Cycle et croissance économique / Business cycles and economic growth

Séminaire organisé par Pascal Bridel (Centre pluridisciplinaire Walras Pareto, Université de Lausanne, Suisse), Muriel Dal Pont Legrand (Université de Nice -Sophia Antipolis et GREDEG-CNRS) du 16 au 21 juin 2014

Participants

Amanar Akhabbar, François Allisson, Richard Arena, Tiziana Assenza, Michael Assous, Roger Backhouse, Pascal Bridel, Olivier Bruno, Katia Caldari, Matthieu Charpe, Muriel Dal Pont Legrand, Domenico Delli Gatti, Rodolphe Dos Santos Ferreira, Harald Hagemann, Maria Cristina Marcuzzo, Andreas Pyka, Alain Raybaut, Hans-Michael Trautwein, Amos Witztum.

Résumé

Avant d’être considérés comme des champs de recherche indépendants, l’analyse des cycles et la théorie de la croissance de la productivité ont longtemps été considérées comme des dynamiques étroitement reliées. Plus précisément, ce n’est qu’au lendemain de la révolution keynésienne que le développement de la macroéconomie s’est véritablement structuré autour de deux champs indépendants avec, d’une part, les théories des cycles qui s’intéressent aux mouvements stochastiques et, d’autre part, la théorie de la croissance qui analyse les conditions d’existence, d’unicité et le comportement d’un équilibre stable de longue période. L’objectif principal de ce séminaire était de réunir des historiens de la théorie économique et des praticiens de la théorie moderne des cycles et de la croissance pour un échange sur l’origine et l’évolution de la relation entre les cycles des affaires et la croissance économique.

Au lendemain de la crise de 2008 et dans le cadre d’une croissance économique anémique, ce séminaire a permis un examen circonstancié de l’évolution de cette importante littérature et plus précisément des différentes tentatives menées par les économistes afin de réconcilier théories des cycles d’équilibre (TCE) et théorie de la croissance. Après avoir identifié une série de tentatives destinées à réconcilier la dynamique des cycles avec celle de la croissance, ce séminaire a permis de clarifier les ruptures et les continuités entre des préoccupations théoriques anciennes et la modélisation récente de cette problématique qui n’est toujours pas appréhendée d’une manière satisfaisante.

Compte-rendu

Dans un premier groupe de communications, plusieurs historiens de la pensée économique ont offert une analyse d’approches anciennes pour en tirer une inspiration et des parallèles avec la problématique contemporaine cycles-croissance. En discutant la contribution de D.H. Robertson (1915), Pascal Bridel montre la parenté qui semble exister avec les chocs exogènes de la théorie contemporaine des cycles réels : les agents réagissent d’une manière rationnelle aux chocs technologiques qui créent à la fois cycles et croissance. Amanar Akkhabar examine le modèle dynamique de Léontief pour illustrer la méthode originale que cet auteur applique à son explication des cycles, de la croissance et des fluctuations. Dans le cadre d’un examen des économises russes du tournant du 20ème siècle (Tugan-Baranovsky, Kondratriev et Boukharine notamment), François Allisson se demande quel rôle jouent les schémas marxistes de reproduction dans leur analyse des cycles et de la croissance. Richard Arena se concentre sur la théorie des changements structuraux dans le court terme et le long terme ; il propose notamment une classification en trois classes des modèles en fonction de leurs capacités à relier cycles et changements structuraux (Pasinetti, Solow et croissance endogène). Ces différences le mènent à distinguer trois types différents de changements structuraux. Michaël Assous propose une nouvelle formulation de la théorie des cycles de Kalecki. Il argumente en particulier que la démonstration de « self-sustaining cycles » dépend presque exclusivement de la valeur que Kalecki donne aux paramètres de son modèle (anticipations et progrès technique notamment). Il examine également les conséquences d’une modification de la distribution du revenu sur la dynamique profit-investissement et donc de la croissance. Roger Backhouse offre une lecture extrêmement intéressante de la disparition apparente de la théorie des cycles dans le programme de recherche des économistes américains dans les deux décennies suivant la publication de la Théorie générale. Utilisant les cas de Hansen et Samuelson, l’auteur démontre cependant que cette transition ne s’est pas faite au détriment total des théories des cycles pré keynésiennes qui conservent un rôle dans le cadre de la macroéconomie naissante des années 1940. Katia Caldari propose une analyse très détaillée de l’extrême difficulté qu’ont les successeurs d’Alfred Marshall à réconcilier son analyse du long terme avec le court terme (et donc des cycles avec la question de la croissance). L’auteur considère que cette difficulté provient d’une incompréhension méthodologique à saisir le rôle joué par le temps dans la pensée marshallienne. Muriel Dalpont (et autres) réexaminent l’ouvrage de Harrod (1936) largement éclipsé par la révolution keynésienne. Tout en se concentrant sur la contribution essentielle de Harrod (le lien entre concurrence imparfaite et cycles), les auteurs la mettent en rapport avec la dynamique du taux d’épargne dans le cadre du modèle dans lequel le taux de croissance effectif s’ajuste au taux de croissance d’équilibre. Dans le même ordre d’idée, Rodolphe Dos Santos Ferreira réexamine l’Essay in Dynamics Theory de Harrod. En particulier, il développe l’intuition principale de l’auteur en modélisant l’interaction entre l’instabilité du sentier de croissance provoquée par les cycles. Il relie également cette tentative précoce avec la démarche récente de Woodford (1992). Harald Hageman (et autres) se sont, quant à eux, intéressés aux parallèles entre les cycles de croissance de Schumpeter et à leur réapparition contemporaine chez Aghion et Saint-Paul. Les auteurs se concentrent sur le concept de récessions productives, une analyse proposée par Aghion et Saint Paul, qui se fonde sur un modèle de coût d’opportunité pour examiner les efforts de réallocation d’actifs effectués par les entrepreneurs à la suite d’une récession, et qui se veut comme étant l’expression moderne de la pensée Schumpetérienne. C’est ce point qui est examiné dans le papier. Christina Marcuzzo offre un survol analytique du débat entre les économistes de Cambridge à propos des équilibres de courte et de longue période. La dispute entre néo-keynésiens et néo-ricardiens est naturellement au centre de cette discussion théorique qui n’a jamais vraiment trouvé de solution parmi la tradition cambridgienne. Hans-Michael Trautwein s’intéresse quant à lui aux contributions de Neisser et Haberler à la transmission internationale des cycles. Après un examen détaillé, ces deux approches sont comparées avec la macroéconomie moderne des économies ouvertes.

