Are inflation statistics even useful?
Ben D. Kritz
LATER today, the inflation data for the month of June will be released. Guessing the result beforehand has become something of a sport; the BSP offers a forecast, the Department of Finance chimes in with one of its own, and there’s never a shortage of local analysts willing to share their predictions.
All of these guesses are duly reported, of course, in as much analytical detail as is made available, which helps to emphasize the significance of the inflation indicator. That inflation deserves close attention because it’s the basis on which monetary policy is formulated, which affects the entire economy, is taken for granted. A question that should be asked, however, is should it be?
The prevailing opinion is that inflation in June was at least a bit higher than it was in May, when it hit a five-year high of 4.6 percent; the average forecast for June of analysts polled by this paper was 4.7 percent, while the BSP, taking its usual shotgun approach, predicted a range between 4.3 and 5.1 percent. My own guess, based purely on unscientific casual observation, is that inflation probably accelerated in June; its rate of growth had been slowing for a couple of months, but with the start of the school year, fuel prices that have remained elevated, and supply and price pressures on commodities like rice, sugar, and animal feeds, that trend has likely reversed. For the sake of putting a possible claim to bragging rights for accuracy later this afternoon on record, I’ll say June inflation clocked in at 4.8 percent.
Bragging rights are about all the forecasts are worth, but the actual official inflation data may not be any more practically useful. The obvious reason for this is that inflation data are unavoidably retrospective. By the time the month’s inflation figure is presented – usually during the first week of the following month – much of the price data that are used to generate it are several weeks old. Even though the inflation result is generally reported accurately as being what inflation was during the previous month, because it’s the latest reported figure, there’s a natural tendency to regard it as describing what inflation is at present, and what it will continue to be until a new figure is reported. In reality, unless we have an army of statisticians equipped with fast computers at our disposal, we never know what the inflation rate is right now.
That is problematic, because any conclusions that are drawn from the published inflation rate – monetary policy decisions, adjustments to wages, price hedging by suppliers, and so on – are inevitably outdated. Over a long period of time the effect is not so significant, but in practical terms the economy is managed in much smaller windows – about six-and-a-half weeks between Monetary Board meetings, quarterly for most businesses, month-to-month for most households. Inflation is a volatile indicator, so over short timeframes, what appear to be trends can change rapidly. For example, over the past three months (March through May), the rate of inflation’s increase appeared to be slowing; March’s inflation rate was 0.5 percent higher than February’s, April inflation was 0.2 percent higher than in March, and May’s inflation rate was just 0.1 percent higher than April’s. The trend was expected to continue and so became part of assumptions, which now have to be revised to catch up to the effects of unexpected changes during the past month.
Not only are inflation figures retrospective, there’s some reason to doubt whether they are even accurate in that context. The consumer price index is generated by a statistical analysis of a “basket” of prices that ostensibly represents average spending, adjusted for the proportion each accounts for in an average household budget, and further adjusted for geographic area (about 26 percent from the NCR and about 74 percent from the rest of the country). The weights of the prices in the CPI basket are not adjusted very often, only when the base year is changed, which it was this year from 2006 to 2012. In reality, however, the weights of expenditures in household budgets change constantly, depending on current economic conditions. For instance, families likely spent more than the average for “education” and “clothing and footwear” in June because of the start of the school year. The static weights assigned to these things in the CPI basket are probably valid as averages across an entire year, but for this one month, they represent a spike that’s not necessarily accounted, therefore throwing off the broad inflation measure by some amount.
Thus, we may be basing a large part of our forward-looking assessments of the economy on an indicator that’s not necessarily accurate, and not certainly indicative of future economic performance. The reality that everyone does it this way and has consistently done so for a long period of time lends a certain regularity to it in terms of comparisons with other economies, but only in the sense that everyone may be uniformly inaccurate. As a tool to determine how this economy should be managed to achieve a certain future result, inflation may not be helpful after all, which may explain why many if not most economic policy initiatives fall short of expectations.
ben.kritz@manilatimes.net
The post Are inflation statistics even useful? appeared first on The Manila Times Online.