1 – Introduction
1Starting in the mid 1950s, and continuing over the succeeding 40 years, Hyman P. Minsky developed his original business cycle theory based on his financial conception of economic fluctuations, more specifically, his ‘Financial Instability Hypothesis’ (FIH) [Minsky, 1975 ; 1982 ; 1986]. This theory is based mainly on two successive phases in the business cycle : a process of transition toward greater financial fragility, which builds up in the expansionary phase ; and the transition from a financially fragile situation to a situation of recession and then major economic crisis.
2Minsky’s commentators usually point to two central features characterizing his business cycle theory. First, his approach is considered a continuation of the theory of economic fluctuations developed by Keynes in his General Theory [Keynes, 1936] and synthesized in Quarterly Journal of Economics article published in 1937. In Minsky’s approach one can recognize the main themes developed by Keynesian fundamentalists : the role of uncertainty and the behaviors it generates, the dynamic instability of market economies, the role of money and liquidity preference. The second feature is the endogenous nature of his fluctuations approach. For most commentators, the FIH is based on the behaviors of private economic agents which endogenously create processes of financial and economic instability, and unsurprisingly, most recent models ‘à la Minsky’  are founded on this idea. The main characteristic of these models is that they introduce financial structure variables (such as indebtedness ratios chosen by private agents) into Keynesian or Kaleckian macroeconomic frameworks. These financial variables are at the origins of unstable endogenous dynamics and produce fluctuations analogous to those studied by Minsky. It should be noted also that Minsky also developed an interest in non-linear models,  able to generate endogenous cycles and complex dynamics. Indeed, ‘these non-linear models are not vulnerable to the criticism that endogenous business cycles are too regular’ [Ferri and Minsky, 1992 : 82]. Moreover, they explain the observed dynamics of market economies which, for Minsky, are not ‘necessarily nice : monotonic explosive, explosive amplitude cycles and even chaotic cycles are possible paths through time’ [Ferri and Minsky, 1989 : 137].
3The aim of this paper is to show that this traditional interpretation of Minsky’s business cycles theory is at least incomplete. However undeniable the influence of the ideas he inherited from Keynes, other theoretical schools and individual authors left their marks on his writings. As a result, these Minsky cycles cannot be considered a purely mechanical and endogenous succession of phases induced by the behaviors of private agents. Other economic agents, such as public authorities through their stabilization policies, also play a key role in the form of economic fluctuations described by the FIH.
4The argument in this paper is developed in three steps. First, we show that Minsky’s conception of business cycles is influenced by nonlinear models with floors and ceilings developed by Hicks in the 1950s. These models provide a relevant theoretical framework for a Minskyan analysis of stabilization policies which act as ‘thwarting systems’ whose purpose is to counteract and contain the naturally explosive amplitude of economic fluctuations (Section 2).
5Second, the central role played by banks in Minsky’s analysis of fluctuations is highlighted, and the behavior of banks guided essentially by a permanent quest for profit opportunities is emphasized, in the tradition of authors such as Tobin or Schumpeter. This feature is crucial for understanding Minsky’s conception of business cycles since it describes how banks’ behaviors undermine the efficiency of stabilization policies, and therefore the form of economic fluctuations (Section 3).
6Section 4 briefly assesses the relevance of Minsky’s approach for understanding recent pre- and post-crisis stabilization policies, and concludes the paper.
2 – Reinterpreting growth models with ceilings and floors : The Hicks-Goodwin tradition revisited
7In the late 1950s, Minsky  proposed business cycle models that drew heavily on the multiplier and accelerator interaction models developed several years earlier by Hicks . Hicks was well aware of both the inherent limits of linear macroeconomic models of the type proposed by Samuelson, and the fundamentally non-linear nature of economic activity. For this reason his business cycle models incorporated constraints, whose role was to act as boundaries to expansion or depression.
8The main assumptions underlying these models can be briefly restated. They are usually composed of a consumption function of the form, Ct = aYt – 1 (where a is the average and marginal propensity to consume), and an induced investment function of the form, It = b(Yt – 1 – Yt – 2) (where b denotes the accelerator). When introduced into the equilibrium condition, Yt = Ct + It, these relations yield a second-order difference equation :
10Equation (1) means that when the value of the reaction coefficients a and b and the values of the initial conditions, Yt–1 and Yt–2, are known, it is possible to determine recursively any solution Yn to the system. In addition, we know that the second-order difference equation (1) solves for :
12where U1 and U2 are the roots of the associated characteristic equation U2 – (a + b)U + b = 0, and A1 and A2 are constants depending on the value of parameters a and b and on the initial conditions. 
