Fiscal policy: post- or New Keynesian?

Gechert, Sebastian Article Fiscal policy: post- or New Keynesian? European Journal of Economics and Economic Policies: Intervention (EJEEP) Provided in Cooperation with: Edward Elgar Publishing Suggested Citation: Gechert, Sebastian (2023) : Fiscal policy: post- or New Keynesian?, European Journal of Economics and Economic Policies: Intervention (EJEEP), ISSN 2052-7772, Edward Elgar Publishing, Cheltenham, Vol. 20, Iss. 2, pp. 338-355, https://doi.org/10.4337/ejeep.2023.0120 This Version is available at: https://hdl.handle.net/10419/284335 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/ Fiscal policy: post- or New Keynesian?* Sebastian Gechert** Chemnitz University of Technology, Germany and FMM Fellow This paper asks whether the post-Keynesian or New Keynesian paradigm provides a more realistic description of the effects of fiscal policy on output and consumption. I establish some macro and microeconomic stylized facts on fiscal multipliers and marginal propensities to consume based on the extensive empirical literature. I compare them to basic and extended modelling approaches within the two traditions. The picture is mixed and points to shortcomings in both paradigms. I outline the elements they could learn from each other as well as from some other strands in the literature to create a coherent and convincing Keynesian macro-micro framework of fiscal policy modelling. Keywords: macroeconomics, fiscal policy, history of economic thought, multiplier, marginal propensity to consume JEL codes: D10, E20, E60, H30 1 INTRODUCTION How and to what extent fiscal policy affects output is a classic and recurring question in macroeconomics. The fiscal multiplier –the measure of how much GDP is created by an additional dollar spent by the government –has always been a central figure to policymakers. This is particularly so during severe downturns, like the Great Recession and the subsequent euro area crisis, as well as the COVID-19 crisis. When it comes to explaining the mechanisms of the multiplier effect, an important element is the marginal propensity to consume (MPC) – the measure of how many extra dollars a household spends within a certain time horizon (for example, one year) after receiving an extra dollar of unexpected income 1 –related to an increase in government spending, transfers, or tax cuts. The multiplier and the MPC have occupied the minds of (among others) Quesnay, Keynes, Kahn, Haavelmo, Samuelson, Friedman, Moore as well as contemporary scholars of various schools of thought, who have described different channels at work. 2 Today, the most comprehensive (and at the same * I would like to thank Jan Behringer, Alejandro González, Jochen Hartwig, Eckhard Hein, Ekaterina Jürgens, Marc Lavoie, Franz Prante and participants of the 26 th FMM conference in Berlin for suggestions and helpful discussions. Of course, they shall not be held responsible for any oversimplification or incomplete reading of the literature. All remaining errors are mine. ** Email: [email protected]. 1. Note that more precisely I should speak of the marginal propensity to spend, since consumption of durable goods and the respective flow of welfare is spread over a longer time span after the moment of purchase. The latter is decisive for my focus on the macroeconomic effects of fiscal policy. Nevertheless, the term MPC is so common that I stick to it. 2. Hegeland (1966) provides a detailed early history of the multiplier, while Gechert (2017) briefly discusses the more recent history. Received 24 March 2023, accepted 31 May 2023 European Journal of Economics and Economic Policies: Intervention, Vol. 20 No. 2, 2023, pp. 338–355 First published online: August 2023; doi: 10.4337/ejeep.2023.0120 © 2023 The Author Journal compilation © 2023 Edward Elgar Publishing Ltd The Lypiatts, 15 Lansdown Road, Cheltenham, Glos GL50 2JA, UK and The William Pratt House, 9 Dewey Court, Northampton MA 01060-3815, USA This is an open access work Invited Article time distinct) macroeconomic frameworks that concern the workings of fiscal policies are those of the New Keynesian (NK henceforth) and post-Keynesian (PK henceforth) paradigms. The main distinction between these two approaches is that PK models focus on macroeconomic relations and are primarily driven by aggregate demand both in the short and long run, while NK models are derived from neoclassical microeconomic utility optimization principles and feature a strong role of aggregate supply, complemented by short-run aggregate demand constraints. This paper asks, which of these paradigms provides a more realistic description of the effects of fiscal policy on output and consumption. To this end, I describe the basic channels at work in respective basic models –a simple PK supermultiplier model as discussed, for example, in Palley (2019) and an analytic description of a NK DSGE model as in Woodford (2011). I then compare them to the empirical stylized facts, both from the macroeconomic and microeconomic literature. The model selection is necessarily arbitrary, given the rich literature in both paradigms, as outlined for example in Hein (2017) and Reis (2018). However, the chosen models are useful