The Post Keynesian alternative to inflation targeting

Asensio, Angel; Hayes, Mark (M.G.) Article The Post Keynesian alternative to inflation targeting Intervention. European Journal of Economics and Economic Policies Provided in Cooperation with: Edward Elgar Publishing Suggested Citation: Asensio, Angel; Hayes, Mark (M.G.) (2009) : The Post Keynesian alternative to inflation targeting, Intervention. European Journal of Economics and Economic Policies, ISSN 2195-3376, Metropolis-Verlag, Marburg, Vol. 06, Iss. 1, pp. 65-79, https://doi.org/10.4337/ejeep.2009.01.08 This Version is available at: https://hdl.handle.net/10419/277152 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Th e Post Keynesian alternative to infl ation targeting Angel Asensio*,*** and Mark Hayes**,*** While mainstream policies may be beyond improvement in the enchanted ›optimizable‹ world, Post Keynesians have to manage without a magic wand in our uncertain world. We discuss the alternative policies proposed in the recent Post Keynesian literature and argue that control of interest rates is too imperfect for such policies to be feasible in general, although they provide useful guidelines and may be successful in favourable circumstances. Consequently, the question of credibility is irrelevant, if this means whether policy-makers will honour their commitment to an unfeasible ideal target. Th e right question is whether policy is convincing enough to make the conventional state of expectation (and the related interest rate) consistent with full employment. It is all a matter of confi dence. Th e basic principles involved in such an approach to monetary policy are discussed. JEL classifi cations: E12, E52 Keywords: interest rate rule, convention, uncertainty, monetary policy * CEPN, University of Paris 13. ** Homerton College, Cambridge. *** Th is paper was prepared for the workshop ›Infl ation targeting: Is there a credible alternative?‹, organised by the Post Keynesian Economics Study Group at Balliol College, Oxford, April 2008. Th e authors are grateful to Vicky Chick, Giuseppe Fontana, Malcolm Sawyer, Geoff Tily and Eckhard Hein for helpful comments. Correspondence Address: Prof. Angel Asensio, CEPN, U.F.R. sciences économiques, Université Paris 13, 99 avenue J.-B. Clément, 93 430 Villetaneuse, France, e-mail: [email protected]. Received 08 September 2008, accepted 16 February 2009 © INTERVENTION 6 (1), 2009, 65 – 79 66 Intervention. European Journal of Economics and Economic Policies »We must remind ourselves that there may be several slips between the cup and the lip.« J.M. Keynes, Th e General Th eory 1. Introduction ›Old Keynesian‹ economic policy recipes have been discarded as no longer credible because they were based upon the degenerate ›hydraulic Keynesian‹ machine, in which the achievement of full employment was an elementary matter of shifting the IS and/or LM curve(s) appropriately. Th e consistent mainstream response has been to develop the idea that rational agents would not make systematic errors in forming their expectations within such a simple economic machine. And here we are: ›infl ation targeting‹ (let us call it ›Non Infl ationary Stabilizing Policy‹) has become the optimal policy response to stochastic disturbances of dynamically stable and therefore predictable and optimizable systems, characterised by the new standard form of modelling, the dynamic stochastic general equilibrium model (DSGE, see Benassy 2007 for a recent stylised version). Echoing the reassessment of monetary and fi scal policy by Arestis and Sawyer (2003a, 2003b and 2003c), a process of collective refl ection has recently been seeking to conceive a Post Keynesian alternative to mainstream economic policy (Fontana/Palacio-Vera 2007, Setterfi eld 2007a, Setterfi eld/Lima 2008, Atesoglu 2007, Palley 2006). Some authors suggest making infl ation targeting more countercyclical, so as to have stronger real eff ects over the cycle and growth path, while others argue for a policy aimed at maintaining the interest rate at a low level. More ambitious proposals aim at designing an integrated monetaryfi scal policy mix (Arestis/Sawyer 2003b, Câmara Neto/Vernengo 2004, Setterfi eld 2007b), sometimes including income policy (Hein/Stockhammer 2007). For example, Arestis and Sawyer (2003b) suggest a ›fi scal Taylor rule‹ so as to compensate for the weakness of monetary policy (see also Setterfi eld 2007b). Although they contain stimulating ideas and provide useful guidelines, these contributions overlook the fact that the central bank’s control over interest rates