The economics of the firm and industry dynamics Michael Dietrich University of Sheffield, UK Summer School: “Knowledge Dynamics, Industrial Evolution, Economic Development”
Objective of presentation: identify a few gaps in the existing literature on the economics of the firm. The firm exists in two separate areas in economics: the “economics of the firm” (the institutional firm) the “theory of the firm” (the technical firm). The economics of the firm: internal structure, organisation and boundaries. The theory of the firm: behaviour in particular market contexts - demand-cost interaction. “Realistic firms” (Coase’s (1937, 1993) terminology): both institutional and technical. Assume knowledge of the “theory of the firm”.
Institutional analysis of the firm assumes (usually) technical aspects exogenous and constant (or ignores them). Technical analysis of firms tends to ignore institutional characteristics. Obvious exceptions e.g. Chandler (1977, 1990). But the “core” analysis of the firm separates technical and institutional analysis. Presentation (Dietrich and Krafft 2011, 2012): separation constrains the analysis of the firm. Analysis of “real” firms needs both institutional and technical factors with the interconnections emphasized. One of the major gaps that I want to analyse.
Title of the presentation: “industry dynamics”. “Empirical background” for analysis of real firms. Industry dynamics = “industry life cycles”. Klepper 1996, 1997; Dietrich and Krafft 2012 Distinction: “early” and “later” stages. Transition or shakeout: important change in the way in which firms and markets function. Early stages: atomistic market structures; high levels of entry and exit based on firm specific knowledge; specific entrepreneur-firm knowledge drives product rather than process innovation. Later life cycle stages: oligopolistic market structures; constrained entry and exit; firm strategies and interaction drive process rather than product innovation.
Rest of the presentation: 1. Some of the existing literature on the firm. 2. Particular conceptualisation of transaction costs. 3. Simple linear oligopoly model of real firms. 4. Monopoly model of real firms with non-linear costs. 5. Conclusion in the form of two conjectures. 3 and 4 are the key sections.
Existing literature on the economics of the firm Two branches (Dietrich and Krafft 2011). Equilibrium approaches: technical aspects provide exogenous constraints to which firms respond. Process approaches: unique firm characteristics channel competence development and competition as a process. Equilibrium approaches: transaction cost economics (TCE) e.g. Coase, Williamson; principal-agent or financial theory e.g Hart; property rights theory e.g. Alchian and Demsetz Concentrate on TCE: comparative static exogenous changes in transaction characteristics lead to efficient institutional changes; behaviour economizing not strategising: Williamson (1991).
Process approaches: use well known literature (e.g. Penrose 1959; Richardson 1972; Nelson and Winter 1982); particularly in business and management writing. Firms have: unique characteristics or competences; based on (largely) tacit and non-transferable knowledge (at least in the short-run).
Two approaches: different accounts of the rationale for the firm. TCE: key (exogenous) factors in transaction complexity are uncertainty and asset specificity. High transaction complexity: large organizational efficiency savings from not using free markets. Institutional rationale for firms. Process approaches: firms group complementary knowledge sets or skills. Activities necessary but not based on complementary knowledge supplied by other firms i.e. the market. Two approaches might be relevant in different circumstances. TCE: change drivers are external and firms adjust. Process reasoning: change drivers internal and the result of unique firm characteristics.
Saviotti (1996) and Nooteboom (2004): different cognitive requirements. Equilibrium based firms: homogeneity of cognition for efficiency. Process based firms: diversity of cognition for internal drivers and competence development. Neo-Austrian combination of two approaches (Langlois 2003; Langlois and Roberston 1995). TCs can be dynamic and static. Dynamic TCs: organization costs of competence development and strategic change. Static TCs exist in equilibrium.
All this literature relevant in the early stages of industry life cycles: 1.With Austrian theory: change is driven by new firm entry and specific entrepreneur knowledge. Without free entry (and new knowledge creation) equilibrium monopoly profits become important. 2.Strategy (and dynamic TCs) a different meaning from strategy based on firm interaction: - shifting firm long-run focus; - oligopoly market functioning. With TCE, the rejection of strategising and emphasis on economising suggests a rejection of both meanings of firm strategy.
Gaps in the literature on the firm: - interaction between technical and institutional factors; - importance of strategic interdependence among firms; - resulting importance of monopoly power. Three points will be developed in two stages: simple linear Cournot-type oligopoly model that incorporates both institutional and technical factors; a monopoly model with non-linear costs. Economics of the firm for later life-cycle analysis.
Conceptualising TCs TCs = costs of organisational effort: search, negotiation and policing involved with the management of firm activity. Functional form, for any single firm: TC = f(p O, e) p O = prices of organisational factors of production; e = organisational effort (e.g. organisational labour hours) to manage search, negotiation and policing. Two measures of firm size: 1. Real output (or inputs into real production). 2. As an organisational unit (TC or e). These two measures are not independent.
