1Banks and Finance in Modern Macroeconomics (2019) by Bruna Ingrao and Claudio Sardoni addresses one of the most fascinating questions raised by the history of macroeconomics after the Great Recession. After 2008, macroeconomists realized that their workhorse DSGE models were unable to explain the depth of the crisis and the slowness of the recovery. The crisis was obviously related to the role that the banking system, in the United States and elsewhere, had played in the massive expansion of mortgages during the previous decades. Its most striking manifestation was the freezing up of interbank markets all over the world. Yet, the main models used in central banks were deprived of any banking sector and assumed perfectly working financial markets. According to Ingrao and Sardoni, this was not new: “The dominant families of macroeconomic models in mainstream macroeconomics substantially avoided incorporating banks and finance in their basic analytical structures.” (2019, p. 2) Contrary to the mainstream, Ingrao and Sardoni assume that a true understanding of the macroeconomy is not possible without taking account of the role of banks and finance.

2For this reason, the book begins with two chapters on the “giants” of pre-Keynesian revolution macroeconomics who gave a crucial role to play for banks and credit markets. Chapter 2 compares Knut Wicksell and Irving Fisher. Both economists considered that it was impossible to understand the fluctuations of the price level and to obtain a correct version of the quantity theory of money without taking account of the banking system. The fluctuations of the volume of bank credit resulting from the gap between the interest rate on credit and the returns on investment modified the quantity of money and thus the price level. Chapter 3 discusses the contributions of Joseph Schumpeter and Dennis Robertson. These economists argued that banks’ intervention in the economy is a pre-condition of innovation and growth. According to Schumpeter, because they are able to create purchasing power ex-nihilo through credit, banks allow innovators to have access to the existing resources of the economy and to use them in new and more productive ways. Banks are not simple intermediaries but expand credit beyond existing savings. This generates inflation with real effects since innovation generates new profits and growth. Bankers play a crucial role in selecting entrepreneurs and in monitoring their activity. Hence, Schumpeter analyzed in depth the organization of the banking system and the conditions of banks’ activity in various historical phases paying attention to bankers’ innovations as a possible source of progress. The authors show that Robertson’s views have a lot in common with the views of Schumpeter. But beyond his tortuous language, it remains difficult to discern what he brings to the subject. Chapter 4 shows how, in the early 1930s, John Maynard Keynes (1931) and Fisher (1933) insisted on the negative effects deflation had on the financial structure of the economy. By increasing the burden of indebted agents, like firms and banks, deflation can lead to waves of bankruptcies that threaten the overall stability of the economy.