is very imperfect, because of the shifting nature of liquidity preference and the demand for money, and because of the undesirable consequences that might result from interest rates adjustments. Th ere is some residual ›hydraulic Keynesianism‹ in the assumption that the central bank can freely set the rate of interest at the ideal level. If it was possible for economic policies to ensure full employment and price stability by means of a set of simple – or even sophisticated – rules, any Post Keynesian policy mix would, at best, do as well as the mainstream’s optimal one. Th e mainstream will always reign supreme in their enchanted ›optimizable‹ world. Post Keynesians must resign themselves to managing without a magic wand in the uncertain and imperfectly malleable real world. Section 2 considers fi rst the methodological roots of the Post Keynesian critique of mainstream policy recommendations and the need for an alternative. Section 3 discusses the alternative interest rules mentioned above, along with the problem of feasibility. Th is sec- Asensio/Hayes: Th e Post Keynesian alternative to infl ation targeting 67 tion also outlines some general principles for improving the eff ectiveness of the Post Keynesian macroeconomic policies. Section 4 concludes that a Post Keynesian ontology has profound implications for the kind of policy that is likely to succeed. Th e mainstream has once again fallen into a ›hydraulic‹ trap. 2. Economic policy as a ›magic wand‹ Th e mainstream view, expressed in macroeconomic terms, is that insuffi cient aggregate demand in the goods market or, equivalently, an excess of planned saving, is a state of disequilibrium, for competitive forces would trigger a decrease in the rate of interest which simultaneously would clear both the market for goods and the market for bonds. In the monetary version of the theory, where the money market is included, the real balance or ›Pigou eff ect‹ and the so-called ›Keynes eff ect‹ contribute to the support of aggregate demand as well. Keynes by contrast stated that, in the face of uncertainty, the tendency of a fall in aggregate demand to reduce interest rates may encounter various obstacles. First, it depends on the behaviour of the banking system: if the money supply decreases along with the demand for money (as in the endogenous money literature), the rate of interest will remain unchanged. Secondly, it may be that the depressive forces harm the state of confi dence so that people try to increase the share of liquid assets in their portfolio (this shift in liquidity-preference would limit or inhibit both the Keynes and Pigou eff ects, for money demand in this case does not fall as much, which limits the Keynes eff ect, and reduces the excess of real balances, compared to money demand, which limits the Pigou eff ect). Furthermore, the worsening business climate could deter investment projects despite any reduction in the interest rate. Th us, without even considering possible destabilizing forces (e.g. the effects of changes in money-wages pointed out in Keynes 1936: Chapter 19, or the debt-defl ation eff ect in Fisher 1933), it appears that, in the presence of uncertainty, the stabilizing forces themselves may fail. According to the mainstream, as competitive forces are assumed to drive the system to a ›natural‹ anchor, macroeconomic policy can at best help to stabilize the economy when rigidities delay the adjustment process. In such a context, automatic monetary and fi scal rules can be formulated, since they aim merely to off set deviations from the anchor (the ›natural‹ value). As such governance principles work symbiotically with the mainstream approach (Dixit/Lambertini 2003), they stabilize the macroeconomic system perfectly. Th e same rules, however, have severe drawbacks if they are implemented in a Keynesian model of the economy (Asensio 2006, 2007a and 2007b, Atesoglu/Smithin 2006, Palley 2007, Sawyer 2007, Setterfi eld/Lima 2008). As Asensio pointed out, in the absence of a spontaneous return towards full employment, a depressed level of unemployment becomes the macroeconomic policy target as soon as it comes to be considered the ›natural‹ rate, with the result that the policy mix ›symbiotically‹ anchors the system away from full employment. Th e persistence of a high ›natural‹ rate of unemployment in that case results from the policy, while it is held by the ›New Consensus‹ macroeconomics to be the result of real wage rigidity. Th is line of 68 Intervention. European Journal of Economics and Economic Policies argument suggests a kind of