because they focus on fiscal policy and represent the PK and NK approaches as simple as possible. Since these pedagogic models lack some realistic features, I refer to extensions that have become a quasi-standard in the respective literatures, if necessary. By this, I intend to strike a balance between a simple analytical description of the paradigmatic differences, an exposition of developments in the NK and PK approaches to fiscal policy, and a clear judgment on their empirical performance. It turns out that both the PK and the NK approaches can explain only some of the multiplier and MPC facts and that the overlap is astonishingly narrow. NK models underestimate the size of the government spending multiplier and the average MPC during times of normal capacity utilization. They do not provide convincing mechanisms for the observed relation between the MPC and household income. Only with far-reaching additional assumptions, NK models can reproduce the observed net-positive multiplier effect of redistributional tax and transfer policies –based on a somewhat extreme bimodal distribution of households into savers and spenders. NK models that feature constraints to monetary policy near zero interest rates can match the strong rise of spending (as compared to tax) multipliers in crisis times, but the latter effect is very sensitive to the persistence of the fiscal shock and rests on questionably large intertemporal substitution effects of household consumption (McKay et al. 2017). Recent NK model extensions can cope with positive effects of fiscal stimulus on potential output during severe slumps, but the mechanisms are still subject to debate. The PK model (including standard extensions) on the other hand, does a good job in matching the sizeable multiplier and MPC in normal times (though ignoring the observed increase in MPCs over time). Several PK extensions cannot convincingly explain the strong non-linearity of spending multipliers (as opposed to tax multipliers) in crisis times. Likewise, hysteresis effects that give way to a strong long-term effect of fiscal policy are a standard feature of the PK framework. However, the data suggest that such hysteresis effects seem to matter only during strong downturns. On the other hand, PK models incorporate a negative relation between the MPC and the income level (likewise based on an extreme bimodal distribution of MPCs out of wage and profit incomes) and thus can explain the positive GDP effects of fiscal redistribution. However, they lack a relation between wealth, debt pressure and the MPC that fits the data. In the conclusion, I outline the elements of a fiscal policy model that draws on the saving and consumption motives in Chapter 9 of the General Theory (Keynes 1936) with extensions from the PK and NK frameworks. This model would feature a (muted) intertemporal substitution channel, a precautionary saving motive, basic needs consumption, liquidity constraints, a constrained monetary policy in downturns and uncertainty about future Fiscal policy: post- or New Keynesian? 339 © 2023 The Author Journal compilation © 2023 Edward Elgar Publishing Ltd income in an economy with an unequal distribution of income. Such a Keynesian fiscal policy model should meet all the stylized facts at the macro and micro levels. This paper speaks to various strands of the literature. It draws insights from the rich empirical macroeconomic literature on the fiscal multiplier (Gechert 2015; Gechert/Rannenberg 2018; Hagedorn et al. 2019; Hall 2009; Hebous 2011; Ramey 2016, 2019; Spilimbergo et al. 2009) and the large microeconomic literature on the size of MPCs (Jappelli/Pistaferri 2010, 2014; Parker/Souleles 2019; Sahm et al. 2010). It is related to several studies that engage in a paradigmatic comparison of macroeconomic models, like Prante et al. (2020), who compare several policy implications of PK versus NK models; Hein (2014), who provides a paradigmatic introduction to different growth theories; or Lavoie (2006) who juxtaposes several heterodox approaches. It speaks to discussions about the development and interrelation of orthodox and heterodox macroeconomics (Colander 2000; Colander et al. 2008; Dullien 2017; Dobusch/Kapeller 2012; Lavoie 2012, 2018; Romer 2016). It relates to recent approaches towards a pragmatic reconciliation of PK and NK models (González 2022; Schoder 2015, 2017) and the endeavor to make NK models more realistic (Auclert et al. 2018; Bilbiie 2020; Hagedorn et al. 2019; Kaplan et al. 2018; Korinek 2018; McKay et al. 2016; Rannenberg 2021). The paper is structured as follows: Section 2 describes the main features of NK and PK workhorse models with a focus on their fiscal policy implications. Section 3 then compares these implications with the macroeconomic stylized facts about the fiscal multiplier and the microeconomic facts about MPCs. The final section then concludes by pointing to major shortcomings in both approaches and outlining the elements of a model that could bridge the gap between the two schools of thought. 