Rather than a public celebration, the recent announcement by the Ghana Statistical Servicethat consumer inflation had made further advances into single-digit territory (8.9% in May 2011, from a peak of 20.7% in June 2009) was greeted by howls of derision from a small but vocal band of skeptics who demanded the underlying inflation data to verify things for themselves.

As an exercise in civic responsibility, the demand is commendable (it’s good to question authority now and then); as a technical matter, it is problematic and unlikely to withstand even the most rudimentary scrutiny.

Many commentators have said, repeatedly, that a fall in the inflation rate does not necessarily mean a fall in prices but rather a fall in the rate at which prices rise. Hence, if the price of an item, for example, rises from GH¢20.00 in 2009 to GH¢30.00 in 2010 and again to GH¢35.00 in 2011 (at the same quality and quantity), the rate of price increase (“inflation”) has clearly fallen from GH¢10.00 (between 2009 and 2010) to GH¢5.00 (between 2010 and 2011), although in both years the consumer paid more for the item.

The consumer then would be right in saying that “prices are increasing all the time”. And it would be equally correct to say that “inflation fell between 2009 and 2010”.

The actual calculation of consumer inflation is based on the prices of 242 items clustered into 12 broad groups and 40 sub-groups. On average, some prices rise, others (such as prices for in-season foodstuff) fall and others remain unchanged. The sum of these changes forms the Consumer Price Index (CPI), which is the basis for calculating inflation. An increase in the CPI from 100 to 109, for example, means an inflation rate of 9.0%. (The CPI comprises only food and non-food household items; it does not cover non-household items like iron rods or roofing sheets, which are tracked separately by GSS).

A sustained slide in inflation is called “disinflation”, and an actual fall in prices over time is known as deflation (or negative inflation). While disinflation to an acceptable level of inflation is good, deflation is considered bad for the economy as it reduces business profits and investment and thus slows down growth; it is good, in the short term, at least, for consumers. The challenge of policy is to attain and maintain low inflation – for sound business and household planning.

In January 1971, for example, the year-on-year inflation was 1.9%; by the end of the year it was 10.9%, and by the close of 1974, it was 40.4%. By December 1983 it was 142.4% and one year later it was down 6.0%! Not surprisingly, this was a period of sustained economic decay.

The disinflation we are currently witnessing is nothing new, the most recent one having taken place between July 2005 and October 2007, when year-on-year inflation fell from 17.3% to 10.1%, before rising steadily to 20.7% in June 2009 (due in part to the global food and fuel crisis). Inflation started falling again from July 2009, breaching the 10.0% floor in June 2010, when it hit 9.3%. It has continued to slide. The only difference between the two periods is that the current one has seen steeper falls (an average of 0.5 percentage point per month, compared to 0.14 points in the previous decline), but the broad trends are not dissimilar (see the nearby graph).

The nuances of calculating inflation from the CPI, however, can create the kind of confusion and cynicism that have greeted the current disinflation. The same CPI can yield different (sometimes even opposite) rates of inflation depending on what periods or policy questions interest the analyst:

Here are a few measures of inflation based on CPIs for 2010 and 2011.

1. Year-on-year inflation (also known as point-to-point): Answersthe question, How did we do this month compared to the same month last year? It is the most commonly used measure in Ghana. It was 8.9% in May 2011 (compared to May 2010). 2. Monthly inflation: Answers the question, How did we do this month against last month? It was 1.8% in May 2011 compared to April, 2011. 3. Year-to-date (YTD) inflation: Answers the question, How have we done so far this year (between December last year and this month? YTD inflation in May 2011 was 8.1%. 4. Quarterly inflation: Inflation in a quarter (or three months) over the quarter immediately preceding it. It was 4.5% in the first quarter of 2011 (Jan.-March) over the last quarter of 2010 (Oct.-Dec.). 5. Quarterly Year-on-Year: Inflation in the current quarter versus the same quarter last year: It was 9.1% (Q1:2011 versus Q1:2010) 6. End-of-period (EOP) inflation: Can apply to any period but typically refers to a year and therefore inflation for December only. (Also known as December-to-December inflation, it is identical to YTD and year-on-year inflation for December). It was about 8.6% in December 2010. 7. Period-average inflation: Typically, the average inflation rate for the entire 12 months of the year, and not just December. (Also known as annual inflation). It was 10.8% in 2010.

A common mistake in Ghana is to confuse EOP inflation with annual inflation. Where there are wide variations in inflation during the year, this can lead to serious miscalculations and a contamination of policy discourse and decision making. In 2000, for example, the December inflation was 40.5% but the annual (for the entire year) was 25.2%, yet many people continue to say that “inflation in 2000 was over 40.0%”. In 1995, annual inflation was 58.5% while December inflation was 70.8%. The folly of confusing the two in any analysis cannot be overstated.

Depending on which measure is used, inflation can be rising and declining at the same time over a given period. Between March and May of 2011, for example, monthly inflation rose from 1.1% to about 1.8%, but year-on-year inflation fell from 9.1% to 8.9% over the same period.

The rise in monthly inflation partly reflected seasonal effects on the CPI; this has been the case for many years caused mainly by weather-induced changes in food supply and hence the food component of the CPI. Between March and May 2007, for example, monthly inflation rose from 0.7% to 2.80%; from March to May 2009, it increased from 1.9% to 3.4%.

The difference between now and the past is that increases in monthly inflation have been progressively smaller than they were in previous years, leading to the fall in year-on-year inflation

All of this is not to say that there are no issues with inflation measurement in Ghana; there are, but they are technical not political as has been suggested by some of those questioning the inflation data.

The CPI, for example, must be adjusted for seasonal effects, as is done by other national statistics agencies, before the inflation rates are calculated and reported. The resulting rates may be higher or lower, but they would be better than the unadjusted rates. The Service should also consider a supplementary measure of “administered prices”, such as petroleum prices, which change less often but yet are critical to the public’s understanding of overall inflation. Lastly, the “arithmetic mean” (averaging method) used by GSS has been known to overstate the rate of inflation. It may consider a switch to the “geometric mean”, which correctly records lower inflation, with a consequent fall in interest rates and of course a boost to businesses and the economy. Countries like Kenya have tried it with dramatic results; there's no reason Ghana shouldn't do likewise.

Credit: Nii Moi Thompson