Recent Posts

Wednesday, October 2, 2019

Reserve Army Of The Hungry

SNAP participation. (U.S. Department of Agriculture)
I have returned from the 3rd annual Modern Monetary Theory conference, and I am now catching up on things. One of the observations that stuck with me was some comments about the Supplemental Nutrition Assistance Program (SNAP) made by the presenter Maggie Dickinson. She noted that the welfare state did not contract completely since the 1980s, as some components have grown, with SNAP being an example of growth. While thinking about this, I realised that the rise in food assistance (such as SNAP, as well as food banks, which have grown in importance in the other "anglo" countries at least) helps explain the breakdown in employment statistics as an indicator.

Noting that the unemployment rate is "broken" as an indicator as a result of the changing structure of the labour market is not particularly novel. For example, Bill Mitchell's academic work has documented the structural changes in employment policies in the OECD since the early 1990s, and the rise of "underemployment." (Bill Mitchell is one of the founders of MMT, and I was on a panel with at the conference.) My concern is somewhat more stark: employment status as a whole has become somewhat meaningless as an indicator.

(Note: the panels are available on video at:

I will first discuss some of the more "technocratic" implications of my thinking first, then dip into political economic ramblings.

Is a Job at Starvation Level Wages a Job?

In the industrialised countries in the post-war era, providing workers with the resources to allow for "food security" was not considered that big a deal. Hunger was an issue for the jobless. Although I am not an expert on the Soviet system, I believe that even the 1980s Soviet Union managed to feed its workers (a major concern was that they were dipping too much into the vodka rations).

Courtesy of the economic policies provided by a highly educated economist class, this is no longer the case. We now have people who are employed, yet need to lean on assistance -- either from charities, or governments -- to get adequate nutrition. Meanwhile, I am unaware of any collapse in productivity in the agricultural sector that could explain this shift.

The question is straightforward: does it matter what the official unemployment numbers are if the people who are working are dependent upon food aid? If we look at the chart at the top of the article (unfortunately, no time for formatting), we can see the explosion of SNAP recipients in response to the Financial Crisis, and the slow rate of decay. (I did not have time to better present the data, but by comparison, the official number of unemployed on the U-3 measure is currently bouncing around 6 million people.)

It seems straightforward that if workers as a class had bargaining power -- which is what the monetary policy hawks have been warning against since 2010 -- SNAP participation would have fallen a lot faster.

As a result. it is no surprise that forecasts of rising inflation based on NAIRU estimates would fail. My concern is that it is unclear that even the underemployment indicators will become misleading in this environment.

No Indicator, 'Mo Problems

Why do we care about this breakdown as an indicator from a technical point of view? We want to forecast the wage inflation rate, based on the plausible argument that higher wages might eventually feed through into consumer price inflation (which is the usual definition of "inflation" that people use).

(As a disclaimer, there is a gap between wage inflation and CPI inflation, and rising wages may just represent an income share shift towards workers. However, it is hard to see how rapid wage rises can be sustained without rising consumer prices. [Thanks to Jerry Brown for spotting typo.] )

We want to be able to forecast wage increases. We typically look at some form of capacity utilisation measure -- like the unemployment rate -- to try to make such a forecast. However, if those measures break down, we have a hard time to measure the "pressures" in the labour market. We have to infer the level of utilisation by looking at wage inflation. That is, we are trying to forecast wage inflation by looking at -- wage inflation.

Luckily, I am not in the forecasting business, so the circularity of that logic is not my problem.

How "Inefficient" is the Job Guarantee?

I will now turn to some political economy issues. I will first start with an argument about the Job Guarantee, that I have seen advanced by Simon Wren-Lewis. Since I do not have a reference handy, I will just paraphrase his argument here.

The argument is that the number of people in the Job Guarantee (the "JG rate") will be higher than the number of unemployed people "all else equal" when we have inflation stability. The premise is based on search theory, that people find unemployment more unattractive than a JG job. As a result, it will take greater wages to coax them out of the JG programme. If we assume that something like NAIRU exists, we need a greater number of JG workers to get the same anti-inflationary effect as a pool of unemployed.

I always had technical complaints about Wren-Lewis' argument. However, my thinking about this SNAP data is causing me to go internet Austrian, on top of my technical concerns.

Under current institutional arrangements, the presumption that private employment is efficient is misguided. We have employers who are incapable of paying their workers a starvation level wage, and this is obviously unsustainable. However, charities and government assistance are bailing them out. These firms are destroying capital, and are undermining the profitability of legitimate firms. Euthanising these firms will raise domestic productivity. A Job Guarantee that pays a living wage will accomplish exactly that, whereas the "reserve army of the unemployed" strategy preferred by neoclassicals leaves those firms in place.

I have some doubts that there are no flaws in this logic. However, it underlines the shakiness of "all else equal" arguments when we are attempting to make structural changes to economic institutions.

(c) Brian Romanchuk 2019


  1. " However, it is hard to see how rapid wage rises cannot be sustained without rising consumer prices."

    Should that be 'can be sustained'?

  2. Welcome back to Blogsville after your MMT conference hiatus.

    The SNAP data opens a window into a puzzling question: In a world where we must eat to live, do workers, when working, merely eat to a better standard?