unemployment trap, which the mainstream calls hysteresis1: when the stabilization of a negative shock works only partially, unemployment increases, and policy-makers then take the actual unemployment rate to be the new ›natural‹ rate. Similar drawbacks may arise with respect to tensions over income distribution. Infl ation depends on factors aff ecting distribution (the mark-up, wages pressure relative to productivity gains, or taxes on profi ts2). Th ese factors may aff ect the unemployment rate indirectly through the monetary policy reaction they trigger. Whatever may be the proximate causes of infl ationary pressures, the central bank can always restrict the actual rate of infl ation by increasing the interest rate and the level of unemployment in such a way that the pressures fade. Indeed, higher interest rates increase unemployment and reduce the workers’ ability to negotiate money-wage increases in proportion to the increase in the price index, therefore reducing infl ationary pressures. Higher interest rates and lower economic activity could temper other sources of cost-push infl ation as well (there is little doubt that the longterm interest rate can always be increased, the problem is in reducing it). Infl ation is always a monetary phenomenon in the trivial sense that it means higher money prices of goods and services, but while mainstream economics blames irresponsible or lax fi scal and monetary policies, the Post Keynesian approach emphasises the dilemma stemming from the distributive tensions: to preserve the value of money and assume higher unemployment, or to preserve employment and let infl ation develop. Th e mainstream holds moreover that reducing monetary infl ation has no permanent cost in terms of unemployment, whereas Post Keynesians claim that it does, insofar as persistent distributive tensions induce the monetary authorities to adopt an ›incomes policy of fear‹ (Davidson 2006).3 Th is argument suggests that New Consensus macroeconomic policy is inadequate, or worse, within a Keynesian understanding of the world; the mainstream attempt to wave a magic wand leads to erroneous targets and the misuse of policy instruments. Th e Post Keynesian perspective demands an alternative approach to policy. 3. Getting rid of the wand 3.1 Activist and ›parking it‹ monetary rules Two kinds of Post Keynesian alternatives to infl ation targeting have been put forward recently (Rochon/Setterfi eld 2007a): ›parking it‹ rules and ›activist‹ rules. Th e advocates of 1 On hysteresis, see the Minisymposium in the Journal of Post Keynesian Economics 15(3), 1993. 2 Taxes paid by fi rms, given the mark up and unit labour cost, reduce distributable profi ts or are shifted onto consumers and workers. Taxes paid by workers also may induce wage pressures aiming at preserving the purchasing power of the money-wage. In an open economy, the prices of oil and imported intermediate goods should also be taken into account as an example of international distributive confl ict. Notice that even the use of seignorage in an infl ationary environment as an alternative to taxation is an example of distributive confl ict. 3 See Palley (1996 and 2001) for an empirical discussion. Asensio/Hayes: Th e Post Keynesian alternative to infl ation targeting 69 activist rules consider that authorities should adjust the interest rate more actively than recommended in the mainstream, so as to take advantage of the real eff ects of monetary policy. ›Parking it‹ rules on the other hand are a response both to the idea that »infl ation is fi rst and foremost the result of confl ict over the distribution of income« (Rochon/Setterfi eld 2007b: 8), so that monetary policy is not the appropriate tool to fi ght infl ation, and to the idea that the wisdom of active monetary policy is questionable, owing to the many uncertainties in the transmission mechanism (Wray 2007, Bateman 2003). ›Parking it‹ rules are therefore to be understood as full policy-mix proposals (Hein/Stockhammer 2007) based on the following principles: – fi scal policy works countercyclically; – incomes policy aims at fi ghting infl ation; – monetary policy parks the interest rate with an explicit distributional objective. Th e philosophy of the Post Keynesian rules therefore diff ers substantially from the mainstream rule philosophy. Furthermore, because of the attention paid to fundamental uncertainty, Post Keynesian rules have to be interpreted as policies intended to achieve desirable goals, rather than strict rules. 