13Solving equation (2) yields different types of dynamics depending on the value of the parameters, namely : (a) monotonic convergence of the income level towards a stable level (if the roots are real and of absolute values of less than unity) ; (b) damped fluctuations, the system is stable and converging towards the long-term equilibrium level by values that are alternately less than and greater than the equilibrium output level (if the roots are complex and their modulus is less than unity) ; (c) explosive fluctuations, the amplitude of fluctuations of national output around the long-term equilibrium level increasing in each period (if the roots are complex and their modulus is greater than unity) ; (d) instability in the form of monotonic divergence or regular explosive growth, the national product diverging increasingly from its equilibrium level (if the roots are real and greater than unity in absolute value).
14Unlike Samuelson’s  analysis, which investigates the whole range of solutions set out above, Hicks  concentrates exclusively on parameter values which, when combined, give either accelerated growth or amplified fluctuations – cases (c) and (d). Hicks therefore situates his approach in a configuration such that both roots U1 and U2 are greater than unity, with U1 > U2 > 1. He further assumes that the economy exhibits a maximum growth rate, a ceiling, which we will denote g. Therefore, when the ceiling is effective. In addition, the dominant root U1 is very much greater than the growth rate of the income ceiling. Finally, both roots U1 and U2 are such that U1 > U2 > g > 1.
15Under these assumptions, the economy evolves as follows. Let us suppose that the economy is initially defined such that it generates two consecutive incomes Y0 and Y1 (Y1 > Y0), and such that both these incomes are less than their corresponding ceilings and U1 > Y1/Y0 > U2. These initial conditions determine the positive coefficients A1 and A2, A2 being much greater than A1, because U1 is assumed to be much greater than g.  Then equation (2), which characterizes the dynamics of the system or rather its unconstrained dynamics, yields an explosive-type evolution over time. However, after a certain time, let us say in period n, income reaches a higher value than the ceiling corresponding to that period. At that point, the constraint becomes binding and the realized income is no longer determined by equation (2). For two successive periods income is determined by equation :
17This takes us back to a situation where Y1/Y0 = g < U2. The prominent feature here is that the sign of the coefficient A1 of dominant root U1 solving equation (2) becomes negative. This change of sign indicates the start of the rebound of the path against the ceiling, that is, the turning point in the cycle. A cumulative depression process follows and the economy henceforth undergoes a downturn that is guided essentially by the negative term of increasing absolute value, A1(U1)t.
18This explosive downward movement of income will only slow if there is a lower limit, or floor that counters the process generated by the unconstrained dynamics, analogous to the effect of the ceiling examined above. The floor may consist of a maximum value set for firms’ divestments. When this maximum value is reached, the realized value Yt is different from that obtained with equation (2). A new equation determines the change taking place in the economy. It features a negative and comparatively large coefficient A2 and a positive and comparatively small coefficient A1. As in the previous case, there is a point where the cycle reverts. A new expansion phase begins, which initially is moderate and then is explosive. It continues until the economy rebounds at the ceiling, as described previously. And so on.
19In the late 1950s, Minsky proposed business cycles models with the same general form as the Hicks model presented above. Like Hicks, Minsky opted to take non linearities into account by introducing constraints on booms and slumps, in the form of floors and ceilings. Beyond this apparent similarity, however, the interpretations differ insofar as these two economists do not attribute the same meaning to the constraints that stabilize the economy dynamics. In Hicks’s model, common sense justifies the existence of these constraints : real investment cannot be negative, hence the existence of a floor, determined by the growth of autonomous investment and the size of the investment due to depreciation. In addition, output, consumption, and investment are limited by the available amounts of natural resources and labor, as well as by the size of the gains in productivity. Hence, the existence of a ceiling.
