IN SEARCH OF A SUCCESSOR TO IS-LM

Transcription

1 IN SEARCH OF A SUCCESSOR TO IS-LM JEAN-PIERRE DANTHINE University of Lausanne and CEPR 1 After discussing the general characteristics that the successor to IS-LM should possess, this article argues that the business cycle research programme initiated by Kydland and Prescott (1992) is beginning to show a promising capacity to incorporate a broad range of modelling features into a logically consistent and theoretically satisfactory framework. This ability and the systematic process of model enrichment it permits make it possible to predict that the dynamic general equilibrium models developed around the neoclassical stochastic growth model but possibly evolving towards friction-prone non-walrasian models will become the platform for a new neoclassical synthesis. Ultimately, a stochastic growth model incorporating [real and nominal macroeconomic] rigidities holds out the promise of a new synthesis in macroeconomics. (Campbell, 1994) I. INTRODUCTION One goal of current macroeconomic theorizing is to propose a successor to the ubiquitous IS-LM model. For much of the post-war period, the IS-LM model by itself or accompanied by a Phillips curve relationship to form the neoclassical synthesis has been the workhorse of macroeconomics. It was the consensus model taught to undergraduates all over the world, the model used by practitioners to analyse the evolution of interest rates and national income, the methodological base for economic forecasting, and even the springboard for macroeconomic research. As eloquently put by Mankiw (1990), the consensus collapsed in the 1970s as a result of two flaws, one empirical and one theoretical. The empirical flaw was that the consensus view could not adequately cope with the rising rates of inflation and unemployment experienced during the 1970s. The theoretical flaw was that the consensus view left a chasm between microeconomic principles and macroeconomic practice that was too great to be intellectually satisfying. The breakdown of the consensus has left the economics profession in a state of schizophrenia where most practitioners continue to resort to the neoclassical synthesis for forecasting and policy analysis, while researchers have almost totally discarded it as a framework for conducting research. Caught in 1 I am grateful to John Muellbauer, John B. Donaldson, and an anonymous referee for usefully critical comments. I alone am responsible for the final product OXFORD UNIVERSITY PRESS AND THE OXFORD REVIEW OF ECONOMIC POLICY LIMITED 135

2 between, teachers often, but not always, present IS- LM to students who are starting economics, while concentrating on its failures at a later stage of the curriculum. Today, the profession is in search of a new consensus or, at least, of an encompassing analytical framework in which irreconcilable ideological stances can be presented. The new construct will have to share some of the IS-LM characteristics, i.e. be a relatively simple and plausible short-run macroeconomic model. It should be convincing enough to be used as a tool for intuitive analysis by economic practitioners, as well as a pedagogical device for introducing students to economics. And, of course, the new model should be in broad accord with the main lessons of the last 20 years of macroeconomic research. In this article, I first discuss the general characteristics that the successor to IS-LM should possess; I then review recent developments in business cycle theory, giving credence to the claim that the neoclassical stochastic growth model is likely to provide the foundations for this successor. As will become clear later, I do not mean by this statement irrevocably to characterize the successor to IS-LM as neoclassical rather than neo-keynesian. Rather, I argue that the research programme initiated by Kydland and Prescott (1982), taking the neoclassical growth model as its main foundation, is beginning to show a promising capacity to incorporate a broad range of modelling features into a logically consistent and theoretically satisfactory framework. Moreover, this promise is the outcome of a disciplined process of model enrichment proceeding less on the basis of ideological priors than on the confrontation of the model s prediction with observations of reality. I also argue that we are not there yet and that the final product will have to be better than current models, in the sense both of reproducing more convincingly the key stylized facts of macroeconomics, and of achieving this goal on the basis of more intuitively plausible economic mechanisms. In the end, the question of the existence of a true successor to IS-LM may well reside in economists ability to summarize in a simple enough model the various lessons of the empirical failures of those early efforts. II. THE IS-LM SUCCESSOR: A COMPOSITE SKETCH Macroeconomic theory aspires to provide an analytical framework which rationalizes the key observations that can be made about macroeconomic reality. The final intent is to