Variation in the Effects of Aggregate Demand Shocks: Evidence and Implications across Industrial Countries.

• Author: Kandil, Magda
• Position: Macroeconomics research

Magda Kandil [*]

Over a sample of nineteen industrial countries, more variable aggregate demand and/or higher mean inflation attenuates (augments) the effect of aggregate demand shocks on real output growth (wage and price inflation) while having no effect on the response of the real wage to such shocks. In all countries examined, aggregate demand shocks are positively (negatively) correlated with nominal variables (real output). Among explanations of the business cycle based on shocks to aggregate demand, this evidence favors the new Keynesian sticky wage explanation over the sticky price and the new classical imperfect information explanations.

1. Introduction

   Much of the research in macroeconomics during the past 20 years has focused on business cycle fluctuations, deviations in real output from the natural growth path over time. Efforts to provide an adequate explanation have focused on the role of demand and supply shocks in generating observed fluctuations. Explanations of demand-driven business cycles fall into two major categories: the new classical approach and the new Keynesian approach.

   Leading the new classical approach is the imperfect information paradigm pioneered by Lucas (1972, 1973). This is a perfectly competitive equilibrium paradigm that assumes imperfect information regarding prices to produce an explanation for economic fluctuations.

   Sticky-wage new Keynesian models, in contrast, have emphasized rigidity in the labor market in explaining economic fluctuations. Labor contracts specify the nominal wage at which firms will be able to purchase labor in advance. Contracts may be explicit formal agreements of the type specified in Fischer (1977), Gray (1978), and Taylor (1980) or implicit informal agreements of the form specified in malcomson (1984). In these models, the response of real output to aggregate demand shocks is dependent on the speed by which wages are adjusted, that is, contractual wage rigidity.

   Sticky-price new Keynesian models examine the behavior of monopolistically competitive firms who face small "menu cost" when they change prices (see, e.g., Akerlof and Yellon 1985; Mankiw 1985; Parkin 1986; Blanchard and Kiyotaki 1987; Rotemberg and Saloner 1987; Ball, Mankiw, and Romer 1988). Given the cost involved in adjusting prices, firms may opt to schedule these adjustments over time. Price rigidity is then introduced in the product market independently of the wage path and conditions in the labor market. The response of real output to aggregate demand shocks is dependent on the speed by which individual firms in the market adjust their prices.

   The leading new Keynesian and new classical alternatives offer powerful analytical tools for the theoretical explanation of business cycles that have made them key components of macro theory textbooks. The empirical validity of these alternatives has become, however, a major source of controversy in the macroeconometric literature (see, e.g., Lucas 1973; Kormendi and Meguire 1984; Ahmed 1987; Ball, Mankiw, and Romer 1988; Bills 1988; Card 1990; Kandil 1991; Gray, Kandil, and Spencer 1992; Kandil 1995, 1996, 1997, 1999; Kandil and Woods 1995, 1997). Efforts to distinguish among these alternatives empirically have often been frustrated by the observational equivalence problem where several of the predictions are shared by the alternative explanations.

   This study contributes to the empirical investigation of competing alternatives of business cycles. The investigation extends the framework employed by Ball, Mankiw, and Romer (1988) to consider adjustments in the labor market and the implications of sticky-wage models in contrast to leading alternative explanations of demand-driven business cycles. Unlike existing studies on the subject, the current investigation will offer a comprehensive evaluation of the various theoretical implications concerning variables' responses to aggregate demand shocks. Using data from 19 industrial countries for the period 1956-1994, the empirical investigation compares the theoretical predictions to fluctuations in response to aggregate demand shocks across countries. Contrary to the theoretical predictions of the new classical imperfect information model, higher trend price inflation increases the effects of aggregate demand shocks on the nominal wage and the price level and decreases their effects on real output. Furthermore , the cyclical behavior of the nominal wage challenges the implications of the sticky-price model. As advocated by sticky-wage new Keynesian models, aggregate demand variability and trend price inflation appear to be consistent with a higher degree of nominal wage flexibility in response to aggregate demand shocks. Nominal wage flexibility increases, in turn, price flexibility. The negative response of the real wage to aggregate demand shocks is, therefore, less pronounced. Subsequently, output fluctuations are smaller the higher the flexibility of the nominal wage in response to aggregate demand shocks across countries.

