Monday, November 08, 2010 

We have always been worried about the approach this particular state institution – the Ghana Statistical Service (GSS) that is – takes to its public duties, and its somewhat cavalier handling of the “middle-income status affair” has vindicated our concerns.  Ministry of Finance better gets cracking. They have just been handed a massive policy burden from the Ghana Statistical Service. So it is time they start doing more to earn their salaries.

(This note has become necessary because recent comments by our affiliates have, in view of airtime constraints, been incomplete)

Public education is one of the core mandates of every public institution.

That is why we are disappointed in the Ghana Statistical Service (GSS) for choosing to announce, without sufficient public education, that Ghana has attained middle-income status.

We have always been worried about the approach this particular state institution – the GSS that is – takes to its public duties, and its somewhat cavalier handling of the “middle-income status affair” has vindicated our concerns.

First of all, rather than creating the impression that a 60% jump in the country’s real Gross Domestic Product GDP figure is routine, the GSS should have been at pains to explain how unusual such a development is, and then gone further to apologise for its unwillingness or inability to abide by the international best practice of rebasing the GDP every 5 years and revamping the overall conceptual and methodological framework every 10 years.

Though the subject matter can be complex, the fundamental issues are easy to appreciate by anyone with a primary school education.

GDP refers to the monetary value of the total production of all goods and services in an economy. Naturally, in trying to sum up the prices of all such production you run into some basic problems.

The first one is double counting. Since there are usually a number of different stages in production, how does one prevent the situation where spending or income already accounted for at one stage gets added again during the next. There are detailed methodologies for handling such complexities, but when approached properly the differences that arise from using different methodologies tend to be less than substantial.

The problem that is of greatest relevance in the ongoing controversy over the middle-income status affair is that of “relative prices”. Very simply, it refers to the issues of inflation and exchange rates.

Let’s say in 2000 the total monetary value of economic output in Ghana was 100 Ghana Cedis. If by 2005, the same output is GHC400, the immediate reaction would be that the economy has expanded by 400% (i.e. quadrupled). But looking critically at the situation one may notice that inflation has averaged 20% throughout this period (for ease of analysis, ignore compound effects) and that the combined inflation rate over the period is 100%. One may also have noticed that the Ghana Cedi was over the same period devalued by 50%. What the two factors – exchange rate and inflation – would do to the GDP figures in this new frame of analysis, crudely speaking, is to reduce the “real” value of GDP for 2005. Hence the real GDP for 2005 measured in 2000 prices is GHC200 as compared to the nominal value of GHC400 in 2005 prices. The base year in this discussion is 2000 and the current year is 2005.

What the GSS is saying in its recent release is that to date it has been using 1993 as the base year but will going forward until further notice use 2006. It would be obvious to you, regardless what year you choose as the base year, that the use of the “deflators” (formulae that irons out inflation effects) as described in the preceding paragraph is to ensure that you obtain an accurate real GDP estimate as possible notwithstanding inflation. There are many international standards available for getting this right. The World Bank’s Atlas Conversion factor, for example, helps you link exchange rate and inflation rate properly through a basket of rates from the world’s foremost economies (known as the “SDR deflator”), while the United Nations (and its partners) System of National Accounts (released in the 60s, revised in 1993 and again in 2008) provides comprehensive mechanisms for ensuring accuracy from year to year.

Therefore, under normal circumstances, as far as the computation of differences between the real and nominal figures of GDP is concerned, the choice of base year should ordinarily be only marginal in its effect.

However, there is a catch. Tracking the effects of inflation and exchange rate fluctuations on GDP from year to year implies tracking the effects of these on the key individual items and sectors that constitute GDP. To give you an example, let us say we know the price of mobile phones today, and we are certain of their contribution in whatever form or shape to GDP. In order to accurately "deflate" their current pricing however, we are constrained by the fact that if our choice of base year is 1993, then we have no real reference since in 1993 there were no cellphones as we know them today.

If however we were to use 2006, as has currently been decided, then the tracking process necessarily improves. Thus, transformations in the economy as a result of quality improvements and the emergence of new industries, goods and services necessarily require that we change the base year periodically.

The point though is not, or rather should not be, that these items had never before been included in the computation of GDP figures in the immediate past, only that the referent for pricing has changed.

If the understanding is that throughout the past decade the GSS has altogether ignored the rise of the telecom industry and the deepening of the financial sector then it is a flawed, self-serving, and quite dangerous way of approaching rebasing.

