---

By Terrence Savundranayagam
There is a controversy with regard to the alleged upward revision of the growth rate of quarterly GDP by 0.5 per cent from 5.5 per cent to 6.0 per cent.
This has created some doubt in the public mind as to the accuracy and reliability of estimates of National Accounts. On the one hand attempts have been made to downplay this issue by making out that the GDP is only an estimate and an approximation. Through this downplaying the more important issue of who benefits from the increase in the growth rate has also been overlooked.

On the one hand attempts have been made to downplay this issue by making out that the GDP is only an estimate and an approximation. Through this downplaying the more important issue of who benefits from the increase in the growth rate has also been overlooked

Firstly let me make a brief comment on National Income Estimates. There are internationally accepted sources and methods used in estimating the National Accounts and in particular the GDP. The “System of National Accounts” (SNA) which is the bible of National Accounts practitioners is an internationally accepted standard set of recommendations on how to compile measures of economic activity. In fact the SNA recognizes the difficulties in estimating National Accounts in less developed countries. While there are criticisms of the GDP notably by Nobel Laureate Joseph Stiglitz  that GDP and GDP growth are not realistic measures of welfare it is nevertheless necessary to have a measure of total production and it’s changes over time in order to evaluate the success or otherwise of Government’s  economic  policies.




Quarterly estimates
I will focus firstly on quarterly estimates of GDP.I am aware of the difficulties in compiling estimates of GDP by quarter having made the first attempt at estimating a quarterly GDP series covering the years 1972 to 1981. (Quarterly Estimates of Gross National Product for Sri Lanka-CBSL Staff Studies April/September 1984 Vol 14 no.1&2).A detailed description of the sources and methods used by me in making these estimates is also given in this study. I found that except in a few instances very little data is available by quarter. While methods of data collection and estimation may have improved since then there are still questions as to how much direct data are available on a quarterly basis. This is true for most production sectors and is particularly true for the services sectors. It appears that data for the quarterly estimates are made using indirect methods based on norms, ratios and other extrapolations. Browsing through the publications of the Department of Census and Statistics (DCS), I cannot convince myself that the DCS has a satisfactory quarterly data base which enables the greater part of quarterly GDP estimates to be made using directly available data.

It is pertinent to note that even the IMF’s Article 4 Report of April 2013 on Sri Lanka observes that “Quarterly indicators are used for compiling quarterly value added estimates.” This also seems to suggest that estimates of production and services are not generated by quarter but are derived using ratios and assumptions based most likely on annual figures.




Usage of consistent methods
While it is possible to point out shortcomings in sources and methods used, the issue is whether consistent methods have been used in arriving at the growth rate of GDP in the relevant quarter. If, when the quarterly estimate of GDP was made the growth rate arrived at was 5.5 per cent an upward revision of the growth rate could have come from a mistake made earlier or if a decision was made to adjust the growth rate by 0.5 per cent to 6.0 per cent.

It must be noted here that the growth rate of GDP (Whether quarterly or annual)   is not obtained as a single figure but is the weighted average of all the items that go into the GDP. The main items are set out in Statistical Table 2 of the CBSL Annual Report of 2012 and these can be broken down further into sub-items. Hence the upward revision of the Quarterly growth rate would have come from a revision of one sector or sub sector. I am not questioning the bona fides of the DCS but the clarity and transparency which it must show when revising the items which are alleged to have been revised. This is important if the credibility of the DCS’s estimates is not to be damaged.

The claim that the National Income estimates have been accepted by International Organizations needs to be taken with a pinch of salt. For instance the Article 4 Report of the IMF for April 2013 while accepting that “data provision has some shortcomings but is broadly adequate for surveillance” goes on to state that “detailed data needed to measure both output and intermediate consumption are mostly unavailable or not collected……As a result the estimates of gross value added are prepared directly relying on outdated fixed ratios ……often with outdated studies and ad hoc assumptions.” Clearly there are serious misgivings about the sources and methods used in the DCS’s estimates. It is relevant to note here that projections of Sri Lanka’s GDP growth in 2014 by international and other organizations are about 1.0 per cent lower than the DCS/CBSL’s stated estimate of GDP growth in 2014.





Timeliness of data
I will conclude with two general comments. Timeliness of data is important where it is urgently needed for analytical and policy use. It is important to provide data by a given deadline if it is critical to analytical and policy uses. The issue relating to the upward revision of the quarterly growth rate clearly shows that there is a tradeoff between providing data by a particular date and providing accurate and firm data.

Up to 2005 both CBSL and DCS made their own (separate) estimates of GDP, with CBSL finalizing its estimates by end April. The CBSL had published a well accepted and recognized series of National Accounts in its Annual Reports from 1959 to 2006.Beginning with 2007 CBSL stopped making its own estimates of National Accounts and decided to use the DCS estimates in its Annual Report. CBSL’s estimates of National Accounts up to 2006 were finalized by end April in time for use in its Annual Report which had to be published by the statutory deadline of 30th April. From 2006 CSD estimates of National Accounts (mainly GDP/GNP and Gross National Expenditure (GNE)) were made available by end March, almost a month earlier. One wonders whether data collection and estimation have been streamlined so as to ensure that accuracy and reliability have not been sacrificed with the aim of producing estimates quickly. In fact the fiasco in relation to the upward revision of the quarterly growth rate would not have occurred if CSD took a little more time to finalize its estimates. Despite this a clear explanation of the reasons for the upward revision of the quarterly growth rate should be forthcoming from DCS. The DCS must take special cognizance of the fact that public confidence in its estimates has been dented.




Unnecessary hype
One reason why the alleged upward revision of the growth rate has become an issue is the unnecessary hype surrounding the growth rate. The (GDP) growth rate is the increase in real terms of Gross Domestic Product (whether annual or quarterly) over the corresponding figure for the previous year. It shows the total increase in the real value (as different from the money value) of goods and services becoming available in the relevant year. In a recent article in the Daily Mirror of 23rd January, Dunstan Perera, former UN adviser on National Accounts explained what the growth rate means and what it does not mean. He pointed out that there is an important difference   between the increase in the growth rate of GDP and the distribution of income (between the size of the cake and how it is sliced).
There is growing evidence that in recent years  increases in economic growth have been accompanied by increasing income inequality. According to data from Household Income and Expenditure Surveys of 2006/2007 and 2009/2010 conducted by DCS the richest 20 per cent of income receivers received 54.1 per cent of income while the middle 60 per cent got 41.4 per cent of income and the poorest 20 per cent received 4.5 per cent of income. More tellingly the bottom 10 per cent got only 1.7 per cent of income.

It is clear therefore that increases in economic growth have been distributed very unequally. This is where the alleged upward revision of the growth rate assumes significance. Given the trend in income disparities set out in the paragraph above it is likely that the greater part of the increase in this income would have accrued to the top bracket of income receivers.

In practical terms the reliability of published statistics can be assessed only in the market place. This is why there is confusion among laymen as to why in a situation where the growth rate is increasing individual incomes are stagnant or falling in real terms (if not in money terms). Similarly when statistics show that inflation is declining this is not reflected in falling prices in the market place.



Conclusions
Two conclusions can be drawn from this article:
Policies should give more importance to reducing income disparities while increasing economic growth.
DCS should take steps to ensure that there is public confidence in its estimates.

(The writer is a Former Director of Statistics of Central Bank and can be contacted via [email protected])