Thursday, February 19, 2009
Stimulus, Multipliers, and The American Recovery and Reinvestment Act of 2009
What creates wealth?
Suppose for a moment an island economy of four agents. Of the first three, each produces a unique good (e.g. one produces clothing, one produces food, and one produces drink), and all are fully employed, i.e. maximizing production. There are no other goods, and the fourth person, unemployed, produces nothing. Suppose too that the people on this island use a fiat currency to facilitate trade (three traders is a sufficient number to ensure that double coincidence of wants isn’t guaranteed).
How might the plight of the unemployed person be alleviated? There seem two answers: 1. Goods could be transferred from the other individuals to the fourth, i.e. redistribution of the total output.* 2. The fourth individual could work and add to the total product.**
Now let’s subject our little economy to a couple of thought experiments. In Experiment 1, suppose we take additional units of currency and give them to the unemployed fourth individual. It doesn’t matter whether we take them from the other three, create them, or borrow them from abroad, just that we stimulate this economy. What happens? The individual purchases goods from the other three, redistributing the total product; if we’ve created or borrowed new currency in the process, prices ought to increase as well. Regardless of how we finance our stimulus plan, the total output is unaffected, and the fourth’s gain is offset by the reduction in the others’ consumption.
Next consider Experiment 2. We take the additional units of currency and give them to the fourth individual, with the proviso that s/he produces something the others will want to consume and sell them in the market as well. In this case, the total output is increased, and the fourth’s gain isn’t accompanied by a reduction in others’ consumption, so long as her/his product now available to the others offsets the reductions in their consumption of food, drink, and clothing.
One might make a nice micro homework problem out of this by adding utility functions and production functions, maybe add some sticky prices or capital market imperfections for a macro problem, but my purpose is simpler: I want to ask "where is the multiplier?" In both experiments, we’ve spent additional units of currency. In case 1, there’s no multiplier, because there’s no increase in production. In the second, there’s a "multiplier" effect because an unemployed input becomes gainfully employed, and production increases. Note that had we conducted the first experiment by having the unemployed individual engage in activity that produced nothing of value (e.g. repeatedly digging a hole and refilling it) the case would be no different, although the measured rate of unemployment would fall.
To explore this further, suppose there’s a fifth person on the island, also unemployed. In either experiment, person four can pass some of the new currency to five so that five also gains claims over the total product. If five must contribute to the product to get the cash, then there can be a multiplier from this too, otherwise no. When four starts producing, and his spending sets five to producing, we start getting something that looks like the classic spending multiplier. But note that if four digs useless holes, and sets five to doing the same, there’s no multiplier at all.
I’m hoping this is starting to sound like Jean Baptiste Say instead of John Maynard Keynes. Keynesian hokum about inadequate purchasing power violates a fundamental law: we cannot possibly consume more than is produced. Total output is the fundamental constraint. Oh, sure, we can consume more today than we produce today, by consuming yesterday’s production (capital), but we’re constrained by the total we produce. The only way to get a multiplier is to produce more - it is not increasing purchasing power, it is not putting people to make-work, it is producing more. And it can only do this by putting unemployed inputs to productive work, and underemployed inputs to higher valued work. And no matter how many individuals we add to our model economy, this remains true.
With this in mind, consider The American Recovery and Reinvestment Act of 2009. I haven’t yet had time to read the full text, but I read a substantial portion of the original H.R. 1, enough to see that too much of it looks suspiciously like a run of Experiment 1...only instead of a thought experiment, it’s being run for real to the tune of USD 750 billion. Of course, it is very possible that some of the spending in ARRA will actually be on productive things, and to that extent "stimulus" might make sense. But I have no idea why Congress or federal agencies should be able to systematically pick productive projects, and I do understand the perverse incentives that lead politicians to select projects that deliver a small amount of concentrated benefits for a large amount of dispersed costs. Hence I tend to doubt that Congress has systematically picked winners in ARRA.
Also, my simple account of Experiment 1 neglected the possibility that when we set person four to a useless task, s/he uses some scarce resource in doing so, reducing the output of persons one through three. It’s a given that ARRA will draw resources from other activities, and hence runs a risk of generating negative net product.***
And aside from whether good projects are selected, the federal government has a particularly poor track record of executing projects in a cost effective manner. There’s no particular reason to think that this has changed. Certainly nothing has changed in the incentive systems government officials face. Even a good project, badly executed, becomes a net destroyer of capital.
In sum, a stimulus can have a positive net effect, a "multiplier," but only by stimulating genuine productivity. The wrong stimulus has zero or negative effects. There’s production, there’s transfer, and there’s waste and destruction. Our crisis was brought on by misallocation of capital into real estate and equity bubbles. Further misallocation doesn’t solve this, but simply plants the seeds for more trouble. ARRA was passed in a white heat, with Barack Obama predicting doomsday if it didn’t. The spending was not carefully considered, and the danger is that we’ve bought a plethora of pet projects, an orgy of pork. If so, it will prove to be our economic version of the PATRIOT (sic) Act.