20However, although Minsky does not deny the influence of these constraints on the level of economic activity, he does not consider them essential for explaining the cyclical dynamics observed in market economies. He argues that floors and ceilings primarily reflect the set of institutional mechanisms put in place by public authorities in order to confine the amplitude of economic fluctuations within reasonable limits. Thus, Minsky terms these institutional arrangements ‘thwarting systems’ [Ferri and Minsky, 1992].
21The original idea developed by Minsky is that the main purpose of these thwarting systems is to modify the initial conditions governing time series during phases of explosive expansion or cumulative depression. As already mentioned, these new initial conditions have the effect of inverting the sign of the dominant root of the oscillator model solution equation. As a result movement in the economy eventually slows and reverses. Depression (or deflation) is thus converted by an institutional-type floor into a moderate and then explosive recovery which in turn, comes up against an institutional ceiling. The time series observed then is the outcome of incessant rebounds between the ceilings and floors, generated by institutional thwarting mechanisms introduced into the economy.
22Minsky’s constrained linear model is interesting in more than one respect. First, unlike Samuelson’s oscillator-type unconstrained linear model, it accounts for the complexity of change affecting capitalist economies over time. Evolution takes the form of ‘steady growth (when U1 > g > U2), cycles (when U1 > U2 > g > 1), booms, or depressions when (g > U1)’ [Minsky, 1959 : 134]. In other words, Minsky’s model ‘exhibits the features of chaotic models, including the sensitivity of the time series that is generated to initial conditions’ [Ferri and Minsky, 1989 : 138]. Second, unlike Hicks’s non-linear model, which ignores institutional thwarting systems, this model accounts for the effect of institutional change and interventions by public authorities on economic dynamics. In Minsky’s model the dependence of floors and ceilings on policy and institutional arrangements can be made quite precise.
23The ‘incoherence’ of economic trajectories – in particular the occurrence of explosive, amplified change leading to either very large (even infinite) values or to negative values of economic magnitude – inherent in unconstrained linear models, can thus be countered by the incorporation of institutional thwarting mechanisms. Under these circumstances, ‘business cycles can result either from the values of the “U’s” being complex, from regular interventions that contain the economy between “floors and ceilings” if the “U’s” are greater than one, and from introductions of energy from outside if the “U’s” are less than one’ [Ferri and Minsky, 1989 : 137].
24In Minsky’s approach, stabilizing economic activity, that is, setting new initial conditions in order to contain the amplitude of time series, is essentially the concern of government, and the Central bank. 
25Minsky views budget deficits and central bank interventions as lender-of-last-resort actions and extremely effective instruments for stabilizing economic fluctuations. Even if full employment is not achieved, these instruments help to limit the drop in income and in liquidity during economic recessions and the onset of financial crises.
26First, let us consider the role of budget deficits. In Minsky’s theory, investment is the essential determinant of economic activity. Investment is influenced largely by aggregate profits (realized or anticipated).  Minsky argues, that it follows that ‘a main aim of policy is to constrain the variability of profits’ [Fazzari and Minsky, 1984 : 107]. Also, Minsky reminds us that in a closed economy aggregate profits are equal to the sum of investment and the budget deficit, as expressed by Kalecki’s  accounting identity.  Consequently, a deficit, by upholding aggregate demand when private investment flags, establishes a lower limit, a floor, to profits, wages, and current production prices. In other words, ‘policy will be stabilizing if a shortfall of private investment quickly leads to a government deficit, and a burst of investment quickly leads to a budget surplus’ [Fazzari and Minsky, 1984 : 107]. Such stabilization of actual and expected profits is crucial for ensuring the continuity of the economic system. It is utilized in particular to maintain the viability of debt structures, and therefore the level of private investment. In fact ‘once rational bankers and business men learn from experience that actual profits do not fall when private investment declines, they will modify their preferred portfolios to take advantage of the stability of profits’ [Minsky, 1992 : 12]. This means in particular that, encouraged by increased confidence in future profits, banks are likely to raise their estimates of the maximum levels of indebtedness (in proportion to the value of the assets both they and potential borrowers hold) to which it is prudent to agree, and will tend to make more loans to businesses.