propose appropriate policies and institutions in the light of a fundamental understanding of the underlying mechanisms. The ultimate objective of policy evaluation points towards the prototype macro model being a microbased general equilibrium model for at least two reasons. First, it is important that alternative policies can be ranked according to a social welfare criterion. Second, the model should be immune to the socalled Lucas critique. In his decisive attack on the old consensus, Lucas (1976) argued that most policy interventions that it is the business of macroeconomics to evaluate are likely to modify agents consumption and investment behaviour as well as other behavioural relationships, in particular because they affect agents expectations about the future. It is thus logically inconsistent to evaluate policies with models based on reduced form equations which will not be invariant to policy changes, and this may be the source of serious errors in policy analysis. This is a strong indication in favour of using micro-based models, where the primitives are preferences and technology parameters plausibly considered to be invariant to the policy alternatives to be studied. In addition, the last 25 years of macroeconomics have clearly stressed that macroeconomics is in its essence intertemporal and dynamic. Intertemporal decisions and trade-offs are the heart of macro. Fundamentally, consumption/ savings/ investment/ borrowing/ lending decisions cannot be discussed in a static context. And once the time dimension is introduced, at the minimum in a now and then framework, the problem of specifying expectations arises. On these scores, the deficiencies of the static IS- LM model are well-known and understood. The IS- LM model builds on reduced-form descriptions of consumption, investment, and money demand and, under fixed price and wage hypotheses, describes a one-shot equilibrium in financial and goods markets. The pure IS-LM model is static and takes the future into consideration only implicitly. As such it gives no 136

3 J.-P. Danthine room for expectations. The neoclassical synthesis, achieved by appending a Phillips curve to IS-LM, provides an intertemporal link. In the expectationsaugmented version, expectations come to play a decisive role. But agents behaviour remains summarized by the same reduced-form equations while the key dynamic equations describing the evolution over time of price and wages are not grounded in micro principles. On both counts, the model is thus subject to the Lucas critique. Moreover, since neither IS-LM nor the neoclassical synthesis are explicit about preferences and objective functions, welfare evaluation in a formal sense is precluded in either context. The earliest contender for the IS-LM succession, the fixed-price general equilibrium (GE) research programme, remedied several of the above deficiencies but not all of them. Fundamentally, it was a static model. While it was later completed with dynamic price and wage adjustment processes, most of those proposed suffered the same defects as the neoclassical synthesis. It thus appears that the new construct will be a significant departure from the old consensus. We want macroeconomics to be a branch of general equilibrium theory, providing not merely a theory of recession or of underemployment equilibria, but principally a theory of cyclical fluctuations, of booms followed by busts, a theory that is inherently dynamic. There is a distinct paucity of dynamic GE models to be used as appropriate platforms for macroeconomics. One can think of two alternative strategies to remedy this deficiency. Following good business tradition, let us label these alternative approaches, bottom up, and top down. One line of attack, the bottom-up approach, consists of starting with the pieces, that is the details of firms and consumers problems, and the nooks and crannies of individual markets, focusing naturally on the labour market, the goods market, and the credit and financial markets. The idea is to be as specific as possible in detailing the partial equilibrium of one or several of the component markets. One can then try progressively to assemble the pieces with the objective of achieving, at some later stage, a legitimate dynamic GE successor to the neoclassical synthesis. For some issues, the exchange economy GE level will be a useful intermediary step. The other approach, which I label top down, consists of starting with full-blown dynamic GE models (i.e. models with a production sector) capable of providing a complete theory of value(s). Initially, the models are taken straight out of microeconomics and boast few typical macro features (in particular, the first models are moneyless) but this is because there such a dearth of dynamic GE models with macro ambitions. Progressively, one can think of adding specific macro features hoping to connect, at a later stage, with the key observations made by those economists who have invested in the bottomup approach. There is no compelling argument