15. This chapter provides a brief overview of the sources and methodology of national accounts data for Nigeria. It then identifies the main limitations of the data and describes the revision of national accounts data made by the Fund staff. It also provides a comparison between the two sets of estimates and analyzes the evolution of the structure of GDP.

Abstract

15. This chapter provides a brief overview of the sources and methodology of national accounts data for Nigeria. It then identifies the main limitations of the data and describes the revision of national accounts data made by the Fund staff. It also provides a comparison between the two sets of estimates and analyzes the evolution of the structure of GDP.

II. Revised National Income Accounts for Nigeria9

15. This chapter provides a brief overview of the sources and methodology of national accounts data for Nigeria. It then identifies the main limitations of the data and describes the revision of national accounts data made by the Fund staff. It also provides a comparison between the two sets of estimates and analyzes the evolution of the structure of GDP.

A. National Accounts: Data Sources and Methodology

16. The primary source of all historical national accounts for Nigeria is the data produced by the Federal Office of Statistics (FOS), generally with a lag of one year. The FOS produces national accounts data both by sector of origin and by expenditure groups at current and constant prices. The FOS values the gross output and intermediate inputs on the basis of periodic surveys of different activities. In many sectors, gross output is converted to value added by applying fixed coefficients observed from the past (base) years, rather than using independent estimates of intermediate consumption. In some cases, like the financial services, the annual accounts of companies are used to estimate the value added of the sector. GDP at market prices is then calculated by adding indirect taxes paid and subtracting government subsidies.

17. On the expenditure side of GDP, government consumption and investment data are obtained from the government budget estimates, and exports and imports data from the Customs Department; gross investment is estimated based on survey data on value added and imports of capital goods, namely, machinery, transport equipment and construction goods. Private consumption is derived as a residual.

18. The FOS also estimates of a price deflator for each sector of activity that is used to arrive at sectoral value added at constant prices. In many cases, sectoral price deflators are not based on producer prices but on consumer price subindices for a particular activity. The cost data for the oil GDP deflator are based on observed data for the early 1990s, extrapolated for subsequent years. The import deflator has been used as the deflator for indirect taxes and subsidies. The implicit price deflator for gross fixed capital formation is estimated as the weighted average of the deflator for the building and construction sector and the deflator for imports. The deflator for consumption is proxied by the consumer price index (CPI). For exports and imports, unit price and exchange rate data are used to derive the deflators. The FOS data for GDP measured at current and constant prices, as well as for the implicit sectoral GDP deflators, are presented in Table 1.

Table 1.

Nigeria: Source Data-GDP and GDP Deflator estimates, 1984-97

(Value in millions of naira)

article image
Sources: Federal Office of Statistics; and staff estimates.

Includes mining and quarrying other than the petroleum and gas sectors.

Base year of the FOS data is 1984. FOS data at constant (1984) naira prices have been rebased to 1990.

Average exchange rate reportedly used by the FOS. For the period, 1984-94, this represents the official exchange rate; for the period 1995-96, this represents a weighted average of the official and the autonomous foreign exchange market (AFEM) rate. For 1997, the rate is the actual AFEM rate.

B. Limitations of the FOS Data

19. Overall, the estimation of the components of the non-oil sector do not suffer from major methodological shortcomings. However, some of the surveys conducted are outdated and there is the usual problem of the quality and timeliness of data. For the oil sector, it has become apparent that FOS estimates of value added suffer from serious methodological limitations. Although there is no clear documentation, a review of existing oil sector data and discussion with FOS officials indicate that for data through 1994, the official exchange rate has been used to convert U.S. dollar estimates of oil sector production and cost into value-added estimates expressed in local currency. This official exchange rate has been used because there is no representative exchange rate for the period 1985-94 that can be used in national accounts estimates.10 The adoption of this methodology has resulted in serious underestimation of oil and total GDP in Nigeria in naira terms, and, a corresponding overvaluation of the U.S. dollar estimates.11

20. The expenditure accounts have several important shortcomings, partly owing to unreliable external trade and government budgetary data. First, the customs cleared import data used by the FOS grossly under-records the actual imports.12 The customs data on exports, particularly non-oil exports, are also lower than the central bank estimates of exports recorded in the balance of payments table. More critically, the FOS uses official exchange rates to convert U.S. dollar estimates into naira, which generates a further downward bias in the estimation of both exports and imports. Second, the government consumption and investment data used by the FOS do not capture the expenditure incurred by all tiers of government, particularly the off-budget accounts and special funds. Third, the private investment estimates of FOS also do not fully capture oil sector investment as well as other components of domestic investment.