De leur côté, les théoriciens et les praticiens de la théorie moderne des cycles ont offert une série de survols liés à l’origine de ces différents modèles. Dans le prolongement de la crise de 2008, Bruno Olivier s’intéresse aux rapports entre les cycles de croissance et la distribution du revenu. L’auteur modélise la manière dont des variations de la distribution du revenu peuvent influencer le rythme de croissance tout en menant, paradoxalement, à une situation de crise économique. Mathieu Charpe réexamine la question des fluctuations et de dependency ; l’auteur montre que le taux d’utilisation des ressources anticipé dans le long terme ajuste l’output gap d’aujourd’hui. En utilisant la technique récente des agent-based models, Domenico Delli Gatti propose de remédier à l’une de leurs faiblesses en offrant une stratégie de formalisation qui réduit la ‘dimensionnalité’ de ces modèles en remplaçant la distribution de l’environnement financier des entreprises par les premier et second moments de la distribution elle-même. Ces moments peuvent alors être utilisés comme des variables macroéconomiques traditionnelles. Parallèlement, Tiziana Assenza propose un modèle qui illustre comment les ‘esprits animaux’ et des anticipations hétérogènes amplifient les cycles et comment une coordination endogène d’anticipations pessimistes renforce la crise et freine la reprise. Utilisant l’approche des chocs structurels, Andrea Pyka montre comment, dans le long terme, l’innovation et les changements structuraux ont influé la croissance économique depuis la révolution industrielle. L’auteur argumente en faveur d’un cercle vertueux entre capital humain, accroissement du revenu et hausse de la demande. De son côté, Alain Raybaut offre un survol détaillé des discussions des trois dernières décennies sur la stabilité, les cycles dans une dynamique complexe dans des modèles de croissance avec monnaie. Finalement, dans un papier très novateur, et en faisant de nombreux aller-retour entre théorie classique et théorie moderne, Amos Witztum examine les rapports entre inégalité et croissance. Sa contribution tente une synthèse entre la question des incitatifs, celle des abilities (capital humain) et l’idée trop souvent négligée de l’utilité sociale de la croissance. En d’autres termes, l’auteur s’interroge avec pertinence sur l’acceptabilité sociale d’une croissance qui serait créatrice d’une plus grande inégalité distributive.

Communications

Akhabbar Amanar : Wassily Leontief’s “Die Wirtschaft als Kreislauf” (1928) as a Contribution to Economic Growth, Cycles, and Fluctuations Analysis
In 1928, while in Germany, Wassily Leontief (1905-1999) published in Archiv für Sozialwissentschaft und Sozialpolitik a paper titled “Die Wirtschaft als Kreislauf.” In 1991 it was partly translated in English and published in Structural Change and Economic Dynamics as “The Economy as a Circular Flow.” Theoretical in nature, this paper was the Ph.D dissertation Leontief “submitted in the Fall of 1927 to the Dean of the University of Berlin” (Leontief 1991) under the official supervision of Ladislaus von Bortkiewicz. One finds only few comments about this early work of Leontief. According to Paul A. Samuelson –Leontief’s former student at Harvard–, this article is to be considered as the “overture to [Leontief’s] Ring of Input-Output.” (1991) Samuelson added that this actually was one of the major pieces of the independent and simultaneous “discovering [of] […] the theory of input-output” by Leontief, Sraffa and von Neumann. Samuelson didn’t define what was the “theory of input-output” that would put together and in a coherent way Leontief’s, Sraffa’s and von Neumann’s contributions. Beyond the fundamental (and intriguing) common reference of the three authors to the principle of the “circular flow” (excluding stocks and so-called primary or original factors of production from economic analysis), important differences exist. Especially, gaps appear when considering that Leontief’s and von Neumann’s works deal with temporal flows into a dynamic intersectoral framework and examine changing physical quantities of commodities, while Sraffa stressed that his study excluded such
quantitative changes and was even beyond static or dynamic analysis since quantities were supposed given. Our article focus on Leontief’s “The Economy as a Circular Flow”. We examine the dynamic framework Leontief developed in his work and we show how it was used to explain economic phenomena like growth, cycles and fluctuations. We aim at clarifying Leontief’s main assumptions and the nature of his explanation of economic change. We finally state the specific connection between cycles and growth in this framework.

Allisson François : From Tugan-Baranovsky to Feldman: Were the Marxian Reproduction Schemes
Well Suited for the Interrelated Analysis of Business Cycles and Economic Growth? The simple and expanded reproduction schemes have been sketched by Marx as a representation of the circulation of capital in the economy. Soon, these schemes became powerful tools to serve for alternative uses.
Notably, they were popularised in the debates on the transformation of labour values to prices of production. Reproduction schemes were also used for the analysis of business cycles, and for the analysis of economic growth. This paper asks whether the Marxian reproduction schemes are well suited for the interrelated analysis of business cycles and economic growth. Our answer will be given for the particular context of Soviet planning in the 1920s, by following the alterations underwent by the reproduction schemes in the hands of a few Russian economists, from Tugan-Baranovsky to Feldman.
In the first part, it is recalled that Tugan-Baranovsky was the first economist to make an analytic use of the Marxian reproduction schemes. He used them in many theoretical debates, on the transformation problem, the dynamics of the rate of profit, and the theory of crises and cycles. His theory of business cycles, the first ever to use reproduction schemes, is analysed. In the second part, the issue of business cycles and reproduction schemes is carried further in Soviet Union. The 1920s were a period with a strong emphasis on economic dynamics (Kondratiev, Pervushin, Yurovsky). At the same time, the notion of equilibrium was actively debated (Rainov, Bogdanov, Kondratiev). Out of these debates, it is possible to sort out the suitability of simple, and expanded reproduction schemes in the analysis of business cycles. In a third and final part, the issue of economic growth and reproduction schemes is analysed. It is shown, through three steps, how expanded reproduction can account for growth: Tugan-Baranovsky’s model of unlimited growth, Bukharin positive/negative growth model, and Feldman growth model. With these theories and models of economic equilibrium, business cycles and growth in mind, the conclusion provides an answer to our question.