why one line of attack should be viewed as dominating the other. In particular, there is no reason to start with one and wait till the corresponding research programme is complete before embarking on the other. The real business cycle (RBC) programme obviously fits into the top-down mode. A lot of what is included in the New Keynesian literature (see Mankiw and Romer, 1991, for a useful collection of the main contributions to that literature) fits into the bottom-up approach. As one would expect, the latter models are more precise and detailed, concentrating on the properties of some of the key components of the future synthesis, but they rarely go to the dynamic GE level which is a prerequisite for a full account of economic fluctuations. III. SUCCESSES AND SHORTFALLS OF THE RBC RESEARCH PROGRAMME Kydland and Prescott (1982) were the first to argue that an appropriately modified version of the neoclassical stochastic growth model had the ability to replicate surprisingly well some of the key stylized facts of macroeconomics (see Pagan s article in this issue for a dissenting view). Simultaneously, and more importantly, these authors have, in effect, proposed a methodology that goes much beyond the (unfortunate) RBC label. While their contribution is frequently identified with the view that business 137

4 cycles are exclusively driven by technology shocks and represent the economy s optimal response to such shocks, it can also be argued that their essential contribution was to operationalize the methodological approach to macroeconomic research hinted at by Lucas (1980). The last 15 years have demonstrated the fruitfulness of the research programme they have thus initiated. It is the contention of this article that the successor to IS-LM will be found in the lineage of the RBC model. Appreciating this claim requires standing back from a literal interpretation of the RBC research programme and recognizing that the approach initiated may well lead to selecting as the successor a fairly distant cousin of the representative agent, perfect market, economy. This contention and the most often cited reasons for scepticism are reviewed below. (i) Real Business Cycles: On the Sources of Business Fluctuations The maximization problem (P) below describes the neoclassical growth model enriched by a labour leisure decision: t max E β u( c, 1 n ) { xt+ 1, nt} t+ 0 t t c + x f ( k, n ) z t t + 1 t t t k = k ( 1 δ) + x t. k t+ 1 t t + 1 n 1, c 0, x 0, n 0, and t t t t 0 given (P) where c t denotes the period t consumption of the representative agent and n t his period t supply of labour (1 n t is thus his leisure in period t). Similarly, k t and x t+1 represent, respectively, the agent s period t capital stock and investment, while u(.,.) is his period utility function and f(.,.) the period production technology which is subject to the technology shock sequence {z t }. The parameter β is the subjective discount factor and δ the period depreciation rate. It is usual to assume that the technology shock follows a highly persistent first-order autoregressive process of the form: z t+1 = ρz t + ε t+1 with {ε t } identically and independently distributed (i.i.d.) lognormally with mean 1 ρ. As presented here, problem (P) describes the decision problem of a central planner in charge of allocating a composite good to be used for consumption or investment across the infinite future. This problem is of interest because it can be shown that the resource allocation of a decentralized economy appropriately formulated coincides with the solution of this central planning problem. This is, in effect, a restatement, extended to the case of an infinite horizon economy, of the first theorem of welfare economics: under perfect market assumptions, the equilibrium of the decentralized economy is a Pareto optimum. The latter is uniquely identified if there is only one representative agent or if all agents endowments and preferences are identical. This equivalence makes it easy to describe the evolution over time of the key macroeconomic aggregates consumption, investment, employment, output as they result from the reactions of individual decisionmakers confronted with and reacting to the external technology shocks. The latter are empirically identified as the Solow residuals, i.e. those changes in aggregate output that cannot be traced to changes in the two aggregate factors of production, capital and labour. Building on the neoclassical stochastic growth model, the RBC label owes its existence to the exclusive reliance of the prototype models on technology shocks as the external source of economic fluctuations. Their proponents often claimed that such a view received strong support from the ease with which a real shock model is able to reproduce some key macro stylized facts. As argued elsewhere (Danthine and Donaldson, 1993), therein does not lie the important contribution of this theory. Kydland and Prescott (1982) do not postulate or imply that there should be aggregate technology shocks only. Fundamentally, the approach they advocate implies that it is only the ex-post assessment of a model s ability to replicate the key observations that allows us to identify the main sources of economic fluctuations, as opposed to prior views on the origin of shocks. Solow residuals are a black box summarizing what we do not know in the relation between factors of production and aggregate output and/or what we do not consider as a priority to model given the state of knowledge and the question at hand. One element of modern business cycle theory is the acceptance that 138

