By W.A Wijewardena –
Lively debate on controversial PCI numbers
My View in this series two weeks ago on ‘Average income of a Sri Lankan: When numbers gathered from top and bottom do not tally!’ has evoked a number of comments from several critical readers.
These readers, including a few reputed economists, had contacted this writer by email and on the telephone to express their reservations on the arguments presented in the article. They had pointed out one important aspect which this writer had omitted to clarify but contributes to make the GDP per capita income or PCI differ from PCI derived from household income and expenditure surveys known as HIESs. That is, GDP PCI includes undistributed corporate profits which are not a part of the household income derived from HIESs.
They also had surmised that in field surveys, there is a normal tendency for people to hide their true incomes. As a result, they have reasoned out, HIES PCI should naturally be on the low side. Hence, on balance, GDP PCI should be higher than the HIES PCI, contrary to what this writer had presented in his article. Some had argued that this gap had been there in all the countries including Sri Lanka implying that it does not pose a serious issue relating to the credibility of the statistics published by Sri Lankan authorities.
The Central Bank too enters the foray
The article had attracted the attention of even the learned Director of Statistics of the Central Bank. In a rejoinder published in Daily FT last week he had asked the question ‘Should GDP per capita be equal to per capita in household income and expenditure survey?’ and gone onto answering the same in the negative.
This writer considers that the generation of a debate on important economic issues among the members of the civil society and also with officials responsible for interpreting economic numbers is a salutary development. Such debates will leave both the debaters and all other interested readers wiser and better informed eventually.
Wrong practices in the past do not justify their continuation in the present
The Director of Statistics of the Central Bank, a reputed economist himself, had raised four points, three of which had been raised by other readers as well. His final point had been that GDP PCI had been higher than HIES PCI even when the GDP numbers had been estimated by the Central Bank during the time this writer had been employed by that glorious institution. Hence, had there been an overestimate of GDP, both this writer and the Central Bank should be ready to take at least a part of the blame.
The implication of this last point has been that if a wrong thing had been practised by the Central Bank in the past, there is nothing wrong in continuing with it in the present as well. This was not what this writer expected by raising the issues involved as conveyed in the last paragraph of his article. He said that “All these plans (that is, plans to double PCI and make Sri Lanka a $ 100 billion economy) will become senseless if the total income and PCI underlying that income have been overestimated and need a correction downward as reflected by the ‘worm’s eye’ calculation of the country’s income and output. This is perhaps a good research topic for researchers and analysts to undertake”.
In fact, in many countries, such research has been undertaken as was also pointed out by Economist Dileni Gunewardena of Peradeniya University relating to India in a comment on a website which had reproduced this writer’s article.
Societies stand to gain by critically appraising the prevailing wisdom
Societies gain wisdom with new information and new knowledge. A learning society always makes a humble attempt at rectifying past mistakes once those mistakes have been brought to its notice. It does not mean an erosion of dignity but a gain of wisdom for further progress. In this context, this writer has not been the only person who has raised the issue relating to the gap between GDP numbers and field survey numbers.
The Princeton University based British economist Angus Deaton, a potential Nobel Laureate according to some, has been in the forefront of research into this issue. In a paper jointly authored with Valerie Kozel and published in the 2005 Fall Issue of World Bank Research Observer under the title ‘Data and Dogma: The Great Indian Poverty Debate,’ Deaton and Kozel have observed that in India in 1990s, national income figures have shown a faster growth in consumption than those revealed in field surveys. The corollary has been that India’s poverty has declined dramatically according to GDP numbers but not according to field survey numbers. Then, it was a vindication of the economic reforms undertaken in the country in early 1990s.
Those who had supported economic reforms have tended to accept national account numbers, while the opponents have believed in survey numbers. But the implication had been that India’s poverty had been underestimated by national account numbers and needed a revision.
