Are Economic Upturns Bad for Military Recruitment? A Study on Swedish Regional Data 2011–2015

ABSTRACT

This paper represents the first effort to explore the relationship between civilian labour market conditions and the supply of labour to the military in the all-volunteer environment that Sweden entered after the abolishment of the peacetime draft in 2010. The paper investigates the effect of civilian unemployment on the rate of applications from individuals aged 18–25 to initiate basic military training, using panel data on Swedish counties for the period 2011–2015. A linear fixed-effects model is estimated to investigate the relationship, while controlling for a range of socio-demographic covariates, unobserved heterogeneity on the regional level, as well as aggregate trends on the national level. The results of the panel-data analysis indicate that the unemployment rate has a positive and statistically significant effect on the application rate. These results are robust to non-linear form specifications, as well as allowing the civilian unemployment rate to be endogenous. As such, the results suggest that the civilian labour market environment in Sweden can give rise to non-trivial fluctuations in the supply of applications to initiate basic military training within the Swedish Armed Forces.

KEYWORDS:

• Military labour supply
• Military labour market
• Recruitment
• Fixed effects model
• Conscription

JEL CLASSIFICATION:

• J01
• J20
• J45
• H56

Acknowledgements

The author would like to thank Olle Westerlund, Katharina Jenderny, Gauthier Lanot and Niklas Hanes, as well as two anonymous referees, for their valuable comments and suggestions. The paper has benefited from comments received from seminar participants at the Department of Economics, Umeå University; at the Swedish Defence Research Agency; and at the Swedish Armed Forces Headquarters. Thanks also to Emma Jonsson at the Swedish Defence University for helping with the data.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. As a comparison, around 32,000 individuals initiated their mandatory military service in 1997 (Olsson et al. 2018, 27).

2. Fisher (1969) and Ash, Udis, and McNown (1983) analysed enlistment supply in a similar theoretical setting.

3. The recruitment test is the mandatory first step of the application process and consists of an online self-reporting questionnaire. The data comprises only recruitment tests that are completed in full and submitted to the Recruitment Agency.

4. Alternative measures of the unemployment rate include: (1) the number of openly unemployed, expressed as a share of the labour force, and (2) the number of openly unemployed, plus the number of individuals in active labour market programs, expressed as a share of the population. Using these measures in the empirical analysis does not alter the main results.

5. An alternative measure of civilian earnings is the total amount of pay per employed worker by county. Using this measure, instead of retail wages, in the empirical analysis does not alter the main results.

6. Table A2 in the Appendix provides results from estimating the main model without population weights. Running the linear fixed effects regression without population weights reduces the point estimate of the unemployment variable somewhat, making in statistically insignificant at the 5 per cent level (however, it’s still statistically significant at the 10 per cent level). Hence, the estimated effect from unemployment appear to be partly driven by the weighting scheme, where more populous counties are assumed to provide more information about the relationship.