Vous consultezThe Dynamics of Price Cost Margins: Evidence from UK Manufacturing

AuteursHolger Görg[**] [**] holger. gorg@nottingham. ac. uk ...
suitedu même auteur

Frederic Warzynski[*] [*] fwarzyns@emp. uc3m. es ...
suitedu même auteur

1 - Introduction

When the Single Market Program (SMP) was implemented in 1992, there were hopes that it would bring increased competition, meaning more productive and innovative companies, and lower prices for consumers. The SMP was therefore targeting improvement in both allocative efficiency and productive efficiency.

2 Estimating the evolution of price cost margins (PCMs, ) basically provides an estimate of change in allocative efficiency, as it measures a change in the difference between the actual price set by firms (P) and the standard of perfect competition where firms should price at marginal cost (c). Deviation from perfect competition has clear welfare implications because it leads to lower output, higher prices, and lower welfare (i.e. the sum of the consumer surplus and of the producer surplus).

3 Has the Single Market Program influenced firms’ pricing behaviour in UK manufacturing? Has it forced the introduction of a tougher price competition in the spirit of Sutton (1991)? Can we talk of a “switch of regime”? In this paper we try to answer these questions by estimating the dynamics of price cost margins in the UK manufacturing.

4 To our knowledge, this paper is the first to estimate PCMs in UK manufacturing using firm level data following the methodology of Roeger (1995). By focusing on the firm as the unit of analysis, we are able to capture the evolution of the average PCM in a given subset. We devote a particular attention to the explanation of the change in intra-industry average PCM.

5 There is a small literature that has estimated the importance of market power in the UK. Using sector level data, Small (1997) estimated the level of PCMs in 16 industries in manufacturing and services over the period 1968-1991. He found evidence of large markups in most industries, especially in services. He also found pro-cyclical markups.

6 Other papers have tackled the issue using firms’ profit margin and accounting margins to check for the presence of market power (Machin and Van Reenen, 1993; Haskel, Martin and Small, 1995). A more recent exercise by Griffith (2001) looked more particularly at the effect of the SMP on the Lerner index. Using the ARD dataset, she found that the Lerner index had declined more in sensitive industries than in the non sensitive industries.

7 Our analysis is complementary to hers in the sense that we use a different methodology that assumes that marginal costs are unobservable and indirectly estimates the deviation from perfect competition by observing the way firms adapt their input demand to changes in output.

8 Some papers have looked at the effect of the SMP in other European countries: Italy (Botasso and Sembenelli, 2001), Spain (Siotis, 2002) and Belgium (Warzynski, 2002), as well as the European Commission (1999), using sector-level data. The main finding is that PCM seem to have declined mostly in large countries, where PCMs were higher before 1992, but stayed relatively constant in small countries.

9 We find that price cost margins declined by 25% after 1990, controlling for cyclical effects, falling from 0.14 to 0.10 on average. This is an extremely large figure and we discuss whether this dramatic evolution can be attributed to the SMP. We also describe the heterogeneity of the evolution across sectors and discuss extensions to this work.

10 Section 2 details the methodology that we follow. Section 3 describes our dataset. Section 4 presents and discusses our results and section 5 concludes.

2 - Methodology

2.1 - Seminal work (Hall, 1986, 1988; Domowitz et al., 1988)

11 Assume a standard production function:

12

13 where i is a firm index, t a time index, Θit is the Hicks neutral technical progress, Kit is capital stock (selected in advance of the realisation of demand), Nit is labour input.

14 For now assume constant returns to scale and competitive pricing for inputs, assumptions that we will relax later.

15 Under perfect competition the firm prices at marginal cost. In a competitive environment, taking logs, using standard rules of derivation and expressing employment and quantities per unit of capital, it can be shown that:

16

17 However, when firms have market power, they will set a higher price that exceeds the marginal cost. In this case. Equation (1) then can be generalised as:

18

19 Incorporating material costs in the production function (Domowitz et al., 1988):

20

21 The debate about this methodology has focused on four issues:

• potential endogeneity problem and validity of instruments: if there are productivity shocks to variable factors, which are observable, the setting of input levels will not be independent of changes in outputs. Typical instruments in studies using aggregate US data are the growth of real GDP, the price of oil, the political party of the president or the growth of military purchases (Hall, 1986, 1988; Domowitz et al., 1988). Blanchard (1986) and Roeger (1995) criticised these instruments on the basis that productivity shocks are likely to be correlated with the instruments as well. To avoid this difficulty, Roeger (1995) and Oliveira Martins, Scarpetta and Pilat (1996) proposed a new method through which they could estimate markups by OLS in a consistent and unbiased way.
• time varying parameters: it is difficult to believe that the degree of market power has remained constant over time. Nevertheless, most studies estimate the average markup over a period. Exceptions are studies using firm level data with a smaller time span and/or trying to capture structural adjustments (Levinsohn, 1993; Konings et al., 2001; Bottasso and Sembenelli, 2001; Siotis, 2002; Warzynski, 2002) and sector studies trying to control for changes in some exogenous parameters like trade (Hakura, 1998) or the nature of antitrust control (Warzynski, 2001). Another aspect is the pro- or counter-cyclicality of the markup ratio, allowing the markup to change from one year to another depending on the economic activity (see e.g. Green and Porter, 1983; Rotemberg and Saloner, 1986; Rotemberg and Woodford, 1999). However, more structural aspects are likely to have changed the nature of competition.
• size of the margin: the original estimates were considered as too high and incoherent for some industries (negative markups). Successive refinements have lowered these estimates and made them more acceptable.
• firm or sector level data: the empirical methodology is based on a model of firm behaviour; yet the literature has mostly used industry level data. Moreover, by doing this, it assumes constant behaviour through time, has difficulty to identify structural change and to find instruments if these are necessary.

2.2 - Coping with endogeneity and time varying parameters

22 A recent paper by Roeger (1995) proposed an alternative method to solve the endogeneity problem presented above. He argues that imperfect competition explains the difference between primal and dual productivity measures. From subsection 2.1, we know that

23

24 We can write a similar expression for the price-based Solow residual:

25

26 where p is the logarithm of price, w the logarithm of the price of labor and r the logarithm of the price of capital.

27 Combining these two equations we obtain an expression where the price cost margin can be estimated with OLS:

28

29 Rewriting the left hand side as Δy and the right hand side as Δx the expression simplifies to:

30

31 Roeger argues that this expression can be estimated by OLS because the error term in this case is not correlated with the regressor.

32 Again including material costs and slightly rewriting the previous equation (Oliveira Martins et al., 1996):

33

34 According to the availability of the data, another rewriting leads to this expression:

35

36 where OR is operating revenue, CE cost of employees, CM cost of materials, all in nominal terms as specified in accounting data. NK is tangible fixed assets net of depreciation and PK is the user cost of capital, defined as:

37

38 where δit is the firm-specific depreciation rate, PI is the index of investment goods prices, r is the real interest rate and t is corporate taxation. PI, r and t are at the country level and time varying.

39 Eq. (6) will be the key equation that we will estimate. To make our analysis econometrically feasible, we need to impose some identifying restrictions. We have at our disposal a panel dataset. We can therefore estimate β for a given time period (βt), for a given industry, (βj) or for a given period and a given industry (βjt). This technique allows much more flexibility that what has been used in the literature.

40 Klette (1999) proposes another method that allows to measure the heterogeneity of the markup within industry. However, he is forced to rely on IV estimation and only provides rough estimates of the within group heterogeneity. While his approach stresses an important issue in the literature, we focus our analysis in this paper on the dynamic evolution of the average price cost margin at the level of the industry.

3 - Dataset

41 Our firm level data comes from the OneSource database, a commercial database derived from the accounts that companies are legally required to deposit at Companies House. The data cover the period 1987 to 1997. This means that we can analyse the dynamics of PCMs on the period 1989-1997, three years before 1992.

42 After dropping firms that were ultimate holding companies or subsidiaries under joint ownership[1] [1] These were dropped as it may lead to double counting if...
suite our dataset contains information on 18,253 firms of which 13,821 are UK-owned and 4,432 are foreign-owned. This yields a total of 124,412 observations implying that, on average, we have at least six observations per firm. However, because our analysis is in first difference and because of panel is unbalanced, the actual number of observations in our regressions will be reduced quite extensively.

4 - Results

4.1 - The dynamics of average price cost margins

43 We start by estimating Eq. (6’) by year considering all firms in the same subset. This provides some descriptive information on the average price cost margin in the UK manufacturing and its evolution. Table 1 shows the results. We observe a dramatic decline especially after 1990, i.e. before the actual implementation of the SMP, what could suggest that firms anticipated its effects but could also reflect more general macroeconomic factors, as we discuss later. The next step is to analyze the level and evolution of the PCM in more disaggregated subsets, like 2-digit industries.

