VIII National Accounts
- Marcello Caiola
- Published Date:
- August 1995
In most developing countries, the only national account statistics available are annual nominal GDP or GNP and the distribution of GDP by sector and by expenditure. In a few cases, quarterly estimates and national income statistics are also available.18 However, the quality of the national accounts is usually poor, and estimates should be used with great caution. In general, it may be assumed that the annual rates of changes of GDP and, therefore, the general trends, reflect reasonably well what has been going on in the country. Hence, even if the level of GDP is incorrect, it would still provide a more or less consistent benchmark against which to compare other economic and financial statistics.
To improve the analysis of a country’s financial situation and developments, it is very important to be aware of the shortcomings and deficiencies of the national accounts. Therefore, as a first step, the desk economist, or the staff following the real sector during a mission, should become familiar with the compiler, as well as the sources used in the country to estimate the national accounts. Usually, real sector statistics are prepared by central banks, planning offices, statistical offices, or certain specific ministries. In certain countries, preparation of national accounts statistics are not centralized, and responsibility is delegated to more than one institution. Also, in certain countries, conflicting national accounts estimates are prepared by different agencies, and no attempt has been made to reconcile these differences.
Quality of Data
From the point of view of the IMF’s analysis, the distribution of GDP by expenditure components is probably the most important. Yet, the quality of these data is often unreliable for several reasons. National accounts statistics are often based on preliminary estimates or are derived as an extrapolation of other variables that have been used as proxies because of the lack of more comprehensive and reliable data. Thus, even though data on public sector operations are either not available or are collected with considerable delay, national accounts usually include estimates for government consumption and for public sector investment. Often, these data have very little in common with actual fiscal estimates, inasmuch as the national accounts may reflect budgeted (rather than actual) outlays or represent, as mentioned above, extrapolation of other estimates. In addition, in countries where information on fiscal operations is available only on a cash basis, adjustments may have been made by the compilers of the national accounts to estimate expenditure on an accrual basis, but these adjustments are often based on preset percentage adjustments or may reflect budgeted rather than actual outlays.
Similarly, data on exports and imports of goods and services sometimes cannot be reconciled easily with balance of payments statistics, because the former may be based on preliminary estimates. An additional problem is the exchange rate used to convert balance of payments estimates into local currency, particularly for countries with multiple currency practices or flexible exchange rates.
Estimates for private investment are often based on imports of capital goods, but little effort may have been made to separate public sector imports from private imports, with the result that investment estimates may be inflated because of duplications.
The coverage of national accounts may also be less than satisfactory, because certain transactions, particularly illegal operations (contraband, narcotics, black market, etc), are usually ignored. As a result, nominal GDP may be underestimated, particularly in countries where the size of the parallel economy is growing rapidly.19 National accounts statistics may also underestimate nominal GDP by excluding transactions of binational enterprises (for example, in Paraguay).
Reconciliation of Data
As mentioned above, national accounts estimates may differ considerably from estimates derived from other sources. Before deciding which data to use, staff should investigate these discrepancies further. If a reconciliation is not feasible because sources of some of the data are vague and/or undocumented, the economist must use his judgment in deciding which figures to use. However, once a decision has been made, adjustments must be made to other sets of statistics (such as nonfinancial public sector, financial system, balance of payments) to achieve consistency of data. In certain countries, the task of staff is complicated by the fact that two official sets of national accounts are available, in which case the staff may have to explain to the authorities the reasons for accepting or rejecting a set of estimates.
The following summarizes the most common sources of discrepancy:
Differences with public sector operations. The main discrepancies between national accounts estimates on public sector operations (general government and/or state-owned enterprises) and data derived from fiscal sources may be attributed to coverage, definition, difference of sources, or valuation.
The coverage of the sector may differ because:
Certain agencies or transactions are not covered in one of two sets of data; for example, national accounts do not cover extrabudgetary operations; or the compiler of the national accounts has estimated the operations of certain government agencies and units, for which actual data are not available;
National accounts data include imputed transactions, which are not covered by fiscal statistics;
National accounts include provisions for depreciation, which are excluded in the fiscal accounts; and
National accounts data are on an accrual basis, whereas public sector estimates are on a cash basis.
