5342.0 - Balance of Payments Statistics, Information Paper on Quality , 1996
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 20/02/1996
|Page tools: Print Page Print All|
2.4. Sampling procedures are used in several collections from which balance of payments estimates are compiled. These include:
2.5. Sample error provides a mathematical measure of the difference between an estimate derived from a sample survey and the true value that would be obtained if the whole population were enumerated. One measure of the likely difference is given by the standard error. Table 1 provides approximate standard errors on Australia's balance of payments statistics for the year 1993-94. When estimating sample error for aggregate series within the balance of payments, independence is assumed between each of the sample surveys used in compiling the balance of payments.
2.6. There are about two chances in three that a sample estimate will differ by less than one standard error from the figure that would have been obtained if all the population were enumerated, and nineteen chances in twenty that the difference will be less than two standard errors. For example, the table shows that income debits (broadly, income earned by non-resident owners of resources from the use of those resources by residents) are estimated as $21,597 million with a standard error of $277 million. Therefore, there are about two chances in three that the true figure lies within the range $21,320 million to $21,874 million and nineteen chances in twenty that it lies in the range $21,043 million to $22,151 million.
2.7. Table 1 shows that there are no standard errors associated with merchandise exports, merchandise imports and unrequited transfers (these data are not based upon sample collections). The relative standard errors for services and income debits transactions are 1% or less. Income credits (principally income earned on Australian investment abroad) has a higher relative standard error of 3%.
2.8. Care is needed in interpreting relative standard errors on net items (net services, net income, net transactions in each of foreign investment in Australia and Australian investment abroad, all the current and capital account balances, and the balancing item). The relative standard error will greatly increase as the value of a net item approaches zero. For example, while the relative standard error on net capital account transactions classified as Australian investment abroad was only 1% in 1993-94, it was 9% in 1992-93, when the net value of transactions was much smaller ($2,959 million). In 1990-91, when the extent of netting was less, the relative standard error was 5%. In the table, relative standard errors have been shown for the balances on the current and capital account, net foreign investment in Australia and net Australian investment abroad because these measures are generally significantly different from zero and are also the focus of analytic attention.
2.9. The ABS keeps sample errors under review and attempts to ensure that they are within acceptable limits. Taking account of the issue of netting in evaluating sample errors, the sample errors shown in Table 1 for annual balance of payments series are considered to be within those limits. The relative standard errors on quarterly data are usually higher for gross measures.