War Bonds and Household Saving in WWII

Household saving increased dramatically during World War II, reaching more than 19 percent of GDP. We study the effects of the war bond program implemented by the U.S. government on the level of household saving during the war.  The bonds were heavily promoted in a series of drives, which encouraged thrift and associated subscriptions with patriotism, and also through a payroll deduction program. Yet as Friedman and Schwartz have noted, the main effect of the program may have been to change the form in which savings were held, rather than to increase saving.  We use county-level data and an instrument for participation in the bond program to estimate the effect of war bond sales on total saving.  We find that for every $100 in war bond sales, bank deposit inflows fell by $70, suggesting that while there was substantial substitution between war bonds and bank accounts, the program did actually increase total saving. A back of the envelope calculation suggests that the bond program increased total personal saving by about 7 percent, in large part driven by the voluntary payroll deduction program.

Time Period(s):   1939 – 1944 (WWII); 1930 – 1950 (US decennial censuses)

Data Type(s):   census/enumeration data

Methodology

Data Source:  View help for Data Source County-level data in the Stata file originate from the following sources: 
United States. Bureau of the Census. County and City Data Book [United States] Consolidated File: County Data, 1947-1977. Inter-university Consortium for Political and Social Research [distributor], 2012-09-18. https://doi.org/10.3886/ICPSR07736.v2

Haines, Michael R., and Inter-university Consortium for Political and Social Research. Historical, Demographic, Economic, and Social Data: The United States, 1790-2002. Inter-university Consortium for Political and Social Research [distributor], 2010-05-21. https://doi.org/10.3886/ICPSR02896.v3

Haines, Michael, Fishback, Price, and Rhode, Paul. United States Agriculture Data, 1840 - 2012. Inter-university Consortium for Political and Social Research [distributor], 2018-08-20. https://doi.org/10.3886/ICPSR35206.v4

WWII production data from the 1947 County Data Book has been supplanted by the data as reconstructed by Gillian Brunet from the original contract micro data files.  The micro data are available in the Harvard data repository for the following publication:
Brunet, Gillian (2024). “Stimulus on the Home Front: the State-Level Effects of WWII Spending,” Review of Economics and Statistics, forthcoming.  
The reconstructed data have far more counties with $0 war spending than the County Data Book version, suggesting that the County Data Book includes noise injections.  However, the choice of which version of the data to use does not seem to affect results.
 
Liberty bond data and liberty bond participation instrument come from: 
Hilt, Eric, Matthew Jaremski, and Wendy Rahn (2022) “When Uncle Sam introduced Main Street to Wall Street: Liberty Bonds and the transformation of American finance,” Journal of Financial Economics, 145 (1), 194–216.

Authors constructed 1940 median wages from full-count 1940 census. 

Price variables are CPI, reindexed so that 1950 = 1(00). 

Additional bank deposit data were provided by Price Fishback and Paul Rhode.  

Unit(s) of Observation:   county