V. Saving behaviour and the effectiveness of fiscal policy

OECD Economic Outlook, Dec, 2004

Several estimators have been used to estimate reduced-form equations such as Equation (3), including the pooled mean group (PMG) estimator, which allows for cross-country heterogeneity in the coefficients, and a variety of GMM estimators, which deal with persistence in saving dynamics and joint endogeneity among the regressors. Although appealing, the possibility of slope heterogeneity may be exaggerated in the case of public saving, the main parameter of interest in this analysis. (13)

Table V.1 (main text) reports the results of the estimation of Equation (3) for a panel of at most 16 OECD countries spanning the period 1970-2002. The data set includes Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Italy, Japan, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, United Kingdom and United States. Country selection was contingent on data availability. The main source of data is the OECD Economic Outlook 74 database. The regressions are estimated using the Arellano-Bond (1991) difference-GMM estimator to take account of inertia in saving behaviour and the joint endogeneity of the explanatory variables. (14) The preferred specification includes proxies for wealth effects (equity market and housing price indices), in addition to the conventional controls. (15) The baseline estimates of a private saving offset of about 50 per cent in the short term and 70 per cent in the long term are discussed in the main text. (16) When the conventional controls are included, the Sargan test of overidentifying restrictions does not reject the orthogonality of the instruments and the error terms, thus underscoring the appropriateness of the model specifications. The tests of first- and second-order serial correlation of the first-differenced error terms also confirm the adequacy of the lags of the explanatory variables used as instruments in the models reported.