Stephen King: Why inflation targeting is an increasingly inexact science

Inflation will be higher because inflation was lower? To my mind, this doesn't sound convincing

One minute, the curse of deflation is upon us. The next, the forces of inflation are being unleashed. As I turned to page 3 of my weekend newspaper - the Financial Times, in case you're wondering - the headlines screamed: "Data raise spectre of inflationary pressures". The story focused on last week's upward revisions to UK GDP data from the Office for National Statistics. It turns out GDP rose by an additional 1.4 per cent in the 1998-2000 period. Assuming no change in the so-called "trend" - or sustainable - rate of economic growth, this implies that we have been using more resources than we previously thought. And, as any fledgling economist will tell you, that points the way to higher inflation.

There is, however, a catch. The reason why growth was revised up was largely because inflation was revised down. Measuring GDP in nominal terms is easy enough but statisticians then have to split GDP into "real" and "inflation" components. Conceptually, this is not so difficult but the practical challenges are enormous. A top of the range laptop may cost today the same as it did four or five years ago, but a top of the range laptop can do a lot more today than it could back then. In effect, the price of the "computing power" has come down. Similarly, today's cars offer many more features than cars of yesteryear so, again, the price of the "car-driving experience" may have come down.

That the revisions reflect new estimates of the split between real growth and inflation doesn't really tell us a lot about inflationary pressures at all. To arrive at this conclusion you have to go through some fairly tortuous logic. Something like this, perhaps: the volume of GDP has been revised up because the deflator - the estimate of the price level - has been revised down which, in turn, implies that the deflator will have to go up again. To my mind, this doesn't sound terribly convincing: inflation will be higher because inflation was lower? No, the revisions may tell us more about the difficulties facing statisticians than about the true state of the UK economy.

And it's not just in the UK that we're seeing these difficulties. This year, the country that has offered the biggest upside surprise on economic growth has been Japan. Does this mean the end of deflation? The knee-jerk answer appears to be "yes". But scratch the surface a little more and it turns out that the main reason for the upside surprise on growth was a downward surprise on inflation: for a given amount of nominal GDP, volumes surprised on the upside only because statisticians estimate that prices fell particularly quickly.

Now, you might think that a lot of these observations are a little obscure and not obviously relevant to the man or woman in the street. However, if we don't really know what's real and what's "imaginary", if we're not sure how much of any given rise in nominal GDP is sustainable, then we really should start to worry about out policy makers' abilities to get things right. Imagine trying to navigate a ship by using the stars but then discovering that the stars keep shifting around from one night to the next. Well, that's exactly the position our policy makers find themselves in. In the past, stronger-than-expected growth might have been a good reason to raise interest rates but if, today, stronger than expected growth results from the same process as lower-than-expected inflation, all our cherished assumptions begin to collapse around us.

One important conclusion from these arguments is that inflation targeting becomes an increasingly inexact science. Inflation targeting isn't really based on current inflation rates at all: instead, it's based on forecasts of future inflation. Most economic models tend to relate future inflation to some measure of spare capacity within the economy. If growth picks up, the usual reaction is to think that there is now less spare capacity and, hence, a higher risk of inflation.

Now go back to 1998. Had the Bank of England correctly forecast economic growth between 1998 and 2000 - taking into account the latest revisions - it might have thought that inflation would have been higher and, hence, that interest rates should also be higher. In the event, though, it turns out that growth was higher because inflation was lower.

Some central banks have recognised this problem and are a little bit more relaxed about pulling the interest rate trigger than unreconstructed inflation hawks would be comfortable with. The Federal Reserve, for example, seems more worried about the medium term possibility of inflation being too low against a background of excessive debt than it is about the short-term strength of the US economy. It seems to have accepted that there is no automatic relationship between growth and inflation. Others, though, seem a bit more doubtful. Chief among them is the European Central Bank.

The incoming new president, Jean-Claude Trichet, could, perhaps, change the ways in which the ECB approaches its economic objectives. For the time being, though, the ECB is still very much focused on the dangers of inflation. The ECB argues - presumably on the basis of better news elsewhere in the world rather than within the eurozone itself - that the worst is now over for the eurozone economy and that activity should look better over coming months. I have no disagreement with this view but I am a lot less sure about the ECB's conclusions on inflation.

Back in June, the ECB published an inflation forecast of just 1.3 per cent for 2004, low enough, one would have thought, to justify more cuts in interest rates. It now appears, though, that the ECB thinks that inflation next year will be a little higher. As Lucas Papademos, the ECB's Vice President, declared at the 4 September ECB press conference, "the latest projection implies that inflation will be below, but not far below, 2 per cent....and that is slightly different from our previous direction..... in an upward direction". Well, if you ever wanted to shut the door - if only temporarily - on rate cuts, that was the way to do it. Invoke the spectre of inflation, even though your economy has spent the past two and a half years growing at a persistently feeble pace.

Of course, the ECB can come up with plenty of arguments to defend its position. As Robert Prior-Wandesforde, my colleague at HSBC, has recently argued, the ECB could claim that interest rates are already very low (see left-hand chart). It could argue that inflation has been rather too high over the last couple of years and therefore it's right to aim for inflation over the next couple of years which is rather too low, thereby ensuring that inflation gets close to target over the medium term (see right-hand chart). It could argue that rate cuts will simply encourage excessive debts, a problem that may eventually bring down both the US and UK economies. And it could argue that rate cuts now would be seen as a reward for politicians who are only too happy to deliver bigger and bigger budget deficits: far better under these circumstances to slap wrists.

Ultimately, though, the ECB comes to its conclusions because it thinks it has a good idea of where inflation is heading. The same applies to the Bank of England. Yet if the relationship between growth and inflation is increasingly less secure - if only because of measurement issues - the inflation targeting approach that appears to have been so successful may increasingly become redundant. We might like to think that policy makers have found the best way of securing macroeconomic stability, but recent events suggest that their navigational equipment might eventually lead them in completely the wrong direction.

Stephen King is managing director of economics at HSBC

stephen.king@hsbcib.com