This is reposted from Medium where it seems to have died the death without any readers:
This sounds like a bit of a waste of money to be honest
Yes, obviously, it’s primary season and promises on the stump don’t have to make a great deal of sense, they just have to gain traction. Yet a proposal to spend some $96 billion to achieve precisely and exactly nothing does seem a bit much even by these standards. This is also what Senator Cory Booker is suggesting we do.
OK, it’s primary season and I am also being a little prejudicial. Spending $96 billion on alleviating rent bills for poor people will obviously have an effect. It’ll certainly alleviate some poverty, may well push up rents a bit and the taxpayer is clearly going to be on the hook for the amount. The problem is that by the system of measurement being used to propose this policy it will achieve nothing. Which is indeed a problem.
I have banged on so much about this that some have taken to calling it Worstall’s Fallacy – to propose a solution without taking account of what we already to do solve the problem under discussion. My ego isn’t large enough for me to have so named it but it is for me to propagate the name. It’s also a large and distressingly common problem in American politics. We should be demanding better from those who would rule us.
The most common form of this is in discussion of poverty. That Official Poverty Measure came from Molly Orshansky in the early 1960s. She noted, correctly, that people spent about one third of their income on food to be eaten in the home. So, if we take three times an economic but nutritious diet then we’ve got a useful, if rough and ready, measure of what is the poverty line.
She was right, it is rough and ready and it is and was useful. It was meant to be very temporary indeed, a holding place, a marker perhaps, while a proper definition of poverty was worked out. That never did happen and we’ve been stuck with the OPM ever since. Milton Friedman’s point about the permanence of temporary government measures seems appropriate here. Ever since then we’ve just upgraded it by the general inflation rate and that’s still how we define poverty today.
There are problems with this, obviously enough. We’re a much richer society and three times that early 60s diet might not be an appropriate measure of those being left behind any more. Adam Smith had something to say on this with his linen shirt. Lacking the funds to buy one doesn’t make you poor. But living in a society where not being able to buy a linen shirt makes you poor, by the standards of those around you, means that you are indeed poor in that society if you can’t afford one. Poverty is, to some extent at least, relative, not an entirely absolute concept. It’s also true that there’s absolutely no one at all in the US living on the global definition of absolute, or extreme, poverty. That being $1.90 a day at current US retail prices. It’s also the value of consumption — so includes home production like growing your own food — not a measure of cash income. There really is no one at all in the US living on this. The homeless do better than that.
There are other problems as well. Food bills — again for home consumption, we’re not talking about eating out here — are more like 10 to 12% of US incomes these days. Meaning that if we used Orshansky’s measure today of 3x the food bill the poverty line would be a third of what it currently is. Or, obviously, we’d have to say that poverty was an income 10x the food bill in order to get to the same cash number we currently use.
But the much bigger problem is what we use as our definition of income here. It’s cash income. This means that it’s market wages, obviously enough. Plus social security income. Any cash that comes through TANF. But what are the major poverty reduction programs in the US? Those would be, in rough order, Medicaid, the EITC, Snap (or food stamps), Section 8 housing vouchers and the tailing off into the plethora of more minor schemes like cell phones and the like. None of which we actually count in our incomes of the poor.
For we count only cash income. We do not count anything that comes through the tax system — thus not the EITC — nor anything that comes as goods or services in kind. Snap is special money that only works for food — not counted. Section 8 only works on housing — not counted. Medicaid is only medical care — not counted. Cell phones are cell phones — not counted. Our measure of poverty doesn’t include most of the things that we do to reduce poverty.
Sure, including Medicaid in here can be argued about but all told those add up to the thick end of a trillion dollars a year. Which reduces a lot of poverty. Absolutely none of which we count when we define how many people are poor when we look at the poverty line and the number of people below it. There are attempts to correct for this and when we do we find that the child poverty rate of some 20% falls to more like 2%. Which is pretty good going for government work.
It also means that we can’t compare the poverty rate over time. For back in the 1960s welfare was pretty much all cash based. As cash flowing in it was included in those incomes used to calculate the poverty line and the people under it. Since the 1970s — the EITC got going mid-decade – we’ve had a bipartisan agreement that whatever it is that we’re going to do for the poor should come either through the tax system or in goods and services in kind. Bill Clinton’s welfare changes were only a part of this.
Our poverty measure has rather changed therefore. Back when it was largely, and not entirely, the number of people still poor after whatever help was on offer. Today it’s largely, and not entirely and completely, a measure of how many people would be poor before the help and aid that is on offer. Recall, we measure poverty before the $70 billion the EITC sends to low wage earners, before the $80 billion on food stamps and before whatever ungodly number it is that Medicaid costs.