27The importance of “Big Government” (i.e. a public sector that constitutes a significant share of aggregate demand) in economic dynamics is so fundamental for Minsky that he divides the performance of the U.S. economy into two periods : a “small government” era from the end of the American Civil War to the Depression, and a “big government” era dating from World War II (Minsky, 1986). He provides a detailed analysis of how the deficit of “Big Government” was crucial for maintaining profits during the 1974-75 recession [Minsky, 1986 : ch. 2]. He points out that the annualized rate of contraction of the economy during the first two quarters of 1975 rivaled that of the worst rate of decline that occurred during the Great Depression. But big government’s deficits prevented profits from declining and set the stage for recovery in the second part of the year.
28Thus, the presence of a ‘big government’ whose fiscal policy is very sensitive to variations in overall profits improves the stability of the economy. Nonetheless, an isolated policy of this sort may prove insufficient in periods of economic turmoil. True public deficits partly offset the reduction in profit flows resulting from a fall in investments, and maintain current production prices and consumer goods prices. However, during an economic crisis, deficits cannot directly counteract the drop in another type of price, and one that is essential in Minsky’s investment theory, the price of capital assets. This price is dependent upon the amount of money in circulation but also on more subjective variables such as liquidity preference, the debt level that is judged acceptable, or the profits economic agents expect.  It is necessary, as a supplementary step, to turn then toward a second type of institutional thwarting mechanism, the role of the Central bank as the lender of last resort.
29The main purpose of this type of intervention is to offset debt-deflation phenomena or the different forms of financial instability that market economies experience.  For Minsky, these phenomena emphasize the need for an extended interpretation of the role of lender of last resort. Thus, he distinguishes three aspects of this type of intervention [Minsky, 1986]. First, when funds are lacking in the money market (a situation generally synonymous with substantial falls in the value of the claims agents exchange for liquidity), the Central bank must intervene by increasing the amount of money in circulation. Second, during the period of financial restructuring that follows a crisis, the Central bank must take care to favor recourse to long-term rather than short-term borrowing by acting on interest rates accordingly. Finally, the Central bank is responsible for guiding the development of the financial system, through both regulation and banking system surveillance, in order to restrain speculative banking (excessive reliance on liability management in particular).
30Including the role of institutional mechanisms implies that Minsky’s theory does not simply describe the cycle as an endogenous mechanical phenomenon, as a succession of the following types of phases : increased financial fragility ? financial crisis ? gradual return to more safe and sound finance. We saw that this is because Minsky considers that financial crises thwarted by institutional mechanisms fail to develop all their effects (debt-deflation, widespread bankruptcies, disappearance of agents with the most fragile financial structures, etc.). In the financially sophisticated economies that Minsky scrutinizes, the natural tendency to generate periods of great instability may be partly offset by thwarting systems.
3 – “Active” banking and difficulty in stabilizing an unstable economy
31In this section we investigate another central idea in Minsky’s approach : that the efficiency of the thwarting systems implemented by public authorities depends strongly on the more or less active behavior of commercial banks. Active behaviors are exemplified by the various practices banks use to modify their balance sheet structure, both the asset and the liabilities sides, and also their capacity to create or to favor the processes of financial innovation. The general idea Minsky develops is that if banks are passive, then they can be ignored in the theoretical reasoning : they do not influence the efficiency of the stabilization policies. On the other hand, if banks are active, then the efficiency of these policies can be questioned.
32Now, a key idea of Minsky’s analysis is that banks are indeed active economic agents because (i) their behavior is guided essentially by the permanent quest for profit opportunities, and (ii) they are also entrepreneurial organizations striving to innovate in order to generate capital gains. The first feature illustrates the influence of Tobin’s analysis, the second feature refers to the influence of Schumpeter’s approach.
33Minsky’s view lessens to a great extent the distinction between banks, other financial institutions, and non-financial agents. Indeed, ‘the line between commercial banks, whose liabilities include checking deposits, other depository thrift institutions, miscellaneous managers of money (like insurance companies, pension funds and various investment trusts), and investment bankers is more reflective of the legal environment and institutional history than of the economic function of these financial institutions’ [Minsky, 1986 : 223]. For similar reasons the distinction between financial institutions and non-financial agents seems not to be crucial for him : ‘banks and bankers are not passive managers of money to lend or to invest ; they are in business to maximize profits. They actively solicit borrowing customers, undertake financing commitments, build connections with business and other bankers, and seek funds’ [ibid. : 229–30].