1 - The evolution of the average price cost margin

44 It is important to note that there are much les observations in the first and the last year of our panel, which means that we should be careful with the composition of our repeated cross section. To control for that we replicated the estimations using observations for those firms present all years (the complete panel). Results were unchanged. Another (and more drastic) way to solve this problem is to limit our attention to the 1990-1996 period.

45 We need to make a few remarks at that stage of the analysis. First of all, we notice that the size of the PCM is relatively small, between 0.096 and 0.141. This reduction in level can be explained by the methodology employed (reducing the upward bias due to endogeneity) and the nature of the dataset. Indeed, previous studies used mostly subsets of large firms, with established market presence, while this dataset has many characteristics of a population dataset, although imperfect. The fact that, by increasing the number of observations, the level of PCMs declines looks logical as small and medium companies are probably less able by definition to benefit from market power. Second, the highly unbalanced nature of our dataset can be explained by the natural evolution of industrial structure. We do not control for this selection effect in our analysis as we simply provide a series of snapshots at various levels that describe the evolution of PCMs over time.

4.2 - Price cost margins by industry

46 We first estimate the average PCM by industry over the period 1989-1997. This documents the heterogeneity of PCMs across sectors (table 2). The highest PCMs are found in the other non metallic mineral products industry (SIC 26), the tobacco industry (SIC 16) and the medical, precision and optical instruments industry (SIC 33). These averages hide the dynamic evolution that might be different depending on the industry. In table 3 we estimate the average PCM by industry and by year. Most industries experienced a decline in PCM after 1990. The evolution would suggest that maybe some cyclical factors are more important than structural reforms. We analyse this aspect in the next subsection.

2 - Average price cost margins by 2-digit SIC industry

3 - The dynamics of average PCM by 2-digit SIC industry

4.3 - Cyclicality and SMP effect

47 We want to determine which part of the evolution of price cost margins can be attributed to a structural change and which part can be explained by cyclical factors. To do this, we first interact ∆x with a sector-specific cyclical indicator CYC. As a first step, we use the growth of real output in the 5-digit industry. We assume a component of the PCM to be constant over the period but we allow another component to vary year by year depending on the cycle:

48

49 When we make this assumption, Eq. (6’) becomes[2] [2] Derivations are straightforward and can be found in Oliveira...
suite:

50

51 We then try to detect evidence of structural change by creating a dummy POST1990 equal to 1 if year > 1990. We interact Δx with this dummy:

52

53 Table 4 shows the result of the cyclicality test without controlling for the structural change. We find a negative and significant effect in 8 industries and a positive effect in 2 industries. In table 5, we control for the structural change. This is important to consider both types of influence. First, we find evidence of counter-cyclical PCM in 11 industries and a pro-cyclical margin in only 1 industry. Second, we find a significant decline in 19 out of the 20 industries where estimation is possible. Thus, controlling for cyclical factors, we can estimate the importance of structural change. And controlling for structural change, we can better understand the cyclical nature of PCMs.

4 - Cyclicality of PCM

5 - Cyclicality and post- 1990 effect

5 - Discussion and Conclusion

54 Our analysis has illustrated a dramatic decline in price cost margins in UK manufacturing after 1990, even when controlling for conjectural factors. The most likely explanation for this impressive decline is the increase in competition generated by the SMP. Therefore, this paper contributes to a growing literature on the effect of increased competition on PCMs.

55 We have used the Hall-Roeger approach and have shown how the methodology can flexibly be adapted to analyze different issues, therefore allowing for straightforward extensions: joint estimation of PCMs and returns to scale (Hall, 1990); joint estimation of bargaining power and PCMs (Crépon, Desplatz and Mairesse, 1999), estimation of within industry heterogeneity (Klette, 1999), conduct, or the estimation of TFP growth corrected for the presence of imperfect competition. This offers new propects for future research in this area.

Bibliographie

References

Blanchard O. 1986: “Market Structure and Macroeconomic Fluctuations: Comments”, Brookings Papers on Economic Activity, 2: 323-328.

Bottasso A. and A. Sembenelli 2001: “Market power, productivity and the EU Single Market Program: Evidence from a panel of Italian firms”, European Economic Review, 45(1): 167-186.

Crépon B., R. Desplatz and J. Mairesse 1999: “A simple way to jointly estimate price cost margins, scale economies and workers’ bargaining power from a panel of French manufacturing firms”, Working paper nr. 9917, INSEE.

Domowitz I., R.G. Hubbard and B.C. Petersen 1988: “Market Structure and Cyclical Fluctuations in U.S. Manufacturing”, Review of Economics and Statistics, 70(1): 55-66.