2 FEATURES OF NEW AND POST-KEYNESIAN WORKHORSE MODELS This section focuses on the main characteristics of typical PK and NK workhorse models related to fiscal policy, avoiding a full-scale and formal description. The NK model I describe here is based on Woodford (2011), as it focuses on fiscal policy. The PK model refers to the neo-Kaleckian variety of the supermultiplier model, for which Palley (2019) provides a workhorse example. Both articles provide a relatively simple comparative-static discussion of the structure of these models, which eases exposition. However, this comes at the cost of complicating a comparison to the later presented stylized facts that rest on dynamic estimations of the multiplier and the MPC, and thus require an explicit time frame. Since the focus is on qualitative conclusions, the benefits outweigh the costs. For the sake of simplicity, I consider a closed-economy setting in this section but will refer to open-economy considerations, if necessary. 2.1 The post-Keynesian model The PK supermultiplier model is derived from the logic of the Keynesian cross by which output and income are determined by aggregate demand. In the PK approach, this principle of effective demand prevails both in the short and long run. Aggregate demand consists of an autonomous demand component (including government spending, baseline consumption, and independent investment), income-dependent consumption and endogenous investment. The endogenous part of investment positively depends on capacity utilization (of the capital stock) in the short run and firms’expectations of demand growth in the long run. These long-run expectations are assumed to equal the growth rate of 340 European Journal of Economics and Economic Policies: Intervention, Vol. 20 No. 2 © 2023 The Author Journal compilation © 2023 Edward Elgar Publishing Ltd autonomous demand, which determines the steady-state growth rate. 3 In the short run, the positive dependence of investment on capacity utilization gives rise to an accelerator mechanism and a crowding-in of investment by additional government spending. Saving adjusts to overall investment via changes in income through the Keynesian multiplier process such that there is never a shortage of saving for investment and the paradox of thrift applies: a higher propensity to save reduces aggregate demand and income. Consequently, it does not lead to higher but potentially lower saving. 4 Income is distributed in the form of wages and profits, whose relative shares are influenced by power structures and determine demand and growth. They influence consumption demand, since the MPCs out of profits and wages are assumed to differ. These MPCs are structural parameters of the model, that is, they are fixed. The MPC out of wages is usually assumed to be close to or equal to one, while the MPC out of profits is smaller and may even be close to zero. A redistribution from profits to wages usually increases the average MPC and consumption demand. The average MPC is a main determinant of the multiplier effect as in the Keynesian cross. It is always positive such that additional government spending crowds in private consumption expenditures. The supply side comes in via a Leontief aggregate production function in labor and capital. Labor supply is usually abundant while the capital stock is underutilized and grows with aggregate demand such that supply constraints normally do not bite. Price adjustments are second order to quantity adjustments and may only become relevant in a boom or depression. That is, price crowding out of fiscal policy is typically irrelevant in normal times. Due to the crowding-in of both private investment and consumption and without retarding factors, the fiscal multiplier in the plain vanilla PK model is in the range of three to four. There are three main extensions that reduce this effect. First, in an open-economy framework, part of aggregate demand leaks out via imports and thus reduces the domestic aggregate demand effect. Second, a tax on income or aggregate demand slows down the propagation of initial government spending in the private sector. At the same time, this allows for the assessment of a tax multiplier. The tax multiplier is lower than the spending multiplier, as it follows the same income-to-spending propagation but misses the initial direct demand effect of government spending. Third, private investment spending could be sensitive to a pro-cyclical interest rate, which might be set by the monetary authority leaning against the wind, or private banks/investors demanding a pro-cyclical mark-up on the monetary policy rate. This boils down to a downward-sloping IS in conjunction with an upward-sloping (‘structuralist’, Palley 2013) or flat upward-shifting (‘horizontalist’, Lavoie 1996) LM curve. However, the interest channel is weak if the interest elasticity of investment is low, which is a standard PK conjecture. 2.2 The New Keynesian model The simple analytic NK model as described in Woodford (2011) starts out from a neoclassical general equilibrium model with labor being the only production factor, assuming away capital and investment for simplicity. 5 The economy produces a single 3. Palley (2019) describes this as the concept of rational expectations within a PK setting. 4. The combined workings of the multiplier and accelerator may create an instability. Stability is usually ensured by assuming that investment reacts less to changes in the capacity utilization than saving (Foley et al. 2019). 