34This aspect of Minsky’s theory is reminiscent of the approach developed in ‘Commercial Banks as Creators of Money’ [Tobin, 1963], an article where Tobin contrasts the ‘old view’ on banking with the ‘new view’. Like Minsky, Tobin and the supporters of the new view make no clear distinction between banks and other financial institutions, or between what Gurley and Shaw call the ‘monetary system’ – comprising the commercial banks and the Fed – and the ‘other financial intermediaries’ [Gurley and Shaw, 1955 : 260–1]. All these financial agents are ‘financial intermediaries’ whose main and characteristic function is to ‘satisfy simultaneously the portfolio preferences of… borrowers [and] lenders’ [Tobin, 1963 : 274].
35Moreover, Minsky insists that banks are entrepreneurial firms whose innovations allow greater profits : new financial instruments, new financial procedures and new financial institutions in particular are created by innovative bankers who receive monopoly rents that disappear as innovations diffuse [Minsky, 1990 ; 1993]. This view is akin to Schumpeter’s analysis of banking. According to the Austrian author, in a capitalist environment, bankers perform an entrepreneurial function, which is in no way less important than that of business entrepreneurs for economic development : creation and destruction occur in the field of finance, as well as innovation – whether in the form of new products, processes or new organization, and whether it involves incremental or radical change. Moreover, new types of financing media may emerge and thereby trigger further process and product innovations. 
36Like Schumpeter, Minsky considered that financial systems evolve not only in response to demands of business firms and individual investors but also as a result of the innovative activity of profit-seeking entrepreneurial financial firms.
37The relationship between “active” banks’ behaviors and the efficiency of stabilization policy was analyzed by Minsky in his first published article, ‘Central Banking and Money Market Changes’ [1957b]. This article highlights the entrepreneurial and innovative behavior of commercial banks, and the effects such behavior has on the efficiency of a monetary
38In this article, Minsky analyzes the case of the Central bank implementing a restrictive monetary policy to prevent an inflationary process during an economic expansion [Minsky, 1957b : 171–3]. This policy causes an increase in the interest rate owing to ‘a vigorous demand for financing relative to the available supply’ [ibid. : 163]. Two possibilities are considered by Minsky. The first consists of reasoning within a stable institutional environment, one where ‘a tight money policy will be effective and the interest rate will rise to whatever extent is necessary in order to restrict the demand for financing to the essentially inelastic supply. This can be represented as a positively sloped curve between velocity and the interest rate’ [ibid. : 172].
39However, institutional stability is not Minsky’s main focus. At the beginning of the 1950s changes began affecting the money market as the markets for Federal funds and for repurchase agreements emerged and expanded. For Minsky these evolutions simply reflected the existence of a form of institutional instability governed by the profit-seeking and innovative behavior of commercial banks.
40The process described in the 1957 article unfolds as follows. Rising interest rates act as a signal which private market operators interpret as new profit opportunities. Higher interest rates imply that greater opportunity costs affect the excess reserves held by commercial banks. Their incentive is therefore to lend these reserves to the Federal funds market. Also, Minsky observes [1957b : 164], the Federal funds rate is always lower than the discount rate b. This is a circumstance that banks short of reserves will exploit. Greater reliance of banks on the market for Federal funds allows a larger volume of demand deposits for a given amount of Central bank money. ‘[A] given volume of reserves now supports more deposits’ [Minsky, 1957b : 171] because now the excess reserves of some banks are available to support the deposits of deficit banks which then are not forced to borrow from the Federal Reserve Bank or to sell securities.
41Rises in interest rates also affect behaviors – and balance-sheet structures – of non-financial firms. The latter are guided away from non-interest-bearing demand deposits toward more profitable investments. New holdings can take the form of short-term liquid assets such as Treasury bills or repurchase agreements issued by government bond houses.  Government bond houses will be further induced to borrow from non-financial firms through the issue of repurchase agreements if the interest rates on these instruments fall below the discount rate charged by the Federal Reserve and the interest rates charged by commercial banks.
42An important consequence of this potential shift of Treasury bills and of lending to government bond houses from commercial banks to nonfinancial firms is that bank resources are freed to finance other activities : for a given volume of deposits, the share of loans to firms in banks assets is likely to increase.
43At the same time, this process can cause a decrease in the liquidity of the economy as illustrated by the following example.