3.1.1 ›Activist‹ rules According to the ›activist‹ rules proponents, the mainstream infl ation targeting rule leads to fl awed interest rate adjustment because it overlooks the real eff ects of monetary policy (see also Sawyer 2007). Th e innovative proposal for a ›fl exible opportunistic approach‹ developed by Fontana and Palacio-Vera (2003 and 2007) seeks to encourage the growth rate of output and employment, besides stabilizing output in the short run and achieving price stability in the long run. Th e standard opportunistic approach (Orphanides/Wilcox 1996) states that, in order to be able to take advantage of a possible exogenous adjustment of the infl ation rate towards the long-run target, the central bank should not adjust interest rates as long as the actual infl ation rate remains within some predetermined upper and lower limits around the target. By contrast, the fl exible opportunistic approach puts forward that the possible long-run eff ects of monetary policy on potential output argue in favour of a policy loosening when the actual rate of infl ation is below the target but above the predetermined lower bound. In a similar way, if the actual rate of infl ation is above the target but below the predetermined upper bound, the fl exible opportunistic approach states that monetary policy should moderately decrease the interest rate so as to take advantage of a possible increase in potential output, which would subsequently off set possible infl ationary pressure. Palley (2007) considers other real eff ects of monetary policy, besides the eff ects on potential output and growth, and argues that infl ation targeting »biases decisions toward low infl ation by obscuring the fact that policy also aff ects unemployment, real wages, and growth« (Palley 2007: 61). Taking these real eff ects into account, Palley calls for setting the rate of interest so as to balance the possible advantages that may follow from accepting an 70 Intervention. European Journal of Economics and Economic Policies increased infl ation rate with the advantage of low infl ation. In his model, when unemployment is suffi ciently high, the only cost to monetary stimulus is increased infl ation. Th e authorities in that case may reduce the rate of interest so that unemployment decreases towards Palley’s MURI (minimum unemployment rate of infl ation, beyond which further infl ation increases would have a counterproductive eff ect on employment)4. However, a trade-off between higher wages and lower unemployment, versus higher infl ation and lower growth, may arise as unemployment falls. Th is is because the economy, under certain conditions,5 becomes profi t-led (as the profi t rate is negatively aff ected by unemployment). Th is interesting feature of Palley’s model suggests that the rate of interest should not be adjusted according to a rigid predetermined rule, for the economy may become wage-led or profi t-led depending on the level of unemployment, which aff ects the terms of the tradeoff facing monetary policy. As they seek to take advantage of the potential real eff ects of monetary policy, Post Keynesian ›activist‹ rules unquestionably improve upon the mainstream’s ›infl ation targeting‹. Th ere is however an important diffi culty related to the conventional character of long-term interest rates which make their feasibility uncertain. Th e feasibility of these rules indeed rests on the questionable assumption that the long-term interest rate (the real rate in the case of the fl exible opportunistic approach)6 can be adjusted so as to reach the ideal target. Th e point is that the shifting nature of the state of confi dence has serious implications for the ability of monetary policy to control the long-term interest rate through operating on the overnight rate. For example, Lavoie (1999: 2), who suggests that ›monetary authorities have the ultimate say on the convention‹, pointed out that the spreads between the longterm rates and the overnight rate vary according to the liquidity preference of the commercial banks and the participants in the fi nancial markets: »As Smithin (1996: 93) puts it, a role for Keynesian liquidity preference can be retained in this scenario, in that liquidity preference considerations may well periodically insert a wedge between those rates of interest which are more or less directly under the central bank control and rates elsewhere« (Lavoie 1999: 2). Such a diffi culty may arise especially in the case of interest rate reductions. When the monetary base is increased as a result of lower short-term interest rates, lower long-term bank rates in principle boost the demand for credit. But if, at the same time, liquidity preference increases, banks may be able to sell more credit without needing to reduce their interest rates, for non-bank loan (bond) rates tend in this case to rise in order to compensate for the increasing liquidity preference. 