5 J.-P. Danthine Table 1 Labour Market Facts, US Quarterly Data Relative Correlation standard deviation (%) with output Employment Average weekly hours Unemployment Duration Unemployment: flow in Unemployment: flow out Note: Correlation (flow in, flow out) = Source: Gomes et al. (1997); see this reference for sources and period length. there are forces driving the economy that are beyond economists duty or ability to explain. Political and climatic shocks are examples of the former, technological progress an example of the latter. Hall (1990) suggests that some of what we include under the label Solow residuals may be too important to be left unattended. Burnside et al. (1993) show that doing justice to variable factor utilization rates indeed leads researchers to hypothesize much less volatile Solow residuals than those simply estimated from the raw data. Evans (1992) and Cochrane (1994) argued persuasively that macroeconomic time series are unlikely to be accounted for on the basis on technology shocks only. Others have emphasized the necessity either in accounting for the volatility of consumption or in explaining observed correlations (between hours and productivity see below) to take exogenous fiscal and monetary shocks into account, an unsurprising development in the process of macroeconomic reconstruction. For most real-world economies, current account shocks are likely to have a role to play as well. To sum up, it is clear that the successor to IS-LM will not be a real business cycle model in the narrow sense of the label. It is more likely to be a dynamic GE model where monetary and fiscal shocks, at the minimum, will play a role along with technology shocks in proportions to be determined by the capacity of the various alternatives to help account for the key observations. (ii) Missing Issues The first RBC models were extremely limited in the list of issues they could address. The missing issues included: money and monetary policy, government spending and fiscal policy, unemployment and stabilization policies, current account, exchange rates, and foreign policy considerations. It is admittedly hard to see how such models could ever substitute for the neoclassical synthesis: most major issues at the heart of traditional macroeconomics could not even be discussed. None of these deficiencies remains today. Money has been integrated, with understandable difficulties, given the non-consensual state of monetary economics: monetary policy can be discussed and, even better, evaluated on a welfare basis (Carlstom and Fuerst, 1995; Cooley and Hansen, 1995; among others). Government spending and taxation have found their way into business cycle models, with McGrattan (1994) and others arguing that this allows a more convincing replication of some key stylized facts. Much attention has been given to the employment variability issue (see next section); more recently, unemployment per se is becoming a major focus of interest in RBC models (Merz, 1995; Andolfatto and Gomme, 1996; Gomes et al., 1997). Gomes et al. (1997), for example, propose for the USA a fairly demanding list of stylized facts (reproduced as Table 1) and then set out to build a dynamic GE model with the objective of explaining these observations. While not fully successful, their work makes clear what the difficulties involved are (keeping track of agents heterogeneity, a function of their past unemployment experience in a world with imperfect unemployment insurance) and helps us think about further modelling advances that might be supported by the data. While all the models cited above are cast in the matching search tradition, nothing prevents testing other candidates, such as 139

6 efficiency wages, for their capacity to explain the characteristics of aggregate unemployment. This section cannot close without mentioning the significant literature that has developed, following Backus et al. (1992, 1995), extending the RBC model to the international context. These extensions have served to underline the substantial puzzles that the international dimension presents to macroeconomists (see also Baxter, 1995; as well as Hess and Shin in this issue). One interesting lesson is that the proposed solutions to the international puzzles point very clearly to the introduction of specific frictions or market incompleteness into the basic model. Thus, Baxter and Crucini (1995) and Kollman (1992) investigate the implications of international asset market