Arena Richard : Structural economic change in the short and the long run
In the two last decades, a revival of the notion of economic structural change happened, which substantially changed the contents of the theory of economic growth (see Acemoglu, Introduction to modern economic growth, Princeton University Press, 2009 and Arena and Porta (eds.), Structural dynamics and economic growth, Cambridge University Press, 2012). Starting from the respective analytical traditions of “scholars associated with the World Bank, including Baumol and al., 1989, Chenery and Syrquin, 1975, Kuznets, 1971 and Rostow, 1971’ (Echeverria, 1997: 431)” and of Luigi Pasinetti’s books (Pasinetti, 1981 and 1993) , to-day three prevailing theoretical approaches emerged.
The first corresponds to the present evolution of Pasinetti’s theory of structural change and to the contributions which tried to develop his analytical views (see for instance some of the papers included in Arena and Porta, op. cit.). The second rooted in Schumpeter’s theory of economic development and therefore called ‘evolutionary’ tries to relate endogeneous business cycles and structural changes of the economic system (see for instance Lordon, 1993; Andersen, 1997 and Lorentz/Savona, 2010). The third corresponds to the development of the post-Solowian literature on economic growth related to the contributions of the theory of structural change developed for instance by contributors as Echevarria,1997; Laitner, 2000; Kongsamut and alii, 2001; Ngaï/Pissarides, 2007 or Acemoglu, op. cit.
Now the main purpose of our contribution is to study the variety of short- and long-run changes in the theory of structural change. Pasinetti’s approach does not deny the influence of short-run changes or even business cycles on structural change and economic growth (Pasinetti, 1981 ). However, the introduction of the so-called ‘separation theorem’ seems to make this influence more uneasy because it creates a natural condition and transforms Pasinetti’s model into a normative one (Pasinetti, 1993); it is then more difficult to articulate this model with economic fluctuations. In the evolutionary approach, both the short-run fluctuations and the verylong run transformations induced by technological change play a crucial role in analysing long-run growth patterns. However, this possibility is clearly related to the existence of an endogeneous growth model with evolutionary micro-founded structural change. The issue is then to understand if structural change is an assumption or a result of the model. Finally, in the post-Solowian literature, business cycles are clearly distinct from structural change. Short-run changes are indeed considered as transitional dynamics which explain how a ‘generalized balanced growth path’ or a ‘partially balanced growth path’ can emerge from any given type of structural change. These differences point out three types of structural change.

Tiziana Assenza (with William A. Brock and Cars H. Hommes) : Animal Spirits, Heterogeneous Expectations and the Emergence of Booms and Busts
We introduce a simple equilibrium model of a market for loans, where households lend to firms based on heterogeneous expectations about their loan default probability. Agents select among heterogeneous expectation rules, based upon their relative performance. A small fraction of pessimistic traders already has a large aggregate effect, leading to a crisis characterized by high contract rates for loans and low output. Our stylized model illustrates how animal spirits and heterogeneous expectations amplify boom and bust cycles and how endogenous coordination on pessimistic expectations amplifies crises and slows down recovery. Taking heterogeneous expectations and bounded rationality into account is crucial for the timing of monetary or fiscal policy.

Assous Michael (and Amitava Dutt) : Kalecki’s theories of the business cycle: a simplified treatment
It is now almost eighty years since Kalecki (1935) published his formal mathematical model of business cycle. This model represents a seminal work that has proven important in the development of macro-dynamics. It is an appealing one because it was able to explain for the first time, in a brief and intelligible way, the movement of a constrained demand economy in a semi-endogenous manner. By presenting a closed-form analytic solution, Kalecki was able to clarify the role for specification of parameters and the use of time lags and initial shock. His theory of cycles went through a number of changes over the years, and it is well known that these changes are related to changes in his theory of investment. Steindl (1981) distinguishes between three such versions: the first developed in the 1930s (Kalecki, 1935, 1937, 1939), the second from the 1940s and 1950s (Kalecki, 1943, 1954), and the third developed in the 1960s (Kalecki, 1968).
Despite the importance attached by Kalecki to his work on the business cycle, compared to his other work on pricing, distribution and effective demand, it has had little effect on the subsequent literature. There are several reasons for this neglect. First, the austerity of Kalecki’s mathematical style might have rebutted his contemporaries. Second, although Kalecki was familiar, from the early thirties, with important authors of both the Marshallian and Wicksellian traditions, as well as specialists in business cycles such Albert Aftalion and Joseph Schumpeter, he barely related his work to the existing literature and most of the time failed to highlight the novelty of his argument. Third, the neglect of his theories is probably connected with the mathematical structure of his analysis, employing as it does mixed differential and difference equations, or high order linear difference equations, and because of the kinds of results they generate.
For these reasons, most of the work done on Kalecki’s cycle theory has been on its mathematical aspects. Ragnar Frisch and Harald Holme (1935) first offered a detailed analysis of Kalecki’s solutions by exploring the characteristics of mixed difference-differential equation. Later, Kaldor (1940), but in a literary and diagrammatic way discussed the importance of non linearities in Kalecki’s models for explaining cycles endogenously. More recently, Stanislaw Gomulka, Adam Ostaszewski, and Roy Davies (1990) reconsidered Kalecki’s post-1943 versions of his theory of the cycle and trend of a capitalist economy (see also Serena Sordi, 1989).
There are two major aspects of Kalecki’s theory of investment which underwent changes over the years but which are present in all his models. The first concerns time lags between different concepts of investment. In the first version Kalecki distinguishes between three such concepts. The production of investment goods is a moving average of investment decisions (and orders) over the period of gestation of investment goods, which is exogenously given. The deliveries of finished investment goods, or plant and equipment, follow orders with the same time lag. The two subsequent versions do not distinguish between the production and delivery of investment goods, and assume that both follow investment decisions with an exogenously given time lag. Second relates to the determinants of investment plans.
The purpose of this paper is to provide a simplified discussion of Kalecki’s business cycle theories, and to suggest some avenues of extending it in fruitful ways. More precisely, the present paper re-examines the Kalecki’s model with the following specific objectives
1) To set up, by adopting a simple continuous-time formulation of his lag structure, a simple dynamic formulation of the model that allow to tract the main differences between the three versions.
2) To prove, under some additional conditions, a theorem establishing the necessary and sufficient conditions for the generation of self-sustaining cycles for each versions. It will be shown that the persistence of cycles depends crucially on parameters values linked to expectation in version 1 and to technical progress and income distribution in version 3.
3) To discuss the implications of changes in income distribution for the profit-investment dynamics.