   The remainder of the investigation is organized as follows. Section 2 reviews the implications of the theoretical models under explanation. Section 3 describes the data and the econometric methodology. Section 4 presents the time-series results. Section 5 provides hypotheses to distinguish among the theoretical explanations and presents the evidence across countries. A summary and conclusion are provided in section 6.

2. Theoretical Models and Testable Hypotheses

   To motivate the empirical investigation of the paper, this section offers a brief background of the alternative developments under consideration. [1] Suppliers in this economy are located in a large number of scattered markets. There are three competing explanations for the behavior of the output supplied in response to aggregate demand shocks in each of these markets. [2]

   A New Classical Explanation

   The first explanation is new classical in nature. It is based on the model developed by Lucas (1972, 1973). In this model, agents in the specific market of the economy wish to change their output in response to changes in their relative prices but not in response to changes in the aggregate price level. An unanticipated price inflation in the specific market leads individuals to mistakenly perceive that the relative prices of the goods they produce are temporarily high. A rise in the price of output relative to that supplier's perception of the general price level represents a rise in perceived real wage, [3] prompting an expansion of the output supplied. The realized real wage is actually falling.

   This explanation implies that cyclical fluctuations in economic variables depend on the deviation of inflation from expected inflation. Because agents anticipate trend price inflation, this trend does not determine cyclical fluctuations in response to aggregate demand shocks. Higher variability of aggregate demand is expected, however, to decrease the response of real output to unanticipated inflation. That is because agents become less willing to interpret unanticipated inflation as relative to their specific market as the variability of aggregate demand increases. Instead, agents respond by increasing the price of the output produced in the specific market. This is equivalent to demanding higher nominal wages for the individual supplier's time, which moderates the countercyclical response of the real wage to aggregate demand shocks. Accordingly, cyclical fluctuations in the nominal wage, the real wage, and price are positively correlated in response to aggregate demand shocks. These fluctuations are negati vely correlated with the output response to aggregate demand shocks.

   A Sticky-Wage New Keynesian Explanation

   The second explanation is Keynesian in nature. It is based on the implications of the model developed by Gray and Kandil (1991). The supply side of the economy consists of a continuum of identical firms. Employment by each firm depends on the nature of contracts as well as the labor market conditions. Following contract negotiation, a positive disturbance to aggregate demand raises the price of the output produced. In the absence of any stipulation for wage indexation, the increased price lowers the real product wage faced by firms, causing it to fall below the market-clearing value. The fall in the real wage, in turn, causes output and employment to rise above their natural (full-equilibrium) levels. The increased supply of output moderates the necessary rise in price following a rise in demand.

   An increase in aggregate demand variability and/or trend price inflation increases the risk of fixing the nominal wage. This is consistent with shorter labor contracts or a higher degree of wage indexation. [4] Nominal wage flexibility moderates the countercyclical response of the real wage and, in turn, output fluctuations in response to aggregate demand shocks. The inflationary effects of aggregate demand shocks are, therefore, more pronounced on price. That is, output fluctuations are negatively correlated with cyclical adjustments of the nominal wage, the real wage, and the price level to aggregate demand shocks. Since conditions in the labor market determine cyclical fluctuations in the output market, wage and price flexibility are highly correlated in response to aggregate demand shocks.

   A New Keynesian Sticky-Price Explanation

   The third explanation is also Keynesian in nature. It is based on the implications of the model developed by Ball, Mankiw, and Romer (1988). Each market in the economy contains imperfectly competitive firms that change prices at discrete intervals rather than continuously because adjustments are costly. Price rigidity exacerbates cyclical fluctuations in real output in response to aggregate demand shocks. A positive disturbance to aggregate demand increases the output produced. Assuming a Walrasian labor market, the increase in labor demand increases the nominal wage, causing a...