Chain-linking in Austrian quarterly national accounts and the business cycle

Abstract

In 2005, European Union member countries began to calculate national account volume estimates using prices from the previous year, rather than from a fixed base year. For quarterly national accounts, the average of the total previous year – and not of the previous quarter – began to serve as the price basis. This allows for the use of a Laspeyres-type quantity index. In order to obtain a time series of absolute values of volume estimates, it is necessary to chain-link growth rates. This is straightforward when calculating annual figures, but when calculating quarterly figures, EU countries can choose from one of three methods. This results in different outputs, time-series properties and, possibly, price-adjusted quarterly national account figures. The current study demonstrates the different results obtained using the three methods, when applied to Austrian quarterly GDP data. I observe the consequences of consecutive time-series-based processing, such as seasonal adjustment and business cycle analysis. Although dating turning points are rather robust using all three methods, seasonal and workday adjustment and detection of outliers based on time-series modelling can be negatively affected, as can business cycle dating.