I suppose my "many regular readers" (I’ve occasionally heard rumors I indeed have readers!) find all this obvious and wonder why the to-do. Unfortunately, there’s a growing chorus of voices in the media, in government, and most unfortunately in the economics profession who claim that objections to government stimulus are just "ideological," a matter of mere tastes for private over government solutions. Anyone who doubts the stimulus multiplier is dismissed as an ideologue. The only "scientific" question is whether the ARRA stimulus is big enough. Paul Krugman is among the most blatantly dishonest here, but even the staid Menzie Chinn joined in, to my surprise. (See my brief exchange with him in the comments.) But there’s nothing ideological in the above. Expending scarce resources on useless projects may "put people to work," but it remains consumption of capital, and is simply more of what Adam Smith called decay. The idea that stimulus is stimulus and spending solves problems of capital misallocation is just crazy nonsense. Waste is waste, and it makes us poorer, not wealthier.
* Note that whether the transfers are voluntary or involuntary is irrelevant for the question at hand.
** As with the previous footnote, whether the unemployment is voluntary or involuntary is not at issue.
***I understand that Paul Krugman is now claiming that there's no such thing as "crowding out," that it's a discredited economic doctrine resurrected by Republican partisans.
Brilliant, Charles. I've been looking for a way to explain this in a simple-to-understand way, and you've nailed it. President Obama ought to give you a job, but somehow I think that's unlikely.
The trouble is, the agent and the fourth individual in reality are missing vital information. What and how much do the others want? If the fourth individual produces because he has been paid by an external agent to produce, the money provided by the agent carries no information. There is no price. Only if the fourth individual is paid by the other three does he gain information about whether he is producing the right amount of what the others want.
It isn't a question of producing goods that have a nominal, hypothetical value (if you believe such a thing can exist separately from subjective valuation). It is a question of producing goods that most accurately satisfy the wants of the consumers. That is why, even if the agent sets the fourth individual to work producing goods that you and I might think are useful or virtuous, the work may still be wasted if the other three disagree.
For this reason, Experiment 2 is sub-optimal. Instead of giving the money to the fourth individual and instructing them to produce something, we should give the extra money to the three producing individuals to allocate to the fourth individual (or each other) as they see fit, in exchange for what that individual produces.
And if we accept that this is the better option, then it is obvious how pointless it is to print more money, as there is no difference in the proportions of the total money supply held by each individual before they start exchanging. We have simply decreased the value of each currency unit in exact proportion to the amount that we increased the total volume of money. We have not changed the amount of the goods that each individual is able to buy with their share of the money supply, only the prices that they must pay and are paid for the goods.
Printing money is no solution to the problems of the economy as a whole. Its purpose is mainly redistributive, because it does not enter the economy spread perfectly to match the existing shares of wealth, but reaches some individuals and businesses earlier and in greater quantities and others later and in smaller quantities. Ensuring that people are rewarded other than in proportion to their ability to satisfy the wants of other consumers (the purpose and effect of redistribution) is unlikely to increase the efficiency and therefore useful productivity of the economy as a whole.
Is that a fair addition to the lessons to be learned from your thought experiment?
Anway, *If* the external agent can systematically assign productive projects, and *if* the factors represented by #4 are indeed un-or-under-employed, then stimulus can work in principle. Once inputs are fully employed, it can't continue to work at all; all it would do is crowd out other activities. So far as I know, this point was first made by Henry Thornton in 1802 when he was attacking inflationist policy. Stimulus is a very limited effect at best, and the rest of his argument reads follows yours on the effects of monetary expansion.
The above "ifs" are crucial. We don't "boost the economy" with useless makework. Now presumably some of the things on which government spends will help satisfy our subjective preferences, and might even do so cost effectively, and hence are productive. But there's little reason to think govt can *systematically* do this -- and plenty of reason to think govt often has incentive to do something quite different.
Incidentally, today's (U.S.) stimulus is financed more by borrowing than money creation. While this still has long term implications for the money supply, it is federal debt that is spiraling out of control just now.
Great to hear from you, BG. I still owe you an essay on Coase; I haven't forgotten
I had. I'm sure he'll come up at some point on your blog, so why don't I settle for holding off for that?
Does this belief linger in people's minds all the time, or is a temporary result of the shock brought by the crisis?
Anyway, I am much troubled by the whole affair, because I know how much the world depends on the USA.
We're witnessing is the fiscal and monetary wrecking of the United States, as well-placed people use government to grab power and wealth. It's intentional.