By W.A Wijewardena –

Dr. W.A. Wijewardena

Household Income and Expenditure Survey of 2012/13

The Preliminary Report of the Household Income and Expenditure Survey, abbreviated as HIES, for 2012/13 has just been released by Department of Census and Statistics or DCS (available here ).

Without doubt, the survey report carries a wealth of information on the socio-economic conditions of the Sri Lankans at this juncture of their history. It supplies feed-material for researchers, analysts, policymakers, commentators and even marketers. Since HIES is conducted periodically – once in every three years now – its numbers can be compared with those in the previous HIESs to assess whether there has been an improvement or a decline. The report has been written in technical terms but it does not reduce its value. All those who had been involved in the survey deserve commendation for the efforts they have made to bring out the survey results.

Disparity in per capita income

However, one area of controversy has been the average annual income of a Sri Lankan, called per capita income or PCI by economists, in 2012. Many including some professional economists have been baffled by the wide disparity in PCI as reported in HIES and in national accounts which estimate the country’s total income. According to National Accounts estimated by DCS and published by both DCS and the Central Bank, PCI in 2012 amounted to $ 2923. But the PCI reported in HIES as monetary income of a household of 3.9 persons works out only to $ 955 in that year. This is not peculiar to this year’s HIES. It has been the trend throughout and PCI numbers reported by national accounts have been two to three times higher than those of HIESs. The following are some examples: in 2005, PCI according to HIES was $ 498 whereas according to national accounts it was $ 1239. In 2006/2007, the two numbers were $ 637 and $ 1422, respectively In 2009/10, they were $ 815 and $ 2064, respectively. Since PCI is considered as an index of wellbeing of people on average, many have questioned whether Sri Lanka has inflated its growth numbers to show a higher prosperity and wellbeing. It is therefore opportune to look at these two numbers and see how they differ from each other and how the disparity can be reconciled.

First of all, these two numbers are derived by adopting two different methodologies.

Bird’s eye view of an economy by national accounts compilers

In national accounts, estimates are done by the estimating agency – in Sri Lanka’s case, DCS – by viewing the economy as a whole from the top. It is like taking a helicopter trip and look at what is happening below. From the air, it assesses how each of the sectors has performed and assigns numbers to those sectors based on the ground data gathered. Hence, national accounts present a “bird’s eye”     view of the economy. Data are gathered from other agencies and therefore they are mostly data reported by those agencies. To ascertain accuracy, it is necessary to cross-check them with other relevant numbers and make necessary adjustments to the “bird’s eye” view. However, an economy is a huge enterprise having many sub sectors and production units. Thus, the estimating agency has a wide leeway to increase or decrease its estimates to give a higher or lower value to growth rate and the total output. A trustworthy estimating agency is expected to place at every point of estimation the appropriate checks and balances to prevent it from being abused. The recent criticism against DCS in Parliament and in the media was based on an alleged instance of abusing this leeway thereby risking its reputation and integrity. If the growth rate is overestimated, the total output and PCI are also overestimated showing a better economic picture.

Worm’s eye view through field surveys

In HIES, the surveying agency – again DCS – visits the people in person to gather details of their income and expenditure. Since it is impossible and too costly to visit all the income recipients, what it does is to select a representative sample and visit them during a given period of the year. There are several factors that have a bearing on the accuracy of the information it gathers. One is that the sample should be large enough to represent the whole country and it should be selected without any personal interests to eliminate biases or prejudices. DCS says that in the current preliminary survey, a sample of 5442 households from a total of 5.2 million households has been used; this works out to be a sample of about 0.1 %. However, when the final survey is conducted, this will be expanded to 25000 households accounting for about 0.4% of the total number of households.  It is the view of this writer that the size of the sample chosen is quite adequate to derive the results which the survey has produced. Second is that the period during which the information is gathered should be an average income-and-expenditure period without seasonal peaks or bottoms. Third is that the officials who fill in the information gathering sheets should have skills to elicit correct information from the income recipients. DCS which has a long experience in conducting field surveys has affirmed that its methods and systems meet with these requirements. Since it is information gathered from the bottom at the ground level, it is considered a “worm’s eye” view of the income received by people of the country. The more accurate method of calculating PCI is this worm’s eye method. That is because it gives information on incomes earned by income recipients themselves and not income estimates made by using secondary sources. Yet, because of the practical difficulties and the costs involved, it is done only periodically, once in every three years today. As such, there are no annual PCI numbers produced by HIES but only with respect to the years in which the particular survey has been conducted.

Field survey PCI should be higher than PCI in national accounts

These two numbers are accordingly derived by using two different methods that draw on two different sources. Hence, they need not be exactly identical. But they should not be too different from each other as well.

There is one other reason that makes these two numbers not comparable in raw form. Hence, if any comparison has to be made, they have to be adjusted taking note of the essential differences in the content of the two numbers. Even after adjustment, PCI derived from HIES should be higher than PCI estimated in national accounts. But PCI numbers in Sri Lanka show the opposite with PCI estimated in national accounts recording a two to three times higher figure than that is PCI derived from HIES.