4 Th is is related to the backward-bending Phillips curve of the model: as real wage resistance increases as infl ation increases, the ›grease eff ect‹ on employment, which is associated with the negative eff ect of infl ation on real wages, erodes as infl ation increases. See Palley (2007) for details. 5 If real wages do not rise too steeply as unemployment decreases, authorities may reach the MURI without encountering a growth trade-off ; otherwise, there may be a trade-off between growth and pushing the unemployment rate to the MURI (Palley 2007: 74). 6 A specifi c problem raised by the control over the real rate is discussed below. Asensio/Hayes: Th e Post Keynesian alternative to infl ation targeting 71 Of course central banks may also intervene directly in fi nancial markets with the aim of infl uencing long-term interest rates. But even if »the monetary authority were prepared to deal both ways on specifi ed terms in debts of all maturities, and even more so if it were prepared to deal in debts of varying degree of risk«, there would be »limitations on the ability of the monetary authority to establish any given complex of rates of interest for debts of diff erent terms and risk« (Keynes 1936: 205 and 207).7 Some of these limitations (see Keynes 1936: 207 – 208 for a detailed discussion) can be considered purely theoretical, insofar as they would only arise in extreme circumstances (virtually absolute liquidity preference when rates are considered too low; breakdown of stability in the rate of interest – owing to a fl ight from the currency or other fi nancial crisis); but others apply in normal circumstances (the intermediate cost of bringing the borrower and the lender together, the allowance for risk required by the lender, including liquidity risk). 3.1.2 ›Parking it‹ rules Feasibility is also an issue for the ›Parking it‹ rules but in a way that diff ers according to the type of the rule considered. ›Parking it‹ rules divide into short-term nominal rate and longterm real rate rules. Let us fi rst discuss the ›fair rate‹ rule, understood in the spirit of Pasinetti (1981) (see also Lavoie 1999), and the ›low real rate‹ proposed in Smithin (2007) (also Atesoglu/Smithin 2006, Hein/Stockhammer 2007). Both are real rate based rules and share the normative purpose of providing economic policy with an ›explicit distributional objective‹. Th e ›fair rate‹ rule consists in equalizing the real interest rate with the productivity growth rate, so that the rentiers’ share in the national income is constant. Smithin’s rule, on the other hand, holds to setting the real interest rate at a low level (a cheap money policy). Th e distribution eff ect here diff ers essentially because »it does not […] guarantee a share for existing wealth holders (as opposed to entrepreneurs or workers) in current productivity increases, as would the notion of the ›fair‹ interest rate […]. Th is omission might be justifi ed on the grounds that it is the latter, rather than the former, who are actually responsible for the productivity increases« (Smithin 2007: 116). Both rules aim at setting the real rate of interest at a target level, and both are therefore subject to the problem of feasibility. Th ere is, however, another diffi culty in this case, for the real rate of interest is not a single variable (as it would be in a barter or ›neutral‹ economy); it is the diff erence between the price of liquidity (the long-term nominal rate) and the expected infl ation rate. Hence how could a central bank go about achieving these twin objectives with only one instrument (the overnight rate)? 7 On monetary policy and debt management, see also Tily (2006). 