Pathological economic phenomena like hyperinflation or chronic inflation have been often attributed to nonfundamental influences, which implies that they are difficult to control by conventional policy instruments. In the literature, such influences have been given different names--for example, bubbles, sunspots, or extraneous effects--but essentially they all represent the homogenous component of the solution to an expectational equation. Hamilton and Whiteman (1985), extending the results by Flood and Garber (1980), have shown that the presence of nonfundamental influences can only be verified by analyzing the difference between the dynamics of the narrow money supply and the price level series, as reflected in their respective orders of integration. If the former is exogenous with respect to the latter, a nonfundamental influence can be assumed to exist.
Evans (1990) criticized this approach, asserting that conventional tests fail to reject the hypothesis that a particular exploding bubble process is nonstationary. Accordingly, he concluded that the Hamilton and Whitman approach has limited practical application.
This paper illustrates the reasons for Evans’ results and explains how his critique can be overcome by using fractionally integrated processes, which allow greater flexibility in modelling long-cycle components of nonstationary time series by extending the standard time - series methodology to noninteger orders of integration.
The empirical part of the paper focuses on the estimation of the fractional orders of integration of money supply and price level in six countries using an exact maximum-likelihood method devised by Sowell (1990). The objective was to test the hypothesis that high inflation is caused by self-fulfilling expectations, as reflected by a homogeneous term in the solution to the expectational equation. For all six countries, the presence of an homogeneous component is rejected. However, in a few cases there is a slight difference in the order of integration, which is attributed to protection from inflation accorded to particular groups, for example in the form of wage indexation.