Nor does the proper accommodation of the right contributions of new items, be they goods and services, sectors or industries, rely so much on the choice of base year as it does on the choice of classification systems and formulae. But as we have said previously, there are international standards for these things, and Ghana ostensibly subscribes to them.

It is therefore not to the credit of a Statistical Service worth its salt to present such huge revisions to the public without adequate explanation. This is a point better put into perspective through a cross-comparison of different national experiences of GDP rebasing.
An extensive review of the literature should show that in the more quality economies, rebasing rarely have the effect of revising final GDP values by more than 10%. For instance, this year, Singapore rebased its GDP from 2000 to 2005 and recorded an upward adjustment of approximately 2%.  In those examples where significant adjustments have occurred (eg. 7.5% for Luxembourg in 2004 and 13.7% for South Africa in 1999) the effect on GDP size remained within reasonable bounds, and the wholesale shifts in market dynamics and industries (such as the financial industry in the case of Luxembourg) were well documented and carefully explained.

The fact that, in places like Guyana, stupendous rebasing effects occur provides the set of exceptions that prove the rule. In properly managed economies rebasing should rarely lead to the result we have seen in the Ghanaian case. And at any rate, national statistical agencies are required to track changes in the economy fairly frequently.

It must be pointed out that rebasing necessarily involves a review of conceptual and methodological models, a re-assessment of data sources, and the updating of classification systems.

It is understandable that in such a process, the relative weights applied to different items as well as their relative contributions to total GDP may change. But the effect is not always that of an absolute change, and in many cases it is also discovered that GDP figures for some items had been overestimated. Nor should the relative change in weights lead to dramatic changes in the size of the GDP “overall”.

Furthermore, it would usually be clarified whether any new figures are based on purchasing power parity (PPP) or not. PPP is an approach that simply helps the economist to iron out the effects of different costs of living across different countries (a dollar in Ghana may not buy you the same goods of the same quality in Kenya, say). For instance, in PPP terms, Ghana’s per capita income is estimated at approximately $1558 by the IMF. This is already higher than the $1300 per capita income implied by the rebased GDP figures released by the GSS. And it is definitely higher than the $900 to $1000 real GDP per capita income that the rebased series estimates have produced for 2010 (there are statistical discrepancies that arise depending on what figure you use for Ghana’s population in 2010 and how you track constant US dollars in 2006 terms).

It will be important to know the views of the GSS on the current PPP base in this country, since this relates directly to the standards of living question many people have raised in the wake of the GSS’ announcement. In the same vein, the GSS is also obliged to share data on income distribution in this country. For instance, how much of real GDP goes to/is produced by/is owned by the top 10% income bracket? These are important measures that should assist the interested public better appreciate the standards of living question. Moreover, the GSS’s use of nominal GDP figures to catapult Ghana into Middle-Income status, when the real GDP figures produce ambiguous outcomes in that regard is unsafe (the World Bank classification system pegs the lower band of the status at $995, but the real GDP per capita figures in the near series could fall short of this mark as explained in the previous paragraph).

As far as the general  socio-economic implications for the nation are concerned, they are obvious. On the downside (some would argue for the better), the new figures may wean us from concessional borrowing from the World Bank’s IDA and place us on the less concessional World Bank IBRD regime in July next year when the Bank’s new fiscal year begins.  This may sound unfair seeing as we are only on the bottom rungs of the “lower income bracket” (i.e. the $1300 per capita income figure puts the country marginally above the lower limit of $1165 for IDA borrowing). Middle income status has an upper bound of more than $12,000.

On the upside, many of our macroeconomic indicators immediately improves. The budget deficit automatically halves and the nominal poverty statistics are dramatically transformed. Even our credit standing with international (private) lenders may be enhanced. Many of the key ratios however remain static. There would be little to no effect on the balance of payments for instance.

Then there are the bizarre implications. Our exports to GDP and imports to GDP ratios dramatically decline. We know this because imports and, especially, exports are quite hard to get wrong. The tradable sector of our economy has, by inference, shrunk. What are the implications for forex management, tariffs, national debt management, and the liberalisation of the capital account, to name but a few policy areas?

The Ministry of Finance better gets cracking. They have just been handed a massive policy burden from the Ghana Statistical Service. So it is time they start doing more to earn their salaries.

Courtesy of IMANI Center for Policy & Education & AfricanLiberty.Org