3Chapter 5 opens the story of how banks and finance were excluded from mainstream models. It shows how, from 1930 to 1936, Keynes progressively excluded commercial banks from his theoretical apparatus and the debates it triggered with economists considering the credit market as an essential part of any macroeconomic model. In his Treatise on Money, Keynes discussed the nature and the role of banks in the economy sharing with Robertson the idea that they could create money ex-nihilo. But when he entered the analysis of the trade cycle, he treated banks as a simple intermediary between lenders and borrowers and put all the emphasis on the liquidity preference of the public. Analyzing the role of banks became a mere refinement of the analysis. This tendency led him to exclude commercial banks altogether from the apparatus of the General Theory in 1936. The debate with Bertil Ohlin is used to show the merits of the loanable fund approach, centered on the credit market, and the errors of Keynes in his attempt to develop his finance motives. Chapter 6 and 7 follow the evolution of mainstream Keynesian macroeconomics from 1937 to the 1970s. This mainstream is related first to John Richard Hicks’ “attempt to expound macroeconomic theory in the context of a general equilibrium model” (2019, p. 114) in “Mr Keynes and the Classics” (1937), an approach that, according to the authors, was pursued further in Value and Capital (1939). But the main focus is on the contribution of Don Patinkin and Franco Modigliani “the two scholars who built the framing of macroeconomic theory by referring to the Walrasian general equilibrium” (2019, p. 145). Chapter 7 analyzes very precisely how Patinkin established a “four market framework” that erased the financial structure of the economy by assuming away distributive effects and that reduced finance to a market for bond without risk of default. At the aggregate level, wealth is reduced to “outside money” and the wealth effect is assumed to automatically restore full employment. A similar simplification of the financial sector is found in Modigliani (1963) who acknowledges that in his model “there is no need to give separate treatment to financial intermediaries” (1963, p. 97 quoted by Ingrao and Sardoni, p. 154). This perspective is related to Modigliani’s collaboration with Merton Miller and the resulting theorem (Modigliani and Miller, 1958) stating the irrelevance of firms’ financial structure. In this context, only two lines of research emerged from the wreckage according to the authors. The first is the contribution of John Gurley and Edward Shaw (1960) showing the importance of financial intermediaries for the process of growth. But their critique of the “net money doctrine” of Modigliani and Patinkin was resisted by Keynesian economists. James Tobin is presented as the only Keynesian economist who seriously addressed the role of banks and finance beginning in the 1960s. He is presented as an “unorthodox Keynesian” who rejected the Walrasian approach of mainstream macroeconomists. Yet Tobin (1982) ended up defending an extended version of IS-LM including equities on top of bonds and outside money, but leaving aside banks and the issues raised by the distribution of private debts. Chapter 8 discusses the contribution of Friedman and ends up with the Real Business Cycle literature which represent the last step in the “disappearance of money”. Chapter 9 surveys the macroeconomic literature spanning the last forty years that considered banks and finance from the perspective of imperfect information. The conclusion of the book explains the authors’ dissatisfaction with the current mainstream. If they credit the post-2008 DSGE literature with a rediscovery of banks and finance they consider that “the general environment within which the analysis is carried out” (2019, p. 243) remains unfit. The Arrow-Debreu intertemporal general equilibrium model that serves as a benchmark for macroeconomics is not consistent with models incorporating money and imperfections. The difference is one of nature and not of degrees. Furthermore, current macroeconomic models stuck with the “equilibrium plus shocks dyad” (2019, p. 256) would blind macroeconomists to the true nature of their object. Ingrao and Sardoni thus end up with a plea for a return to the insights of the giants of the inter-war and to practices less tied to mathematical modelling and more open to the complexities of history, the role of institutions and the reality of behaviors characterized by bounded rationality.

4To my knowledge, Ingrao and Sardoni’s book is the first attempt to explore systematically the attention that macroeconomics has paid to banks and finance since its beginnings around 1900. It is history of ideas at its best, a practice of history that takes the time to assess the consistency of the theories under scrutiny and to discuss their limits preparing the reader to “study the present state of economics from the standpoint of past authors” (Kurz, 2006, p. 468). In this respect it will probably remain a landmark in the history of macroeconomics historiography. It provides simultaneously a big picture of the subject and a myriad of subtle case studies to which the above summary does scant justice. The book is a must-read because of its breadth. But this breadth goes along with a lack of comprehensiveness in various places that leaves a number of questions open for further research.

5In all the book, Ingrao and Sardoni stick to the premise of their inquiry: banks and finance have been excluded from the mainstream. But this is not what a first exploration of the subject suggests and this is not what the evidence that they themselves gather suggests. This is particularly problematic when it comes to the 1960s and to the 1980s.

