Markup, returns to scale, the business cycle and openness: evidence from Australian manufacturing

Economic Papers (Economic Society of Australia), March, 2004 by Michael Olive

This paper aims to measure markup and returns to scale for eight Australian manufacturing industries, for the period 1971-72 to 1984-85, and to explore the relationship between markup, the business cycle and openness to the international economy in this period. A Hall type model is used for this purpose, where allowance is made for non-stochastic time variation in the contribution of technical change to output growth and intermediate materials are included in the production function.

Keywords: Business cycle, Manufacturing, Markup, Openness

JEL Codes: C5, L1

1 Introduction

The Solow residual is a measure of the contribution of technical change to economic growth using an aggregate production function approach (Solow, 1957). In a series of papers, Hall (1986, 1988, 1990) concludes that it is a flawed measure, as it does not take into account imperfect competition, and that market power is a major reason for the empirical observation that the Solow residual is pro-cyclical. In the process, Hall devises a method for estimating the industry markup (price divided by marginal cost) as a parameter in a single equation regression, thus avoiding the need to directly measure marginal cost. He finds that markups are significantly greater than one in most two-digit US industries.

Developments on the Hall method include adding intermediate factors into the production function, simultaneously measuring returns to scale and markup in the one estimating equation, and estimating the impact of cyclical and structural variables on markup. Domowitz et al. (1988) find that the estimated markups in the US are not as great when intermediate inputs are included, however they are still significantly greater than one. This is the conclusion of most studies of this type when taken over a range of countries, although Norrbin (1993) finds that markups are not significantly different from one in nearly all US industries when non-wage compensation is added to labour costs.

Returns to scale, measured as average cost divided by marginal cost, can also be estimated as a parameter using Hall's (1990) method. Haskel et al. (1995) and Linnemann (1999) find constant returns to scale in UK and German manufacturing industries, respectively, while Basu and Fernald (1997) and Klette (1999) find a mix of constant and decreasing returns to scale for US and Norwegian manufacturing industries, respectively.

In this paper, we estimate markup and returns to scale for eight Australian manufacturing industries, for the period 1971-72 to 1984-85, using a Hall type model. The novelty of our approach is that we use fixed cross-section and time effects to try and account for the contribution of technical change to industry growth. Most of the authors mentioned above assume that this contribution can be represented as a constant and a random error term. However, the Solow residual may be pro-cyclical, even when imperfect competition is accounted for. One reason is that exogenous changes in technology may be driving the business cycle and the Solow residual. An alternative reason is that technology may be biased toward a particular input factor that is itself pro-cyclical, resulting in higher total factor productivity at the top of the cycle.

Extending the analysis further, we examine the impact of the business cycle and openness on markups. Macroeconomists have been interested in the cyclical nature of markups in order to support arguments about the cyclical nature of real wages and also to understand the extent to which markups affect booms (for discussions see Bils, 1987; and Rotemberg and Woodford, 1999). However, there is no consensus on the direction of the cyclical influences on markup. Pro-cyclical markups are found by Domowitz et al. (1988) for US industries, Haskel et al. (1995) for UK manufacturing industries, Beccarello (1996) for G-7 country manufacturing industries and Bloch and Madsen (2001) for Australian industries. Counter-cyclical markups are found by Oliviera Martins et al. (1996) for 14 OECD countries and by Linnemann (1999) for German manufacturing industry.

Olive (2002) derives an industry pricing equation that includes real manufacturing demand as an independent variable. Using Australian manufacturing data for 24 industries, markups are positively related to manufacturing demand when costs are held constant. Using a similar method at the four-digit ISIC level, Bloch and Olive (1999) find that aggregate demand only impacts on markup for low import share industries. This suggests that markups could become less responsive to aggregate demand the more open industries are to the international economy.

Openness to the international economy is seen to represent the extent of foreign competition, rather than the level of competitiveness in a market (Feinberg, 1986). However, markups are still found to be negatively related to openness in most applied studies (see Feinberg and Shannon, 1994; Katics and Peterson, 1994; Lopez and Lopez, 1996; and Ghosal, 2000). This is interpreted as trade increasing competition in the domestic market and thereby reducing domestic market power. As such, it is often used as an argument to support tariff reduction. Freedman and Stonecash (1997) outline the importance of this argument in the development of Australia's competition and trade policies.