J. Atsu Amegashie*
March 27, 2015
On March 24, 2015, Dr. Mahamudu Bawumia delivered a lecture titled "IMF bailout: will the anchor hold?" in which he challenged the credibility of inflation data by the Ghana Statistical Service (GSS). He pointed out discrepancies between the GSS' food inflation figures and the corresponding figures computed by the Statistics Research and Information Directorate of the Ministry of Food and Agriculture (MOFA).
On March 26, 2015, the GSS responded to Dr. Bawumia's query as follows: http://business.peacefmonline.com/pages/news/201503/236897.php
"With regard to the difference in food prices quoted by MOFA and GSS, we wish to state that GSS produces inflation rates for food (which include nonalcoholic beverages) and non-food. MOFA, on the other hand, compiles information on average food prices only. In addition, GSS' food basket contains 82 food items while MOFA collects information on 24 food items, which excludes non-alcoholic beverages. Therefore, the food basket of GSS cannot be compared to that of MOFA. Furthermore, GSS collects prices from fixed outlets in 42 selected market centres across the country for the computation of the weighted CPI and inflation with 2012 as the base year.
... The Ministry of Food and Agriculture, on the other hand, collects average prices of 24 food items from the 10 Regional capitals and five other market centres, namely Techiman, Obuasi, Ejura, Mankessim and Tema.
If Dr. Bawumia had taken some time to check the facts from GSS or MOFA, he would have known that the coverage and the methodologies used by the two institutions differ and do not have a direct correspondence. Thus, statistically they are incomparable. We wish to urge interested individuals and institutions to contact the Ministry of Agriculture or GSS to seek clarification if they have any difficulty." END OF QUOTE
The GSS' response is defensible with respect to the aggregate food inflation figures quoted by Dr. Bawumia. But it is on very shaky grounds with respect to the single food items in Bawumia's lecture: yams, rice (local), groundnut (shelled), oranges, maize, and smoked herrings. Clearly, when it comes to, for example, inflation in the price of yam, the number of items in the food basket and weights should not matter. In fact, since the consumer price index is computed in this case for a single food item, the MOFA and the GSS must use the same weight (100%). It is therefore not quite accurate for the GSS to argue that "... the methodologies used by the two institutions differ and do not have a direct correspondence. Thus, statistically they are incomparable" because this is not applicable to the case of single food items.
A possible source of discrepancy in the inflation rates may be the base year used in computations. But as, I show in a technical analysis, this is should not matter in the case of single items. The proof is available at the following link:http://www.uoguelph.ca/~jamegash/Bawumia_vrs_GSS_Appendix_2015.pdf
However, the logic is very simple and does not require a mathematical analysis. The base year matters if we have several items (goods and services in the basket) because in that case, the various items will have different weights depending on the expenditure pattern in the base year. In contrast, if there is only one item in the basket (e.g., only yam or rice), the weight must be 100%, regardless of the choice of the base year. The simple mathematical analysis at the above link shows that, for single items, the percentages changes in the Consumer Price Index (CPI) are equal to the percentage changes in prices.
Therefore, it is very surprising that there is a huge discrepancy between the GSS' inflation figure for each of the six food items relative to MOFA's corresponding figure. In fact, in the cases of rice and yam, the GSS' figures show that prices have been falling from January 2014 to December 2014 while MOFA's figures show that prices have been rising or that inflation is positive. The discrepancy for rice in December 2014 was 16.8 percentage points while the corresponding figure for yam was 10 percentage points. For each of the six food items, the discrepancies are too big and cannot be explained by the reasons given by the GSS. As a result of the differences in markets sampled, there can be differences in the GSS' and MOFA's figures but they cannot be too big nor should the trends show negative correlations. Either the GSS' computations are wrong or MOFA's computations are wrong.
In its statement, the GSS said that:
"To the best of our knowledge, MOFA does not compute consumer price indices or inflation as stated in the lecture and the figures cannot be traced to MOFA."
Surely, MOFA has a Statistics Research and Information Directorate and Dr. Bawumia claims that is the source of his information. I will be surprised if MOFA’s statistics research unit does not compute a simple statistic such as percentage changes in the prices of food items. Alternatively, Dr. Bawumia could easily have computed these percentage changes from MOFA’s price level data. This is no rocket science. I have not been able to obtain inflation figures for any of the six food items in Bawumia’s paper from the GSS’ website. The data at the GSS’ website are aggregate food inflation data.
It is important to state that the fact that the GSS is officially responsible for computing inflation does not mean that MOFA’s food inflation figures cannot be used to verify the credibility of the GSS’ food inflation figures. After all, the GSS’ response suggests that if the two methodologies are comparable or similar, then there are legitimate grounds for comparisons. I have argued that the two methodologies ought to be the same in the case of single items.
The GSS assured the public that it "... is prepared to offer further clarifications to any individual or group of persons on the work we do, especially on issues relating to inflation and GDP."
The ball is in your court, GSS and MOFA.
*J. Atsu Amegashie teaches Economics at the University of Guelph..
---------------------------
J. Atsu Amegashie
Department of Economics
University of Guelph
Guelph, Ontario
Canada N1G 2W1