by Michael Poyker (UCLA)
The United States has the largest prison population in the world. More than half of the convicts are engaged in some form of convict labour, and inmates employed in prison industries convicts constituted 4.5% of total US manufacturing employment in 2005. They earn substantially below minimum wage, ranging from nothing to $4.90 per hour in state prisons. Such low labour costs may have effects on free labour.
Despite numerous examples of malevolent competition, no research has been done on the possible effects of how convict labour affects free labour. In addition, contemporary public policy research typically considers convict labour as purely beneficial for society (through rehabilitation of prisoners and alleviating budgetary expenses on corrections). Finally, US prisons are built in economically depressed counties under the assumption that they will provide jobs (such as guards or janitors) in the local labour market.
My research addresses the previously unstudied question of how convict labour affects local labour markets and firms that employ free labour by studying convict labour in the United States in the nineteenth and twentieth centuries.
Data on contemporary convict labour output is not available, and as prisons are strategically located in economically depressed areas, it could confound the results. I use the historical context of when convict labour first appeared in 1870s.
First, very detailed data are available. Second, the rule of prison location was different – prisons were located in large urban areas (with higher wages) to save money on transport of prisoners. Third, the introduction of convict labour system was a nation-wide movement uncorrelated with the local economic conditions.
I have collected and digitalised archival data on US prisons and convict labour camps to construct county-level exposure to convict labour for the period 1886-1940. I find a significant negative effect of convict labour on wage growth and manufacturing employment, and a positive effect on patenting.
The magnitude of convict labour output was enormous: for each manufacturing worker with an average annual wage of $242, there were at least $18 per worker of prison-made goods. Regarding convict labour exposure comparing counties at the 25th and 75th percentile, the one more exposed to convict labour experienced 12.6% slower wage growth. In terms of counterfactuals, the introduction of convict labour in the 1870s accounts for 16% slower wage growth in the period 1880-1900 (when wage growth was 7.2%).
Competition with convict labour affected firms. Firms in affected industries couldn’t compete with prison-made goods in terms of labour costs. (The unit labour cost of prison labour was 4-50% of that of free labourers) Thus, they had to innovate-away in product-space or upgrade their technology to decrease costs or substitute labour with capital. I find that the introduction of convict labour accounts for 6% of the growth in patenting in affected industries.
I also shows the effects of convict labour on other economic outcomes. Convict labour gave police the incentives to arrest more people. Counties more exposed to convict labour had higher incarceration rates. There is also suggestive evidence that convict labour adversely affected intergenerational mobility in the long run.
Nowadays, as transport costs have decreased over time, competition with prison-made goods may spread farther from the prison. Thus, the overall effect of convict labour on contemporary manufacturing wages could be smaller around the prison but more substantial overall. Moreover, the number of convicts has soared from approximately 80, 000in 1920 to more than 2.5 million today.
Overall, these findings may be important for public policy research on convict labour since it may worsen local labour market outcomes, thus overshadowing any possible positive effects from rehabilitation or jobs provided by prisons. Finally, as the state is a beneficiary of the convict labour, welfare redistribution may be necessary to offset adverse effects on those affected by competition.