5. Alternatively, the standard assumption in a model with capital would be to have a Cobb–Douglas production function with unitary elasticity of substitution between labor and capital, which is however questioned by the data (Gechert et al. 2022). Fiscal policy: post- or New Keynesian? 341 © 2023 The Author Journal compilation © 2023 Edward Elgar Publishing Ltd output good which can be used for private or government consumption (the demand components). Private consumption is paid out of accumulated lifetime labor income, subject to a lump-sum tax by the government. Government consumption is paid for by these lump-sum taxes. It could in principle also be financed by debt issuance, but the transversality condition, a crucial element to make the intertemporal structure of the model solvable, requires that the government redeems the debt by tax revenues eventually. The model features perfect competition in goods and labor markets (that is, no extra profits, no price setting power, labor being paid its technically determined falling marginal product) and a perfect financial market (allowing borrowing or lending over time at a unique interest rate). A representative household engages in intertemporal optimization of utility as determined by private consumption and leisure (time endowment minus work time). The government expenditures provide no utility to the household. Labor supply depends on the net wage and the relative utilities of consumption and leisure. Consumption demand in a period is subject to the very same relative utilities, an intertemporal budget constraint, an intertemporal substitutability of consumption, and a discount rate for future consumption (the Euler equation). Today’s consumption thus rises with a higher lifetime net labor income, a lower interest rate, a higher discount rate, and a higher relative utility of consumption over leisure. This consumption function is in line with the permanent income hypothesis (PIH) (Friedman 1957; Hall 1978). The MPC after a transitory increase in income is close to zero in the short run, as the shock adds little to permanent income and the additional consumption is spread out over the entire remaining lifetime. However, the MPC after an unexpected permanent income increase is close to one instantaneously. Due to the forward-looking skills of the household in the model, this is even the case for anticipated future income changes. Yet, these mechanisms, which arise for example after an increase in the marginal productivity of the household’s labor supply, should not be misinterpreted to apply to a government spending increase. Aggregate demand is supply-constrained by the production function and hence higher government spending crowds out private consumption. Moreover, since the additional government spending requires higher taxes at some point, the household’s lifetime wealth is reduced, forcing it to consume less and save more today (that is, Ricardian equivalence holds). However, this negative wealth effect increases the marginal utility of consumption over leisure and stimulates the household to work more today (and in the future), thus increasing aggregate supply. All this creates a supply-sided fiscal multiplier effect, which in the simplest model only depends on the relative marginal utility of consumption over leisure (increasing the multiplier) and the degree of intertemporal substitutability of private consumption (decreasing the multiplier in the short run). In such a purely neoclassical world, the (supply-sided) multiplier is positive, but necessarily smaller than one and likely closer to zero (Woodford 2011). When adding capital to the model, the positive supply effect on working hours would also increase the marginal product of capital and lead to crowding-in of investment, thus increasing the overall multiplier effect to some extent (Baxter/King 1993). How do things change in a NK extension? As a first element, monopolistic competition introduces a mark-up (or a wedge) betweenpricesandlaborcosts.Thisalonedoes not change the size of the multiplier, let alone its supply-sided logic. Augmenting the model with sticky prices, however, endogenizes the wedge. Then, additional government demand reduces the wedge, leading to higher real wages, higher production, and labor demand. 6 This allows for government spending to have a direct effect on 6. A similar effect with different channels can be achieved by sticky wages. 