European Commission 1999: “Report on Economic and Structural Reform in the EU”, European Economy, Supplement A.

Green E.J. and R.H. Porter 1984: “Noncooperative Collusion under Imperfect Price Information”, Econometrica, 52: 87-100.

Griffith R. 2001: “Product market competition, efficiency and agency costs: an empirical analysis. IFS Working Paper 01/12.

Hakura D. 1998: “The Effects of European Integration on the Profitability of Industries”, IMF Working Paper 98/85.

Hall R. 1988: “The Relation between Price and Marginal Cost in U.S. Industry”, Journal of Political Economy, 96: 921-947.

Hall R.E. 1990: “Invariance Properties of Solow’s Productivity Residual”, in Diamond P. (ed.), Growth, Productivity, Unemployment: Essays to celebrate Bob Solow’s Birthday, Cambridge, MA: MIT Press, 71-112.

Haskel J., C. Martinand and I. Small 1995: “Price, Marginal Cost and the Business Cycle”, Oxford Bulletin of Economics and Statistics, 57(1), 25-41.

Klette T.J. 1999: “Market Power, Scale Economies and Productivity: Estimates from a Panel of Establishment Data”. Journal of Industrial Economics, 47(4): 451-476.

Konings J., P. Van Cayseele and F. Warzynski 2001: “The Dynamics of Industrial Mark-ups in two Small Open Economies: Does National Competition Policy Matter?”, International Journal of Industrial Organization, 19: 841-859.

Levinsohn J. 1993: “Testing the imports-as-market-discipline hypothesis”, Journal of International Economics, 35: 1-22.

Machin S. and S. Van Reenen 1993: “Profit Margins and the Business Cycle: Evidence from UK Manufacturing Firms”, Journal of Industrial Economics, 41(1) 29-50.

Oliveira Martins J., D. Pilat and S. Scarpetta 1996: “Mark-Up Ratios in Manufacturing Industries: Estimates for 14 OECD Countries”, OECD Economics Department Working Paper Nr. 162.

Oliveira Martins J. and S. Scarpetta 1999: “The Levels and Cyclical Behaviour of Mark-Ups Across Countries and Market Structures”, OECD Economics Department Working Paper Nr. 213.

Roeger W. 1995: “Can Imperfect Competition Explain the Difference between Primal and Dual Productivity Measures?”, Estimates for U.S. Manufacturing, Journal of Political Economy, 103(2), 316-330.

Rotemberg J. and J. Saloner, J. 1986: “A Super-Game Theoretic Model of Business Cycles and Price Wars During Booms”, American Economic Review, 76: 390-407.

Rotemberg J. and M. Woodford 1999: “The Cyclical Behavior of Prices and Costs”, NBER Working Paper Nr. 6909.

Siotis G. 2002: “Competitive pressure and economic integration: An illustration from Spain, 1983-1996”, Mimeo, Universidad Carlos III.

Small I. 1997: “The cyclicality of mark-ups and profit margins: Some evidence for manufacturing and services”, Bank of England Working Paper Nr. 72.

Warzynski F. 2001: Did tough antitrust policy lead to lower markups in the US manufacturing industry?, Economics Letters, 70: 139-144.

Warzynski F. 2002: “The Dynamics of Price Cost Margins in Belgian Manufacturing: 1989-1995”, Mimeo.

Notes

[ *] fwarzyns@emp.uc3m.es

[ **] holger.gorg@nottingham.ac.uk

[ 1] These were dropped as it may lead to double counting if firms have consolidated accounts.

[ 2] Derivations are straightforward and can be found in Oliveira Martins and Scarpetta (1999). We thank Joaquim Oliveira Martins and Werner Roeger for clarifying this point.

Résumé

PLAN DE L'ARTICLE

• 1 - Introduction
• 2 - Methodology
  • 2.1 - Seminal work (Hall, 1986, 1988; Domowitz et al., 1988)
  • 2.2 - Coping with endogeneity and time varying parameters
• 3 - Dataset
• 4 - Results
  • 4.1 - The dynamics of average price cost margins
  • 4.2 - Price cost margins by industry
  • 4.3 - Cyclicality and SMP effect
• 5 - Discussion and Conclusion

Holger Görg et Frédéric Warzynski « The Dynamics of Price Cost Margins: Evidence from UK Manufacturing », Revue de l'OFCE 5/2006 (n^o 97 bis), p. 303-318.
URL : www.cairn.info/revue-de-l-ofce-2006-5-page-303.htm.
DOI : 10.3917/reof.073.0303.