C. Revision of the National Accounts Data

21. The revision of the national accounts data undertaken recently by the Fund staff had two objectives: (a) to correct the downward bias in oil sector value added, as well as total GDP data provided the FOS, and (b) to derive a set of national accounts data consistent with other macroeconomic aggregates that could form the basis of staff projections of macroeconomic variables.

22. The existing IMF database is based almost entirely on the primary data generated by the FOS (Tables 1 and 3). Some adjustments are made when additional sources of data become available and when there are strong reasons for revising the FOS data. With a few exceptions, FOS sectoral and aggregate value-added data are taken as they are reported. In contrast, the accounts by expenditure category are based mainly on the staffs own estimates. Export and import data are based on balance of payments estimates, which for recent years are identical to the central bank estimates; government consumption and investment data, however, are more comprehensive than the estimates used by the FOS and include all identifiable off-budget accounts and special funds, as well as state and local governments.13 Aggregate investment and consumption are estimated as ratios to domestic absorption. Private sector consumption and investment are then derived residually. Apart from these adjustments, which are described below, the staff has recently undertaken a major revision of the oil GDP series to correct the strong downward bias associated as indicated earlier, with the use of the official exchange rate. The methodology for revising the national accounts is described below:

Revised national accounts at current prices

23. The first task in the revision exercise was construction of a representative exchange rate series. The parallel exchange rate could be a good proxy for a free market rate. However, during the period under consideration (1984-94), a large part of the transactions were actually carried out at the official rate. The staff opted for a representative exchange rate measured as the weighted average of the official and parallel exchange rate for the period 1984-94. For 1995-96, the central bank estimates of weighted average of the AFEM and the official rate (used also by the FOS) have been used as our representative exchange rates. For 1997, the AFEM rate is accepted as the approximate measure of a free market exchange rate. In general, the parallel (or the free market) exchange rate was given a two-thirds weight with only one-third weight given to the official exchange rate. The representative exchange rates are found to be significantly different from the official rates in all years except 1990 (Table 2).

Table 2.

Nigeria: Revised IMF GDP and GDP Deflator Estimates, 1984-97

(Value in millions of naira, unless otherwise indicated)

article image
Sources; Federal Office of Statistics; and staff estimates.

Adjusted for exchange rate; official exchange rate is used to derive the U.S. dollar estimate and then representative exchange rate is used to get back the naira estimates.

Weighetd average of the official and the parallel exchange rate estimated by the staff for the period 1984-94. For 1995-96, the rate represents the weighted average of the official and the autonomous foreign exchange market (AFEM) rate used by the FOS. For 1997, it represents the actual AFEM rate used by the FOS.

Gross national income is derived by adding net factor income from abroad and net transfers to the GDP measured at market prices.

24. The representative exchange rate series is then used to revalue the oil sector value-added data, expressed in current naira. This revaluation is done by converting the oil sector value-added data (reported by the FOS) into U.S. dollars at the official exchange rate, and then converting back the resulting figures into naira, using the representative rate series. This revaluation raises the naira value of GDP and also changes its composition, with the oil sector accounting for a much higher share of GDP in all years. For the period 1995-97, the FOS estimates of oil sector value-added, which are already based on weighted-average exchange rates (AFEM rate in 1997), have been kept unchanged.

25. Once the revised naira GDP series is constructed the U.S. dollar GDP series is derived simply by converting the data into U.S. dollars, using the representative exchange rate series. The revised GDP estimates measured in current prices, constant (1990) prices and in current U.S. dollars are presented in Table 2.

Revised national accounts at constant prices and implicit GDP deflators

26. The revision to the FOS naira GDP series expressed at constant (1990) prices was carried out by correcting the FOS constant price oil GDP series in naira terms by a factor representing the ratio of the representative to official exchange rate in the base year. Taking the FOS estimates of oil GDP at constant prices as the point of departure (Table 1), the exchange rate factor is used to convert the original naira figures into revised naira estimates. No revision was done for non-oil GDP. As a consequence, although there was no change in underlying oil and non-oil sector growth rates, the relative share of oil sector in GDP went up, causing the growth rate of aggregate GDP to differ from the official estimates (Tables 1 and 2).

27. The revised implicit naira GDP deflator series are then derived directly as the ratio of the revised GDP estimates measured at current prices to the revised GDP estimates measured at constant prices. The overall GDP deflator estimates differ significantly from the official deflator estimates, owing to large changes in the oil GDP deflator (Tables 1 and 2).

