Comments on Employment, Interest, and Money: Oh, No, Not Again

>any policy that tries to use even some of the ...

2009-01-21T19:57:00.000-08:00

>any policy that tries to use even some of the idle resources is liable to result in serious inflation, just because any amount of stimulus might by chance turn out to be too much

There are certanily idle resources now, but there may not be idle resources in the future, and the market may take this into account when evaluating any stimulus, since the stimulus will not just affect the immediate future but also the mid-term future (like the next year or two). If the market judges this mid-term impact of the stimulus to be excessive, then even the short-run impact of the stimulus may be contractionary, even though there are currently idle resources in the system.

In particular, if inflation expectation jump as a result of the stimulus, then the dollar might collapse. This might have no immediate effect on export quantities, but would have an immediate effect on import costs, and thus would probably cause a reduction in aggregate demand. If the Fed responded to the dollar collapse by doing nothing, then long interest rates would skyrocket, causing a further collapse in housings, bonds, stocks and business investment, all of which will be contractionary. So the Fed would be forced to raise overnight rates, which is contractionary in itself.

Perhaps this is the underlying line of reasoning motivating Fama and the other critics of a big stimulus now. It is important to recall that the federal government is now set up for ongoing fiscal stimulus from now to eternity due to exploding Medicare costs, high payments on existing federal debt, and insufficient taxes. The only reason we haven't had serious inflation in the past few years is those 6% of GDP trade deficits we've been running.

Anonymous, it sounds like you are essentially maki...

2009-01-21T11:49:00.000-08:00

Anonymous, it sounds like you are essentially making the classical argument, but with just enough of the Keynesian Phillips curve thrown in to allow for the existence of idle resources. ("The goods/services extracted can potentially be used to pay unemployed workers to produce something of value, so the loss to consumers is not necessarily total.")

I have a couple of objections. First of all, nowhere does Prof. Fama suggest that this Phillips curve exists at all. He seems to think that the loss to consumers is necessarily total. Second, while your argument might apply under ordinary circumstances, it doesn’t carry much weight when the country is on the verge of deflation. Under these circumstances, inflation is a good thing.

“The social cost to fixed income retirees et al can easily exceed the benefit to the unemployed.” Perhaps, but aren’t the fixed income retirees getting an unfair subsidy when the inflation rate turns out lower than expected, and especially if it turns out to be negative? And personally, I don’t have a whole lot of sympathy for retirees that are rich enough not to have to live mostly on Social Security, which is inflation-adjusted, or for retirees that were imprudent enough to put all their eggs in the fixed income basket.)

The only way your argument works is if you think either (1) that there is extreme uncertainty, so that any policy that tries to use even some of the idle resources is liable to result in serious inflation, just because any amount of stimulus might by chance turn out to be too much, or (2) that there is a knife-edge equilibrium such that, if we don't hit it precisely, we have to choose between very low inflation or very high inflation (or some combination of weaker versions of these two premises). My opinion is that, while neither of these premises is 100% wrong, they are not true enough to make an empirically relevant agument.

Andy..."In general, for an argument like Professor...

2009-01-21T09:59:00.000-08:00

Andy..."In general, for an argument like Professor Fama's to work, there has to be some finite resource that is being fully utilized..."

The binding constraint is the quantity of goods/services that can be efficiently extracted from consumers via inflation. An infinite quantity of money can be created, but a finite quantity of goods/services can be extracted via this method. If more than a small quantity is extracted, a wage/price spiral stagnates the economy. If an attempt to extract even more is initiated beyond the wage/price spiral, hyper inflation makes the economy dysfunctional, and GDP spirals downward.

The goods/services extracted can potentially be used to pay unemployed workers to produce something of value, so the loss to consumers is not necessarily total. However, inflation is a very blunt and regressive extraction tool, and results in very real social costs to vulnerable citizens. The social cost to fixed income retirees et al can easily exceed the benefit to the unemployed. Fiscal policy offers far better ways to extract resources from targeted citizens for redistribution to the unemployed.

"And it's no answer to say that the extra savin...

