John McCombie

The CES Production Function, the accounting identity, and Occam's razor

This paper reconsiders the argument that empirical estimations of aggregate production functions may be interpreted merely as statistical artefact. The reason is that Occams razor, or Herbert Simons principle of parsimony, suggests that the aggregate production function, together with the side equations derived from the usual neoclassical optimizing conditions, simply reflect the underlying accounting identity that value added definitionally equals the wage bill plus total profits. This argument is illustrated with respect to the empirical evidence presented by Arrow, Chenery, Minhas and Solow (Review of Economics and Statistics, XLIII, 225-50, 1961) and which led them to derive the Constant Elasticity of Substitution aggregate production function. It is shown that their results are more parsimoniously explained with reference to the underlying accounting identity than to any technological relationship.

Returns to Scale and Regional Growth: The Static-Dynamic Verdoorn Law Paradox Revisited

It has long been an article of faith amongst regional economists that increasing returns to scale are necessary to explain the punctiform location of economic activity and population. However, there is no consensus in the empirical literature over whether returns to scale are constant or increasing. A notable example of this lack of agreement is provided by the static-dynamic Verdoorn law paradox. While the dynamic Verdoorn law (specified using growth rates) yields estimates of substantial increasing returns to scale, the static Verdoorn law (specified using log-levels) indicates only the presence of constant returns to scale. In this paper, we explain the static-dynamic Verdoorn law paradox by showing that estimates of returns to scale obtained using the static law are subject to a spatial aggregation bias, which biases the estimates towards constant returns to scale. We illustrate our arguments by means of simulation exercises. The results obtained hold general lessons for applied economic analysis using spatial data.

The Tyranny of the Identity: Growth Accounting Revisited

Abstract It has been argued in the literature that growth accounting may be undertaken by directly differentiating the national income and product accounts identity where total income equals labour’s total compensation and total profits. This paper shows that this is simply an exercise in the manipulation of an accounting identity without necessarily having any theoretical foundation. Simulations show that the estimates of total factor productivity growth resulting from growth accounting performed with aggregate monetary data are not equivalent to the true rate of technological progress implied by the micro‐data. This suggests that results from the orthodox growth accounting approach may be very misleading.

NEW ESTIMATES OF RETURNS TO SCALE AND SPATIAL SPILLOVERS FOR EU REGIONAL MANUFACTURING, 1986-2002

Esta investigacion aborda el fenomeno de la desindustrializacion de la economia colombiana en el periodo 2005-2017 desde un marco de analisis kaldoriano; se centra en indagar el papel de la industria manufacturera como motor de crecimiento. El modelamiento parte de las Leyes de Kaldor usando dos paneles: uno de tipo balanceado con datos agregados para 23 areas metropolitanas y el otro desbalanceado agregando datos desde el nivel de firma con la muestra obtenida de la Encuesta Anual Manufacturera para el periodo. El analisis demuestra que el debilitamiento del sector manufacturero se expresa en el declive sostenido de su participacion en el empleo total y el PIB, y en el vinculo nocivo entre la dinamica del empleo manufacturero y el crecimiento del producto. Las evidencias de la desindustrializacion identificadas admiten concluir que el aporte de la manufactura al crecimiento economico se sustenta en rendimientos crecientes a escala estaticos.

Estimating technical change in aggregate production functions: a critique

This paper extends the approach of Simon and Shaikh and demonstrates that the estimates of aggregate production functions can provide no independent evidence of the underlying technology of the economy. It is further shown that estimates of labour and capital augmenting technical progress will merely be equal, by definition, to the growth rates of real wages and of the real rental price of capital respectively. This is illustrated by reference to recent studies that estimate technical progress for the Australian manufacturing industries.