Backhouse Roger : Alvin Hansen, Paul Samuelson and the transformation of American business cycle theory
The way economists viewed business cycle theory changed profoundly during the 1940s. For most of the 1930s and 1940s, business cycle theory was the framework within which economists tackled the problem of unemployment; in contrast, by the 1950s, business cycle theory had become a specialized topic within macroeconomics, its place as an integrating framework replaced by the theory of income determination, whether that was conceived on Keynesian lines or as part of a short-run general equilibrium model. This was, of course, closely connected with the Keynesian revolution, but the story of how it happens remains to be told, for there has been a tendency to see a sharp break as having happened with the publication of Keynes’s General Theory in 1936, and not to consider the 1940s as part of the process of transition.
This paper will examine this transition through considering the work of two of the main actors in this process — Alvin Hansen and Paul Samuelson. Hansen was one of the prominent American business cycle theorists of the 1920s and 1930s, but by 1953 he was arguably America’s leading Keynesian, publishing a widely read A Guide to Keynes, helping to establish the IS-LM model–static, short run, and devoid of cyclical considerations — as the main analytical device in macroeconomics. Working very closely with Hansen, Samuelson made the same transition: someone who started as a business cycle theorist wrote a textbook, Economics (1948), one of the major works in economics after the Second World War, in which “The business cycle” was simply the sixth out of seven chapters dealing with what later came to be known as macroeconomics.
The aim of the paper will be to illuminate this transition, placing it in the context of policy problems arising during the 1940s, reinforcing the point previously made by Perry Mehrling, that ideas from “pre-Keynesian” American business cycle theory had more effect on thinking about the cycle in the 1940s than is sometimes believed. The Keynesian revolution was more complex than it is usually pictured as being.

Bridel Pascal : Robertson’s Industrial Fluctuations (1915): An early real business cycle approach?
Modern real business cycle theory is a class of macroeconomic models in which trade cycle fluctuations to a large extent can be accounted for by real (in contrast to nominal) shocks. In opposition to other business cycle theories, RBC theories consider recessions and economic growth as the ‘efficient’ response to exogenous shocks in the real environment. Such theories differs markedly from other trade cycle theories such as Keynesian and monetarist economics that consider recession as the failure of some markets to clear and in which monetary variables have a central part of play.
As is well known, this long and rich tradition of market failure cum monetary factors approach finds its roots in Cambridge early macroeconomics in which Marshall, Pigou, Keynes and Robertson play a central role. However, and in clear opposition to the traditional Marshallian trade cycle theory, Robertson’s 1915 Study of Industrial Fluctuations suggests a non-monetary overinvestment theory of the trade cycle. Crises and cycles are seen to be caused by structural maladjustments resulting from overinvestment; the factors generating these fluctuations are non-monetary in nature and can be associated with inherent characteristics of the capitalist mode of production. In his 1915 book, Robertson always tried “consistently and thoroughly to dig down behind money appearances to real facts” (Robertson to Pigou, 1913). In short, real forces only can set the cycle on its way. Productivity shocks on capital goods provide “a rational inducement to the producers … to restrict their production” (1915, p. xiii).
This paper is an attempt to present, discuss and explain this a-typical approach to trade cycle (soon abandoned by Robertson after the first world war). Explicitly linked with Continental (as opposed to Cambridge) economists (like Aftalion, Spiethoff, Schumpeter, and above all, Tugan-Baranowski), the entire logical structure of Robertson’s cycle theory is shown to be devised to demonstrate the recurrent succession of booms and recessions in terms of the rise and fall of the productivity of investment goods. The link with an acceleration principle (which he opposed later vehemently to Keynes multiplier) is the other element to take pride of place in the cycle. The connection with a particular ‘real’ theory of interest brings eventually Robertson back to a Marshallian-type of modelling and not, like modern RBC theorists, to a general equilibrium approach. The real forces primarily represented by the gestation period of investment, but also by its durability, its imperfect divisibility and, allied with this, its intractability. These features of investment lead to excessive outlays upon capital investment that ultimately depresses their marginal productivity. The inevitable and rational result is a downturn in the capital goods industries and the onset of a cycle.

