fb-pixel Skip to main content
edward l. glaeser

Towns have too much power, but not the power they need

What good is there in having scores of publicly managed micro-hedge-funds dotting the Commonwealth? In Massachusetts, while state law mandates the terms of public pensions paid by local governments, local investment boards still manage their portfolio decisions. This makes little sense. Local investment boards have little upside, and have historically earned lower returns. There is certainly a place for local experimentation in pension matters, but the management of pension funds isn’t it.

We’re in the middle of a much bigger debate on political centralization. Europe must decide whether countries will cede further fiscal and regulatory powers to a central authority, or whether individual nations like Greece will regain control of their currency. Within our Commonwealth, Boston chafes under the limits that the state places on its ability to tax restaurant meals and issue liquor licenses.


The same tension exists in public-sector pensions, which present a particularly precarious balance of state and local control. Since 1945, the terms of public-sector pensions — for instance, how large a pension a teacher can expect after a certain number of years of service — are determined by state law, and a state commission determines how much communities must set aside for pensions. But localities can run their own investment funds, as long as their returns aren’t too low. This is a bizarre set-up, since local investment boards are unlikely to out-invest the state system.

Good investments require expertise, and there is no advantage from having little laboratories of financial engineering. Using data provided by the state pension authority, Harvard graduate student James Mahon has found that since 1986, the systems outside of the main state fund have, on average, earned lower returns during every five-year period. This doesn’t mean that the local systems are boobs — some earn lower returns because they take on less risk — but the sheer number of funds is illogical.


Local governments should still have a role to play in the pension system. It’s just not the one they’ve been assigned. Even as the state gives them too much power over investments, it gives them too little power over pension terms. In fact, they could help in addressing our big pension problem: that we pay public workers too little current salary and too much retirement income. Young teachers and policemen need cash today, not large pensions. Research by Maria Fitzpatrick of Cornell shows that the typical public employee is willing to give up just 19 cents of current compensation for each dollar of expected future compensation.

We pay public workers with pensions not because workers want them, but because pension costs are easy to hide. Before 1983, local systems only paid for pensions when they came due, which was a disaster. Even today the money we set aside will suffice to cover our costs only if our investments earn a whopping 8.25 percent annual return — far more than you would assume if you were saving for your kids’ college tuition.

Local control would allow more experimentation with defined contribution plans, and more bargains that could increase workers’ wages today in exchange for lower future pensions. But communities shouldn’t just be left to their own devices either. Before the 1945 law that unified pension conditions across the state, communities were either too generous — giving pensions to the politically powerful without requiring any employee contributions — or too stingy — providing no pensions whatsoever.


The core dollar amount that localities now contribute to worker’s pensions should be preserved, and there should also be a strict lower bound on the generosity of the traditional defined-benefit plan. Our public workers need that cushion, because they are not in Social Security. Above that base, though, local communities should be free to offer different deals, such as allowing workers to take 401(k)-style defined-contribution plans, instead of defined-benefit plans. This experimentation may help us explore lower-cost options that deliver more of what workers actually want: to be paid now, not in the distant future.

We need a combination of state and local control over pensions, but currently we have the tasks allocated in exactly the wrong way. The state should manage every community’s investments, but communities should have more power to negotiate innovative terms for their workers’ pensions.

Edward L. Glaeser, a professor of economics at Harvard University, is author of “The Triumph of the City.’’ His column appears regularly in the Globe.