Does “inflation targeting” work?

November 20, 2005


Does “inflation targeting” work?


The short answer is: no. Let’s see why.


It is no surprise that the notion of “inflation targeting” is getting more attention, as it reflects the widespread desire for a monetary system that works “automatically.” The alternative, of course, is a monetary system that works “manually,” which is what we have now, with some appointed central bank manipulator of dubious talent who, we all hope, can string together a streak of lucky guesses. The anxiety over this system is so great that when a central bank governor makes fewer mistakes than the global average, like Alan Greenspan has, accolades shower down from all sides as disaster has been averted for another year. After all, what’s to prevent the next guy (Bernanke in this case) from being worse than average?


The fact of the matter is that the United States, and the world, did have a system that worked “automatically,” and it worked quite well. It was known as the gold standard, and its most recent incarnation perished at the relatively late date of 1971. But the gold standard, in its various forms, dates back into the mists of time. The Greeks and Romans used gold and silver as money, just as businessmen did during the Kennedy administration. There are ancient gold mines in the Zambezi river basin in Africa that date from 100,000 B.C. Here’s how the system worked: gold (and silver) were used as money, or chits (paper banknotes) that could be redeemed in gold, or in any case whose values were maintained to be equal to a specified amount of gold. No central bank governor. No CPI. No interest-rate targeting. No guessing.


So why wouldn’t “inflation targeting” work? The basic problems are twofold: first, the difference between “inflation” and changes in the value of a currency, and second, the lag effect of inflation. To this must be added a third issue, which is not an inherent problem but is in effect another barrier today: the central bank mechanisms popular today, namely the manipulation of an interest-rate target, are not effective in managing either “inflation” or currency value. Thus while “inflation targeting” can be likened to driving while looking in the rear-view mirror, trying to implement “inflation targeting” with today’s interest-rate manipulation schemes is like driving while looking in the rear-view mirror, with the steering wheel and brake pedal disconnected. The likely outcome is a car crash.


The first problem with using “inflation” as a guide is: what is “inflation”? (Note that I keep using quotes around “inflation” and “inflation targeting”. This is a sort of quarantine device to prevent readers’ minds from being polluted by taking today’s poisonous economic concepts at face value.) As we have seen recently, so nicely outlined by Pimco’s Bill Gross among others, official “inflation” (i.e., CPI statistics) can be jiggered by all sorts of statistical tricks. But beyond that, prices may rise or fall for all sorts of reasons. In some countries, when the government passes a sales tax hike, which of course raises effective prices, the central bank goes into a tizzy as it tries to figure out how to deal with the “inflationary consequences.” Are you beginning to suspect that central bankers are, perhaps, not exactly the brightest bulbs?


It is also widely recognized that “inflation” is a lagging indicator. First you have some sort of inflationary impetus, and then the “inflation” begins to appear in statistics, perhaps twelve months later. To this is added the notion that the effects of central bank policy also appear a year or so later, which may or may not be true. Thus, the “inflation targeting” system can be said to have a lag of about two years, though this is somewhat of a theoretical notion, since the system doesn’t work in the first place. The inflation targeting advocates (mostly academics with little or no experience with how markets and economies function in real time) typically try to get around this problem by having the central bank react to expected future inflation. In other words, they guess. If reacting to this month’s inflation statistics is like driving while looking through the rear view mirror, this is driving by predicting what the view in the rear view mirror will look like two years from now. The “solution” to this is to use some sort of forward-looking indicator like the spread between “inflation-protected” government bonds and regular government bonds, or even surveys of economists. The fact of the matter, however, is that the inflation expectations supposedly reflected by bond markets simply don’t turn out to be the case, while economists’ track record is so bad that a statistically effective money-making strategy is to expect the exact opposite of what surveys of economists say is likely to happen.


In any case, it should be apparent that the “automatic” system, which is supposed to replace the guy trying to make lucky guesses, itself involves an uncomfortable amount of guessing. From that we have the problem of implementation: either a system is automatic, or it isn’t. In practice, what would likely happen is that people would feel so uncomfortable with this “automatic” system (for good reason too) that they would expect some sort of “manual override” in the form ofäthe central bank governor’s seat-of-the-pants judgment. In other words, the system of automatic guessing would be tempered by our present system of arbitrary guessing. Not to mention, once again, that even if the guesses are right, the present system of targeting interest rates is an ineffective way of doing anything about it! Remember: even if your guess about what will be in the rear-view mirror two years from now is right, the steering wheel and brake pedal are still disconnected.


This, it appears, is the best the academic economists can come up with. The real goal is stable currency value. Stable currency value is desired because when a currency is stable, then an economy does not suffer the monetary distortions which have gained the casual labels “inflation” and “deflation.” The traditional way of ensuring stable currency value (in practice, as stable as can be achieved), is by pegging the currency’s value to gold. This system works so well that it was in continuous use for well over two thousand years, up until 1971. The proper tool of a currency manager is the one that affects currency value directly, namely, by adjusting the supply of base money. There is no lag when judging currency value. If it takes $400 U.S. dollars to buy an ounce of gold on Wednesday and $410 on Thursday, then the dollar’s value has fallen against the benchmark of value, which is traditionally gold. No guessing. No two-year lag, No forecasting. No interpretation. Under a gold standard in which the dollar was pegged to gold at $400/oz., the system would remove base money, boosting the value of the currency back to its $400/oz. parity. Automatically.


Notice that this memo is entitled “Does inflation targeting work?” not “Would inflation targeting work?” That’s because “inflation targeting” is already in use. The experiment has already been run. The system already in use is about as orthodox as any the most ardent advocate could ask for. The central bank actually surveys economists to come up with an “expected inflation” value to guide policy. The guinea pig is Brazil. Brazil has been, and continues to be, one of the most macroeconomically unstable among the world’s major economies, at times spiraling into disaster, and at other times, like today, making a dazzling recovery. Either way, nobody would call Brazil’s experience one of “stability.” Nor would anyone be particularly thrilled by the 19% short-term interest rates produced by this experiment in “inflation targeting.”