6Chapter 7 of the book claims that only Gurley, Shaw and Tobin seriously studied the role of financial intermediaries at the macroeconomic level in the 1960s. Besides, they are presented as mavericks. Actually, we have good reasons to suspect that commercial banks and the financial sector were a major topic of interest in the 1960s and that the three economists previously mentioned were influential. In the last pages of his 1963 article, Modigliani discussed what economists were beginning to call credit rationing on the part of banks. He considered this behavior to be important because it meant that monetary policy could influence investment even when it did not modify the rate of interest. Ingrao and Sardoni downplay this section of his paper because Modigliani himself admits that his version of IS-LM leaves no room for banks. But Modigliani’s interest for the “monetary mechanism” did not stop there. What Ingrao and Sardoni fail to consider is the fact that in the 1960s, the mainstream macroeconomic models were macroeconometric models. From 1966 to 1970, with Frank De Leeuw of the Division of Research and Statistics of the Board of Governors and Albert Ando, Professor at the University of Pennsylvania, Franco Modigliani headed a research team that developed the FMP model, later to become the first macroeconometric model of the Federal Reserve Board (Acosta and Rubin, 2019) [2]. The main objective of Modigliani and his team was to model and quantify the effects of monetary policy on income and prices. To do so, they developed a very detailed representation of the financial sector of the economy. The core of this sector were the equations representing the portfolio choices of banks or what they called the “money supply mechanism”. Their aim was to measure how a modification of the various instruments in the hands of the central bank would affect banks’ supply of deposits hence the short term rate of interest. Another very important issue for the participants of the project was the modelling and the measurement of credit rationing. Two Ph.D. students of Modigliani, John Hand and Dwight Jaffee, worked on the subject and discussed the already rich contemporary literature. In other words, it is not possible to have a clear picture of the role given to banks and finance in 1960s macroeconomics without taking into account macroeconometric models. In this light, I suspect that Tobin was not the quasi-heterodox Keynesian described by Ingrao and Sardoni. Like the other mainstream macroeconomists of the time, he was engaged in efforts to develop good macroeconometric models allowing to analyze the impact of monetary policy on the economy. Ingrao and Sardoni discuss Monetarism in a separate chapter and focus on Milton Friedman. But in the 1960s, two key figures of monetarism, Karl Brunner and Alan Meltzer, published a string of papers in which they discussed an enriched version of IS-LM included various asset markets and different types of financial intermediaries with ideas very close to those of Tobin. Hyman Minsky, whose work is discussed only briefly in chapter 9 because he was not a mainstream figure, was also working on the role of banks in the 1960s. A more integrated picture of this moment of history would have shown the 1960s as marked by a boom of research on banks and finance in macroeconomics. These topics were not absent from the mainstream. Still, the banking sector remained underdeveloped in textbooks and in presentation of their vision of the economy by leading figures like Modigliani (1977) or Tobin (1980). The reason why is still an open question.

7Chapter 9 on contemporary macroeconomics is problematic for the same kind of reason. Ingrao and Sardoni tend to group together all the literature spanning the period from 1980 to the post-2008 years. Besides, they consider that: “Initially, during the 1980s and 1990s, the New Keynesian concern for imperfections has essentially focused on the goods and labour markets. Although with exceptions, some of which are considered below, New Keynesian economists did not pay much attention to financial and monetary issues.” (2019, p. 215) In contradiction to this statement, one can argue that credit market imperfection was among the core issues of the New Keynesian literature. The number of seminal papers appearing in the early 1980s in the main journals is striking. In 1981, Joseph Stiglitz and Andrew Weiss published their paper on “Credit Rationing in Markets with Imperfect Information”. In 1981 and 1983, Ben Bernanke published two papers discussing how banks’ failures could deepen a recession. This nascent credit view was also defended by Alan Blinder and Stiglitz (1983). In 1983 also, Douglas Diamond and Philip Dyvig published their famous paper on bank runs. In the ensuing years, many works attempted to introduce financial intermediaries in macroeconomic models. Bernanke and Blinder developed an augmented IS-LM framework to present their “credit view”. With Mark Gertler, Bernanke tried to introduce banks in overlapping generations models, giving birth to the financial accelerator mechanism (Bernanke and Gertler, 1985, 1989). Business Cycle theoreticians like Robert Townsend (1983) or Stephen Williamson (1986) introduced credit in equilibrium business cycle models. In the 1990s, the literature went further by introducing the insights on the relation between banking and asymmetric information in DSGE models (Bernanke, Gertler and Simon Gilchrist, 1999; Nobuhiro Kiyotaki and John Moore, 1997). This rich literature, of which we only show the tip of the iceberg, laid the ground for the models that were developed after 2008. The end the 1980s also saw the beginning of a new stream of research on finance and growth mixing theoretical modelling and more empirical works with recurring references to Gurley and Shaw (1960) (Levine, 2005). The presentation of Ingrao and Sardoni prevent important questions to emerge. For instance, why did banks and credit markets attract so much attention starting at the very end of the 1970s? And if the financial accelerator and credit market imperfections made their way at the center of the mainstream in the 1990s, why did banks and financial frictions disappear from the models used in central banks in the 2000s? Henri Sneesens (2017) offers an empirical explanation: the financial accelerator did not improve the empirical performance of the models. More can certainly be said on the subject.

