Unemployment and international shadow economy: gender differences

ABSTRACT

Adding to the body of research on cross-country determinants of the informal economy, this article studies how unemployment differences across gender drive the propensities to work in the informal sector. Results, accounting for possible simultaneity between unemployment and the shadow economy, show that unemployed males, rather than unemployed females, were more likely to operate underground. These findings about greater male propensities are generally robust to alternative considerations.

KEYWORDS:

• Gender
• informal economy
• shadow economy
• unemployment

JEL CLASSIFICATION:

• K42
• J16
• E26

Acknowledgement

A preliminary version of this work was done with Ummad Mazhar who has withdrawn from the project. We thank him and a referee for their insights, and Shabana Tabassum and Ashley Wessman for research assistance.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ We use the terms informal economy, shadow economy, and underground economy interchangeably.

² In a related study, Allen and Curington (2014) show differences in motivations between men and women in entering self-employment.

³ Frederiksen, Graversen, and Smith (2005) provide some corroborating evidence from a developed nation (Denmark). Also see, Dell’Anno, Gómez-Antonio, and Pardo (2007).

⁴ Manes, Schneider, and Tchetchik (2016) examine the dynamic relation between unemployment and the shadow economy while noting that there may be decreasing returns to shadow activities and there may be congestion effects in job search.

⁵ GDP per capita may also be seen as proxying for individual wealth, which would affect the propensity to operate in the underground sector. Goel and Saunoris (2014) further consider the relative impacts of defense and non-defense government spending on the shadow economy.

⁶ The International Labour Organization (ILO) (2013) provides somewhat more detailed data on shadow prevalence across gender. However, the sample of nations is less than 50 in the cross-section and even more limited over time.

⁷ Due to the difficulty in locating suitable external instruments (i.e. those correlated with the endogenous variables and orthogonal to the error term), it is imperative to test these instruments for their validity and relevancy. To this end, we conduct three tests: (1) Kleibergen-Paap rk LM test; (2) Kleibergen-Paap rk Wald F test; and (3) Hansen’s J test. The first test, the Kleibergen-Paap rk LM test, is used to test for the relevancy of the variables (i.e. correlated with the endogenous variables) where a rejection of this test indicates that the instruments do not adequately identify the model. While the Kleibergen-Paap rk Wald F test is used to test if the variables are only weakly correlated with the selected instruments. This statistic is compared to the critical values (based on bias and size) from Stock and Yogo (2005). Finally, a test of overidentification is conducted using the Hansen’s J test. Failure to reject the null hypothesis in this case signifies that the instruments are correctly excluded from the equation and orthogonal to the error term (for more details on these tests see Baum, Schaffer, and Stillman (2007)).

⁸ These dimensions can be overcome with more detailed data on employment decisions.

⁹ Economic prosperity can alternately be accounted for by including the growth rate of GDP (e.g. see Dell’Anno, Gómez-Antonio, and Pardo (2007)).

¹⁰ The diagnostic tests initially revealed problems with the original instruments, therefore we use own lags dated 3–5 periods back as instruments.

¹¹ While the use of this alternate measure of the shadow economy is instructive, the problems with adequately capturing the shadow sector are well documented (see Schneider and Buehn (2013)).

¹² However, the insignificance of the Kleibergen-Paap LM statistic and significant Hansen J statistic suggests caution when interpreting these results (see Kleibergen and Paap (2006)).

¹³ Not considering gender aspects and focusing only on the United States, Goel and Saunoris (2017) find complementarity between three white-collar crimes: corruption, shadow economy, and uninsured motorists. Consideration of linkages with other white-collar crimes, however, is beyond the scope of current work.