Garett Jones on Stimulus

Episode Summary

Executive Summary: Russ Roberts and Garrett Jones examine Keynesian stimulus theory and Jones’s Mercatus study of the 2009 Recovery Act. They argue that wartime and stimulus spending often crowds out private activity, that many stimulus-funded jobs did not go to the unemployed, and that labor is less fungible than Keynesian models assume. The episode emphasizes supply-side crowding out, targeting problems, and the limits of aggregate multipliers.

Main Topics: How Keynesian stimulus is supposed to work (Priority: 5/5): Roberts and Jones discuss the standard Keynesian claim that government spending can utilize slack resources during recessions, producing a multiplier and helping the economy bridge a short-term downturn. Natural experiments and war-related spending (Priority: 5/5): Jones explains the use of wars as exogenous shocks to government spending, especially Valerie Ramey’s work using anticipation dates from Business Week to measure when firms expected spending increases. Mercatus study of Recovery Act recipients (Priority: 5/5): Jones describes the mixed-methods project: 85 interviews plus a mailed survey of stimulus recipients to determine whether stimulus dollars actually hired unemployed workers and how firms responded. Crowding out of private-sector activity (Priority: 5/5): The conversation stresses that government spending may increase measured GDP while shrinking private activity, especially when scarce labor and inputs are pulled toward the public sector. Labor market heterogeneity and matching frictions (Priority: 4/5): Roberts and Jones argue that workers are not interchangeable; stimulus money often went to skilled workers or firms needing specialized labor, not to broadly unemployed workers in distressed sectors. Tax cuts versus spending increases (Priority: 4/5): Jones cites recent literature suggesting tax cuts may have larger multipliers than spending increases, and that spending cuts may be less harmful than tax hikes when reducing deficits. Survey methods and the limits of policy reporting (Priority: 4/5): The episode discusses the value and limitations of interviewing recipients directly, as well as the shortcomings of official reporting that counts jobs as created or saved without adequate counterfactuals.

Key Arguments: Keynesian stimulus assumes recessions leave idle labor and capital that government can temporarily employ while markets adjust, but this assumes a level of substitutability that may not exist. War-related spending studies, especially Ramey’s, suggest that when government demand rises, private-sector activity often falls rather than multiplies strongly. Jones’s interviews found that recipients often used trusted existing contractors and organizations, implying the stimulus favored established firms rather than newly unemployed workers. The study’s survey found only 42% of workers hired with stimulus funds came from unemployment, less than the roughly 50% share typical in normal hiring. Many stimulus positions went to highly educated, skilled workers; this suggests the Recovery Act did not primarily target the workers with the highest unemployment rates. When government expands demand for specialized labor, it can bid workers and inputs away from private firms, raising costs and creating supply-side crowding out. Fiscal policy debates should compare not just spending multipliers above or below one, but whether the multiplier is positive or negative and whether benefits go to the intended recipients. Tax cuts often appear to have larger short-run effects than government spending increases, according to several recent econometric studies cited in the discussion.

Data Points: Podcast date: September 14, 2011 - EconTalk episode introduction Interviews conducted: 85 organizations - Face-to-face interviews with stimulus recipients for the Mercatus study Survey mailed: 8,000 surveys - Two-page surveys sent to selected stimulus recipients Survey responses: 1,300 responses - Returned survey pieces of paper from recipients and workers Estimated stimulus recipients in screened universe: about 13,000 - Approximate number discussed before screening Workers from unemployment: 42% - Share of workers hired through stimulus funds who reported being unemployed beforehand Workers already employed before hire: 47% - Share of workers who said they were working somewhere else immediately before Job-switching in healthy labor markets: about half - Roberts and Jones note that in normal times roughly half of hires come from unemployment and half from job switchers Benefits improved: about 40% better, 40% same, 20% worse - Workers’ comparison of benefits at the stimulus-funded job versus their previous job Employer view of hiring difficulty: about 15% harder; about 37% no easier; about 47% easier - Survey responses on finding good workers compared with before the recession Romer and Romer tax multiplier: 3 - Cited as a short-run estimate that one dollar of tax cuts raises output by about three dollars Ramey/Shapiro finding: multiplier between 0 and 1 - War-spending natural experiment suggested government spending increased GDP but reduced private-sector activity Stimulus bill size: $820 billion - Approximate size of the Recovery Act mentioned in the discussion Unemployment threshold prediction: over 8% - Reference to a famous pre-stimulus forecast if the bill had not passed Observed unemployment: over 10% - Unemployment rate after the stimulus passed, as discussed Job churn: about 4 million jobs created and destroyed every month - John Haltiwanger’s labor market dynamics research referenced in the discussion

Pivotal Quotes: "There are a lot of workers, a lot of firms that just aren't going to be very useful, very productive for a while." — Garrett Jones: Explaining the Keynesian view of recession and idle resources "What we found evidence of here is that when workers are pulled toward the government sector, especially when the best workers are pulled over to the government sector, those are workers that aren't available to the private sector anymore." — Garrett Jones: Summarizing the supply-side crowding-out interpretation of the stimulus "There’s no such thing as shovel-ready projects." — Russ Roberts: Discussing President Obama’s later description of stimulus implementation problems

Implications: The episode suggests stimulus programs may be much less effective at hiring the unemployed than advertised, especially when they rely on specialized labor and trusted contractors. For future policy, better targeting and more realistic assumptions about labor mobility are crucial.