Explaining GDP Quarterly Growth from its Components in the Context of the Annual Overlap Method: A Comparison of Approaches

The use of chain-linked methods reduces significantly the problem of price structure obsolescence present in fixed base environments and it has been, therefore, adopted by many countries to measure GDP. The price updating it involves introduces a dimension new to those accustomed to the fixed based methodology that may produce confusion if not accounted for. Probably the most notorious difficulty generated by the introduction of chain-linked indices to the measurement of GDP has been that the aggregate is not the direct sum of its components, thus making it harder to explain its behaviour in terms of the specific sectors. To alleviate this problem most countries publish sector contributions in conjunction with aggregate GDP growth, however, there is no consensus on a single way of calculating these contributions when the annual overlap method is applied. In this context, this document compares a number of different ways of calculating contributions that have been suggested in the relevant literature and highlights their strengths and weaknesses. The results show that the outcomes of using different measures may vary considerably under certain circumstances, such as high price volatility, and that some of the measures do not fulfil certain desirable properties. In an application to Chilean GDP we find that the differences are negligible between the measures that do account for the chain-linking but significant when compared to the traditional fixed base measure.