44Consider a very simplified monetary system composed by a bank B, a non-financial firm F, and a government bond house GBH.
45The government bond house holdings are composed of Treasury bills (TB) and longer-term government debt (LTGD). These assets are financed by their own capital (C), sales and repurchase agreements with firms (RAF,) sales and repurchase agreements with the Federal Reserve System, at the discount rate, (RAFed), and finally loans from commercial banks (LB). The holdings of the non-financial firm F are made up of demand deposits (DD), Treasury bills, and sale and repurchase agreements with government bond houses (RAGBH = RAF). The assets of the commercial bank B are made up of Treasury bills, loans to the non-financial firm (LF), and loans to the government bond house (LGBH = LB).
47Suppose a situation similar to the one prevailing in the US money market from 1954 to 1957. During this period the following two phenomena are observed [Minsky, 1957b : 163 and 167–8] : (a) The interest rate pattern relevant to the operations of GBH becomes TB < RAF < RAFed < LTGB < LB ; (b) Short-term interest rates rise.
48These changes appear as new profit opportunities for private economic agents which react by modifying their balance-sheets as follows :
49The GBH reacts to this new interest pattern by increasing its borrowings from the non-financial firm, that is, by issuing more (+100) repurchase agreements. Simultaneously, the GBH decreases both its bank and Federal Reserve borrowings by 50.  This is facilitated especially because rises in short-term interest rates lead the firm to move away from non-interest-bearing demand deposits (which decrease by 50) and to invest its cash in Treasury bills (which increase by 50) and repurchase agreements issued by the GBH (which rise by 100). These changes to the firm’s and the GBH balance sheets in turn affect the portfolios of the commercial banks and Treasury bills (–50) and loans to the GBH (–50) are replaced by loans to firms (+100).
50The above example highlights an important outcome of the profit-seeking behavior of economic agents. Liability and asset management induced by changes in the interest rate pattern generate a greater risk of illiquidity in the economy since both financial (commercial banks) and non-financial (firms) agents now have less liquid assets in their portfolios.
51The institutional changes in the money market, and the behaviors of private economic agents in this market, induce another central result : they increase the lending ability of commercial banks. Indeed, for a given quantity of reserves and money, the amount of loans to firms that can be supplied by commercial banks tends to increase. In other words, the process of institutional innovation tends to increase the velocity of the circulation of money. This results in rightward shifts in the interest rate – velocity curve. An upward-stepped curve obtains, similar to the curve depicted in the figure 1.
Active commercial banks and the efficiency of monetary policy
Active commercial banks and the efficiency of monetary policy
52The increasing portions of the curve describe the effect of a restrictive monetary policy on the interest rate when the institutional environment remains stable. However, these increases are not permanent ; an increase in the interest rate (for instance, from r0 to r1) creates profit opportunities, money market innovations and thus institutional instability, as described by the curve shift from I to II. A plateau, a – b, is reached, which characterizes the period during which institutional innovation is spreading.
53During this plateau period, the impact on the interest rate of a restrictive monetary policy is completely counteracted, while the velocity of circulation and the lending ability of commercial banks appear to be infinitely elastic. As stressed by Minsky, in such a context the effectiveness of monetary policy based on the surveillance of monetary aggregates weakens, with the result that, in order to fight inflation, the Central bank has no choice but to act directly on the liquidity of commercial banks. The Central bank will try to diminish their reserves to an extent great enough to compensate for the increase in velocity. This reaction to the profit-seeking and innovative behavior of commercial banks then has the effect of pushing interest rates up even more, thus recreating the conditions for the whole process to recur. Hence, the succession over time of the increasing and horizontal portions of the curves represented in the figure.
54To sum up, Minsky’s first published article is very instructive as to his conception of the ability of a monetary policy to be an efficient thwarting system.
55The efficiency of a restrictive monetary policy aimed at fighting inflation during a period of boom is severely reduced by the behavior of commercial banks because they are not only eager (because of profit opportunities), they are also able (thanks to asset management and innovation) to respond to firms’ larger credit demands.