Revised GDP by expenditure

28. In recognition of the problem with the FOS GDP data on the expenditure side (Table 3), the staff has made its own estimates of expenditure components, namely, exports, imports, consumption and investment. The staff has used the balance of payments estimates of exports and imports of goods and services and converted them into naira, using the representative exchange rate (Table 4). Domestic absorption (sum of aggregate consumption and aggregate investment) is then calculated by subtracting the external balance (exports less imports) from the revised GDP series. The public (government) consumption and investment data are obtained from the consolidated government accounts calculated by the staff. The aggregate consumption and aggregate investment data are derived as the ratios of domestic absorption, using the trends in FOS data as a guide. The actual ratios, however, have been adjusted to correct for the downward bias in investment observed in the historical data provided by the FOS.14

Table 3.

Nigeria: Source Data, GDP by Expenditure Groups at Current and Constant Prices, 1984-97

(Value in millions of naira)

article image
Sources: Federal Office of Statistics; and staff estimates.
Table 4.

Nigeria: Revised IMF GDP Estimates by Expenditure Groups, 1984-97

(Value in millions of naira)

article image
Sources: Federal Office of Statistics; and staff estimates.

29. The GDP expenditure aggregates at constant prices are then estimated by using the data at current prices and implicit expenditure deflators. For imports and exports, the trade-weighted unit price deflators reported by the World Economic Outlook database for Nigeria have been used. However, for consumption and investment, the original FOS deflators, rebased to 1990, have been employed to derive the constant price estimates. For convenience, the residual balancing items (small, in most years) were included in the estimates of private consumption. The revised GDP by expenditure groups at current and constant prices are presented in Table 4.15

D. GDP Comparisons

30. Table 5 provides a comparison between the levels of revised GDP and those of the FOS GDP estimates. It indicates that, in current naira prices, the revised GDP is about 6-38 percent higher than the FOS GDP estimates, owing to the revaluation of the oil sector GDP discussed earlier. The four years in which revised GDP estimates are about one-third higher than the original data are 1984, 1985, 1993 and 1994. These are also the years when revised oil sector GDP estimates are about 106-219 percent higher than the original data, resulting directly from the exchange rate distortion measured by the differential between the representative rate and the official rate.

Table 5.

Nigeria: GDP Comparisons and Evolution of the Structure of GDP, 1984-97

article image
Source: Data provided by the Federal Office of Statistics; and staff estimates.

Includes mining and quarrying other than the petroleum and gas sectors.

Representative exchange rate.

E. Evolution of the Structure of GDP

31. An important question relating to national income accounts in Nigeria is whether the economy has undergone a structural shift during the last decade or so. The revised current price estimates indicate that from a high level of nearly half the size of GDP, the oil sector share has stabilized to about 40-44 percent level in recent years. Looking at the period 1984-97, it appears that the share of non-oil output peaked in 1986-88, declined during 1989-93, and recovered during the last three years largely as result of a shift in intrasectoral shares within the non-oil sector. Agriculture, whose share declined during 1989-93, recovered most of the lost ground while the share of the services sector declined; industry’s share remained unchanged in recent years but has declined somewhat over the long run (Tables 5 and Figure 1).

32. In constant 1990 prices, the sectoral shares have undergone only small changes over the period under review. The share of the oil sector has fallen somewhat from a high level in 1984-85. Unlike the trend, identified in the case of current prices, the share of agriculture remained nearly unchanged while the share of the service sector has risen. Industry, however, appeared to have suffered a mild decline.

9

Prepared by Shahabuddin M. Hossain.

10

For 1995 and 1996 data, the FOS uses weighted-average exchange rates to derive the naira estimates; beginning 1997, it uses “autonomous foreign exchange market” (AFEM) rate for the purpose.

11

In addition, as noted earlier, the FOS has, in the absence of actual data, used fixed ratios of value added to gross output to derive value added for the oil sector. Gross output is estimated from the oil production and export price data.

12

The FOS data do not capture many non-oil exports or non-dutiable imports, particularly for the oil sector. In general, estimates of imports of goods based on central bank exchange transactions data and expressed in U.S. dollars are substantially higher than the customs estimates.

13

For a description of the consolidation of the fiscal accounts undertaken by the Fund staff, see Chapter IV.

14

On average, as a share of domestic absorption, staff estimates of aggregate investment are about 10 percent higher than the FOS estimates; staff estimates of aggregate consumption are, as a consequence, about 10 percent lower than the FOS estimates of aggregate consumption.

15

Once the structure of GDP at current (and constant) prices has been derived, the estimation is a straightforward exercise. The representative exchange rate series can be used to convert naira values into U.S. dollar estimates.

Nigeria: Selected Issues and Statistical Appendix
Author: International Monetary Fund