Please don’t start shouting that those numbers aren’t accurate. I know they’re not, they’re there just to illustrate the point, not to try and give a reading of the current budget. About the right number of zeros after the first digit is all I’m aiming for here, along with getting that first digit close enough.
So, to start arguing that we must spend more on child poverty alleviation by pointing to the official poverty measure — that 20% of the kids — is to commit Worstall’s Fallacy. It’s to propose a solution without accounting for what we already do to try to solve the problem. Our actual remnant number, after what we already do, is closer to 2% of children still in poverty.
Sure, we can use other poverty measures. Nearly everyone else in other countries uses 60% (or in international comparisons, 50%) of median household income adjusted for household size. That’s really a measure of inequality, not poverty, but see above about Adam Smith’s linen shirt. It also includes all income, including that which arrives through the tax system, credits like local equivalents of the EITC. And goods and services in kind although not, often enough, medical services. Foreign measures of poverty are of what remains to be solved, American largely of what exists before partial solution.
We can even use the Supplemental Poverty Measure which is a newish thing from the government. It’s a relative measure and it includes much but not all of what is done to reduce poverty. Because the US is a more unequal place it’s always going to be higher than foreign measures.
But the point to take ahold of here. If we measure US poverty by that normal measure, the official poverty measure, we get wildly different answers if we look at it in the raw or after the things we do to alleviate it. And yes, I’m sorry, this matters. Say, for example, that we thought we’d go and reduce working poverty by doubling the EITC. I think that would be a good idea. It’s true that some of it becomes a subsidy to low wage employers but best estimates are that that’s about 30%. The other 70% does indeed flow through into raising low end incomes. This is a good thing. But what does this do to the number of people we define as being in poverty? Nothing. By our usual and official measure we’d be spending another $70 billion a year and we’d not have moved one single person out of poverty.
See the problem?
Which brings us to Cory Booker and his suggestion about housing and rents. It’s here, on Medium. The idea is that some half of all renters are rent stressed by paying 30% or more of their income to put a roof over their heads. We’d like to alleviate this. So, a tax credit which would benefit some 20 million families by about $4,800 each. That’s where the $96 billion comes from.
Which is to fall prey to Worstall’s Fallacy. We’re not including the effects of what we already do to house the poor. And we’re also not going to move the number, as we measure it, of those who are rent stressed by one single percentage point nor even one household nor person.
I am one of those very sad creatures that reads footnotes to these sorts of proposals. Even looks up the source reports to see what they’re saying. These numbers on rent stress come from The Joint Center for Housing Studies at Harvard University. The latest version is here.
About which my first comment is that I must have something wrong. Because they really can’t be measuring income in the way I’m reading they do. I’m getting the impression that they’re not recording household income at all. Rather, they’re looking at what jobs pay. Then at what rents are. And saying that this job can’t pay that rent.
But household incomes aren’t based upon one job. We can, as some people do, insist that the one minimum wage job should be able to raise the traditional nuclear family of two adults and two kids. But even if we do that’s an aspiration, not a measure of what people are actually doing.
I’m also very puzzled by the idea that a single worker two adult family faces significant childcare bills. A two adults working household, yes, but then we’d need to have two incomes there, not just the one.
Perhaps I’ve misread that and if I have please do correct me. But one bit I’ve not misread and I went and checked it. The answer I got:
…the above surveys record self-reported income measures for persons and households from a wide variety of income sources (wages and salaries, ssi, rental, interest, and dividend income, etc.). In general, income measurements:
Are pre-tax (so excludes benefits administered through the tax code)
Exclude most in-kind transfers (like housing vouchers, Medicaid/Medicare, and usually SNAP)
Include direct cash assistance like TANF
The definition of income we’re using here doesn’t include what we’re already doing to alleviate housing poverty. Section 8 alone is some $20 billion and there are no doubt varied state programs which add to that Federal amount. Hey, sure, maybe that’s not enough and we should do more. But we’ve got to start from what is the situation out there on the ground, not what it would be if we’re not doing anything at all. The stress is also before whatever effect the EITC and Snap have and so on.
But note the next point as well. Next year, after we’ve gone and spent $96 billion on tax credits to aid with the rent, those 20 million households will be recorded as not having been helped at all. Our measure of housing stress will not have changed by one single person.
Now do you see why Worstall’s Fallacy is important? We can’t possibly know what we should be doing unless we account for what we’re already doing.
Now, that The Donald is currently in office is proof perfect that inattention to policy detail isn’t a disqualification for that highest one. But of those asking to occupy it we should perhaps ask for a little greater knowledge of what is being said, no? And a plan to spend $96 billion to achieve precisely and exactly nothing seems worthy of a little questioning.
Perhaps one of the journalists covering the Democratic primaries would like to ask the Senator about it?