incompleteness, while Kehoe and Perri (1997) justify the constrained allocations (relative to a complete market setting) that emerge from the hypothesis that international loans are not perfectly enforceable. They conclude that such endogenous market incompleteness might prove more productive in solving some of the outstanding puzzles than exogenous restrictions on asset markets. Summing up, all the issues at the heart of short-run macroeconomics in the days of the neoclassical synthesis have now found their way into dynamic GE models, where they can be the subject of a considerably richer discussion than was feasible with the neoclassical synthesis. (iii) Anomalies and Puzzles While reproducing an important set of stylized facts, the bench-mark RBC model fails to explain other equally important macro observations. The first such puzzle to attract attention was the volatility of employment. While in fact (for the USA) employment is about as volatile as GDP, it is much less variable in the artificial economy of the bench-mark model. This failure, which was viewed as reassuring given the purely Walrasian modelling of the labour market in the prototype models, has led to considerable work and discussion centred on the labour leisure decision. Some of the proposed alterations, such as the indivisibility hypothesis of Hansen (1985) and Rogerson (1988), or the home production hypothesis (Benhabib et al., 1991; Greenwood et al., 1995) have remained, in spirit at least, in the Walrasian tradition; others have led to incorporating into the dynamic GE model important pieces of the New Keynesian programme. Thus dynamic GE models with efficiency wages and labour contracts have appeared, often with significant explanatory success (see Boldrin and Horvath, 1995; Danthine and Donaldson, 1995; or Dow, 1995). At the labour market level again, Christiano and Eichenbaum (1992) have argued that an equally significant failure of the bench-mark RBC model was its inability to reproduce the near-zero correlation between employment and real wages emphasized by Dunlop (1938) and Tarshis (1939). This observation remains a significant challenge for most pure RBC models, the solution proposed being the inclusion of another, demand side, source of external shocks. One definite strength of current micro-based models is the soundness of the key consumption savings decision at the heart of macroeconomics. While IS- LM was bound to be unsatisfactory in its attempted description of consumption and saving in a static framework, the dynamic feature of current business cycle models remedies this defect in a way which is simple enough while being fully consistent with the permanent-income hypothesis. While theoretically fully satisfactory, it appears, however, that consumption in artificial economies is often too smooth relative to the observed series. The so-called Deaton paradox (Deaton, 1992; see also Muellbauer, 1994, for a discussion) suggests that the problem may well be with the identification of the stochastic process for output. Another frequent interpretation is in terms of a measurement problem, since actual consumption data include durable consumption purchases and not only the flow of consumption services from durables. Another response is to generate increased consumption volatility by introducing a government sector with variable expenditures and tax rates (McGrattan, 1994). Consumption is more volatile in her model as a result of agents adjusting their consumption to quarterly changes in tax rates and government expenditure. Taking the opposite view that observed consumption volatility reflects the variability of agents permanent income, Danthine et al. (1997) investigate the possibility that it may signal some form of expectational instability about future rates of economic growth. 140

7 J.-P. Danthine In this section we have discussed, without attempting to be exhaustive, a set of puzzling observations, i.e. observations about reality which are in conflict with the predictions of the standard business cycle model. There is no dispute that these puzzling observations can be taken as falsification of the pure neoclassical growth model. The uncovering of these puzzles has, however, been used profitably to modify and enrich the model context, beyond what previously appeared to be formidable technical barriers. This systematic process of model enrichment following the confrontation with observations is precisely what makes it possible to foresee the dynamic GE models developed around the neoclassical stochastic growth model but possibly evolving towards friction-prone non-walrasian models growing into the successor to IS-LM. (iv) The Strength of the Propagation Mechanism The strength of the mechanism by which exogenous non-technology shocks are propagated and the persistence of the fluctuations induced by this mechanism is of importance if only because of widespread scepticism about technology shocks. If the mechanism is too weak, large exogenous shocks, both positive and negative, are necessary