlimited duration, or, do you have no contract?”.

Despite that previous research has shown that subjective health, marital status, having children are affected by temporary employment, we use them as controls as their inclusion improves the model fit. However, to eliminate the possible so called “overcontrol bias” from these endogenous variables we tested the models for the sensitivity of our results with respect to the inclusion of those variables (see Elwert and Winship 2014).

There is no data on family income for Portugal. Moreover, income variables usually have a high non-response rate. In our case, it would reduce the sample to 25,000 respondents and that would require multiple imputation techniques to deal with this problem. Another problem here is that the close relationship between family income and employment status of a person that might also bias the results. We excluded income from our final model in order to keep the number of countries and number of cases. However, we did estimate the models with family income to check for robustness and the tested effects for precarious work remained the same.

We also checked the results for their robustness by applying OLS regressions with clustered errors for the same specifications. The main effects as well as all others for control variables were same in terms of coefficients size and significance.

We have chosen HDI because it reflects not only the material wealth of the country but also the level of development of human capital. This aspect is especially relevant in respect of research focused on labor markets. The HDI increases the individual levels of happiness and life satisfaction, which is in line with previous research by Easterlin and others (Easterlin 1995; Sarracino 2013).

EPL remains one of the most influential factor in terms of regulations of job creation and job opportunities in the labor market, and it mainly defines the choices between permanent and temporary employment. The limitation of the study might be that we did not include into the models any indicators for active or passive labor market policies, which are much more important for the unemployment outcomes. Since unemployment is not the focus of the paper and we can include only a few country-level variables into the analysis in order to provide a better model fit for multilevel modeling, we limited the country level characteristics to the three of them: HDI, EPL_dismissals and EPL_temps.

We can interpret the main effect from the second interacted variable only when the first of the interacted variables equals zero (Jaccard 2001). Thus, we estimate the impact of temporary/informal employment on SWB for countries with rigid EPL and countries with liberal EPL via the calculation of conditional effects in order.

Along with the main results, we can claim that younger or older age groups (compared to 36–45 year olds), tertiary education, good health and religiosity positively influence subjective well-being. At the same time, men and residents in big cities show lower levels of SWB. We underline that these results are consistent and stable in all further models we estimated. Moreover, they are in line with previous findings (Bardasi and Francesconi 2004; Brereton et al. 2008; Andersson 2008; etc.).