The weighting scheme has been incorporated into the indicator by scaling the volume indicator to match the 1987 weights. The ratio of the sums of agricultural output in current and constant dollars can then provide the implicit deflator for agricultural output. It should be emphasized, however, that use of single base year joined in a chain-linked series raises problems over a period encompassing profound structural changes and significant movements in relative prices, as certainly occurred during a period of 20+ years. The problems exist in indicators for a single country and are simply more apparent when country indicators are agglomerated.
It is debatable whether an analytical ratio, say, "agricultural output to GDP," should be a weighted average of country ratios or a ratio of separately aggregated numerator and denominator. The results will be different, sometimes significantly so. World Tables first estimates the numerator and denominator for the group and then calculates the group ratio.
World Tables strives for group time series that retain the same country composition over time and across topics. It does so by permitting group measures to be compiled only if the country data available for a given year account for at least two-thirds of the full group, as defined by 1987 benchmarks. So long as that criterion is met, uncurrent reporters (and those not providing ample history) are assumed to behave like the portion of the group that does provide estimates.
The same technique applies to regional groupings in the main part of the topical pages. The benchmarking procedure underlying group measures requires that some weight be assigned to each economy within the group. For an economy reporting inadequate data in every year, Bank staff must choose some arbitrary 1987 base value within a broad range of plausible estimates. Readers should keep in mind that the purpose is to maintain an appropriate relationship across topics, despite myriad country problems, and that nothing meaningful can be deduced about behavior at the country level by working back from group indicators. In addition, the weighting process may result in discrepancies between summed subgroup figures and overall totals.
For example, if agriculture has a 9 percent growth rate and accounts for a third of GDP, it contributes 3 percentage points to GDP growth. If industry and services remained unchanged, GDP's growth rate would have been 3 percent. In practice, contributions are obtained by expressing the year-to-year change in a component, for instance, agriculture, as a percentage of GDP in the earlier year.
National accounts in constant 1987 prices are used. Regional aggregation requires that country estimates be expressed in 1987 US dollars and then summed. Even so, regional aggregation raises an index number problem closely related to the partial rebasing issue arising at the country level (see Sources and Methods), which is resolved in a similar manner. Moreover, regional measures of the contribution to GDP of resource balance are actually the difference between separately compiled measures for exports and imports (of goods and nonfactor services); regional measures for private consumption are then derived as a residual (GDP less government consumption, investment, and the net export of goods and nonfactor services).