Aggregate demand, uncertainty and oil prices: the 1990 oil shock in comparative perspective

BIS Economic Papers No 31
August 1991

Introduction

With the doubling of oil prices and the greater potential for a supply disruption following the outbreak of the crisis in the Persian Gulf the general economic picture deteriorated significantly in the second half of 1990 for most industrial economies. By the end of 1990 the rise in oil prices was associated with slowing output growth or deepening recession and somewhat higher inflation rates. The slowdown continued into 1991 despite the decline of oil prices to around their pre-crisis level. In some respects, this was not surprising. For the oil-importing countries, the rise in oil prices represented a terms-of-trade loss and an associated fall in real income, as well as an important negative "supply shock" reflecting the central role of oil as an intermediate input in the production process. Earlier dramatic increases in the price of oil, in 1973-74 and 1978-80, had been followed not only by higher inflation, higher interest rates and falling output in most industrial countries, but also by major restructuring in production and consumption.

In contrast to actual economic developments, however, most forecasts immediately following the outbreak of the Gulf crisis had predicted fairly moderate adverse effects from higher oil prices - much more moderate than direct analogies with previous oil shocks would have seemed to indicate. Not only were the effects predicted limited in magnitude, they were also short in duration. For example, Graph 1 presents estimates which are illustrative of the results typically found in more complex simulations of econometric models underlying published forecasts. The net cumulative effect on output and the price level (percentage change) one year after a 1% rise in the price of oil for the United States, Japan and Germany had been declining for a number of years prior to the oil shock last August. The graph shows the estimated net (reduced form) effect of an oil price shock on output and prices, together with 90% statistical confidence intervals, from the mid-1970s until 1990. Each data point on the solid line shows the estimated cumulative impact of an oil price rise evaluated at the date shown, and the broken lines show the statistical confidence intervals, holding constant the effects of monetary and fiscal policies. The net effect of oil on both prices and output levels is seen to decrease over time in five of the six panels, i.e. the solid line moves closer to zero and the confidence boundaries tend to encompass zero (no significant effect) at the end of the sample period. This pattern suggests that the major industrial countries were in a better position to weather an oil shock in mid-1990 than had previously been the case.

Statistical results of this nature are also consistent with several empirical developments, which in turn had helped to underpin initially optimistic views about the limited effects of an oil price shock. More resilient supply-side features of economies are one aspect, as production profiles have become less energy intensive and presumably more flexible in adjusting to energy price changes.In addition, the development and deregulation of financial markets during the 1980s should in principle also have allowed economies more flexibility in responding to oil shocks without major output disruptions or inflationary consequence Finally, the extent to which aggregate demand was affected by the oil shock - working through real income and wealth effects - should have been more limited because of the reduced share of imported oil in total production.

Why then did economic conditions deteriorate so quickly after the oil shock? In this paper we argue that three related aggregate demand factors led to weaker-than-expected output performance following the mid-1990 oil shock, all of which had also been evident at the time of the 1973-74 oil price hike. Specifically, the increase in uncertainty due to the Gulf crisis - over oil supplies, price hikes and regional conflict - led to a sharp fall in consumer and business confidence and related weakness in domestic demand in a number of countries. In some respects this is similar to the uncertainty in 1973 over the Yom Kippur war in the Middle East, the oil embargo and the rise in oil prices. In addition, we argue that the momentum and intensity of the business cycle downturn in mid-1990, much as in 1973, had reduced the resilience of many economies to an adverse oil shock. The oil price shock and associated economic uncertainty in 1990 came at a time when a downturn was already under way and greatly weakened business conditions in several large industrial economies. Finally, we show that a very low level of business confidence existed in several large industrial economies even before the beginning of the Gulf crisis, which not only directly contributed to the subsequent weakness in investment and consumption expenditure, but may also have magnified the decline in confidence and fall in demand growth by the end of 1990.

Placing the 1990 oil shock in comparative perspective, this study investigates the role of aggregate demand factors in determining the response of the economy to oil shocks. We provide a non-technical and eclectic review of the demand-side issues involved, with particular attention devoted to contrasting the 1990 shock with the events and consequences associated with the oil price hikes of the 1970s in the major industrial economies. The objective is to gain an insight into the demand-side channels of oil shock transmission and how these factors have affected the output and price responses following an oil shock. In our review of past oil shocks we find that declines in output growth and increases in inflation around the time of previous oil shocks were in large measure related to demand factor. We also argue that a focus on demand factors provides the clearest insight into why the 1990 price swing had such adverse, and largely unanticipated, effects on the real economy.

The structure of this paper is based on the three basic channels through which aggregate demand factors may influence the response of the economy to an increase in oil prices. First, we identify factors which influence the magnitude of the direct aggregate demand effect of an oil shock. By this we mean factors such as the degree of dependence on imported oil and the ability to substitute other non-oil energy sources, which determine the ultimate terms-of-trade and wealth effects of an oil shock on aggregate demand. Secondly, we consider the initial (cyclical) demand conditions at the time of the oil shock and evaluate their likely effects on the transmission of oil shocks to real activity and inflation. In particular, we compare demand conditions prevailing in mid-1990 - factors such as capacity constraints, inflation and wage pressures and business conditions - with those prevailing prior to the two oil shocks in the 1970s and assess whether they tended to be more or less favourable. Finally, we consider the aggregate demand policy responses to previous oil shocks, in their fiscal and monetary dimensions, and evaluate the role they played in determining the ultimate pattern of output and inflation. Our concentration is on broad differences in policies between the periods and, for that purpose, we do not attempt a detailed country-specific analysis.

The analysis is organised as follows. Section II considers changes in the structural features of the major industrial economies which influence the real income and wealth effects of an oil price shock, and thereby directly affect the demand response. Section III examines cyclical conditions from the same perspective, and Section 1V reviews previous policy responses to oil shocks and the state of policy immediately following the 1990 oil shock. Section V summarises the findings.