56This is not to say that such a policy is a totally inefficient tool to curb inflation. The upper turning point in Minsky’s cycle is explained partly by the increases in interest rates induced by a restrictive monetary policy : ‘in terms of the financing of investment… an infinitely elastic supply of finance will exist if and only if the banking system is willing and able, to finance any amount of investment… at unchanging interest rates. For a variety of reasons – the limited equity base of banks, internal and foreign drains of bank reserves and, in modern times, Central bank (Federal Reserve) actions to restrain the money supply – the supply of finance from banks eventually becomes less than infinitely elastic’ [Minsky, 1986 : 195 – emphasis in original]. This implies that a restrictive monetary policy contributes to the existence of a ceiling to aggregate activity. However, the main message in Minsky’s approach is that this policy is relatively inefficient during a boom phase if implemented in isolation. As underlined above, to be a really efficient thwarting system, such a policy must be complemented by other types of Central bank actions bank (such as prudential regulation) and by an active anti-cyclical fiscal policy with a government budget that goes quickly into surplus when investment, debt, and inflation are high.
57Another consequence is that a restrictive policy may also be destabilizing since, as we have seen, it generates a greater risk of illiquidity in the economy when liquidity-decreasing money-market innovations occur. In other words, the innovative process described earlier is likely to lead to a situation where the economy becomes less liquid although the quantity of credit supplied to firms is rising. Minsky [1957b : 173] describes it thus : ‘the reverse side of the coin to the increase in velocity is that every institutional innovation which results in both new ways to finance business and new substitutes for cash assets decreases the liquidity of the economy’.
4 – Concluding comments
58For more than two centuries, there have been two opposing conceptions of the evolution of economic activity. In his memorial to Wesley Mitchell, Schumpeter [1951 : 252] distinguishes between those economists who consider that ‘the economic process is essentially non-oscillatory and that the explanation of cyclical as well as other fluctuations must be sought in particular circumstances (monetary or other) which disturb that even flow’ and those who believe that the ‘economic process itself is essentially wavelike – that cycles are the form of capitalist evolution.’
59The analysis developed by Minsky in his earliest works, while not questioning this distinction, shows that it must nevertheless be reconsidered if institutions are to be embodied explicitly within business cycle analysis. More precisely, his approach highlights that taking into account the interactions between public (government and Central bank) and private institutions (commercial banks) opens space for a third family of business cycle theories. Fitting somewhere in between purely exogenous conceptions and endogenous explanations of business cycles, it depicts the economic process as not “essentially non-oscillatory” just as it is not “essentially wave-like” to use Schumpeter’s terms. This approach encompasses both steady growth and regular business cycles as possible transient features of an economic time series but it also does not exclude the emergence of the potentially unstable dynamics resulting from the interactions between private and public institutions.
60Thus, Minsky’s conception of business cycles in the 1950s revives the treatment of the relations that arise between public institutions and economic fluctuations. In traditional macroeconomic models, it is often assumed that economic policies are totally exogenous and public institutions can only disturb the operations of markets that otherwise clear through the interplay between supply and demand. In contrast, Minsky’s approach shows that the public authorities react in an endogenous fashion to the behavior of private agents. Thus, they participate in creating a genuinely institutional dynamics that interwines with the real and the financial dynamics of the economy and modifies the unconstrained outcomes that would occur in the absence of institutional thwarting systems.
61Finally, scrutiny of Minsky’s earliest works provides a better understanding of his vision of an efficient stabilization policy as necessarily based upon an active anti-cyclical fiscal policy. This is considered necessary (in the tradition of Kalecki) to avoid the economy sinking into a never-ending depression because budget deficits increase global profits and validate the financial structure of private economic agents. To be efficient, this stabilization policy has to be associated with a monetary policy that remains relatively accommodative to whatever economic situation in order to maintain interest rates at a low level, and to stabilize asset prices. The efficiency of other kinds of monetary policies aimed at regulating economic activity, is very weak in a context where banks adopt active and innovative behaviors which annihilate a great part of the Central bank’s actions.
62The conclusions reached by Minsky in the 1950s appear relevant today ; only recall the inability of the Fed to prevent financial instability problems during the 2000-2007 pre-crisis period, because of the active behaviors of U.S. banks. Although the Fed succeeded in maintaining inflation stability, it failed to prevent the rise of financial fragility and instability induced by innovative and riskier banking behaviors. 