Bruno Olivier : Income Distribution and Growth Cycle
The financial crisis of 2007 shed a new light on the relationship between income distribution and economic growth. It is more and more acknowledge that the structure of income distribution matters in explaining economic growth in the long run (Berg, Ostry, and Zettelmeyer, 2012; Berg and Ostry, 2011). Moreover, some papers show that disturbances in income distribution giving rise to income inequality are not only detrimental for those at the bottom: it is also bad for economic growth as a whole (B.Z. Cynamon and S.M. Fazzari (2014) for instance).[1]
In this paper, our aim is to contribute to the theoretical literature showing how income distribution could impact long run economic growth and lead to economic crisis. In a growth model with Keynesian features in the spirit of Kurz (1990), Lavoie (1992) or Bruno (1999), we show that multiple long-run growth equilibria may emerge according to the structure of income distribution. Two main results are obtained from our analysis. First, assuming exogenous income distribution, we show that three long-run equilibria situations are possible. When the real wage is low, there is only one “low level” equilibrium in the long-run (low equilibrium). In that case, workers’ demand is weak and entrepreneurs are incited to produce a low level of output. On the contrary, high real wage is a signal for high activity and high level equilibrium is reached in the long-run (high equilibrium).
Multiplicity of equilibria emerges for an intermediate value of the real wage. In that case, low and high equilibria coexist, and the coordination on one particular equilibrium depends on entrepreneurs’ expectations (“animal spirit”). The shift from one kind of equilibrium to another depends on two critical values of the real wage. When the real wage reaches one of these two values, the economy can switch, for instance, from high equilibrium to low equilibrium. In that case, the switch can be interpreted as an economic crisis. Moreover, due to the hysteresis characteristic of the model’s dynamics, this change in the long-run growth is permanent, even if the change in the real wage is temporary. Consequently, we show that temporary exogenous small change in income distribution (higher or lower wages for instance) may affect permanently the long-run growth equilibrium of an economy and lead to hysteresis dynamics. Second, we render endogenous the dynamics of income distribution. We assume that the level of demand affects the determining of the rate of margin of entrepreneurs and the level of price. Moreover, we assume that the dynamics of quantity and the dynamic of income distribution do not follow the same time frame (fast-low dynamic process, Zhang (1993)). Whereas the dynamics of quantity is a fast one (current adjustment to equilibrium), the dynamics of income distribution is a low one (structural modification of income distribution and of long-run equilibrium). As a result of the combination of these two dynamics, economic crisis and recovery (growth cycle) are endogenous.
[1] See also Kumhof and Rancière (2010) for the link between inequality, leverage and crisis.

Caldari Katia : Dealing with time: short run and long-period analyses
Though classical economists recognized, at least implicitly, the distinction between short and long periods, it was Alfred Marshall to put it explicitly into economic analysis. The main problem for the Cambridge economist was how to deal with time: his solution, especially at the analytical level, was severely criticized and his main problem – the question of time – was either neglected or highly misunderstood by the following economists. The aim of this paper is to inquire into the way in which Marshall dealt with time and to underline how much of the criticism of his analysis is, in fact, grounded on the failure to properly and fully understand his thought and methodological approach.

Matthieu Charpe : Fluctuations and Growth in a Goodwinian Model
This paper presents a model in which a traditional Goodwinian prey-predator mechanism is combined with an endogenous growth mechanism. Goodwinian model such as Chiarella and Flaschel (2000) studies the cyclical properties of the model around a unique equilibrium. In this model the steady state is given by the rate of utilization of productive capacities in the long term. It follows that these models do not address the question of endogenous growth. Contrastingly, the stock-flow consistent approach in the post-keynesian tradition focuses their analysis entirely on long term growth disregarding the dynamic properties of their model. This paper attempts to fill-up this gap by analysing the properties of a Goodwinian model with a path dependency effect. In this model, the rate of utilization of productive capacities in the long term adjust to the output gap today.

Dal Pont Legrand Muriel (with Michael Assous and Olivier Bruno) : The Saving Rate Dynamics in Harrod’s 1936 Trade Cycle Essay
Harrod’s reflections on imperfect competition are at the heart of his inquiry on the theory of business fluctuations (Besomi 2003). Harrod formulated the principle of instability in a two step process of which his 1934 paper on “Doctrines of imperfect competition” form the first and his 1936 book The Trade Cycle the second.
Harrod initially emphasized the importance of imperfect competition for the trade cycle theory because of its compatibility with decreasing cost. “The key which the doctrines of imperfect competition provide for solving the mystery [concerning movements away from the general equilibrium of output] is that […] industries may be subject to the law of decreasing costs (in the long and short periods)” (Harrod, 1934: 465). The exploration of the implications of decreasing cost brought him to the conclusion that imperfect competition can provide an equilibrium approach to business cycles relying upon the existence of multiple long period equilibria. Instability may then result from alternative producers (correct) expectations leading successively the economy in position of high and low equilibia.
During 1935, Harrod gathered together the analytical components of his 1936 Trade Cycle book: the accelerator and the multiplier and devised a new mechanism capable of generating an equilibrium model of advance, yet unstable enough to give rise to cyclical growth. By resorting on imperfect competition, Harrod now explained turning points may result from changes in the value of the multiplier and the accelerator coefficients. It is in that context that the Law of Diminishing Elasticity of Demand – asserting that people become more affluent, they pay less attention to price changes – comes into play. By indeed establishing that changes in demand elasticity are likely to change income distributives shares, Harrod explicitly defined a mechanism allowing describing the adjustments of the actual and warranted rate of growth and eventually stabilizing the economy. By the way, he killed two birds with one stone. Beyond the stability issue, he succeeded – without resorting on the Law of Diminishing Returns – in explaining a stylized fact according to which “the commodity price fluctuation has
greater amplitude than that of (money) rewards to prime factors” (Harrod, 1936: 84).
The first readers of The Trade Cycle were generally aware the relevant coefficients determining the warranted rate of growth were not assumed to be constant and that their variation was actually central in Harrod’s growth cycle analysis. Oddly enough, most of them mentioned in passing the importance of Harrod’s Law of Diminishing Elasticity of Demand as regards its implication on the dynamics. Most of the time, comments were limited to the relevance of Harrod’s law for explaining the behaviour of price during the business cycle. Examining this episode of the history of macroeconomics requires proceeding in three steps. Section 2 restates Harrod’s 1936 basic model of imperfect competition and clarify the role of the Law of Diminishing Elasticity of Demand in Harrod’s analysis of income distributive shares. Section 3 considers the possible way by which changes in demand elasticity are likely to stabilize the economy. It is then argued that the reconsideration of that law has strong implications on Harrod’s explanation of business cycle turning points. Then, a brief survey of the reviews of Harrod’s trade cycle model published in the 1930s is provided in section 4.

Delli Gatti Domenico : “Growth and Fluctuations in a Hybrid Macroeconomic and Agent Based Model”
From the point of view of the average macroeconomist, agent based modelling has an obvious drawback: It makes impossible to think in aggregate terms. The modeler, in fact, can reconstruct aggregate variables only “from the bottom up” by summing the individual quantities. As a consequence the interpretation of the transmission mechanism of shocks is somehow arbitrary. In this paper we propose a modelling strategy which reduces the dimensionality of an agent based framework by replacing the actual distributional features (in our model: the distribution of firms’ financial conditions) with the first and second moments of the distribution itself. These moments are then dealt with as macroeconomic variables, along the usual lines. We put this strategy at work in a model of growth and fluctuations in which firms’ heterogeneous degree of financial robustness affect (optimal) investment in a bankruptcy risk context (à la Greenwald-Stiglitz).

