RDP 9209: Financial Liberalisation and Consumption Behaviour 7. Conclusions

According to the econometric evidence presented in this paper, aggregate consumption in the United States, Japan, Canada and Australia seems to be less responsive to fluctuations in income in the 1980s than in the 1970s or 1960s. The favourite candidate for explaining this phenomenon is the combined and reinforcing effects of financial liberalisation (the progressive lessening of the extent and intensity of official regulations in both national and international capital markets) and innovation. Further support for this conclusion was also evident in the findings for countries where a priori information suggests a lack of financial liberalisation over the sample period. While inter-country comparisons of precise λ values are unreliable, for reasons discussed in section III, there is nevertheless little evidence to suggest declines in this parameter in the cases of Germany, France and Italy.

These conclusions are based on empirical tests which hinge on an equation derived on the basis of several maintained hypotheses. A finding of excess sensitivity of current consumption to disposable income could be the consequence of a failure of one or more of those hypotheses. However, the broad pattern displayed by the excess sensitivity parameter, both across countries and over time, is best accounted for by the changing pattern of deregulation and financial innovation in most cases. This conclusion obtains further corroboration from tests that reveal some asymmetric household consumption behaviour in response to increases and decreases in income combined with a tendency for the magnitude of these asymmetries to fall systematically over time. Nor are the results altered when allowance is made for changing intertemporal substitution effects. One of the main contenders for the rejection of the random walk consumption model is a failure of the rational expectations maintained hypothesis. But for many countries the results obtained would imply that consumers were becoming systematically less myopic over time. That this could occur independently of financial liberalisation seems unreasonable.

To the extent that greater consumption smoothing possibilities are associated with increasing financial liberalisation, there are likely to be important implications for policy. Household demand will depend much more on the behaviour of real interest rates, financial prices and expectations, which operate via wealth and intertemporal substitution effects, than on any shift in liquidity constraints that can be manipulated easily by the authorities. Perceptions about the longer-run goals of monetary policy actions which might influence permanent income are therefore more likely to be important influences on household demand than any fluctuations in interest rates perceived to be temporary. Furthermore, short-run fiscal policy measures may be a relatively less useful tool for demand management in circumstances where liquidity constraints are not binding. This is because transitory income is less relevant for spending decisions as the proportion of the population who are liquidity constrained declines.