Military spending and jobs: what the research says

Military spending and jobs: what the research says

In short: the most cited research on military spending and jobs comes from economists Robert Pollin and Heidi Garrett-Peltier at the University of Massachusetts Amherst. Their 2011 study estimated that $1 billion of military spending supports about 11,200 US jobs, against about 16,800 for clean energy, 17,200 for health care and 26,700 for education. The studies also note that military jobs tend to pay more, and the estimates depend on input-output models with real limits.

The claim that civilian spending creates more jobs than military spending was central to the old Jobs Not Wars campaign and to budget proposals it promoted, such as the Congressional Progressive Caucus’s Back to Work Budget. This guide explains where the numbers come from, how they changed between editions, and what they can and cannot tell you.

Who produced the research?

The studies come from the Political Economy Research Institute (PERI) at the University of Massachusetts Amherst. Robert Pollin and Heidi Garrett-Peltier published versions in 2007, 2009 and 2011. According to PERI’s page for the 2009 update, that edition was produced with the Institute for Policy Studies, and a summary sheet says it was commissioned by the Institute for Policy Studies and Women’s Action for New Directions. Heidi Peltier later continued the work at Brown University’s Costs of War project.

A large aerospace manufacturing plant floor with partly assembled aircraft fuselages
Defence manufacturing jobs tend to be well paid. Illustration.

What are jobs per $1 billion?

Jobs per billion dollars is an estimate of how many full-time jobs a given amount of spending supports across the economy. PERI’s figures combine three kinds: direct jobs in the industry that receives the money, indirect jobs in its suppliers, and induced jobs created when those workers spend their wages. The estimates come from an input-output model, which traces how spending in one sector flows through others using government data on the structure of the economy.

Military spending and jobs: the headline numbers

The three editions used different data years and categories, so their figures are not identical, but the ranking is consistent.

Spending area2007 study2009 update2011 update
Military8,55511,60011,200
Tax cuts for personal consumption10,77914,80015,100
Clean energynot included17,10016,800
Health care12,88319,60017,200
Education17,68729,10026,700
Mass transit19,795not includednot included

Sources: the 2007 working paper, the 2009 updated analysis and the 2011 update. All figures are total jobs per $1 billion, including induced effects.

A hospital corridor with nurses seen from behind walking away
Health care spending creates more jobs per dollar in the PERI models. Illustration.

Why do civilian sectors show more jobs?

The 2011 study gives two main reasons. Civilian sectors such as education and health care are more labour-intensive, so more of each dollar goes on wages rather than equipment. And a larger share of civilian spending stays within the US economy. In the 2009 breakdown, military spending supported about 7,100 direct jobs per $1 billion, against 16,900 for education.

The 2007 working paper also found that 3,902 of the 8,555 jobs from defence spending, about 46 percent, were in government itself, with professional and business services next at 1,748.

Workers seen from behind installing solar panels on a roof
Clean energy was added to the comparisons from 2009. Illustration.

What about pay?

The studies acknowledge that military jobs tend to pay relatively well, which is part of the reason fewer are created per dollar. The 2007 paper estimated average compensation of about $65,986 for defence-supported jobs, against $56,668 for health care and $74,024 for education. The 2011 study argued that civilian spending still produced more jobs in mid-range and high-paying categories, but readers comparing the figures should note that a higher job count can partly reflect lower wages per job.

City buses parked in a transit depot at dawn
Mass transit produced the highest job estimate in the 2007 study. Illustration.

Limits of the input-output model

Input-output estimates rest on assumptions. They take the current structure of each industry as fixed, assume the economy has spare capacity to absorb new spending, and measure jobs supported rather than jobs that would not exist otherwise. They do not weigh the purposes of the spending, which is a political judgment rather than an economic one. The authors present their figures as comparisons of alternative uses of the same money, not as forecasts.

The PERI papers do not address national security value, and their findings are most useful for the narrow question of which spending supports more employment per dollar. Readers on all sides use them; the figures themselves are the same.

The Costs of War research

Heidi Peltier extended the analysis at Brown University’s Watson Institute. Her paper The Job Opportunity Cost of War, dated May 24, 2017, applied the same approach to post-2001 war spending. The project page summarises an earlier 2014 paper as estimating that war spending over 14 years cost between 1 million and 3 million jobs compared with spending the same money elsewhere.

An economist's desk with printed bar charts and line graphs
Input-output models estimate direct, indirect and induced jobs. Illustration.

How the numbers were used

The PERI figures became a staple of budget campaigns in the 2010s. The old site promoted the Congressional Progressive Caucus’s Back to Work Budget for fiscal 2014, which proposed lower military spending and more domestic investment. The same figures were cited during the sequestration debates, when both defence and domestic programmes were being cut. In defence-dependent states such as Connecticut, covered in our Connecticut guide, the jobs argument also cuts the other way, because existing defence jobs are concentrated in a few communities.

For the scale of defence spending these studies apply to, see world military spending: SIPRI data, and for the policy idea built on them, economic conversion explained.

Key takeaways

  • PERI’s 2011 study estimated about 11,200 jobs per $1 billion of military spending.
  • Clean energy, health care and education each showed more jobs per $1 billion in all editions.
  • Estimates include direct, indirect and induced jobs from an input-output model.
  • Military jobs tend to pay more, which partly explains the lower job count.
  • Military spending and jobs research compares alternative uses of money; it does not judge security needs.

Common questions

Does military spending create jobs?

Yes. PERI’s 2011 study estimated that $1 billion of military spending supports about 11,200 jobs in the US economy. The study’s point is comparative: the same $1 billion spent on clean energy, health care or education supported more jobs in its model, between about 16,800 and 26,700.

Why do the numbers differ between studies?

Each edition used a different base year of economic data, different spending categories and updated model assumptions. The 2007 study found 8,555 jobs for defence, the 2009 update 11,600 and the 2011 update 11,200. The ranking of sectors stayed broadly the same.

Are military jobs better paid?

The PERI studies say military jobs tend to pay relatively well. The 2007 paper estimated average compensation of about $65,986 for defence-supported jobs, compared with $56,668 for health care. Education was higher still, at $74,024.

What is an induced job?

An induced job is one supported by the spending of workers whose jobs depend directly or indirectly on the original spending. For example, defence workers buying groceries support jobs in shops. PERI’s totals include these induced effects alongside direct and indirect jobs.

What did the Costs of War project find?

Heidi Peltier’s work at Brown University’s Costs of War project applied the same method to war spending. The project summarises a 2014 paper as estimating that 14 years of war spending cost between 1 million and 3 million jobs compared with alternative uses of the money.

How we put this together: we read the three PERI studies and the Costs of War paper and compared their figures in one table, noting where categories changed. We describe the limits of input-output models in general terms and do not endorse any budget proposal. Return to the home page.

Sources