Dos Santos Ferreira Rodolphe : Unstable growth and the trade cycle : Harrod’s “Essay in Dynamic Theory”
In his “attempt to give the outline of a ‘dynamic’ theory”, transposing to “the study of change” the “laws of supply and demand” which apply to the “study of rest,” Harrod arrives to the crucial conclusion that “in the dynamic field we have a condition opposite to that which holds in the static field. A departure from equilibrium, instead of being self-righting, will be self-aggravating”, so that the “moving equilibrium” is in fact “a highly unstable one” (Harrod, 1939, pp.14 and 22). And he comments: “Of interest this for trade-cycle analysis!” Shackle emphasizes “the paradoxical nature of this instability, and its sharp contrast with the stability normally ascribed to ‘demand and supply’ relations in a static model” (1967, p. 259). However, the instability is not only paradoxical, it seems in fact to be unwarranted: “the instability or otherwise of the system depends on the exact error-adjustment assumptions made. Some formalisations of the model support Harrod’s main conclusion, while others do not, and yet others conclude that it depends on the exact values taken by the parameters. […] It is also possible for the model to yield oscillations; the assumptions are indeed very close to those used by Hicks and others as the basis of cycle models” (Hahn and Matthews, 1965, pp.27-28). In spite of being almost a half century old, this assessment still holds: Harrod’s “instability principle can lead to various dynamic patterns such as a stable path, growth cycles or a corridor of stability,” according to the parameter values (Bruno and Dal-Pont Legrand, 2012; see also Yoshida, 1999, and Sportelli, 2000).
Why then Harrod’s strong conviction that moving from statics to dynamics is enough to entail the reversal of equilibrium local stability? Is it the consequence of having adopted a peculiar dynamic approach, namely “instantaneous analysis”, confined to an infinitesimal period t, as Besomi (1995) seems to suggest in his discussion of the Harrod-Keynes correspondence (1937-1938)? Consider however the way Besomi himself presents the rationale for Harrod’s instability: “Insufficient expenditure on investment would cause, through the multiplier process, unforeseen accumulation of stock that would be interpreted by entrepreneurs as a signal of excess investment” (Besomi, 1995, p.313). Hence, “they would further reduce investment, thus reproducing and enlarging the initial difficulties” (l.c.). In other words, disequilibrium, resulting here from an excess of saving over ex ante investment, would thus be interpreted as excess investment, rather than as excess production. In his Treatise on Money, Keynes assumes a Marshallian error correction mechanism working in the case of an increase in investment uncompensated by increased savings, where we “have, under the influence of the windfall profits accruing from the price rise consequent on the primary phase of the credit cycle, a secondary stimulus to an increased volume of production” (Keynes, 1930, I, p.258, my emphasis): a situation of aggregate excess demand is thus rightly corrected by an increase in supply. Harrod refers to the Treatise on Money, recalling that Keynes “said that if investment exceeded saving, the system would be stimulated to expand, and conversely” (Harrod, 1939, p.19). However, this “stimulus to expansion” is interpreted by Harrod as follows: “Firms finding themselves short of stock or equipment will increase their orders” (op. cit., p.21, my emphasis). Hence, a situation of aggregate excess demand is now assumed to be wrongly corrected by an increase in (investment) demand, a self-aggravating move.
Harrod’s instability thus seems to result neither from the adoption of a particular error correction mechanism nor from the choice of a dynamic approach, whether peculiar or not, but simply from the assumption that a discrepancy between ex ante investment and saving triggers a response in investment demand rather than in output supply. To illustrate, the recent formalisation of Harrod’s “knife-edge” by Yoshida (1999) is based on a “Harrodian investment function”, with the rate of growth responding positively to the excess of the required over the actual incremental capital-output ratio. The novelty of Harrod’s approach lies in his assuming that, in presence of, say, an excess of aggregate ex ante investment over aggregate saving and of the concomitant signals (upward price movements, depletion of stocks, overutilization of equipment), entrepreneurs respond as investors, by amplifying demand, rather than as producers, by curtailing supply.
Now, it is true that Keynes had already considered in the Treatise a destabilizing response of stock accumulation to an upward price movement expected to persist (a case of extrapolative expectations which was going to play an important role in the analysis developed in chapter 19 of the General Theory): “in so far as a further rise of prices is expected, there may set in a tendency to hoard liquid goods, which aggravates the excess of investment over saving, and so precipitates the very rise in question” (Keynes, 1930, p.259, my emphasis). Extrapolative demand expectations are at the very core of the acceleration principle, which is one of the main ingredients of Harrod’s Essay. They are not explicitly assumed but – and here we join Besomi – they are so to say implicit in Harrod’s “instantaneous analysis” according to which on a period of regular advance “it is proper to take a crosssectional view, assuming that the immediately preceding and succeeding periods yield similar developments” (Harrod, 1934, p.478).
Contrary to Harrod, Keynes confined the destabilizing effect of extrapolative expectations through investment decisions to a secondary position, even if he acknowledged it in the Treatise and promoted it to a more significant status in the General Theory. For Keynes, “the essence of the nature of [expectations] is their soapbubble fragility” (Shackle, 1968, p.xxi), responsible for “the artificial, ephemeral basis of the inducement to invest, and its consequent proneness to unpredictable and unpreventable collapse” (op. cit., pp.xxii-xxiii). Harrod’s and Keynes’s approaches to the trade cycle are illuminating sketches of the “two rather different kinds of models of endogenous fluctuations” contrasted by Woodford (1992): “models in which equilibrium is determinate but unstable on the one hand, and ‘sunspot’ models on the other.”