Survey data are not discredited at all times as claimed by some
Many readers including the Central Bank’s Director of Statistics had expressed doubts about the validity of field surveys. The general belief is that surveys fail to elicit true information and are characterised by a high sampling error. Those who have these doubts should read a paper published in 1988 by B.S. Minhas, Indian economist and one time member of the Indian Planning Commission in the Indian Journal Sankya under the title ‘Validation of Large Scale Sample Survey Data: Case of NSS Household Consumption Expenditure’.
In this paper, Minhas has carefully laid out the case that there is no basis for such doubts. Minhas has been very critical of the practices of the Indian Planning Commission, but the very same Planning Commission had in all humbleness changed its policy of adding up to the survey data numbers to match the high GDP estimates on the basis of critical comments made by him.
So clearly, instead of defending the current practices tooth and nail, Sri Lanka should go into new research on the controversial numbers as this writer had suggested. Now back to the comments made.
Many differences in estimating GDP and deriving survey statistics
One of the comments that the gap has been there in all the countries is related to another comment that GDP PCI includes undistributed corporate profits whereas HIES PCI does not. This writer has drawn the attention of the readers to the methodological differences between the two types of numbers: One is ‘bird’s eye’ while the other is ‘worm’s eye; one is indirect estimates while the other is direct data gathering; one is value added while the other is all cash and non-cash receipts; one is one-time counting while the other is multiple counting.
Based on these differences, this writer argued that HIES PCI should be higher than GDP PCI. However, one important element omitted by this writer is the inclusion of undistributed profits of corporate entities in GDP data. Accordingly, the issue boils down to a single element. That is, undistributed corporate profits make GDP PCI bigger. Similarly, multiple counting and inclusion of all types of receipts make HIES PCI bigger. Since a household is a household in both national accounts and field surveys, what should be bigger will finally depend on the relative size of the two separate additions.
Government and NGO expenditure too mostly end up in household income
The Central Bank’s Director of Statistics has referred to two other areas that might make GDP PCI bigger than the HIES PCI. They are the operations of non-profit NGOs and those of the government in public services like education, health and national defence. He has pointed that these two operations are included in national account calculations and not in the household incomes. But according to the UN System of National Accounts, of which DCS follows the 1993 edition and not the latest 2008 edition as disclosed by the Director of Statistics, both these operations are valued at cost, that is, without a profit element, for GDP purposes.
All expenses incurred by NGOs and the government by way of paying salaries and wages, purchasing of services and materials and so forth will become household income as well. For instance, if the government pays salaries to a teacher or a doctor or a soldier in providing common services like education, health or defence, they are both in the GDP and the household income. As such, they do not make a difference either number. What is disputed is when the government or NGOs make payments to corporate entities.
There again, those entities use what they get from the government or from NGOs to pay salaries, wages and purchase materials and services from households and unincorporated business entities. They too therefore become household incomes. Even when they make profits, to the extent those profits are distributed, they still become household incomes. What is not in household income and only in GDP is the undistributed portion of their profits. But that should be a negligible part of the total government or NGO expenditures.
In the case of governments, the almost entirety of the government expenditure is paid out to households by way of salaries, pensions, compensations, direct subsidies and purchase of services and so forth. All these are included in household income because it is made up of all receipts, similar to the credit side of the cashbook of an individual as clarified by this writer. But not all government expenditures come into GDP numbers. One good example is pension payments and subsidies which are called unrequited transfers and therefore not a value addition to the economy.
If profits are added, losses should be subtracted
Thus, of the corporate profits, as argued above, only the undistributed portion will make the GDP PCI bigger than HIES PCI. What is referred to as corporate profits here is defined in UN SNA as ‘operating surplus’, as per the value added concept used in estimating GDP. It is arrived at by valuing the output at the prevailing market price called the basic price and deducting from it the value of inputs used and payment of salaries and wages to workers.
According to UN SNA 2008 (p 132), “It measures the surplus or deficit accruing from production before taking account of any interest, rent or similar charges payable on financial assets or natural resources borrowed or rented by the enterprise, or any interest, rent or similar receipts receivable on financial assets or natural resources owned by the enterprise”. Accordingly, profits made by corporate entities will enhance GDP while losses will do the opposite.