to explain observed fluctuations. If the propagation mechanism does not generate persistent variations in macro aggregates, the burden is increased in that the exogenous shocks themselves have to be strongly autocorrelated. Attempts at providing the standard model with a stronger engine follow several directions. The imperfect competition route has been well explored (Rotemberg and Woodford, 1995; Gali, 1995; Hornstein, 1993); aggregate increasing returns is another possibility advocated by Hall (1988) among others, while multiple equilibria, possibly in combination with one or the other above, is a third alternative (Benhabib and Farmer, 1994). Finally, the financial accelerator, by which is meant situations where financial frictions amplify the effects of exogenous shocks, has many proponents (see Gertler (1988) for a survey, and Tinguely (1997) for a dynamic GE model where the assumption of reorganization costs indeed results in a stronger propagation mechanism). In general, features able to generate endogenous cycles are by their very nature likely to reinforce the propagation mechanism of a model hit by exogenous shocks. Keynesian properties, such as fixed price or wage contracts, forcing some of the adjustments to be made in quantities rather than through prices, are also available to strengthen the propagation mechanism and thus diminish the size of the shocks needed to explain observed fluctuations. Finally, combining multiple sources of shocks is another natural way to relieve the pressure on technology shocks, although diversification effects do not guarantee that more volatility will necessarily follow. Critics often complain that the propagation mechanism of the neoclassical growth model generates very little persistence, forcing its proponents to rely on highly correlated technology shocks (Cogley and Nason, 1995). While a complete assessment of this criticism may have to await better evidence on the effective degree of persistence of the external shocks themselves, several modelling features have been identified that are likely to provide a more persistent propagation mechanism. Some, such as slowly diffusing information about external shocks and time to build, were part of the seminal article of Kydland and Prescott (1982). More recent developments include, for example, multi-period and overlapping contracts. Chari et al. (1996) argue that, in themselves, the latter are not sufficient to generate plausibly persistent responses to transitory monetary shocks. Summing up, while it is too early to tell which of the modelling features mentioned above will, in the end, be deemed important enough to remain associated with the prototype macro model, it is quite safe to predict that the latter will possess a substantially stronger and more persistent propagation mechanism than the neoclassical stochastic growth model. (v) On the Instability of Aggregate Demand and the Plausibility of Underlying Mechanisms At the heart of the opposition to the current RBC models, one finds priors on the issue of the inherent stability or instability of decentralized economies. The demonstration of such an instability, potentially founded on coordination failures or multiple equilibria, would restore a useful and significant role to macroeconomic stabilization policies. It would reinforce 141

8 the neo-keynesian view that business cycle fluctuations are necessarily the manifestation of a market failure on a grand scale (Mankiw, 1990). In any case, the answer to the question, Is aggregate demand unstable? is of paramount importance. Built on a dynamic version of the standard GE model, current dynamic equilibrium macroeconomic models appear to support the notion of stable, if not Pareto optimal, equilibria. But the RBC approach also provides a framework in which to test, as a residual explanation, the question of instability: suppose a fully coherent, plausible, dynamic macro model with well-measured exogenous shocks generates the right variability for the major macro aggregates, while simultaneously reproducing the main stylized facts; then surely this would be prima facie evidence of the stability of modern economies. However, if, with well-measured shocks, the most plausible models cannot account for observed volatilities, then this should be taken as evidence of a missing element which could plausibly be some form of aggregate instability. It is, of course, also legitimate to reverse the direction of the argument and to build multiple equilibria dynamic macro models and test their ability to reproduce the key stylized facts of macroeconomics. This is what Benhabib and Farmer (1994) and Farmer and Guo (1994) set out to do, thus providing evidence supporting the aggregate instability view. How should the debate proceed from here? It must proceed in two directions. It should try to break the tie between the two observationally equivalent models by checking their respective ability to pass the test of an enriched list of stylized facts. At the same time, it should discriminate