National accounts reckons only incomes earned in the market

In national accounts, income is simply the total income earned by people by selling their resources in the market. These resources have been categorised by economists into four main categories, namely, labour, land, capital and enterprise. Hence, income consists of the income derived from these four sources in the form of wages, rents, interest and profits, respectively. The total income is therefore the total of these four types of incomes earned in the process of producing various goods and services in a market set-up. Hence, income in national accounts comprises only the income received in money form in the market by providing an equivalent service. A benefit received in non-monetary form such as the self-services rendered by people to themselves is not included in income. For instance, if I go to supermarket and buy a gas cylinder for the household it provides a service to the family. But, since I am not paid in money for that benefit, it is not included in the family income. By the same reasoning, services rendered by housewives and househusbands to their families are not included in the income of a nation when it is estimated in national accounts.

Problem of multiple-counting of incomes

There is another problem faced when calculating national income from the top. That is the problem of eliminating the multiple counting of incomes that arises in different stages of production. This is because one man’s purchase of an input from another man is an income for that second man and his purchase of inputs from a third man is an income of that third man. Hence, if all incomes are added together to derive the total income of a nation, it overestimates the income.

Value addition in tyre industry

This can be illustrated with an example. Suppose in the sale of motor car tyres, there are four different production stages, production of rubber, manufacture of tyres, distribution of tyres and sale of tyres to car owners. Suppose a tyre priced at Rs 1000 is sold and the seller gets an income of Rs 1000. However, in this income are included several other incomes earned by others involved in the process. Suppose the rubber producer sells his rubber for Rs 200. The manufacturer uses that rubber and manufactures a tyre which he sells for Rs 800. The distributor sells the same tyre to the sideway tyre shop for Rs 900 which the shop owner sells to the car owner for Rs 1000. If all these incomes are added together, the total incomes received by the four different people come to Rs 2900 (that is, 200+800+900+1000). But that is not the total income earned by a nation because in each stage income, the incomes earned in the previous stages are also included. Thus, it leads to an overestimation of national income.

National accounts use only value added incomes

To avoid this multiple counting of incomes, economists take into account only the incomes added to the total income at each stage of production. This is called the addition of value at each stage or in economists’ language, the value added. To calculate the value added, only the new incomes created at each stage of production is considered and that is calculated by deducting the costs of inputs purchased from outside by each seller. In this case, it assumed that the rubber producer does not buy any input and therefore he creates a total of new income amounting to Rs 200. But the tyre manufacturer has to deduct what he has purchased from the rubber producer and therefore he creates only a new income of Rs 600. Similarly, the distributor and the sideway tyre shop create a new income of Rs 100 each making a total of Rs 200 jointly. Thus, the value addition is simply the total of these new incomes which come to Rs 1000 or the value of the final commodity sold by the sideway tyre shop. Since there are four people involved, the income on average earned by them, or PCI in our jargon is Rs 250. If the multiple-counted income of Rs 2900 is used, PCI becomes Rs 725 which is an overestimated figure.

Thus, the value added income is simply the surpluses made by each party and is equal to the aggregation of all the profit figures in each person’s profit and loss account before charging depreciation.

HIES income is simply the credit side of a cash book

The income figures reported in HIES include many items that are not included in the income estimated in national accounts.

First, they include income received in both money form and non-money form. Hence, for comparison with national account figures, the non-money income has to be ignored. In Sri Lanka, historically, this amounts to about 16% of the total income earned by an individual.

Second, even in the case of income in money form, HIES considers all incomes and not just the value additions. Hence, HIES income numbers which have this multiple-counting feature should be pretty much higher than the value-added income numbers in national accounts.

Third, for HIES, even for income in money form, incomes received in all forms are taken into account. But in national accounting it is only incomes received for providing an equivalent service are reckoned. As a result, HIES incomes include lottery gains, repayment of loans by debtors, cash received as proceeds of loans borrowed, sale proceeds of assets, free gifts received in money form, and subsidies and pension receipts. These are not considered as income in national accounts. Because of this reason as well as the inclusion of multiple-counted incomes, HIES income numbers should be pretty much higher than the income figures estimated in national accounts.

HIES income numbers are like the summation of the credit side of a cashbook that records all types of receipts in cash. All of them are not incomes according to definitions used in national accounts. Hence, PCI numbers produced in HIES should naturally be higher than those estimated in national accounts.

Both PCIs not inflation adjusted

But a common feature in both numbers is the inflation of incomes due to increases in prices. For instance, suppose that the price of a coconut goes up from Rs 25 to RS 50. If a coconut is sold for this price, both HIES and national accounts treat it as a receipt of money income amounting to Rs 50.

It is therefore clear that HIES income numbers should in theory be pretty much higher than the income figures reported in national accounts. However, in Sri Lanka, in practice, it is the opposite that has taken place: PCI numbers in national accounts are two to three times higher than the HIES numbers. This wide disparity has several implications for assessing the welfare levels of the people in Sri Lanka.

Is Sri Lanka’s PCI overestimated?

For all policy making and international comparison purposes, it is PCI estimated in national accounts that is being used by all. Sri Lanka is considered as having elevated to a lower middle income country because its PCI is reported to have surpassed the threshold of $ 2000 now. The implication of this elevation is that Sri Lanka is no longer eligible to receive concessionary funding from donors and multinational lending institutions. As a result, the country now has to borrow from commercial markets and countries like China at commercial rates. Even the government has set itself with a target of doubling this inflated PCI from $ 2000 to $ 4000 by 2016. It has even asked everyone – banks, private chambers and government institutions – to get ready for a $ 100 billion economy in a few years’ time. All these plans 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.

*W.A Wijewardena, a former Deputy Governor of the Central Bank, can be reached at waw1949@gmail.com