The redistributive effects of pandemics: Evidence on the Spanish flu

Abstract

Which are the effects of pandemics on the returns to factors of production? Are these effects persistent over time? These questions have received renewed interest after the out-burst of deaths caused by Covid-19. The Spanish Flu is the closest pandemic to Covid-19. In this paper, we analyze the impact of the Spanish Flu on the returns to labor and capital in Spain. Spain is an ideal country to perform this exercise. First, the "excess death rate" was one of the largest in Western Europe and it varied substantially across regions. Second, Spain was transitioning towards industrialization, with regions in different stages of development. Third, Spain was developed enough to have reliable data. We identify the effect of the Spanish Flu by exploiting within-country variation in "excess death rate". Our main result is that the effect of the Spanish Flu on daily real wages was large, negative, and broadly short-lived. The effects are heterogeneous across occupations and regions. The negative effects are exacerbated in (i) occupations producing non-essential goods like shoemakers and (ii) more urbanized provinces. Quantitatively, relative to pre-1918, the decline for the average region ranges from null to around 30 percent. In addition, we fail to find significant negative effects of the flu on returns to capital. Whereas the results for dividends are imprecisely estimated (we cannot reject a null effect), the effect on real estate prices (houses and land), driven by the post-1918 recovery, is positive. Experts on inequality have argued that pandemics have equalizing effects especially in a Malthusian setting, due to real wage increases. Our findings suggest that, at least, for a developing economy like Spain in the early 20th century, this result does not apply. Indeed, we document that the flu pandemic was conducive to a (short-run) reduction in real wages. In addition, we interpret our heterogeneous results as suggestive evidence that pandemics represent a demand shock.

Keywords: Pandemics; Real wages; Returns to capital; Spanish flu.

Fig. 1

Evolution of Number of Deaths in Spain, 1910–1930 Sources: Death data obtained from INE Base Hist́orica.

Fig. 2

Evolution of Gross Domestic Product and Deaths, 1910–1930 Sources: Death data obtained from INE Base Historica and Gross Domestic Product (GDP) in 1995 constant prices from Prados de la Escosura (2017).

Fig. 3

Excess Death Rate, 1910–1930 Notes: Excess death rate obtained as deviation from the predicted linear trend (normalized by population), by province. Death data obtained from INE Base Hist́orica and Census Data.

Fig. 4

The Flu: Excess Death Rate in 1918 Notes: The map represents excess mortality rate in 1918. Excess death rate obtained as deviation from the predicted linear trend (normalized by population), by province. Includes the 10 largest capitals of provinces (1910). Death data obtained from INE Base Hist́orica and Census Data.

Fig. 5

Effect of Spanish Flu on Real Wages (by Occupation) Notes: Each point is the coefficient of the interaction between excess death rate in 1918 and the time dummies. The dashed line is the 95 percent confidence interval. Pre-1918 was excluded from the regression and it is the reference group. The coefficients are reported in Table 4.

Fig. 6

Dynamic Effect of Spanish Flu on Real Wages Over TimeNotes: Each point in line is the coefficient of running the interaction between “excess death” in 1918 on different time dummies. The area represents the 95 percent confidence interval.