between different variants on the grounds of plausibility. In that vein, Aiayagari (1995) has suggested that the multiple equilibria route followed by Benhabib and Farmer (1994) and Farmer and Guo (1994) does not pass the test of replicating the key stylized facts on the basis of plausible economic mechanisms. Specifically, he demonstrates that the ability of these particular models to reproduce the observations is predicated on the twin assumptions of an upward-sloping labour demand curve and a downward-sloping labour supply curve. The issue of plausibility is an interesting one. A problem with most current models of the RBC class is that they look and sound implausible to a large fraction of the profession, especially among its European component. Plausibility is likely to be a decisive criterion for success in the contest for the IS-LM succession. Views of plausibility, however, are subject to change: what is considered implausible today may well become part of the consensus tomorrow. Indeed, how are we supposed to react, if not by modifying our priors, if, repeatedly and convincingly, the models best able to match the stylized facts are those based on what we view, a priori, as implausible mechanisms, while the more plausible models systematically fail to account for some major observations? This said, it is conceivable today to build and put to the test a true Keynesian dynamic GE model. What one would like to see is a dynamic model economy in which, because of price rigidities, a negative demand shock, originating in macro policies or the current account, or possibly consumer confidence, is met with quantity adjustments across sectors rather than pure price adjustments as implicitly hypothesized in the standard business cycle model. In other words, we would like to see Keynesian spill-overs reappear in dynamic GE models and being confronted with the stylized facts. Gali (1996) sketches such a model with imperfect competition, sticky prices, and variable efforts, arguing that such a model stands a better chance of explaining what he sees as a negative (conditional) correlation between shocks identified as technology shocks and employment than more traditional RBC models. Most recently (in fact, too late to be given a full account in this article), Goodfriend and King (1997) also work out a similar model, not hesitating to give it the label the New Neoclassical Synthesis. IV. CONCLUSION It is evident that the demands placed on the successor to IS-LM are considerably heavier than what IS-LM itself or the neoclassical synthesis were able to deliver. It is also clear that the future workhorse of macroeconomics will not be able to do everything we would like it to do and still retain one major trait of IS-LM, its simplicity. A compromise will have to be found between performance and simplicity. 142

9 J.-P. Danthine Indeed, the key issue may well turn out to be whether the research programme that has evolved from the first RBC model will be able to deliver a sufficiently simple, transparent, and theoretically satisfactory model to satisfy the requirements of the IS-LM successorship. It is clear from our discussion that the successor will not be as simple as IS- LM. Could this lack of simplicity prevent a reconciliation between economic practitioners favourite model and the macroeconomic model put forward by academic economists? The widespread use of complex option pricing formulas by finance practitioners affords some optimism. It is not inconceivable that the standard tool for macroeconomic analysis, for practitioners and undergraduate students alike, will, in the future, be a palm-top computer version of a sophisticated, friction-prone, dynamic GE model. REFERENCES Aiayagari, S. R. (1995), Comments on Farmer and Guo s The Econometrics of Indeterminacy: An Applied Study, Federal Reserve Bank of Minneapolis Staff Report 196. Andolfatto, D., and Gomme, P. (1996), Unemployment Insurance and Labor-Market Activity in Canada, Carnegie Rochester Series on Public Policy, 44, Backus, D. K., Kehoe, P. J., and Kydland, F. E. (1992), International Real Business Cycles, Journal of Political Economy, 100, (1995), International Business Cycles: Theory and Evidence, in T. F. Cooley (ed.), Frontiers of Business Cycle Research, Princeton, NJ, Princeton University Press, Baxter, M. (1995), International Trade and Business Cycles, NBER Working Paper Crucini, M. (1995), Business Cycles and the Asset Structure of Foreign Trade International Economic Review, 36, Benhabib, J., and Farmer, R. E. (1994), Indeterminacy and Increasing Returns, Journal of Economic Theory, 63(1), Rogerson, R., and Wright, R. (1991), Homework in Macroeconomics: Household Production and Aggregate Fluctuations, Journal of Political Economy, 99(6), Boldrin, M., and Horvath, M. (1995), Labor Contracts and the Business Cycle, Journal of Political Economy, 103(5), Burnside C., Eichenbaum, M. S., and Rebelo, S. T. (1993), Labor Hoarding and the Business Cycle, Journal of Political Economy, 101(2), Campbell, J. (1994), Inspecting