RDP 9807: Inflation Targeting in a Small Open Economy Appendix A: Data

Nargis Bharucha and Christopher Kent

July 1998

Real gross domestic product, GDP(P)
Definition: Production measure of GDP by industry (1989/90=100, seasonally adjusted). Components of this series are used to construct non-traded GDP.
Source: ABS National Accounts Cat. No. 5206.0.
Underlying consumer price index (CPI)
Definition: Components of the Treasury underlying consumer price index (1989/90=100) are used to construct our measures of traded and non-traded inflation (see Kearns 1997). Based on analysis of input-output tables, the findings of Dwyer (1992) and Dwyer and Groeger (1994) are used to classify industries as traded or non-traded according to the 10 per cent threshold discussed in Section 3. A concordance between these industries and expenditure classes is made. A Laspeyres index of traded and non-traded prices is then constructed.
Source: Consumer Price Index, ABS Cat. No. 6410.0.
Unit labour costs
Definition: RBA's measure of underlying unit labour costs per wage and salary earner.
Source: Reserve Bank of Australia, unpublished data.
Tariff-adjusted import prices
Definition: Tariff-adjusted implicit price deflator for merchandise imports, excluding exogenous imports, computers and other lumpy import items.
Source: The constant and current price series for merchandise imports less exogenous items are taken from Balance of Payments ABS Cat. No. 5302.0. Data prior to September 1981 are taken from the SITC database. Constant and current price series for computers are unpublished data provided by the Australian Bureau of Statistics. Taxes on international traded are drawn from ABS Cat. No. 5206.0.
Aggregate and non-traded output gaps
Definition: Components of GDP(P) are classified as traded or non-traded based on Dwyer and Groeger (1994). Output gaps are constructed for both aggregate and non-traded GDP(P) by taking deviations of the log level of the series from a Hodrick-Prescott filter (with a smoothing parameter, lambda, set to 1,600).
Source: ABS National Accounts Cat. No. 5206.0.

In Section 4 of the paper we argue that a non-traded output gap may be more relevant to the determination of non-traded inflation than an aggregate output gap. Aggregate and non-traded output gaps move together quite closely (in part because the non-traded sector accounts for about two-thirds of total output). Replacing the aggregate output gap with the non-traded output gap in the regressions reported in Table 1 did not alter the results significantly. In part this is because the econometric specification captures the effect of the output gap on margins, but does not capture the important effect of the output gap on inflation via its effect on wages. Also, the Hodrick-Prescott methodology allows potential output for the traded sector to move towards actual output gradually, which is part of the reason why the aggregate and non-traded output gaps behave so similarly. Calculating potential output by linking local peaks in the log level of output did not alter the results.