342 European Journal of Economics and Economic Policies: Intervention, Vol. 20 No. 2 © 2023 The Author Journal compilation © 2023 Edward Elgar Publishing Ltd output via aggregate demand. The demand-side channel, however, does not add to the supply-side effect but partly mutes it such that the fiscal multiplier need not be much higher in the NK world. With these extensions, monetary policy can influence real activity via its command over the nominal interest rate and its indirect influence on the real interest rate. Since the real interest rate has a strong influence on current and future consumption via the Euler equation, the multiplier effects of fiscal policy strongly depend on the monetary policy reaction. Woodford (2011) shows that under fairly general assumptions about the marginal utilities of consumption and leisure and the degree of price stickiness, a transitory increase in government spending has a multiplier exactly equal to one if monetary policy maintains a constant real interest rate. 7 So, there is neither crowding-in nor crowding-out of consumption. Notably, the effects do not directly hinge on the amount of slack in the economy. However, a higher capacity utilization (of the labor force) increases marginal labor costs and prices. This makes it less likely that the central bank maintains a constant real interest rate in upturns. After all, the case of the constant real interest rate is merely hypothetical. The standard case is that the central bank is able and willing to steer the business cycle with interest rate changes (for example, following some sort of Taylor Rule). Any fiscal interventions change the effort by the central bank and lead to a different real interest rate. That is, expansionary fiscal policy during a downturn implies less expansionary monetary policy and a rise in the real interest rate relative to a scenario without the fiscal intervention. This crowds out private consumption and investment. Therefore, the multiplier effect is smaller than one under standard assumptions. Depending on the reaction function of the central bank, it could even be smaller than in the purely neoclassical version above. 3 STYLIZED FACTS AND THE MODELS After having laid out the main features of the PK and NK approaches to fiscal policy, I now compare their implications with the empirical stylized facts. Three caveats are in order: first, the list of stylized facts is notoriously incomplete, but space is limited. Second, their generality can be debated. Nevertheless, they are based on a broad reading of the literature and seem to be fairly robust. Third, some of these stylized facts have no counterpart in the basic models just described. In this case, I will refer to established extensions from the literature. To qualify as an ‘established extension’,itshould be incorporated by a considerable share of the literature or by highly cited reference papers. 3.1 The macro stylized facts According to several overview studies (Gechert 2015; Gechert/Rannenberg 2018; Hall 2009; Hebous 2011; Hagedorn et al. 2019; Ramey 2016, 2019; Spilimbergo et al. 2009), I point out six central stylized facts about fiscal multipliers on the macro level (Ma-I to Ma-VI) that will be discussed one-by-one in relation to our two models. 7. The effects would revert to the neoclassical case and therefore be smaller for a persistent increase in government spending. They would be larger for a credibly announced future spending reversal, which boosts current private consumption in the NK model (Corsetti et al. 2012). Fiscal policy: post- or New Keynesian? 343 © 2023 The Author Journal compilation © 2023 Edward Elgar Publishing Ltd Ma-I: During times of normal capacity utilization, the spending multiplier is roughly one. In the baseline representative agent NK (RANK) model according to Woodford (2011), this result is an extreme case, which could only be achieved if the central bank maintains a constant real interest rate and the government spending shock is short-lived. Under normal parameterizations, the spending multiplier would rather be around 0.5, featuring partial crowding-out of consumption and investment. While crowding-out of investment can be found in a number of empirical studies (Blanchard/Perotti 2002; Barro/Redlick 2011), most studies report a crowding-in of private consumption (Blanchard/Perotti 2002; Fatás/ Mihov 2001; Galí et al. 2007; Gechert/Mentges 2018). As shown by Galí et al. (2007), to bring the standard NK model in line with this evidence, requires a substantial fraction of non-Ricardian ‘Hand-to-Mouth’(HtM) households 8 in what is now called a two-agent NK (TANK) model (Bilbiie 2020). HtM households will play a decisive role to match the NK model with several of the stylized facts discussed below. This extension shares similarities with the PK model. Still, typical calibrations of a share of HtM households of ca. 30 per cent (Gechert 2015) would leave the multiplier far below one. An additional assumption of non-competitive labor markets, where wages would be entirely controlled by unions, based on which firms chose labor demand, would be necessary to reach a multiplier of one (Galí et al. 2007). However, this would take the model away from the baseline case in several dimensions. The PK supermultiplier model in Palley (2019) would clearly transgress the boundary with a spending multiplier in the ranks of three or four, as it entails strong crowding-in effects of both consumption and investment. However, adding some realistic features to the model, like an open-economy setting with import leakage, a tax leakage by the government itself, or interest-elastic investment in conjunction with pro-cyclical interest rates will be sufficient to bring the multiplier down to one. 9 The interest channel would also be relevant to enable the PK model to mirror the observed crowding-out of investment in linear estimations. Ma-II: During times of