Monday, December 12, 2005

If men were angels, no gold standard would be necessary...

And since men are not angels, no gold standard will be sufficient.

Jane Galt blogs on the flaws of the gold standard. I think they can be summed up briefly. Obviously, it's inconvenient to peg a currency to a commodity like gold; it exposes our wealth to random swings in the market price of gold. Of course, with a fiat currency our wealth is exposed to random swings in the market price of dollars, but at least we have some control over that. And it's unlikely that some enterprising Chinese or Indian businessman will discover a new dollar mine or develop some revolutionary new technology for harvesting dollars—though if printing technology improves enough, all bets are off.

More seriously, there's only one source of unexpected supply shocks to the dollar market: the Federal Reserve Bank. Most gold standard advocates oppose fiduciary currency for just this reason: they want to deny the Fed the power to play games with our money supply. And I'm all in favor, as far as that goes. I advocate something like giving the Fed a money supply target—e.g. "keep growth of the money supply between 0% and 2% a year," although I confess those numbers were made up on the spot and have exceptionally little contact with reality—and leave it to run. No cyclical policy, no counter-cyclical policy, no interest-rate pegging, no watching the CPI, nothing. We don't need supply shocks to our currency system.

But the gold standard will only prevent monkey business from the Fed if the board of governors are inclined to behave themselves anyway. The gold standard merely requires the Fed to obey a set of rules:
  1. Exchange X dollars for Y amount of gold to whoever asks, in either direction.

  2. Manage the money supply and the government's gold reserves so that obeying 1 won't bankrupt us.
But if we have a Fed that's going to nicely obey our rules, and not do anything screwy, why do we need a gold standard to tie its hands?

On the other hand, if the Fed is determined to play games—say, trying to peg the exchange rate with foreign currencies without altering the money supply—it can do that on a gold standard as easily as it can off. But the consequences are much worse.

Part of me would really like a gold standard. But I know that I want a gold standard because I think other people are going to make poor decisions. The gold standard would be a rule for them to follow to keep the decisions good. But if they were going to follow whatever rules I gave them, I could come up with much safer and more effective ones. And if they're not going to follow the rules, the gold standard will just make our economy more volatile and more vulnerable.

6 Comments:

Blogger McAfee said...

Why a money-supply target instead of an inflation target? And which measure of money?

December 14, 2005 1:07 PM  
Blogger McAfee said...

This is not my area at all, but I sometimes eavesdrop on people whose area it is and am led to think that measuring money is really tricky. We don't use just cash to pay for things--checks work almost as well, for example, and actually better if you're paying for a Toyota--but even if banknotes and coins were the only money we had, a lot of them are in mattresses, penny jars, or being swapped around overseas and probably don't have much to do with the price of eggs in Peoria. And even if you agree what money is, navigating by a money target seems like deciding how many logs you want on the fire instead of how warm you want the room to be.

December 17, 2005 12:15 PM  
Blogger Jadagul said...

Mearuing the money supply is indeed far more complicated than it sounds, even if you agree on what counts as money. We generally use four different measures of money—M0,M1, M2, and M3 (although I don't think M0 is actually used very much)—with increasing amounts of liquidity: M0 is actual physical currency, while M3 contains pretty much any financial asset that isn't related to a physical asset (money-market accounts, long-term CDs, eurodollars, etc). I read a very good article a while back that argued that none of these are good measures of money, and advocated a new measure, but I can't find it at the moment. I suspect that M1 is probably the proper measure of money to use for this; but that's just a guess. Which particular measure you use doesn't affect the broad outlines of the rest of the argument. As for the actual method of measurement—I realize that's very sticky, and that's why I'd give a loose target instead of saying "grow at exactly 1% a year." There's just no way to do that.

As for why a money-supply target instead of an inflation target: first of all, I don't like the way we define inflation today; I think that 'inflation' should refer to changes in the money supply. Inflation right now is a measurement that combines two independent phenomena: roughly speaking, the rate of change of the money supply and the rate of change of output. If output and money supply are both constant, inflation is zero. If output and the money supply are both growing by 20% a year, inflation is zero. But the two circumstances are obviously very different.

The attempt to monitor and regulate inflation can lead to a couple of problems. First, measurements like the CPI are fundamentally flawed. It's impossible to do hedonic adjustments properly--as in, ten years ago you could get a thousand-dollar computer, and today you can get a thousand-dollar computer, but the two computers are very different. Obviously, in some sense, the price of a computer is fallen, but trying to factor effects like that into actual numerical price changes—for every good in the CPI—is impossible. Further, the CPI requires someone to decide 'what the typical family buys,' and has to deal with substitution effects (if the price of apples goes up, I buy more bananas), and has other problems of that sort. Economists in recent years have engaged in heroic efforts to fix these problems, and I'm sure the CPI has grown more accurate, but the flaws are incorrectable (in much the same way that we can't exactly measure M1).

Further, inflation measures send mixed messages since they combine two fundamentally different phenomena. When inflation goes up it could indicate that the money supply is growing too fast, or it could indicate that output growth is down. The Fed would raise interest rates (and cut the money supply) to bring inflation down, even if the problem wasn't in the money supply; mainstream theory suggests this would decrease investment and further cripple output growth. My plan would let the Fed focus in solely on the money supply, and let credit markets arbitrate interest rates.

And yes, this will probably lead to deflation, as currently defined. But we had that pretty much throughout the 19th century, and seemed to do okay. As I understand it, most of the fear of deflation comes from the onset of the Great Depression. But, as I argued in the original post, that was a result of a sudden, large contraction of the money supply at the wrong time, forced by stupid policies in the twenties. If we don't try to peg the interest rate or the exchange rate, I think we'll be okay.

December 18, 2005 3:35 PM  
Anonymous Anonymous said...

Jay is assuring the fed is still around to work for after college.

December 31, 2005 10:14 PM  
Blogger McAfee said...

A worthy goal; let's hope he succeeds.

January 02, 2006 8:31 AM  
Blogger David Friedman said...

I suggest that competing private currencies, using a commodity bundle to control their value, provide an incentive compatible solution to your problem. For details see:

http://www.cato.org/pubs/pas/pa017.html

January 30, 2006 12:13 AM  

Post a Comment

<< Home