“Make a wish”: How a de-institutionalised care programme shaped the dynamics of women’s aggregate labour market outcomes in Croatia

The article by Marko Lucić, Research Assistant at the Institute for Social Research in Zagreb, and Jakša Puljiz, PhD, of the Institute for Development and International Relations, analyses the effects of the "Zaželi" ("Make a Wish") intervention between 2018 and 2023. The intervention engaged women registered as unemployed with the Croatian Employment Service who had completed no more than upper secondary education to provide in-home support and care to older people, persons in need of assistance, and persons with disabilities. The article examines the effects of the intervention on the share of long-term unemployed women within the target group, women’s employment outside the intervention, and women’s labour force participation rates in the municipalities covered by the intervention.

The analysis applies a variant of the quasi-experimental difference-in-differences method designed for settings in which the intervention begins at different times across the observed units. The apparently large and statistically significant, but temporary, effects on reducing the share of long-term unemployed women within the target group were largely attributable to the dynamics of intervention implementation. More specifically, women returned to the unemployment register after completing their work assignments, thereby increasing the denominator used to calculate the share of long-term unemployed women and temporarily reducing that share. Nevertheless, more stable effects were identified in the final observed year following the local introduction of the intervention.

An analysis of effect heterogeneity according to municipalities’ pre-intervention development and ageing indices found significant reductions in the share of long-term unemployed women only among municipalities in the lower half of the development-index distribution, as well as somewhat stronger effects in municipalities with older populations. Regarding women’s employment in other jobs, no positive or negative effects—including displacement effects—were found in quarters during which the intervention was not being implemented locally. The same applies to women’s labour force participation rate.

In addition to presenting the results, the article discusses lessons learned regarding the possibilities and limitations of quasi-experimental evaluations of aggregate-level labour-market effects in Croatia and elsewhere. This is particularly relevant because the study represents one of the few evaluations conducted to date of the aggregate-level effects of large-scale public works programmes.