normal capacity utilization, the tax multiplier is on average slightly smaller than the spending multiplier. While the selective literature review of Ramey (2019) comes to a different conclusion, the bulk of empirical estimates in more comprehensive studies (Gechert/Rannenberg 2018) finds that tax multipliers are on average smaller than one. Even if the baseline models in Palley (2019) and Woodford (2011) do not cover tax shocks, one can derive what the respective approaches would imply. Again, the PK model (with the above-mentioned extensions) makes it easy to meet this stylized fact. The logic of the Keynesian cross has it that the tax multiplier will always be smaller than the spending multiplier. Also, simplified NK models can produce tax multipliers smaller than spending multipliers: if the tax is modelled as lump-sum, the tax multiplier is quite small due to Ricardian Equivalence. Yet, if the tax is assumed to be distortionary, which is a standard conjecture in NK models, the tax multiplier can easily exceed the spending multiplier. 8. The NK literature has developed various forms of such HtM households under the names of ‘non-Ricardian’,‘Keynesian’,‘myopic’,‘rule-of-thumb’,or‘liquidity-constrained’households. While the underlying assumptions differ, they exhibit similar MPCs close to one. 9. While these factors could be easily calibrated to match a spending multiplier of one, in a fullyfledged model they would need to be matched to further empirical regularities about import propensities, elasticities of the tax system, and the sensitivity of investment to interest rates. This would go beyond the scope of the present paper. 344 European Journal of Economics and Economic Policies: Intervention, Vol. 20 No. 2 © 2023 The Author Journal compilation © 2023 Edward Elgar Publishing Ltd Ma-III: During times of considerable recessions and crises, the spending multiplier is much higher, more in the range of two. This is a very robust finding in the literature. The difference to normal times multipliers is qualitative and not gradual. Since the outfalls of the Financial Crisis, many NK models have been extended to incorporate situations of a deep recession, where monetary policy is constrained by a zero lower bound (ZLB) of the nominal interest rate (Christiano et al. 2011; Eggertsson 2011). In this case, the Taylor Rule would advise an interest rate in negative territory, which the central bank cannot reach. A transitory fiscal expansion would then not be met by a nominal interest rate hike and the relatively higher inflation would even reduce the real interest rate. Due to the strong sensitivity of private consumption to the expected real interest rate, such passive monetary accommodation would considerably increase current consumption. The fiscal multiplier could then easily exceed two and is highly non-linear. While this result fits well to the empirical facts, it is highly fragile, hinging on the duration of the fiscal stimulus. If the stimulus persists longer than the ZLB period, the forward-looking consumption effects are turned upside-down: the expected crowding out and reduced inflation after the return to normal times strongly dampens current consumption and can even imply strongly negative short-run multiplier effects (Woodford 2011). Within the PK model, it is hard to produce a non-linear multiplier effect. Since the models generally assume underutilized resources, there should not be a strong difference between the spending multiplier in normal times and crisis times. It requires extensions of the baseline model that render demand components endogenous to the level of capacity utilization. One possible channel for such a non-linearity of the multiplier could be a pro-cyclical marginal propensity to import in an open-economy framework. However, according to the calculations for the PK model in Charles (2016, table 5), this can still only explain around 10 per cent of the empirical multiplier difference, even for a huge recessionary shock of around 8 per cent of GDP. A second channel may arise from a pro-cyclical profit share, implying an endogenous redistribution from low-MPC capitalists to high-MPC workers during downturns. Yet again, Charles (2016) finds that the impact on the multiplier is even an order of magnitude smaller than for the import channel and thus negligible even for large changes in the functional income distribution. An amplification of the redistributional channel might occur when a reduction in the profit share also reduces wage inequality (high managerial wages that are correlated with profit incomes versus low bluecollar wages). This would increase the MPC out of wages in the ‘normal case’(Hein 2023, ch. 4.6). However, even if this would decuple the effect, it would remain small. 