With constant p O : f’(e) = > 0i.e. change in TC with extra effort. Size of derivative depends on transaction complexity: increasing complexity increases this derivative. For traditional TCE complexity is the result of (in particular) transaction uncertainty and asset specificity.
A Cournot-type duopoly model of real firms Dietrich and Krafft (2012). Linear model of real firms: incorporates strategic interdependence. Assume two (identical) firms with inverse demand functions: p i = α 0 – α 1 X i – α 2 X j + α 3 e i 1. Product differentiation exists when α 1 and α 2 are not equal. 2. Organisation effort (e) in the demand function reflects (e.g.) greater search and negotiation that increases p without a reduction in X. Firm average production costs (AC Pi ): - constant returns technology - respond to organisational effort: more effective internal firm management or more effective search, negotiation or policing in input markets: AC Pi = β 0 – β 1 e i > 0 Because of the linearity assumption: organisational capacity constraint e i < β 0 /β 1.
Two choice variables: X, e. Specify profit functions for the two firms. Assume Cournot conjectures for both X and e. Maximise profits: firms choose X and e given the Cournot assumption. Equilibrium X 1 (there is an equivalent function for firm 2): Equilibrium firm output depends on the organisational effort of both firms. This is different from a “normal” Cournot model where X 1 is a function of X 2.
To solve for firm effort: formulate effort reaction (or best response) functions. For firm 1 e 1 depends on e 2 (equivalent function for firm 2): Analyse effort levels with diagrams. Firms are followers, equilibrium effort defined where the two effort reaction functions intersect.
Effort reaction functions Zero effort = no organisation i.e. no firm. In oligopoly markets the existence of firms depends on (a) market as well as firm functioning i.e. strategic interaction and (b) a required degree of monopoly power (product differentiation). With e i > 0 can solve for positive real output. Organisational activity precedes output production.
Monopoly model of real firms with non-linear costs (work in progress) Linear inverse demand function for a single firm p = α 0 – α 1 X + α 2 e Generalised unit production costs: AC P = f(X, e) Define profit function. Maximise profit with choice variables X and e. Equilibrium output is a function of organisational effort:
Equilibrium effort is a (complex) function of f’(e): From earlier: f’(e) > 0 depends on transaction complexity: increasing complexity increases this derivative. In traditional TCE: complexity is the result of (in particular) transaction uncertainty and asset specificity.
Use a specific form for unit production costs: AC P = X 0 e ε ε < 0 Constant returns in production Constant elasticity on the effect of effort: ε determines the impact of e reducing AC P. Traditional TCE assumes ε = 0 separable organisational and technical aspects of the firm. We assume two possible values: ε = -0.1 (i.e. “close” to traditional TCE) and ε = -2 (i.e. a “large” impact of effort). With ε = -0.1 also assume α 0 = 10 (i.e. a “small” total market). with ε = -2 also assume α 0 =100 (i.e. a “large” total market). For all results α 1 =0.5 and α 2 =1.
Simulation results: Horizontal axis: extent of transaction complexity. Vertical axis: equilibrium effort. Increasing e → higher TC
Comments: LHS diagram: minimum f’(e) before positive e and a positive slope. With f’(e) less than the threshold minimal transaction complexity: use markets not firms. Above this threshold increasing complexity increases TCs and firm size. But a viability constraint limits max e. Features are consistent with TCE (n.b. track traditional theory as a special case). Small absolute ε (technical and organisational factors are almost independent). RHS diagram: threshold also exists but at higher f’(e) (because of α 0 ). Non-monotonic relationship: both arms are profit maximising. No unique relationship between transaction complexity and organisational response: departure from traditional TCE (n.b. large absolute ε). Small firm equilibrium has a maximum f’(e). Large firm solution: upper limit on e determined by profit constraint. Small firms are viable because transaction complexity is absorbed by the effort effect on production costs, as long as there is not too much complexity to absorb.
Two conjectures – conclusions From the oligopoly model: In later stages of industry life cycles there is a strategic rationale for the existence of firms. Positive organisational effort requires a degree of monopoly power (product differentiation). This is similar to competence theory that emphasises the importance of unique firm characteristics – but the analytical logic is different. From the monopoly model: Early in industry life cycles markets are atomistic. Consistent with a universal small firm equilibrium. Firms gain rationale from production costs responding to effort - minimises any requirement for process innovation. Some firms “jump” equilibrium during the life-cycle shakeout. Allows growth in organisational size and the development of oligopoly. Coexistence of small and large firms is possible with large (absolute) ε. If industry evolution reduces absolute ε this will eliminate small firms.