Mr Medlock and the Classics: Part II

Welfare positivity!

Welfare Positivity

Having figured out how we raise revenue, what are we spending it on? Welfare is the obvious answer - but what is welfare actually for? It is primarily a form of social insurance - so the question we’re really asking is why social insurance is needed over private insurance and why it achieves functions that cannot be replicated by charity.

Firstly, there is a moral argument that people who should not be punished for things outside their control - for example, it is reprehensible that there are children without healthcare coverage through no fault of their own. Secondly, there are improvements in economic efficiency by correcting for market failures and incompleteness - social insurance allows the pooling of risk, it means people can take entrepreneurial risks and drive the process of creative destruction, it creates more efficient labour matching and it helps people engage in income-smoothing over their lifetime. Thirdly, it prevents the rise of reactionary backlash caused by creative destruction. Fourthly, it redistributes money towards those who are likely to get more utility out of it. Fifthly, it frees people from familial dependence.

There are some who argue that we need work, not welfare. And intuitively, this sounds nice - why give handouts when we can give people jobs? Isn’t this more dignified and more cost-effective? The problem with this is that non-workers make up a significant portion of the population - around 49% of Americans work 0 hours a year, and about 86% of them should not be working i.e. they are children, elderly, disabled or students. And since the poverty rate for non-workers is thrice that of workers, it becomes apparent that the way to deal with poverty is welfare, not work.

Recognise that most of the aforementioned five benefits are best achieved with universalism. For one, universalism ensures horizontal equity, where those with similar levels of income and wealth are treated equally. Universalism also means a reduced level of bureaucratic burden, which results in fewer poor people being inadvertently excluded from benefits. It creates popular buy-in, generating political sustainability for these programs while reducing the stigma of using them by delinking feelings of redistribution from the welfare state - insofar as welfare budgets are endogenous to the targeting regime, this can increase the total amount available to be redistributed. And it makes sure there aren’t welfare cliffs, where there is a high marginal tax rate that disincentivses work. Furthermore, universalism has institutional benefits - its simplicity requires less state capacity and allows a better division of labour, where the welfare side only has to worry about underpayment instead of obsessing over clawing back overpayment, since tax authorities are the ones focusing on collecting revenue.

Why does means-testing suck? For one, it often comes with all sorts of ordeals that are designed to make life more difficult, which themselves create deadweight losses. An example would be the incredible list of conditions in order to be eligible for SSI disability payments, all of which makes applying and getting help nigh impossible. Another would be the 47 page Medicaid renewal application form in Tennessee that has caused 250,000 children to lose coverage. In fact, means-testing can basically act as a tax, exemplified by the existence of various deductions for food stamps. The consequence of all of these restrictions is it suggests we don’t trust people. Not only does this make people’s lives difficult, it reduces social trust. And it isn’t even justified, because the narrative of “welfare queens” is not based in reality - by and large, welfare programs lead to people spending the benefits well.

A second issue with means-testing is that income volatility makes this targeting difficult - within a year, half of those below 200% of the poverty line will have seen a shift in their eligibility for Medicaid. That’s 28 million people who will be fluctuating in and out of just one welfare program. And the result of falling off these eligibility cliffs, even when there is a household income increase, is of significantly increased food insecurity and healthcare unaffordability. While means-testing within programs would be an improvement compared to means-testing into programs i.e. we could let those over the eligibility limit pay a small fee to stay on the program, universalism is a much easier option. A third harm is that means-testing results in a smaller tax base, meaning higher marginal rates. And since deadweight losses are not linear to the tax rate, we will get a larger amount of deadweight loss overall.

And fourthly, the actual process of means-testing can often result in very badly designed welfare programs. The most prominent problematic mechanism is that of trapezoids - essentially, this is where the benefits increase with income for a bit, then flatten out and even are phased out. The problem with this is not only that the very poorest are excluded, but also that this can incentivise bad practices - for example, the trapezoidal Earned Income Tax Credit may have encouraged mothers to spend less time caregiving, resulting in worse child test scores. Another problematic structure arises from the Child Tax Credit, which is actively regressive, excluding 90% of children in the bottom income decile while including nearly all children in the top half of the income distribution. More generally, means-testing results in chaos, due to the need to claw back overpayments, the delay in payments causing liquidity problems and hardship for families, as well as the uncertainty in incomes making budgeting a nightmare.

Incidentally, universal welfare provision isn’t just better than means-tested approaches. It is superior to the rent seeking that can occur when welfare is segmented on an occupational basis, as those often result in spending being directed towards pensions instead of more poverty-reducing measures. And it’s also better than benefits which are tied to employers, as they reduce labour mobility and exclude non-workers.

But for all of this analysis, is welfare actually the most important factor in stopping poverty? Yep - there’s a strong negative relationship between public social spending and post-transfer poverty. And we can also see this in the relationship seen between Social Security expenditures and elderly poverty or in the results of the War on Poverty or in the CARES Act. In fact, we can see that the US, despite leading the pack in post-tax-and-transfer poverty, is only the middle of the pack pre-tax-and-transfer. That is, it is the welfare state that makes the difference.

So should this welfare be provided in kind or in cash? Insofar as providing cash of the same value will always provide at least as much utility (and grants more choice), this is fundamentally a question of trust. Because poor people are very much financially literate, cash transfers generally reduce the consumption of temptation goods and pay for themselves. So the party we should distrust isn’t those we are distributing to, but the people who decide the restrictions on in-kind benefits, producing ridiculous policies about sodium-content in tofu and legume types in the WIC program. Indeed, this regulatory overreach is how we got a shift to purely digital payments for WIC which resulted in fewer stores participating in the program.

Now if we are providing welfare as money, how should this be dispensed? As a Universal Basic Income? Or as a Negative Income Tax? There are many people who support a NIT, because it is progressive in a way that the UBI isn’t. But once you recognise that progressivity is a term that only makes sense in reference to one’s income, you realise aUBI is progressive. In fact, a NIT and a UBI are basically equivalent. The difference comes down to implementation, where we’ve already established that a universal program like UBI is likely to provide a better administrative division of labour.

In practice, we can see these principles playing out in ideas like auniversal child allowance. This has all the trappings of universalism, but is helpfully targeted in a way that is most effective at poverty reduction i.e. for children. And as with tax policy, there are lots of administrative fixes to be made such that welfare brings joy. One of the biggest changes we could get is a far better system of cash transfers via public banking, but even within programs, there are lots and lots of possible minor tweaks.