Gaffard Jean Luc : Novelty, Evolution and Growth
As a matter of fact, and as pointed out by Schumpeter, there is no growth without innovation. Novelty and hysteresis are the main engines of economic evolution. However, they are at the origin of co-ordination issues, insofar as consequences of any innovative choice can never be fully expected. Thus, there is no sense to analyse economic change as an intertemporal equilibrium with rational expectations. As a consequence, not only, growth and fluctuations cannot be dissociated, but there is no long-term trend that would be independent from what happens in the short term. Therefore, any growth analysis should focus on the time dimension of production and consumption decisions, and should underline why and how money cannot be neutral in the long term as well in the short term. J.R. Hicks and N. Georgescu-Roegen, among others, gave the key insights that should permit going further along this way. In any way, they can be considered as evolutionary economists. This paper is dedicated at presenting their contributions to the analysis of different aspects of a growth process conceived as an out-of-equilibrium process, with the perspective of providing new foundations to economic policy.

Hagemann Harald (and Dal Pont Legrand Muriel) : Schumpeter and Neo-Schumpeterian growth cycles analysis: (how) can recessions be “productive”?
Joseph A. Schumpeter is one of the grandmasters in economics who never ceased to inspire new generations of economists. One of his major contributions concerns his view of business cycles and economic development as closely interrelated dynamics, a position which led him to be highly critical of most of the econometric studies (Schumpeter 1931) and theoretical approaches which tried to identify and then distinguish these two dynamic issues. Among the different lines of research proposed by Schumpeter in order to allow economists to capture how growth and cycles dynamics intertwine (section 1), one can find the analysis of the entrepreneurs’ behavior during recessions. This line of research has been developed in the 1990s by Aghion and Saint Paul (1993, 1998) and Saint Paul (1994), the so-called “productive recessions” analysis. These authors build on an opportunity cost model in order to analyze the (re-)allocation issue entrepreneurs are facing during recessions (section 2). Our objective is to describe precisely the mechanism involved in the modern literature mentioned above as well as in Schumpeter’s original writings. We can then determine the necessary conditions for this mechanism to be at work and evaluate what was the importance Schumpeter attributed to such a mechanism (section 3). Finally, in the conclusion we analyze the continuity as well as the differences between Schumpeter’s contributions and the modern approaches questioning the ‘Schumpeterian character’ of some elements of the latter. Some focus will be on potential differences with regard to economic policy.

Marcuzzo Cristina : Short or long period equilibrium? Competing views in the Cambridge tradition of economics.
Marshall transformed the Ricardian conception of natural vs market prices into a distinction between short and long period equilibria, determined by the same forces, i.e. supply and demand. Kahn and Keynes objected to the idea that the short period was a temporary equilibrium, by showing how expectations and insufficient effective demand may render the level of income below full employment a permanent equilibrium. Sraffa reintroduced the fully adjusted ( long period) positions as those which solely could make possible the reproduction of the system. In the late 1970s Garegnani and Joan Robinson brought to the fore the Keynesian and Neo-Ricardian approaches as two competing views in the Cambridge tradition of economics.

Pyka Andrea (and Paolo Saviotti) : Competences, Income Development and Demand: A long run Virtuous Circle
In our contribution we explore how innovation and structural change affected economic development in the long run, by which we mean a period such as the one between the industrial revolution and the present. We separate the period since the industrial revolution into two sub periods, which we call ‘necessities’ and ‘imaginary worlds’ and focus on three trajectories, increasing productive efficiency, increasing output variety, and increasing output quality and differentiation. In the paper we show how a combination of the three trajectories gave rise to the transition between ‘necessities’ and ‘imaginary worlds’ and propose a mechanism of economic development which could have given rise to the type of economic system which we can observe today. To create growing output quality and differentiation higher competencies were required. These higher competencies required higher levels of education and demanded higher wages, which contributed to raise consumers’ purchasing power. These phenomena, combined with the income effect of the creation of new sectors, generated the disposable income with which consumers could purchase the new, higher quality, non-necessities, goods and services generated by innovation. In the paper we study the impact of several model parameters on the stability of the virtuous circle previously described.

Raybaut Alain : Stability, cycles and complex dynamics in growth theory with money
This contribution is dedicated to the interaction between short run and long run dynamics in the money growth literature. We chose to concentrate on deterministic dynamics. From this standpoint we discard the literature dedicated to the relation between fluctuations and growth in the RBC and stochastic frameworks with money. To begin with, we briefly recall how the issue was initiated in the neoclassical and then the Keynes-Wicksell growth models with money developed in the 1960s. Indeed, the latter approach constitutes the first attempt to fill the gap between short run dynamics and long run growth. However, stability and dynamic results are often ambiguous and this literature did not succeed to integrate properly cycles and long run growth. The second approach refers to the ”nonlinear” Keynes-Wicksell tradition which shows how endogenous growth cycles can arise through the interaction of monetary growth dynamics and policy rules in aggregated nonlinear models with budget constraints and local instabilities. The variations between the different models mainly concern the specifications of the nonlinearities (labor market, investment behaviors) and the treatment of monetary policies. Finally, we examine more closely several recent contributions extending the new synthesis framework to long run growth and the analysis of complex dynamics. Some contributions focus on the effects of different monetary rules on indeterminacy and learning mechanisms with multiple equilibria. Another interesting branch of the literature dedicated to the new open economic model with growth shows that this framework may allow a large variety of qualitative dynamic configurations under different policy rules.

