There is, or at least can be, value in treating economic variables in the way econometrics does for the purposes of understanding generalized behavior. The problem for Economists, these statisticians, is that they’ve turned stylized lessons drawn from regression analysis into literal rules defining their worldview.

By 1957, Milton Friedman had already been busy publicizing just those. Positive Economics was meant to become the scientific standard for how economists might better understand – by testing their knowledge – the complex systems of the real economy and how they often interacted. To attempt to make this “soft” social science into a hard one like Physics required, Friedman reasoned, rigorous mathematics.

In the area of consumers and spending or income, how does any scientist reduce human behavior – all kinds – into one or a few variables purposed into an equation? The answer: tons of subjective assumptions, the origins of our 21st century Greek tragedy.

Friedman in ’57 published his account of the Theory of the Consumption Function, contributing his Permanent Income Hypothesis (Chapter 3) to it. To test his theory(ies) required compacting and standardizing such behavior into “consumer units” before then proposing how the “average” or “typical” unit might act. Once decided by statistical analysis, then economists could divine further implications by solving for X.

As it related to the possible disparities between consumption and income arising from the complexities of real world dynamics, Friedman required an exhaustive explanation and series of calculations for what otherwise might seem a simple knowledge problem:

For example, if Mr. A’s measured income fluctuates widely from year to year while Mr. B’s is highly stable, it seems reasonable that Mr. A’s measured income is a poorer index of his permanent income than Mr. B’s is of his.

The plain truth is that not all of our income is steady or near-fixed, nor are our expectations for it over any timescale you might choose (long run or short). The key issue, then, is how an economist might be able to piece together a more accurate puzzle of the economy presented by such complications for what may be its major piece: consumption.

Put another way, would we expect “transitory” contributions to income to raise consumers’ expectations for their “permanent” income treating, as we might, permanent income as an average? After all, any increase whether temporary or steady raises the average; the only question is by how much which becomes a function of time more than anything.

As Friedman wrote in his first paragraph:

The magnitudes termed “permanent income” and “permanent consumption” that play such a critical role in the theoretical analysis cannot be observed directly for any individual consumer unit. The most that can be observed are actual receipts and expenditures during some finite period, supplemented, perhaps, by some verbal statements about expectations for the future.

For him, there seemed a crucial difference between them; that not all or even much “transitory” income might contribute to expectations of, and for, permanent income leading toward heightened (or curtailed, in the case of transitory declines in income) consumption.

This is easily understood as the windfall problem.



The evidence had shown more than sixty years ago, and still shows today, that consumers treat transitory income differently than they might changes, or expected changes, in the more permanent (better characterized as durable?) sources, jobs chief among them.

While both intuitive and uncontroversial, the problem for Economists (those who use this sort of mathematics to masquerade as social engineers) is simply this:

The purpose of these remarks is not to demonstrate that a zero correlation is the only plausible assumption—neither evidence like that alluded to nor any other can justify such a conclusion. Its purpose is rather to show that common observation does not render it absurd to suppose that a hypothesis embodying a zero correlation can yield a fairly close approximation to observed consumer behavior.

The “zero correlation” Friedman referred to was between transitory income components and consumption for the typical consumer unit. In other words, a zero correlation would imply no change in consumption should be expected whenever an average consumer registers a windfall (or a temporary furlough).

On the contrary, there is evidence historically that there is some relationship between “unexpected” and short run changes in income when related to proclivities to spend (or invest in financial assets). Into such gray area pours all sorts of mathematical assumptions as to just how much of a relationship there might exist – and what sorts of current conditions and future perceptions might contribute to that relationship.

Friedman argued that it wasn’t much; later especially neo-Keynesians like to fatten their multipliers considerably more despite the fact that, all these years later, while still not a zero correlation the world economy tends to come out much closer to it than especially fiscal agents might wish (which was Milton’s central warning).

Another good example is playing out right now right before our American (and European) eyes. The federal government, as you know, is kicking off massive payments in helicoptered bunches; three so far, with two fully completed. The question on everyone’s mind, especially in the inflation debate, is whether the correlation between such transitory payments (as are undoubtedly understood in consumer expectations) and consumption (both goods and services) while not huge could be further from zero than otherwise understood.

In one sense, what other choice do millions of consumers have since as workers near ten million of them no longer have jobs therefore nothing other than Uncle Sam’s transitory payments to depend upon; they’ll surely have to spend every penny, or so it might seem. Yet, we have to recognize that though so many tens of millions are depending on the government now for so much, at the margins, this monumental multitude still might not act in the same way they would have had the income been drawn from steady work.

The data is pretty compelling if not actually conclusive, particularly for Helicopter #2 ($600):



The red line (Real Personal Income excluding Transfer Receipts) is a well-documented proxy for mostly labor but also regular financial income (including proprietors’ receipts). In other words, permanent components rather than transitory. As you can see above, spending has followed far more closely the red line rather than the orange – which is overall income from any and all sources, the key difference since last year the “transitory” payments from the government.

In January 2021, there obviously had been a positive effect on spending (see: retail sales) due to the second of the helicopter drops – but, nearer zero correlation, maybe not as much of one (see: services) as might have been hoped/expected from fiscal social engineers.

Or markets, many of which appear to have been picturing therefore pricing an almost-too-easy road out of this mess underwritten by such government-funded largesse, much of that optimism pinned on the next one ($1400) in particular.

If Friedman’s hypothesis instead stands with more data and helicopter payments further in hand, then expectations for further consumption contributions from the same moving forward would, at least should, be downgraded accordingly; a reason, a good and historic one, for possibly pausing all the inflation hysteria and reflation certitude that had been running rampant unchecked almost everywhere (mainstream media most of all).



Further inflation data (lack of inflation) merely confirms these more grounded suspicions; as the CPI had already shown up to the month of February 2021, the PCE Deflator estimates which accompany the personal income and spending data again display just how little inflationary pressures are evident in the real economy right now (and, to reiterate, it does not matter that M2 was up 27% year-over-year in February, adding more to historic gains lasting for an entire year). Starting next month’s inflation base effects notwithstanding. 

This is, and remains, a highly sick economy (not from COVID any longer) that cannot be fixed by “transitory” contributions – even if well-intended. This is not an argument against aid and stipends; on the contrary, the nature of the sickness, and the degree of the malady as well as its type, practically demand as much or more.

The overall issue, then, is entirely “stimulus” and what to expect from it. Quite simply, not much. Never much. As the data above shows, once again, until the labor market is fixed (and it remains more than an entire Great “Recession” worse off at the one-year mark) all else is just transitory conversation as well as small consumption.