THE EFFECTS OF PANDEMICS ON INEQUALITY: Evidence from the Spanish flu

12 Apr 2021

Mortality rates in Spain during the influenza pandemic of 1918 were unequal across several dimensions, according to new research by Sergi Basco, Jordi Domènech and Joan Rosés. First, they were highest among 15-40 years old; second, women had higher pandemic-induced rates; third, the pandemic was less deadly in the largest cities; and finally, high-income occupations had much lower excess death rates than low- or middle-income occupations, probably because many poor workers did not have the means to keep social distancing.

In terms of the economic effects of the pandemic on inequality, the study documents sharp, negative but short-lived effects of the pandemic on real wages for most occupations, exacerbated in occupations like shoemakers or tailors, which produce non-essential goods. In some cases, the pandemic reduced real wages by as much as 30%. The fall was also larger in more urbanised provinces, which points at a decline in demand driving the reductions on wages.

In contrast, there appears to have been no negative effect of the pandemic on the returns to capital, measured by dividends and real estate prices. The redistributive effect of the pandemic must have been regressive.

More…

In many influential studies, pandemics have been considered ‘big levellers’, which can trigger positive effects in the affected economies. According to this view, despite the tragedy of high mortality rates, pandemics affect all social classes equally, can compress income differentials, increase wages, and promote institutional and technological changes.

The Spanish influenza was the closest historical case comparable to the current Covid pandemic. This study focuses on the effects of the Spanish flu on income and mortality inequality in Spain.

Spain is an optimal testing ground to measure the impact of the pandemic. Its mortality rates were not affected by the First World War. It had reasonably good and disaggregated data on mortality across regions and occupations. Finally, it was a developing economy with a large agrarian sector, but also with fast growing urban sectors linked to industry and services.

The researchers first describe the main contours of the pandemic in Spain. They do not use the causes of death in the mortality statistics, which may bias the real deaths attributed to the pandemic (or example, since no flu-test existed, it was based on the criterion of the physician or local authority which may vary across time and locations).

Instead, they use deviation of actual monthly mortality from the trend to compute pandemic-related excess deaths. Using this methodology, they show that there was one big peak of mortality in October 1918, almost quadrupling the typical levels of mortality in October, and two much smaller peaks in June 1918 and January 1920.

These mortality rates were unequal across different dimensions:

  • First, mortality rates were highest among 15-40 years old.
  • Second, women had higher pandemic-induced mortality rates.
  • Third, the pandemic was less deadly in the largest cities.
  • Finally, there was substantial heterogeneity across occupations. High-income occupations (rentier and liberal occupations) had much lower excess death rates than low- or middle-income occupations (rural labourers, miners, retail or trade-related workers). The researchers argue that many poor workers did not have the means to keep social distancing.

How was the burden of the shock distributed? Looking at longitudinal data at the equivalent of a NUTS 3 region from 1910 to 1930, the research documents sharp, negative, but short-lived effects of the pandemic on real wages for most occupations, exacerbated in occupations like shoemakers or tailors, which produce non-essential goods. In some cases, the pandemic reduced real wages by as much as 30%.

In addition, the fall is larger in more urbanised provinces, which points at a decline in demand driving the reductions on wages. Wages fell more in areas with more social distancing and reduced consumption of non-essentials, rather than in high mortality regions (typically more rural).

In contrast, there appears to be no negative effect of the pandemic on the returns to capital, measured by dividends and real estate prices. The redistributive effect of the pandemic must have been regressive.

Aggregate data show a decline in GDP in 1918, a recovery in 1919 and strong growth in 1920 and subsequent years. Investment fell by 34.2% in 1918 and recovered in the 1920s. Consumption also collapsed for non-essential goods in 1918, but food and other essential consumption were not affected.

Thus, the Spanish flu represented a temporary negative demand shock, which increased inequality in the short run, but did not have a generalised persistent effect on the economy.

Sergi Basco

Universitat de Barcelona | sergi.basco@gmail.com

Jordi Domènech

Universidad Carlos III de Madrid | jdomenec@clio.uc3m.es

Joan R. Rosés

London School of Economics | j.r.roses@lse.ac.uk