Criticisms and Defences of the Balance-of-Payments Constrained Growth Model: Some Old, Some New

This paper assesses various critiques that have been levelled over the years against Thirlwall’s Law and the balance-of-payments constrained growth model. It starts by assessing the criticisms that the law is largely capturing an identity; that the law of one price renders the model incoherent; and that statistical testing using cross-country data rejects the hypothesis that the actual and the balance-of-payments equilibrium growth rates are the same. It goes on to consider the argument that calculations of the “constant-market-shares” income elasticities of demand for exports demonstrate that the UK (and by implication other advanced countries) could not have been balance-of-payments constrained in the early postwar period. Next Krugman’s interpretation of the law (or what he terms the “45-degree rule”), which is at variance with the usual demand-oriented explanation, is examined. The paper next assesses attempts to reconcile the demand and supply side of the model and examines whether or not the balance-of-payments constrained growth model is subject to the fallacy of composition. It concludes that none of these criticisms invalidate the model, which remains a powerful explanation of why growth rates differ. JEL Codes: E12, O41

Does the aggregate production function imply anything about the laws of production? A note on the Simon and Shaikh critiques

The aggregate production function (especially the Cobb–Douglas) is widely used in both applied and theoretical work, in spite of the large number of criticisms of it that have been advanced since its inception. Two related criticisms by Simon and Levy (1963) and Shaikh (1974) are particularly damaing since they have shown that if factor shares are stableanyunderlying technology will generate a Cobb–Douglas relationship. This argument seems to have been largely ignored in the literature or, by implication, not seen as particularly important. The purpose of this paper is to assess and extend these criticisms. New empirical evidence is presented which, it is argued, illustrates the former. Furthermore, it is shown that a Kaleckian mark-up model will equally give rise to a Cobb–Douglas even though, of course, no neoclassical assumptions are invoked. It is concluded that the Cobb–Douglasper se can give no independent corroboration of either the marginal productivity theory of distribution or the assumption of ...

The Aggregate Production Function: ‘Not Even Wrong’

Abstract The foundations of the aggregate production function were long ago thrown into doubt by problems of aggregation and the Cambridge capital theory controversies. Yet the aggregate production function, whether in the familiar form of the Cobb-Douglas, the CES, or the translog, continues to be widely used in both theoretical and applied analysis. The reason for its continued use rests on the instrumental position that ‘it works’. The aggregate production function sometimes yields good statistical fits with plausible estimates of the coefficients. However, for some time, it has been realised that the existence of an underlying accounting identity can explain the regression results, even if the aggregate production function does not exist. This argument has been widely ignored. This paper, drawing on a rhetorical approach, assesses why this is the case. It shows that the few criticisms that have been made of the critique involve fundamental misunderstandings that represent a failure of the economic method.

PRODUCTIVITY, EFFICIENCY AND TECHNOLOGICAL CHANGE IN EUROPEAN UNION REGIONAL MANUFACTURING: A DATA ENVELOPMENT ANALYSIS APPROACH*

Using data envelopment analysis, we calculate indices of total factor productivity (TFP), efficiency and technological change for the manufacturing sectors of 68 European NUTS1 regions over the period 1986-2002. We subsequently examine these indices using exploratory spatial data analysis (ESDA) techniques, before considering tendencies towards convergence in both TFP and technical efficiency levels. While the ESDA reveals significant spatial autocorrelation, the convergence analysis uncovers no tendency for regions with initially lower TFP to catch up with regions with initially higher TFP. However, convergence is found in levels of technical efficiency, although towards a relatively lower average level. Copyright Blackwell Publishing Ltd and The University of Manchester 2006.

‘Are There Laws of Production’: an assessment of the early criticisms of the

This paper traces the development of the Cobb-Douglas production function from its inception in 1927 and critically assesses its early hostile reception. Further econometric evidence is also presented on these issues. Some of the criticisms were easily dealt with, but other more serious ones remained and, although equally relevant today, have been all but forgotten. The original regressions of Cobb and Douglas using time-series data produced some spectacularly good fits, with the estimates of the output elasticities being virtually identical to the relevant factor shares. (This was erroneously argued by Douglas, and others following him, as providing strong empirical support for the neoclassical marginal productivity theory of distribution.) It is shown that these results collapse once account is taken of the existence of either outliers or technical change, or both. There is some evidence that Douglas himself realised this and his emphasis subsequently shifted to cross-industry regressions. However, an important critique by Phelps Brown in 1957, formalised later by Simon & Levy, demonstrated that all that was being estimated was an accounting identity. This criticism was later generalised by Shaikh to time-series estimations. These critiques have been largely brushed aside and ignored in the literature. If they had not been, there would perhaps be a greater appreciation of just how flimsy is the theoretical basis of the production function.