10 A different argument for an increased multiplier during recessions comes in the form of induced investment spending (or less crowding-out of it). This could be implemented by a kink in the LM curve, which would be flat during a recession and upward sloping or 10. Charles et al. (2015) make another argument: in the face of a recession and falling incomes, capitalists would want to keep up their consumption level by resorting to accumulated wealth and thus increase their propensity to consume (relative to their falling income). This reasoning rests on a misunderstanding of the concept of the marginal propensity to consume. If capitalists reduce the dependency of their consumption on current income, they in fact reduce their MPC, while the measured propensity to consume (consumption-to-income ratio) rises. More consumption smoothing logically implies a lower MPC which should apply to both negative and positive income shocks. When the government raises its spending during a recession, capitalist households would symmetrically react less to the additional income created. So, the fiscal multiplier effect would be smaller, not larger! The model in Charles et al. (2015) indeed imposes a higher MPC out of capitalist income and thus creates a larger multiplier during recessions, but this is inconsistent with their own reasoning of consumption smoothing. Fiscal policy: post- or New Keynesian? 345 © 2023 The Author Journal compilation © 2023 Edward Elgar Publishing Ltd channels outlined above. Some of them might not work together or some might be redundant. Also, the resulting multiplier effects are yet to be quantified. I leave such an attempt for future research. REFERENCES Aiyagari, S.R. (1994): Uninsured idiosyncratic risk and aggregate saving, in: The Quarterly Journal of Economics, 109(3), 659–684. Auclert, A., Rognlie, M., Straub, L. (2018): The intertemporal Keynesian cross, NBER Working Paper, No 25020. Anzoategui, D., Comin, D., Gertler, M., Martinez, J. (2019): Endogenous technology adoption and R&D as sources of business cycle persistence, in: American Economic Journal: Macroeconomics, 11(3), 67–110. Auerbach, A.J., Gorodnichenko, Y. (2012): Fiscal multipliers in recession and expansion, in: Alesina, A., Giavazzi, F. (eds), Fiscal Policy after the Financial Crisis, NBER chapters. Chicago: University of Chicago Press, 63–98. Bachmann, R., Sims, E.R. (2012): Confidence and the transmission of government spending shocks, in: Journal of Monetary Economics, 59(3), 235–249. Barro, R.J., Redlick, C.J. (2011): Macroeconomic effects from government purchases and taxes, in: Quarterly Journal of Economics, 126(1), 51–102. Baxter, M., King, R.G. (1993): Fiscal policy in general equilibrium, in: American Economic Review, 83(3), 315–334. Bilbiie, F.O. (2020): The New Keynesian cross, in: Journal of Monetary Economics, 114, 90–108. Blanchard, O., Perotti, R. (2002): An empirical characterization of the dynamic effects of changes in government spending and taxes on output, in: Quarterly Journal of Economics, 117(4), 1329–1368. Browning, M., Lusardi, A. (1996): Household saving: micro theories and micro facts, in: Journal of Economic Literature, 34(4), 1797–1855. Candelon, B., Lieb, L. (2013): Fiscal policy in good and bad times, in: Journal of Economic Dynamics and Control, 37(12), 2679–2694. Carroll, C. (1998): Why do the rich save so much? NBER Working Paper, No 6549. Carroll, C.D. (2001): A theory of the consumption function, with and without liquidity constraints, in: Journal of Economic Perspectives, 15(3), 23–45. Charles, S. (2016): An additional explanation for the variable Keynesian multiplier: the role of the propensity to import, in: Journal of Post Keynesian Economics, 39(2), 187–205. Charles, S., Dallery, T., Marie, J. (2015): Why the Keynesian multiplier increases during hard times: a theoretical explanation based on rentiers’saving behaviour, in: Metroeconomica, 66(3), 451–473. Christiano, L.J., Eichenbaum, M., Rebelo, S. (2011): When is the government spending multiplier large?, in: Journal of Political Economy, 119(1), 78–121. Colander, D. (2000): The death of neoclassical economics, in: Journal of the History of Economic Thought, 22(2), 127–143. Colander, D., Howitt, P., Kirman, A., Leijonhufvud, A., Mehrling, P. (2008): Beyond DSGE models: toward an empirically based macroeconomics, in: American Economic Review, 98(2), 236–240. Corsetti, G., Meier, A., Müller, G.J. (2012): Fiscal stimulus with spending reversals, in: Review of Economics and Statistics, 94(4), 878–895. Deaton, A. (1991): Saving and liquidity constraints, in: Econometrica, 59(5), 1221–1248. Demirel, U.D. (2021): The short-term effects of tax changes: the role of state dependence, in: Journal of Monetary Economics, 117, 918–934. Demyanyk, Y., Loutskina, E., Murphy, D. (2019): Fiscal stimulus and consumer debt, in: Review of Economics and Statistics, 101(4), 728–741. Dobusch, L., Kapeller, J. (2012): A guide to paradigmatic self-marginalization: lessons for post- Keynesian economists, in: Review of Political Economy,24(3),469–487. 