State enterprise sector is one big loss maker
The overall position of Sri Lanka’s State Owned Enterprises or SOEs has been a massive ‘operating deficit’ in the past. As reported in the Fiscal Management Report of the Ministry of Finance for 2014 (pp 67-8), the non-bank SOEs have made operating losses of Rs. 113 billion and Rs. 154 billion in 2011 and 2012, respectively. However, the State owned financial institutions have made operating surpluses of Rs. 45 billion and Rs. 44 billion respectively in those two years. But the overall position has been that of operating losses amounting to Rs. 68 billion and Rs. 110 billion in the two years under reference.
What this means is that SOEs have paid out more to the households for purchasing inputs and labour in these two years than retaining anything with them. Thus, the operations of SOEs have resulted in a contraction of GDP but an enhancement of the household incomes. Thus, as far as the SOEs are concerned, there is no reason for GDP PCI to be higher than the household PCI.
Private corporate entities too do not make big profits
It is therefore, the operations of the private corporate entities that should make a difference. Though Sri Lanka has some 30,000 limited liability companies, only a handful of them publish their final accounts on time. Unlike in developed economies where the statistical agencies conduct comprehensive surveys, Sri Lanka has to do limited surveys to gather data in view of the constraints in funding. Hence, its surveys do not cover the entirety of the corporate sector of the country.
One may therefore use a representative number coming from the Colombo Stock Exchange. Accordingly, the total profits, not the operating surplus as defined in UN SNA, made by the companies registered with the Colombo Stock Exchange numbering 291 as at the end of 2013 should serve as a guide to any layman to make judgment of the operations of the corporate private sector in the country. The overall profit levels of these companies in 2011 have been Rs. 134 billion and in 2012 Rs. 136 billion. These are the total profits and not the retained profits. When the losses of SOEs are set off, the value addition to GDP contributed by this so-called retained profits amounts to Rs. 66 billion in 2011 and Rs. 26 billion in 2012.
The undistributed part of those profits must still be lower. Then, GDP PCI could be higher than HIES PCI if and only if the gross receipts of households constituting multiple recording of all receipts are lower than the undistributed component of these numbers. But that has to be verified through further research and not by making sweeping statements.
The widening gap between GDP PCI and HIES PCI
It appears that many are of the view that since there is a gap between the GDP PCI and HIES PCI and the former is larger than the latter, there is no problem for Sri Lanka. This gap has been due to, as claimed by the Director of Statistics of the Central Bank, the undistributed profits of the corporate sector which are included in GDP. In other words, it is due to the non-payment of dividends by the corporate entities to their shareholders.
According to the numbers in the past few years, the gap between the GDP PCI and HIES PCI has been rising. In 2005, the gap was 2.4 times and in 2012 it had increased to 3.1 times. Thus, corporate entities appear to be retaining more and more profits with them instead of distributing the same to shareholders.
Government policy is to promote dividend pay-outs
This trend is contrary to the government policy on the payment of dividends. In the Budget Speech 2007, the Minister of Finance having been dissatisfied with the non-payment of dividends by companies in adequate amounts, had insisted that companies should make a dividend pay-out of at least 50% of their distributable profits; as a deterrent to the practice of not paying even one third of distributable profits, the Minister had proposed to impose a tax of 15% of the difference between the rate paid and the minimum rate of one third which they should pay (p 13).
However, if GDP data are a fair estimate as claimed by official sources, they denote an unsavoury development. That is, their rising faster than HIES data shows that companies are keeping more and more profits with them without distributing among shareholders thus frustrating the wishes of the government.
Once again it is emphasised that these facts present rich raw materials for further research and studies by concerned economists and researchers.
*W.A. Wijewardena, a former Deputy Governor of the Central Bank, can be reached at email@example.com