8For Ingrao and Sardoni the “‘troubled marriage’ of macroeconomics with general equilibrium theory since Patinkin’s early efforts in the 1950s” (2019, p. 244) explains why the mainstream fails to account for the role of banks and finance. Macroeconomics uses as its benchmark a “model of perfectly competitive equilibrium” in which money and, a fortiori, banks are not only inessential but cannot be integrated. They also stress the dead end in which the research program of general equilibrium theory ran in the 1970s after the results of Debreu, Mantel and Sonnenschein. Like the authors, I believe the reference to the Walrasian theory of general equilibrium in its various incarnations has been a major driving force of the history of macroeconomics. The unsolved problems of the Walrasian research program necessarily had consequences for the evolution of the field. But this interaction is complex and needs to be more carefully discussed.

9According to Ingrao and Sardoni, Hicks (1937) is responsible for putting the straightjacket of Walrasian theory onto macroeconomics: “theoretically, the sketchy ‘apparatus’ [the IS-LM model] mixed up Wicksellian with Walrasian influences.” (2019, p. 145) In other words, they suggest that IS-LM was a Walrasian model from the start. This claim ignores all the literature on the subject which leads to the opposite conclusion. IS-LM was invented by Keynes himself who presented the “sketchy apparatus” to his students in 1933 (Dimand, 2007). After the publication of the General Theory, the system of equation first reappeared in papers written by two students of Keynes (David Champernowne and Brian Reddaway) and two economists who participated to the making of Keynes magnum opus (James Meade and Roy Harrod). In his 1937 paper, Hicks correctly identified IS-LM as a Marshallian framework. As he was working on Value and Capital, he knew very well the difference between a Walrasian model and the model of Keynes (Barens, 1999). Oskar Lange (1938) was the first to spread the confusion by presenting Keynes’ model as a simplified version of the models of Walras and Pareto (Rubin, 2016). Following his lead, some macroeconomists, like Modigliani and Patinkin, considered IS-LM as a Walrasian model. Others, like Tobin and Solow, did not [3]. To discuss the influence of Walrasian general equilibrium theory on macroeconomics before the 1970s is to discuss how various Walrasian models (the models of Walras, Hicks and Arrow and Debreu) were used to develop a different and non-Walrasian set of models (IS-LM and macroeconometric models). The situations changed in the 1970s when explicitly microfounded models were developed by disequilibrium theoreticians and New classical economists. But even then, and this is highlighted by the authors themselves, the Walrasian nature of macroeconomic models remained questionable. What is the relation, they ask, between a representative agent RBC model and the Arrow-Debreu model? I would add that many New Keynesian models were typically Marshallian being partial equilibrium and characterized by assumptions concerning information and price formation that were totally at odds with a Walrasian setting (De Vroey, 2016, p. 352–3). In this context, it is not enough to point out the absence of money in the Arrow-Debreu model. Macroeconomists produced many models incorporating banks and that were part of the mainstream. How did the Walrasian benchmark influence the way banks were defined and modelled or the way they were excluded when it was the case? Was general equilibrium theory really the prime influence here? The argumentation remains to general and cursory to be convincing. The discussion of the DSGE literature in particular is wanting. Ingrao and Sardoni consider the general DSGE methodology introducing shocks in the context of steady state economies as inadequate to deal with serious financial problems. But they do not discuss how this methodology constrains the way banks are defined and introduced in the models. Instead, in the case of the financial accelerator of Bernanke and his co-authors (1999) they only note that they “do not explicitly consider banks but only generic financial intermediaries” (2019, p. 220). Why is it so and what does it mean? It was the project of Bernanke and Gerlter (1985) to model banks in a general equilibrium setting. Why did they abandon this project? How did Gertler introduce less trivial banking sectors in more recent papers with Karadi and Kyotaki? Is their modelling of the banking sector satisfactory? If not, is it a consequence of the Walrasian benchmark or of other considerations like tractability or their understanding of the financial system?

10To conclude, the book by Ingrao and Sardoni is inspiring and a perfect starting point for all scholars interested in the subject.