Means testing is always bad
My preferred formal definition of politics is the authoritative allocation of resources. Resources — money, healthcare, oil — are subject to scarcity, and thus politics becomes the authoritative allocation of scarce resources.
On the optimal distribution of these scarce resources, different people hold different moral values and political judgments. Should society provide assistance to the homeless in securing housing? Should we levy additional taxation on those who contribute to climate change, to make felt the hidden social cost? Should a child die because their parents cannot afford healthcare? (Yes; yes; no.) Nonetheless, because of the diversity of political opinion, categorical statements on the optimal allocation of scarce resources that can be broadly agreed upon are hard to come by.
But there’s one belief that I think should be accepted by anyone who, in good faith, believes in trying to maximize the efficacy of benefits-based public policy programs. (Yes, this caveat may excise a sizable portion of American politicians). The idea is: means testing benefits is always bad.
Means testing is the checking of eligibility for government assistance programs through income or wealth based criteria. In the United States, it is most predominantly associated with eligibility for Medicaid, food stamps, and public housing. More recently, eligibility to receive pandemic stimulus checks was capped at a certain level of income ($75,000 for single earners) And right now, Senator Joe Manchin is demanding additional means testing as part of the Democrats’ social safety net bill.
The problems with mean-testing are myriad and have been broadly written about elsewhere. To summarize some arguments:
Means testing creates bureaucratic hurdles that exclude eligible people. Not everyone who is theoretically eligible for a government program will know to apply, will have the required documentation, and will have the time to apply. Anyone familiar with statistical modeling will know about precision vs. recall tradeoffs — there is a similar effect here, where in order to effectively limit the benefit to those “eligible,” a broad portion of those eligible may effectively be excluded from receiving the benefit. According to an article in Jacobin, “the overall participation rate of [those eligible] for the food stamp program is 85 percent and is only 75 percent for the working poor who likely have a harder time proving their eligibility to the welfare office.”
Means testing makes programs vulnerable to cuts. Once an income threshold to receive a benefit is in place, lowering that threshold becomes a popular target in order to reduce government spending. Universal benefit programs are more popular than means-tested programs, and the implementation of means testing can create social stigma surrounding programs like food stamps.
Means testing creates high marginal tax rates that can trap people in poverty. Because increasing one’s income can result in a benefit cut under means-tested programs, for many people increasing one’s earnings can result in a net loss of earnings plus benefits. A recent study found that one in four low-wage workers face effective marginal net tax rates above 70 percent, locking them into poverty.
The counterargument for means testing is that it is efficient: benefits are targeted directly to only those who need them. If everyone including the wealthy is eligible for a program such as free college tuition, as the argument goes, then this is not an efficient use of government resources.
But who cares if the wealthy get some benefit from universal programs? If someone is dissatisfied with the level of program usage by the wealthy and uses that to argue for means testing, then simply raise marginal tax rates for high earners and have the increased burden to pay for these programs on the wealthy. That is, if one is controlling the levers of policy by implementing means testing thresholds, they can just as simply raise marginal tax rates in the right spots to accomplish the same effect of targeting a program to the poor. But in effect they have now created a universal, more popular program.
The perceived problem that means testing purports to solve — that those who don’t need assistance might take advantage of the program — isn’t actually a problem. Using other redistributional levers through taxation accomplishes the same income-distributional effect of targeting a program — but does so in a way where the resulting program has greater popularity and efficacy.
Note: Considering this blog often concerns product analytics, apologies to any reader who saw the title concerning “means testing” and thought this post concerned some kind of statistical test about the average of a variable.