2009-01-21T09:07:00.000-08:00
"And it's no answer to say that the extra savings comes from the increased income (even if it's true, in the new equilibrium). Eugene Fama would legitimately complain that that's just begging the question: assuming what we seek to prove, that income does in fact increase"

I disagree. I think, provided one is careful about how one wants to define “income,” that one can show in an example that income initially increases, and the question will be whether some other income will subsequently have to decline. And I define income as when somebody pays you some money that you don’t have to pay back.

So here’s the scenario: the government issues a T-bill. Jane Saver, who has $1,000,000 of savings in the bank, buys the T-bill. So far, no effect on private savings (Jane just changed the form of her savings), and no effect on income. Now, the government sends a $1,000,000 tax rebate check to Joe Billionaire, which he deposits in the bank. Effect on private income: +$1,000,000. Effect on private savings: +$1,000,000. Now Joe hires Betty Builder, previously unemployed, to build a house for him for $500,000, which (just for simplicity) he pays in advance and she deposits in the bank. Effect on private income +$500,000. Effect on private savings: nothing. (Joe’s savings is reduced by $500,000; Betty’s is increased by the same amount.) Then Betty, being an honest craftsperson, builds the house instead of absconding with the money. Voila: you have productive activity that would not otherwise have occurred. And total private savings has risen by exactly enough to offset government dissaving. The money that was in the bank when the process started is still in the bank, and any firm that wanted to borrow it to make a capital expenditure can still borrow it. And the whole process involved no changes in the total quantity of money: $1,000,000 worth of bank deposits were simply passed around from one entity to another. Where is the flaw in that argument?

Granted, that argument is only about what could happen and not about what will happen, once we start thinking about how our characters are likely to behave. The monetarists will say, “Why ever would they hold so much of their savings in the form of money, when it could be in interest-bearing instruments? Surely they will only hold enough money to facilitate their day-to-day transactions, and that amount will be proportional to their incomes.” And then it proceeds into our argument about money demand. But Prof. Fama’s claim, if you take him at his word, has nothing to do with behavioral assumptions. He says that the national income identity, in and of itself, implies that the stimulus cannot be successful. I think my last paragraph is sufficient to disprove that contention. If he is actually sneaking in behavioral assumptions without telling us what they are, then the onus is on him to defend his argument by making those assumptions explicit.

Also, if he is using the classical model, as Greg Mankiw suggests, then he needs to be more careful about using the national income identity as the basis of his argument. The national income identity is inherently in nominal terms, because there are no implicit deflators for the income side. If he wants to say that it must hold in real terms, there is no empirical substance to his argument unless he specifies how incomes should be deflated.

Andy: thank God! Finally someone's understood what...

2009-01-20T21:12:00.000-08:00
Andy: thank God! Finally someone's understood what I was saying!

And don't apologise for having a third go at this question. It's important. It's important both for policy reasons, and for understanding some very basic issues in macroeconomics. Take a fourth go if you feel like it!

Nobody else addressed Eugene Fama's central question: where will the extra loanable funds come from? Will they be taken away from other investments? (If so there is no increase in aggregate demand, just a switch from private investment to government spending). Will they come from increased savings? (If so, why will savings increase, and even if they did increase, that would just mean a decrease in consumption, so there's no increase in aggregate demand, just a switch from private consumption to government spending.)

And it's no answer to say that the extra savings comes from the increased income (even if it's true, in the new equilibrium). Eugene Fama would legitimately complain that that's just begging the question: assuming what we seek to prove, that income does in fact increase.

Too many people have been critical of Eugene Fama for getting it wrong. I would be equally critical of those who failed properly to explain to him why he was getting it wrong. (And Fama's at least got the excuse that he's not a macroeconomist). Hey, we all get stuff wrong, anyway. It's not enough to just point to another model which shows different conclusions to Fama's. You have to be a teacher, and get inside his head, figure out why the approach/model he is using is wrong, figure out his implicit assumptions, and then explain to him why it's wrong. He needed to be reminded that we are talking about a monetary exchange economy, where an excess demand for goods is a result of an excess supply of money, then explained, via interest rates and money supply and demand, where that excess supply of money comes from. (After many years teaching ECON 1000, I've developed practice at this!)

This is a sorry episode, for all of us. We so often don't understand intuitively the basic models we teach.