The fact that national accounts data for general government consumption are net of purchases of goods and services from agencies of the general government, is an example of discrepancy attributed to definition. Discrepancies may be due also to the different sources used by compilers. For instance, national accounts estimates may have been based on budgets or on preliminary data, whereas fiscal data are based on actual operations; or public investment in the national accounts has been estimated by using a standard formula related to import of capital goods, rather than on the basis of actual data. Valuation differences may reflect adjustments to data due to exchange rate fluctuations, or in order to agree with certain other sources (such as external donors).
Differences with the balance of payments. The main discrepancies between national accounts and balance of payments statistics are attributable to coverage (such as different estimates of transactions) and valuation (exchange rate used to convert external transactions in local currency).
Rule of thumb. In countries where national accounts cannot be reviewed in depth, the desk economist should follow certain basic rules to ensure that national accounts data for expenditure are by and large consistent with other available information. These rules are summarized as follows:
Review the overall coverage of GDP and make adjustments, if possible, for transactions that have been left out (for example, binational enterprises, parallel economy, extra-budgetary operations);
Revise data for exports and imports of goods and services to agree with the balance of payments statistics. At this stage, the issue of the exchange rate used to convert balance of payments transactions in the national accounts should also be addressed;
As a result, total domestic demand would be obtained as a residual;
Review the quality of investment data, including their sources. If no revision is made, national accounts data for investment should be accepted and consumption revised to become the final residual entry;
Use fiscal statistics to estimate government consumption, if the former are either considered more reliable than national accounts data or are used for negotiating purposes. Private consumption would be a residual; and
To ensure the overall consistency of data, some additional checks can be made, including comparing the rate of growth of nominal GDP with the rate of growth of private sector financial assets; comparing consumption with receipts from domestic indirect taxes; and comparing the rate of growth of GDP and imports.
Measurement of Savings
Domestic savings are defined as the difference between disposable income and final consumption expenditure, and they are also equal to the difference between investment and external financing. The latter is defined to include external capital movements, including capital transfers, and is equal to the current account of the balance of payments with a reversed sign. Since in many developing countries data on disposable income are deficient, in practice the staff estimate savings as the difference between investment and the current account of the balance of payments. However, staff should be aware that until the recent revision of the balance of payments manual in 1993, the current account of the balance of payments included capital transfers. As balance of payments statistics are still in the process of being converted to the new manual—a process that is likely to take some time—caution should be taken to exclude capital transfers from the current account.
Domestic savings are also classified between government and private savings, and usually the latter is estimated as a residual item. In this context, it is of great importance to determine whether government operations are on a cash or on accrual basis. In particular, whenever a government accumulates arrears in its current operations (for example, nonpayment of interest and of purchases of goods and services), a cash presentation would underestimate the government’s current account deficit and overestimate its savings, while underestimating, as a result, private savings.
IMF documents refer to gross domestic product (GDP), gross domestic savings (GDS), gross national product (GNP), gross national savings (GNS), nonfactor services (NFS), and net international reserves (NIR). Since some of these concepts are not recognized in the System of National Accounts, Table 9 presents a simplified national account scheme for the purpose of stressing the differences among these definitions, which refer to the treatment of income received from or remitted abroad. At sectorial level, the difference would refer to the treatment of total, that is, domestic and external income. In the example in the table, domestic savings, as well as government and private savings, differ considerably, depending on the definition (GDS or GNS) used.
|1.||Gross Domestic Product (GDP)|
|Consumption||600||Net domestic product||552|
|Investment1||100||Foreign income, net||–30|
|Exports2||100||Current external transfers||20|
|2.||Gross National Product (GNP)|
|3.||Balance of Payments|
|Foreign income, net||–30|
|Net international reserves (incr. –)||–10|
|4.||Flow of Funds||Total||Government||Rest of Economy|
|Current external transfers||20||0||20|
|National savings (GNS)||70||–40||110|
|Domestic income, net||0||–15||15|
|Foreign income, net||–30||–17||–13|
|Domestic savings (GDS)||40||–72||112|
|Capital external transfers||30||20||10|
|Net international reserves (incr. –)||–10||0||–10|