72 Intervention. European Journal of Economics and Economic Policies Post Keynesians reject the belief in a natural real rate of interest, and it is therefore often assumed that the determination of this variable is left to the central bank, as suggested in Smithin (1994: 172 – 173) and Lavoie (1996: 277 and 1999: 2). However the discussion of how the central bank could manage to control the long-term real interest rate is not totally convincing, for it rests on the idea that, provided the central bank is able to adjust the long-term nominal rate, it can easily adjust the nominal rate to take account of the expected infl ation rate. But this requires that the expected rate of infl ation is independent of the nominal rate of interest, which is not self-evidently true. Even if authorities intend to anchor expectations by committing themselves to an infl ation target, the offi cial target would not anchor expectations if agents thought the nominal interest rate was inconsistent with the target. Th us, either the central bank anchors the expected infl ation but cannot set the nominal rate independently, or the central bank sets the interest rate but cannot anchor the expected infl ation rate independently. In either case, the central bank can hardly be said to control the real interest rate. Short-term nominal rate rules are not subject to such a limitation, for the short-term nominal interest rate is very closely related to the central bank’s overnight rate (outside periods of crisis). Th e ›Kansas City‹ rule calls for the ›euthanasia of the rentier‹ by means of a zero short-term nominal rate (Wray 2007).8 Let us fi rst consider Keynes’s views on the issue. ›Th e social philosophy towards which the General Th eory might lead‹ (Keynes 1936: 374 – 377) focuses on our ability to manage the rate of interest so as to raise the inducement to invest to the level where, given the aggregate propensity to consume (including the State), there is full employment. Insofar as the accumulation of capital decreases the marginal effi - ciency of capital, a decrease in the interest rate will be necessary in the long run. Th at is the essence of Keynes’s prediction of the euthanasia of the rentier. According to his argument, the ideal policy is not to maintain the interest rate at a low fi xed level unconditionally; it is to adjust the interest rate to the level that ensures full employment, given the marginal effi ciency of capital and the aggregate propensity to consume. As these variables may change in response to changes in the rate (and the state) of capital accumulation, in productivity growth or in the government’s propensity to consume, among other factors, it would be imprudent to adopt a rule that could not take account of such developments. Th e ›Kansas City‹ version of the short-term nominal rate ›parking it‹ rule would work as well as is possible against unemployment, but in the face of distributive tensions aimed for example at increasing the share of profi ts, or wages, or government revenues, it would allow for monetary accommodation of the resulting infl ationary pressures. Hein and Stockhammer (2007: 17) suggest that low real interest rates rather reduce infl ationary pressures and that it is, on the contrary, high interest rates that fuel infl ation, based on a cost push argument. Although such a mechanism must of course be considered, notice that there are many cost push channels which could feed distribution confl ict even when interest rates are low, and that in this case, the monetary accommodation induced by the ›parking it‹ rule 8 Câmara Neto and Vernengo (2004) also advocate a low interest rate policy so as to make it easier for the government to implement a sound countercyclical fi scal policy. Asensio/Hayes: Th e Post Keynesian alternative to infl ation targeting 79 Sawyer, M. (2007): Seeking to reformulate macroeconomic policies, 3rd bi-annual conference of the CEMF: Post Keynesian Principles of Economic Policy, Dijon (France): University of Burgundy, December 2007. Setterfi eld, M. (2007a): Is infl ation targeting inicial to employment?, University of Vermont conference paper, URL: http://www.uvm.edu/~econ/keynesian_conf/Setterfi eldI_ paper.pdf. Setterfi eld, M. (2007b): Is there a stabilizing role for fi scal policy in the New Consensus?, in: Review of Political Economy, 19, 405 – 418. Setterfi eld, M., Lima, G.T. (2008): Infl ation targeting and macroeconomic stability in a Post Keynesian economy, in: Journal of Post Keynesian Economics, 30, 435 – 461. Smithin, J. (1994), Controversies in Monetary Economics, Aldershot: Edward Elgar. Smithin, J. (1996): Macroeconomic Policy and the Future of Capitalism: Th e Revenge of the Rentiers and the Th reat to Prosperity, Aldershot: Edward Elgar. Smithin, J. (2007): A real interest rate rule for monetary policy?, in: Journal of Post Keynesian Economics, 30, 101 – 118. Tily, G. (2006): Keynes’s theory of liquidity preference and his debt management and monetary policies, in: Cambridge Journal of Economics, 30, 657 – 670. Wray, R. (2007): A Post Keynesian view of central bank independence, policy targets, and the rules versus discretion debate, in: Journal of Post Keynesian Economics, 30, 119 – 141.