By Javier Gardeazabal & Eduardo Polo-Muro


The 2003 Spanish Academy award winning film Mondays in the sun portraits the daily lives of three men who were laid off from the local shipyard. Directed by Fernando León de Aranoa and starring Javier Bardem, Luis Tosar and José Ángel Egido, this social drama gem exemplifies how unemployment can affect individuals and their families. In our paper Cultural expenditure of those who enter (or exit) unemployment, we assess the impact of transitions from employment to unemployment on household expenditure on cultural goods and services.

Entering unemployment is one of the largest income shocks a person can experience and therefore is bound to have a large and significant effect on household expenditure, and particularly so for cultural expenditure which is not as necessary for ordinary life as food or clothing. However, those who become unemployed not only suffer income losses, they also have more time, which is a required input for cultural consumption, and a fact highlighted in the movie, where the unemployed men have time to lay in the sun on Mondays.

The most reliable way to assess the effect of unemployment on cultural consumption would be to conduct an experiment. However, carrying out an experiment in this case is certainly unethical and probably illegal. Despite this impossibility, suppose for a moment we were allowed to assign unemployment status randomly to a group of individuals and not to another group to be used as a control. We would then compare the cultural expenditure of those who received the treatment (unemployment) and those who did not. The difference between the average cultural expenditure of both groups would be an estimate of the effect of unemployment on cultural expenditure. To understand why the experimental method works, notice that, in the experiment, unemployed (treated) and employed (untreated) are randomly selected, that is to say, any individual would have the same probability of being assigned to the group of employed. This assignment mechanism is likely to result in groups of individuals with similar characteristics, making them comparable. So if we find that the unemployed spent less than the employed on cultural goods and services, we can conclude it was caused by their employment status, as there are no differences in the characteristics of the two groups that could be driving the result.

In reality, we can observe cultural expenditure of employed and unemployed people, but their comparison is not a good estimate the impact of unemployment on cultural expenditure, because employed and unemployed people are different in many respects. To tackle this problem, we use two different approaches: matching and weighting. The matching approach is very intuitive. As employed and unemployed individuals are dissimilar, a feasible option is to select a subgroup of employed individuals with similar characteristics to the unemployed, and hence comparable. Proceeding in this way, the impact of unemployment on cultural expenditure would then be estimated as the difference in the average cultural expenditure of the unemployed and the selected (matched) group of employed. The weighting method is a bit more involved, but also easy to understand. Assessing the impact of unemployment on cultural expenditure is estimated in this case as the difference between the average cultural expenditure of the unemployed and a weighted average of the cultural expenditure of the employed. The weights used to construct the average are inversely proportional to the (ex-ante) probability of becoming unemployed, thus weighting up those whose probability is small and down those whose probability is high. As a result, in the weighted sample all observations have approximately the same probability of becoming unemployed, thus mimicking an experimental random assignment setup where the probability of being treated would have been exactly the same for all subjects.

These quasi-experimental methods construct a comparison group that resemble the characteristics of the treated group as regards to all observable characteristics. However, employed and unemployed individuals might also differ in dimensions that we do not observe or measure, such as work-attitude, self-confidence, diligence and the like. The matching and weighting methods cannot deal with these unobserved differences. Nevertheless, we can account for them using repeated observations of subjects’ cultural expenditure, provided those unobserved differences in, say, work-attitude or self-confidence, remain constant over time. Availability of repeated observations allows us to compare the change in cultural expenditure of subjects who become unemployed and those who remain employed.

Combining the matching and weighting methods with the availability of repeated observations of the same individuals, our article provides estimates of the effect of transitions from employment to unemployment on cultural expenditure using data from the Spanish Household Budget Survey from 2006 to 2015.

What are the takeaways from this quasi-experimental study? When one household member becomes unemployed, on average, Spanish households reduce their cultural consumption by 16% to 23% and this effect increases to 19 to 26% when households have a second unemployed member. What about cultural participation? Unemployment barely affects household cultural participation, a finding in favor of the taste for cultural goods hypothesis, according to which cultural consumers with a large stock of cultural capital would continue consuming cultural goods and services even after becoming unemployed. What happens when a member of the family finds a job? This has a symmetric quantitative impact but in the opposite direction on cultural expenditure, although this effect fails to be statistically significant. Are there gender differences? The reduction in cultural expenditure is larger when the unemployed household member is a man relatively to when it is a woman, even after accounting for the different income losses. Does the level of education affect the change in cultural consumption? Those with university studies reduce their cultural consumption to a lesser extent, perhaps because they expect to experience shorter unemployment spells. Are the results affected by the business cycle? The reduction in cultural expenditure is larger in harsh economic times, despite the fact that the effect of unemployment on household income does not fluctuate that much over the business cycle. Hence, during a recession, people might moderate their consumption as a result of social norms, and particularly the consumption of cultural goods and services.

Finally, it can be argued that the reduction in cultural expenditure as a result of unemployment is somewhat tempered as a consequence of more disposable time. However, Spanish full-time workers report having sufficient time for leisure and personal care (OECD 2017). Therefore, time does not seem to be a binding restriction in Spain as far as cultural consumption is concerned, although it might be for laying in the sun on Mondays.

About this article:

Gardeazabal, J., Polo-Muro, E. Cultural expenditure of those who enter (or exit) unemployment. J Cult Econ 46, 571–596 (2022).

About the authors:

Javier Gardeazabal is Professor of Economics at the University of the Basque Country.


Eduardo Polo-Muro is Postdoctoral researcher at the Department of Economic Analysis of the University of the Basque Country.

About the image:

Converse Sneakers on Feet of Person Laying in Field by Image Catalog is marked with CC0 1.0.

January 25, 2023
(h)om(e)nibus Culture At A Time Of Lockdown
January 10, 2023
Counteracting Hate By Encouraging Participation In Cultural Activities
December 27, 2022
Economists Talk Art Wishes You A Happy Holiday And A Wonderful New 2023
December 13, 2022
Cultural Expenditure Of Those Who Enter (or Exit) Unemployment

Become a member

Members of the ACEI will be part of a network of scholars, researchers and practitioners interested in advancing cultural economics.

Join Today