    I guess we could extend the thought to miles driven, kind of car used, and housing. Hmmm, maybe working primarily relates to daily living standards?

  3. Thanks Brian. This is an interesting inquiry - really a bit of a different look at SNAP, correlating or connecting those data to the unemployment underemployment numbers. It will take me longer to think about your thesis than your other readers.

    one thing that has bothered me is that as I recall, the unemployment rate calculations were redesigned (?) during the (I think) Obama administration. I think the real unemployment rate after including those who had simply given up (excluded from the official (now modified) unemployment rate) was 4-7% higher. Perhaps this is captured with the SNAP recipient count.

    and one more thought: there are about 15+ other nutrition programs across 3 federal agencies. I would expect a similar trend line in those programs.

    Thanks again!

    1. The only thing I can think of during the Obama administration was that states changed the rules for unemployment insurance. This caused the amount of jobless claims to collapse to unprecedented lows.

      Earlier, there were changes made, possibly in the Clinton admin, if not in the 1970s or early 1980s. Working from memory, the choice of which unemployment rate is the “headline” number was switched.

      To explain, there are multiple definitions of unemployment, ranging from U-1 to at least U-6. (Working from memory, there...) The “headline” is U-3, which is somewhat strict. The term “headline” refers to the fact that it is the number that is used for “unemployment rate” in the BLS news release, and how it is reported. It does not capture things like underemployment (working part time, but want full time), but underemployment shows up in U-6. U-6 did give a more realistic picture than U-3, but my argument here is that it might be missing some key changes to the economic structure.

  4. No one picked up on my question "In a world where we must eat to live, do workers, when working, merely eat to a better standard?" so I will see where I can take it.

    First, the scene. In Brian's chart SNAP is reported in terms of unadjusted population. SNAP currently takes the form of a debit card that can be used to purchase food, thereby extending the food budget of eligible recipients. For 2018, the average SNAP benefit was $126 per person per month. The program targets low income persons but includes working persons.

    A federal government program, the payments were larger than possible if the program was limited by tax receipts. The surplus benefits were funded by borrowing without intention to reduce future debt or by outright creation of money.

    With scene set, what does this SNAP data say about unemployment?

    Not much in my opinion. The data speaks more to expected standards of nutrition compared to income received no matter what the source of income.

    Who benefits from this program? The most direct benefits go to recipients. Next in line: food producers and purveyors. Farther down the line: all other producers, vendors and sales tax collectors. Last in line: Long term savers (individuals, pension funds) who delay spending until they have need for the newly created money.

    SNAP spending into the economy should increase the number of jobs available but nothing is said about what wage levels might be. In all probability, some of the jobs that are held (maybe part time) by SNAP beneficiaries would be due to the SNAP program itself.

    Brian mentions the MMT Job Guarantee program with better paying jobs for participants. Presumably these JG jobs would include a pay rate that would make SNAP benefits unnecessary for participants. I can't help but wonder what JG workers will be doing. Would they displace present low income workers or compete with them by spending more newly created money in the market place?

    It would be nice if the SNAP program was not needed as a result of work related income that was more evenly distributed. That said, many of the SNAP recipients cannot work for some reason. An even distribution of work related income would not be an instant solution.

    I will end this by saying: I really don't see much relationship between SNAP numbers, unemployment numbers, and the MMT JG program.

    1. Roger, most descriptions of the Job Guarantee rely on wages for the work to be set at some socially determined standard of decency rather than being determined by 'the labor market'. Hopefully society would set a standard that allowed wages to be adequate to feed and house and clothe oneself if you are working.

      Brian has pointed out that some current private sector jobs do not even pay enough for that minimal decency. And that those employers are, in a real sense, being subsidized by society through our various social programs, without which, the employee would not be able to maintain even that minimal standard. And this allows their employer to be more competitive because of the lesser labor costs. And undermines the ability of firms that pay decent wages to stay in business.

      One way of understanding the Job Guarantee is to consider the wage paid as the de facto minimum wage plus the opportunity to always have work at that wage if desired. Hopefully it would compete with crap jobs that don't pay enough to support oneself and drive their wages up.

    2. Not sure how to answer. My argument is that we ended up in a situation where being employed is not a meaningful indicator for well-being. This is very different than it was in the 1950s-1980s, where the conservative mantra was that getting a job (and getting married) was the best anti-poverty programme. Now, you can get a job and still be living in poverty. This greatly changes the bargaining position of workers, and hence the breakdown of NAIRU.

      In an ideal world, the JG wage would be a living wage, and return the situation closer to that earlier era. However, a JG has not been enacted, so it is far too early to speculate on what the level of the wage will be.

    3. It may be too early to speculate what level of wage of the JG would be if we ever even got something like it into actual policy, but it is not too early to think about the range of wages that would allow it to fulfil the economic purposes that MMT says it will. I mean it seems to me that there is some range of wage levels that would enable a JG to act as a buffer stock of labor that might modify inflationary tendencies like MMT says it will. And that there could be wage levels that were either too high to do that because the private sector could not entice people out of the program, or too low that people would ever use the program in the first place.


Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.