the Mechanism An Analytical Approach to the Stochastic Growth Model, Journal of Monetary Economics, 33, Carlstom, C. T., and Fuerst, T. S. (1995), Interest Rate Rule vs. Money Growth Rules: A Welfare Comparison in a Cashin-Advance Economy, Journal of Monetary Economics, 36(2), Chari, V. V., Kehoe, P., and McGrattan, E. R. (1996), Sticky Price Models of the Business Cycle: Can the Contract Multiplier Solve the Persistence Problem?, Federal Reserve Bank of Minneapolis, Staff Report 217. Christiano, L., and Eichenbaum, M. (1992), Current Real Business Cycle Theories and Aggregate Labor Market Fluctuations, American Economic Review, 82, Cochrane, J. (1994), Shocks, Carnegie Rochester Conference on Public Policy. Cogley, T., and Nason, J. M. (1995), Output Dynamics in Real-Business-Cycle Models, American Economic Review, 85(3), Cooley, T. F., and Hansen, G. D. (1995), Money and the Business Cycle, in T. F. Cooley (ed.), Frontiers of Business Cycle Research, Princeton, NJ, Princeton University Press. Danthine, J. P., and Donaldson, J. B. (1993), Methodological and Empirical Issues in Real Business Cycle Theory, European Economic Review, 37, (1995), Non Walrasian Economies, ch. 8 in T. F. Cooley (ed.), Frontiers of Business Cycle Research, Princeton, NJ, Princeton University Press. Johnsen, T. (1997), Productivity Growth, Consumer Confidence and the Business Cycle, Columbia University, mimeo. Deaton, A. (1992), Understanding Consumption, Oxford, Clarendon Press. 143

10 Dow, J. P. (1995), Real Business Cycles and Labor Markets with Imperfectly Flexible Wages, European Economic Review, 39, Dunlop, J. (1938), The Movement of Real and Money Wage Rates, Economic Journal, 48, Evans, C. L. (1992), Productivity Shocks and Real Business Cycles, Journal of Monetary Economics, 29, Farmer, R. E., and Guo, J. T. (1994), The Econometrics of Indeterminacy: An Applied Study, UCLA Department of Economics Working Paper No Gali, J. (1995), Real Business Cycles with Involuntary Unemployment, CEPR Discussion Paper 1206, London. (1996), Technology, Employment and the Business Cycle: Do Technology Shocks Explain Aggregate Fluctuations?, CEPR Discussion Paper 1499, London. Gertler, M. (1988), Financial Structure and Aggregate Economic Activity: An Overview, Journal of Money Credit and Banking, 20(2), Gomes, J., Greenwood, J., and Rebelo, S. (1997), Equilibrium Unemployment, University of Rochester, mimeo. Goodfriend, M., and King, R. G. (1997), The New Neoclassical Synthesis and the Role of Monetary Policy, University of Virginia, mimeo. Greenwood, J., Rogerson, R., and Wright, R. (1995), Household Production in Real Business Cycle Theory, ch. 6 in T. F. Cooley (ed.), Frontiers of Business Cycle Research, Princeton, NJ, Princeton University Press. Hall, R. E. (1988), The Relation between Price and Marginal Cost in US Industry, Journal of Political Economy, 96, (1990), Invariance Properties of Solow s Productivity Residuals, in P. Diamond (ed.), Growth/ Productivity/ Unemployment: Essays to Celebrate Bob Solow s Birthday, Cambridge, MA, MIT Press. Hansen, G. D. (1985), Indivisible Labor and the Business Cycle, Journal of Monetary Economics, 16(2), Hornstein, A. (1993), Monopolistic Competition, Increasing returns to Scale and the Importance of Productivity Shocks, Journal of Monetary Economics, 31(3), Kehoe, P., and Perri, F. (1997), International Business Cycles with Endogenous Incomplete Markets, Federal Reserve Bank of Minneapolis, Research Department, mimeo. Kollman, R. (1992), Incomplete Asset Markets and International Business Cycles, University of Montreal, mimeo. Kydland, F., and Prescott, E. C. (1982), Time to Build and Aggregate Fluctuations, Econometrica, 50(6), Lucas, R. E., Jr (1976), Econometric Policy Evaluation: A Critique, in Brunner and Meltzer (eds), The Phillips Curve and Labour Markets, Carnegie Rochester Conference Series on Public Policy, Vol. 1, (1980), Methods and Problems in Business Cycle Theory, Journal of Money, Credit and Banking, 12, McGrattan, E. R. (1994), A Progress Report on Business Cycle Models, Federal Reserve Bank of Minneapolis Quarterly Review, Mankiw, G., (1990) A Quick Refresher in Macroeconomics, Journal of Economic Literature, 28, Romer, D. (ed.) (1991), New Keynesian Economics, 2 vol., Cambridge, MA, MIT Press. Merz, M. (1995), Search in the Labor Market and the Real Business Cycle, Journal of Monetary Economics, 36, Muellbauer, J. (1994), The Assessment: Consumption Expenditure, Oxford Review of Economic Policy, 10(2), Rogerson, R. (1988), Indivisible Labor, Lotteries, and Equilibrium, Journal of Monetary Economics, 21, Rotemberg, J. J., and Woodford, M. (1995), Dynamic General Equilibrium Models with Imperfectly Competitive Product Markets, ch. 9 in T. F. Cooley (ed.), Frontiers of Business Cycle Research, Princeton, NJ, Princeton University Press. Tarshis, L. (1939), Changes in Real and Money Wage Rates, The Economic Journal, 49, Tinguely, O. (1997), Reorganization Costs, Business Cycle, and Asset Prices, University of Lausanne, mimeo. 144