63The same reasoning can be applied to the current post-crisis situation in Europe where commercial banks have huge excess reserves at their disposal thanks to the very low interest rate three-year loans granted by the European Central Bank (ECB) in December 2011. However these reserves are not being activated by the banks : crisis-hit euro zone banks are parking record amounts of cash in the ECB for overnight storage, despite low interest rates. In normal times, banks shy away from depositing cash at the ECB, preferring to lend any overnight surplus to other banks and earn a higher rate of return. But the crisis has spawned a lack of trust among the banks with the result that they are opting to store their money with the ultra-safe ECB rather than take the riskier route of lending it to counterparts. So, commercial banks which are adopting behaviors founded on a greater liquidity during the current post-crisis period are tending to privilege more liquid and less risky assets. In doing so, they are contributing to reducing the quantities of money and credit in circulation and reducing the efficiency of the monetary policy, just as Minsky described more than 50 years ago.
University of Nice Sophia-Antipolis–ISEM. E-mail : Eric.Nasica@unice.fr
Arena and Raybaut  ; Delli Gatti et al. [1993a ; 1993b] ; Delli Gatti and Gallegati , Asada , Charles , Isaac and Kim , Lima and Meirelles , Sasaki and Fujita . For models à la Minsky which explicitly integrate stabilization policies, see e.g. Keen  and Nasica and Raybaut .
See Ferri and Minsky  which explicitly refer to Day’s models [1982 ; 1986].
Minsky [1957a ; 1959].
If m is the effective growth rate of income, then for any two successive dates chosen as initial conditions, Y1 = mY0. This therefore gives (since Y1 = A1U1 + A2U2 and Y0 = A1 + A2 : A1 = [(m – U2) + (U1 – U2)]Y0 and A2 = [(Ut – m)/(U1 – U2)]Y0. Assuming that values a and b are such that U1 > U2 > 1 (i.e. in the case of explosive time series in an unconstrained system), it follows that : U1 > U2 > m, therefore A1 < 0 and A2 > 0 ; whereas U1 > m > U2 implies that A1 > 0 and A2 > 0.
See fn. 2. This characteristic implies that during the early periods (t small), the weight of U2 is predominant in determining the dynamic evolution, whereas during subsequent periods, it is the root U1 that tends to dominate. It follows that the income growth rate converges towards U1 when t tends towards infinity.
In a few papers, Minsky examines other thwarting systems such as labor market institutions and market power. See e.g. Ferri and Minsky  where the authors study these stabilizing mechanisms in some length.
See in particular Minsky [1986 : Ch. 7 and 8].
If Y represents the national income, C the total consumption, W total wages, I the investment, ? the aggregate profits, G the government spending and T the taxes, we have : Y = C + I + G and Y = W + ? + T. If, as Kalecki does, we suppose that C = W, we obtain : ? = I + G – T.
Minsky [1975 ; 1986]
In the case of the U.S., to which Minsky gives precedence, he observes that the Fed’s function as lender-of-last-resort has extended constantly to new institutions and new instruments. At the end of the 1960s the Fed intervened to sustain the municipal bonds market. In 1970 it acted in order to avoid the collapse of the commercial paper market. In the 1980s it stepped in during the foreign debt crisis, the Continental Illinois bankruptcy crisis, and the financial market crash. In each of these events the Fed (believed, as it was, to be following a monetarist policy) provided liquidity and was compelled to validate to some extent many risky financial practices.
Though rarely stressed by commentators, this feature of bankers was emphasized by Schumpeter. He noted that ‘financial institutions and practices enter our circle of problems in three ways : they are “auxiliary and conditioning” ; banking may be the object of entrepreneurial activity, that is to say, the introduction of new banking practices may constitute enterprise ; and bankers (or other “financiers”) may use the means at their command in order to embark upon commercial and industrial enterprise themselves (for example John Law).’ [Schumpeter, 1947 : 153] See also Festre and Nasica .
Government bond houses are private financial institutions that invest their funds in government debt (Treasury bills, longer-term government debt). They finance their position partly by their own resources, and partly through borrowing from commercial banks, the Federal Reserve, and nonfinancial firms.
Here we ignore the liability of banks and firms whose potential changes are not central to the current reasoning.
As Minsky explains, non-financial firms became the main source of finance for government bond houses in mid 1956.
Innovations took different forms : securitization, CDO, CDS, prime brokerage operations while banks’ leverage strongly increased during this period.