Most people understand a basic truth: you get what you pay for. Skip maintenance on your roof, and you shouldn’t be surprised when leaks appear.
The same is true of government. If we want a high-functioning public sector—and we should—there is no shortcut. It requires sustained investment in the people and capacity that make government work. Starve it of resources, and its performance will inevitably suffer.
In a recent New York Times essay, academics Nicholas Bagley and Robert Gordon argue otherwise. In their telling, government underperforms because public-sector unions have too much power, driving up costs and resisting efficiencies. Their solution is simple: rein in unions and invest less—largely by cutting pay for public-sector workers. It’s a tidy story that promises an easy fix.
It is also economically incoherent.
The central constraint on public-sector performance is not the power of unions—it is chronic underinvestment. For decades, policymakers have allowed public-sector pay and prestige to fall behind comparable private-sector jobs and have outsourced key functions that should have been performed by skilled civil servants, not profit-maximizing private contractors that are the real source of excess costs for state and local governments. The predictable results have been staffing shortages, uneven service quality, and degraded state capacity—not because we are paying too much, but because we have been trying to get government on the cheap.
Start with the most basic implicit claim Bagley and Gordon return to again and again: that public-sector unions have extracted excessive compensation and resisted efficiencies at every turn. If that were true, we would expect to see the total compensation of public-sector workers rising as a share of the overall economy. In fact, the opposite has happened—the combined compensation of public employees has shrunk noticeably as a share of national income for the last quarter century.
To be sure, policymakers should always aim to deliver value for taxpayers. And—just as in the private sector—there are surely instances where some public employee’s pay is out of line or workers resist useful improvements. But if overpayment for services delivered inefficiently was a general feature of the public sector, their aggregate compensation wouldn’t be shrinking sharply over time.
Bagley and Gordon support their claims with a shotgun blast of anecdotes about public-sector unions able to muscle excess pay out of colluding Democratic politicians that are almost laughably context-free. L.A. Mayor Karen Bass gave larger-than-normal raises to public-sector employees in 2024? I’d hope so—prices had risen 23% in the previous five years (this inflation had made some news) and private-sector pay for non-supervisory employees was up 28% over that time. The suggestion by Bagley and Gordon that these raises were untoward only makes sense if you actively want the desirability of public employment to crater relative to the rest of the economy.
Bagley and Gordon also note darkly that “more than half” of local government expenditures are paid to employees. So what? Local government spending is not like federal government spending where the overwhelming majority of it is simple transfer payments—sending checks to people (Social Security) and medical providers (Medicare and Medicaid). Local governments must directly deliver public goods and services, like public education. That’s going to be done by people who need to be paid. The private sector, too, devotes the majority of its spending to labor (in the corporate sector, labor’s share is well over 70%.)
Even the data they cite for this irrelevant point show that compensation—including the benefits that Bagley and Gordon decry—in state and local jobs is lower than for similar workers in the private sector. That gap matters. Public-sector employers must compete in the same labor markets as everyone else, and low relative pay for skilled workers in the public sector compromises the ability of public-sector employers to attract and retain highly effective workers.
This ignorance of how labor markets in the private and public sectors interact is the root of many of Bagley and Gordon’s economic misunderstandings.
Consider their discussion of education spending, where they note that California spends more per pupil than Mississippi. California does spend more per pupil in nominal dollars, but prices in California are far higher than in Mississippi. Even more importantly, private-sector salaries for college-educated professionals in California are much higher than in Mississippi—and those are the jobs that set the outside options that talented college graduates weigh when deciding whether to enter and remain in teaching. Put another way, it is competition from the private sector that determines how high pay must be to attract and retain high-quality teachers. Education researchers know this, and that’s why the generally accepted way to assess the sufficiency of education spending is not nominal dollars spent per pupil, but per pupil spending scaled to per capita GDP in a state. In forthcoming work we show that on this measure, California ranks 36th in the nation—lower than Mississippi.
This also shows why the Bagley and Gordon claim that “…blue states and cities often also pay state and local government workers more than similar jobs pay in red jurisdictions, even after adjusting for the cost of living” misses the point so spectacularly. State and local governments are embedded in their local economies and public-sector pay has to rise in line with private-sector pay in the economy around them, or the quality and quantity of available public employees will suffer.
The big problem over recent decades is that public-sector pay has not kept pace with the surrounding economy, which has made it harder to recruit and retain qualified workers. Teacher shortages, for instance, stem directly from the huge gap that has emerged in recent decades between what public school teachers earn and what comparable private sector workers earn, even in the highest-spending states. How would making these jobs lower-paying and lower-prestige add excellent new teachers and improve educational outcomes?
Another common complaint about the public sector is that it slows infrastructure projects. The public is often invited to imagine huge teams of paper-pushing bureaucrats gleefully stamping “no” on planning documents. But the clearest finding in empirical research about the drivers of higher-cost infrastructure is that costs have risen fastest where states reduced the number of transportation department employees. Fewer public-sector workers means that more of the planning work has been outsourced to more expensive private consultants.
Bagley and Gordon claim that when policymakers bargain with public-sector unions, there is no constraint on their incentives to grant union demands in exchange for electoral support. In reality, there is a crushing countervailing constraint—the overwhelming perception that voters are rabidly anti-tax. This results in a deep reluctance by policymakers to call for the level of revenue needed for public sector excellence. It is a far bigger structural problem today than any supposed excess power of public-sector unions.
Public-sector workers don’t just bear the brunt of underinvestment, they are also one of the few consistent voices arguing for robust financing of state and local governments, bargaining directly for the public good. They advocate for libraries to remain open in rural communities so that everybody has at least some access to the internet, for higher levels of K–12 education spending, and for proper training for EMTs and other first responders to ensure public safety.
Despite these efforts, public sector financing has been throttled in recent decades, and the results have been a predictable degradation of services. Even worse is coming, as the Republican tax and spending megabill will impose crushing cuts to safety net programs that states administer and jointly fund.
For decades we have been relying on the admirable intrinsic motivation of public employees to shield us from some of the damage of underinvestment—nurses, first-responders, and teachers going above and beyond the strict demands of their jobs to provide services they feel called to perform. But we’ve already asked too much while paying too little. If we want a truly excellent public sector—and we should—we need to pay for it.