Trautwein Hans-Michael : Some International Aspects of Business Cycles (and Structural Change): Neisser, Haberler and Modern Open Economy Macroeconomics
Despite the transnational character of the Great Depression, there are very few works in the inter-war literature that deal in depth with the propagation of business cycles across national borders, and even fewer that take the border-transcending structural effects of depressions into account. Two notable exceptions are Hans Neisser’s Some International Aspects of the Business Cycle (1936) and chapter 12 in Gottfried Haberler’s Prosperity and Depression (1937), which carries the heading “International Aspects of Business Cycles”.
Neisser’s book was published by the University of Pennsylvania Press in 1936, when Neisser worked at the Wharton School after his escape from Nazi Germany. It reflects Neisser’s background in the economic thinking of the Kiel School, blending classical perspectives on the long-term dynamics of capital accumulation, technical progress, output growth and (un)employment with his specific structural approach to the quantity theory of money and empirical research on international channels of the transmission of real and monetary impulses (followed up after the Second World War by Neisser and Franco Modigliani in National Incomes and International Trade, 1953).
Part II of Haberler’s Prosperity and Depression, the central part of the book, is designed as a synthetic exposition of the most relevant business cycles theories of the time (described in a survey in part I). Chapters 9 – 11 contain indeed an attempt at finding a consensus view in which the influences of different theories are clearly traceable by key concepts and references. The last chapter in part II, which deals with the international aspects of business cycles, is different. It appears to be a rather independent analysis of the role that transport costs, market-inherent and political restrictions of capital mobility as well as different exchange-rate regimes play in the cross-border spreading of cyclical movements.
This paper presents and analyses the different approaches by Neisser and Haberler in a two-stage comparison. At the first stage the two approaches are compared with each other. At the second stage they are compared with the actual state of open economy macroeconomics. As Haberler’s account accentuates the role of transport costs, capital flows and monetary policies, it is used as a catalogue of criteria for checking what modern attempts to link international trade and finance have got (back) in sight and what they have achieved, whereas Neisser’s approach serves to identify what the moderns have lost out of sight.

Witzum Amos : Distribution and Growth in Context: A Classical Perspective
Two main themes seem broadly to dominate the current debate about the relationship between inequality and growth. First is the question of incentives, or the relationship between inequality, re-distributive policies and entrepreneurial activities. Second is the question of ability, or the accessibility of better skills (human capital).
The implications of the two for the role of income distribution are quite distinct. However, there is a third dimension which seemed to have been completely omitted from the current debate about growth and distribution, it is the question of the social relevance of growth. Is growth a social objective? If so, why and how does it relate to the social state as is captured by the distribution of income?
On the incentive side the view seems to be that inequality is essential for growth as it allows entrepreneurs to keep the benefits from their innovative activities [1]. Inequality may be problematic to growth only if it raises the prospect of re-distributive policies which will deprive the entrepreneurs of their anticipated reward. Here, a complex literature has developed to explain the relationship between income distribution and voting mechanisms which may help in understanding the first element in the triangular relationship between income distribution, redistributive policies and entrepreneurial behaviour [2]. The second line of research—on the acquisition of skills, or ability– is much more straightforward and follows Meade’s work which suggests that better skills will be readily available for people who have wealth rather than the poor. This, together with imperfection in the capital market, may generate an equilibrium where some people in society will never be able to acquire better skills. From this point of view, inequality seems bad for growth.
However, at the heart of these arguments lies the presumption of the almost unique relevance of specific skills to the question of growth. Namely, what matters to growth is innovation which requires both skills and incentives and not necessarily the overall level of productivity. According to standard analysis, equality will dampen incentives to innovate though it could provide for greater provision of skills. Yet what is not clear is how much return one can expect from their skills if all people have similar skills. But there is an omission here of the contribution of the majority of workers in society who do not have specific skills. Surely an increase in their productivity is bound to have an impact on the ability of the economy to grow. How would inequality affect this and could the loss in productivity from people without specific skills not offset the gains from innovation? The theoretical underpinning of the incentive theory is closely associated with the property right approach to the theory of the firm. Here, in the presence of missing markets, the allocation of property rights which secures the return on effort explains both corporate structures as well as entrepreneurial activity. It is, of course, also a source of inequality. One of the most important determinants of the level of entrepreneurial innovation is, of course, the notion that agent’s productivity depends on the proximity of their operation to the asset they own. Naturally, to create an association between actions and assets the presumption is that we are talking about specific skills but what about all the agents who do not have specific skills? They too work with assets and their productivity may vary too. A disgruntled workforce may completely offset whatever gains achieved through the protection of the returns of those with specific skills. Theory tells us that the labour market will discipline the workforce but this is quite an odd position to take given the underlying inefficiencies of markets which gave rise to the need to secure the property rights of the innovatore. Surely the labour market is not immune to these kinds of inefficiencies.
Strangely, Classical economics seemed to be ahead of modern economics on all these accounts for a very simple reason: the firm and the behaviour of individuals were viewed as part of the question of social organisation and not only economic organisation. I will show in this paper that both Smith and Mill were acutely aware of the significance of the proximity of ownership to the productivity of the operator. However, they were also aware of the function of the larger social context within which economics operated. For Smith, this meant a complete diffusion of ownership. This, of course, will ensure inequality as people have different abilities. However, this is where the role of society becomes crucial. If society provides education then the gaps between individuals will diminish. In the meanwhile, we need growth to ensure that basic moral principles are not violated through the poor distribution of income which emanates from the facts that (a) ownership is not diffused and (b) people do not have similar educational background and skills.
Mill too recognises the importance of ownership to the productivity of agents but he also acknowledges the social and moral aspects which may affect productivity. For instance, while he did recognise the technological benefits of larger operations he also recognised that in an environment where the system repeatedly fails the expectation of the larger body of workers, the gains from large enterprises may be completely eroded. To a great extent, Mill’s take on Smith’s division of labour lies at the heart of the classical view on productivity. He draws our attention to two types of division of labour which are hidden in the Smithian idea. Firstly, there is the simple division which we all understand as the break up of the production process. But there is also the complex division of labour according to which we put our faith in fellow members of society to ensure our life’s necessities now that we have given up on procuring them ourselves. This latter part is of course dependent on social institutions and in particular, on the distribution of income which is generated by particular institutional structures. If the latter does not work, claims Mill, nor would the former. Given Mill’s general belief in the progress of humanity and in the possible improvement in the characteristics of the individual members of society, he thought that in the future, such large enterprises may work well if ownership was given to all those engaged in the production process. In the ideal world there will be no growth but stagnation and co-operation where ownership and income are such as to allow people to lead a life which is other than the endless pursuit of more material well-being. So here too, growth is a way of covering over the failing of the system. Promoting growth with more inequality is therefore, an augmented social failure.
[1] This line of research culminated in Acemoglou and Robinson (2012) where the thesis about the importance of non-extracting institutions for growth generation.
[2] See, for instance, Piketty (1995) and Benabou (2005).