352 European Journal of Economics and Economic Policies: Intervention, Vol. 20 No. 2 © 2023 The Author Journal compilation © 2023 Edward Elgar Publishing Ltd Dullien, S. (2017): How to promote alternative macroeconomic ideas: are there limits to running with the (mainstream) pack?, in: European Journal of Economics and Economic Policies: Intervention, 14(2), 238–249. Eggertsson, G.B. (2011): What fiscal policy is effective at zero interest rates?, in: NBER Macroeconomics Annual, 25, 59–112. Engler, P., Tervala, J. (2018): Hysteresis and fiscal policy, in: Journal of Economic Dynamics and Control, 93, 39–53. Fatás, A., Mihov, I. (2001): The effects of fiscal policy on consumption and employment: theory and evidence, CEPR Discussion Papers, No 2760. Fatás, A., Summers, L.H. (2018): The permanent effects of fiscal consolidations, in: Journal of International Economics, 112(C), 238–250. Fazzari, S.M., Morley, J., Panovska, I. (2015): State-dependent effects of fiscal policy, in: Studies in Nonlinear Dynamics & Econometrics, 19(3), 285–315. Foley, D.K., Michl, T.T., Tavani, D. (2019): Growth and Distribution, 2nd edn, Cambridge, MA; London: Harvard University Press. Friedman, M. (1957): A Theory of the Consumption Function, Princeton, NJ: Princeton University Press. Galí, J., López-Salido, J.D., Vallés, J. (2007): Understanding the effects of government spending on consumption, in: Journal of the European Economic Association, 5(1), 227–270. Gechert, S. (2015): What fiscal policy is most effective? A meta-regression analysis, in: Oxford Economic Papers, 67(3), 553–580. Gechert, S. (2017): On theories and estimation techniques of fiscal multipliers, FMM Working Paper, No 11. Gechert, S., Havránek, T., Irsova, Z., Kolcunova, D. (2022): Measuring capital-labor substitution: the importance of method choices and publication bias, in: Review of Economic Dynamics,45 (July), 55–82. Gechert, S., Horn, G., Paetz, C. (2019): Long‐term effects of fiscal stimulus and austerity in Europe, in: Oxford Bulletin of Economics and Statistics, 81(3), 647–666. Gechert,S.,Mentges,R.(2018):Financial cycles and fiscal multipliers, in: Applied Economics, 50(24), 2635–2651. Gechert, S., Paetz, C., Villanueva, P. (2021): The macroeconomic effects of social security contributions and benefits, in: Journal of Monetary Economics, 117, 571–584. Gechert, S., Rannenberg, A. (2018): Which fiscal multipliers are regime-dependent? A meta-regression analysis, in: Journal of Economic Surveys, 32(4), 1160–1182. Gechert, S., Siebert, J. (2022): Preferences over wealth: experimental evidence, in: Journal of Economic Behavior & Organization, 200, 1297–1317. Godley, W., Lavoie, M. (2007): Fiscal policy in a stock-flow consistent (SFC) model, in: Journal of Post Keynesian Economics, 30(1), 79–100. González, A. (2022): Post-Keynesian growth: a neoclassical interpretation, mimeo. Hagedorn, M., Manovskii, I., Mitman, K. (2019): The fiscal multiplier, NBER Working Paper, No 25571. Hall, R.E. (1978): Stochastic implications of the life cycle-permanent income hypothesis: theory and evidence, in: Journal of Political Economy, 86(6), 971–987. Hall, R.E. (2009): By how much does GDP rise if the government buys more output?, in: Brookings Papers on Economic Activity, 2009(2), 183–249. Hebous, S. (2011): The effects of discretionary fiscal policy on macroeconomic aggregates: a reappraisal, in: Journal of Economic Surveys, 25(4), 674–707. Hegeland, H. (1966 [1954]): The Multiplier Theory, New York: Kelley. Hein, E. (2014): Distribution and Growth After Keynes. A Post-Keynesian Guide, Cheltenham, UK and Northampton, MA: Edward Elgar E-Book Archive. Hein, E. (2017): Post-Keynesian macroeconomics since the mid 1990s: main developments, in: European Journal of Economics and Economic Policies: Intervention, 14(2), 131–172. Hein, E. (2023): Macroeconomics After Kalecki and Keynes. Post-Keynesian Foundations, Cheltenham, UK and Northampton, MA: Edward Elgar Publishing. Jappelli, T., Pistaferri, L. (2010): The consumption response to income changes, in: Annual Review of Economics, 2010(2), 479–506. Fiscal policy: post- or New Keynesian? 353 © 2023 The Author Journal compilation © 2023 Edward Elgar Publishing Ltd Jappelli, T., Pistaferri, L. (2014): Fiscal policy and MPC heterogeneity, in: American Economic Journal: Macroeconomics, 6(4), 107–136. Kaplan, G., Moll, B., Violante, G.L. (2018): Monetary policy according to HANK, in: American Economic Review, 108(3), 697–743. Kaplan, G., Violante, G.L. (2022): The marginal propensity to consume in heterogeneous agent models, in: Annual Review of Economics, 14(1), 747–775. Keynes, J.M. (1936): The General Theory of Employment Interest and Money, New York: Harcourt, Brace & Company. Korinek, A. (2018): Thoughts on DSGE macroeconomics: matching the moment, but missing the point?, in: Guzman, M. (ed.), Towards a Just Society: Joseph Stiglitz and 21 st Century Economics, New York: Columbia University Press, 159–173. Kueng, L. (2018): Excess sensitivity of high-income consumers, in: Quarterly Journal of Economics, 133(4), 1693–1751. Lavoie, M. (1996): Horizontalism, structuralism, liquidity preference and the principle of increasing risk, in: Scottish Journal of Political Economy, 43(3), 275–300. Lavoie, M. (2006): Do heterodox theories have anything in common? A post-Keynesian point of view, in: European Journal of Economics and Economic Policies: Intervention, 3(1), 87–112. Lavoie, M. (2012): Perspectives for post-Keynesian economics, in: Review of Political Economy, 24(2), 321–335. Lavoie, M. (2018): Rethinking macroeconomic theory before the next crisis, in: Review of Keynesian Economics, 6(1), 1–21. McKay, A., Nakamura, E., Steinsson, J. (2016): The power of forward guidance revisited, in: American Economic Review, 106(10), 3133–3158. McKay, A., Nakamura, E., Steinsson, J. (2017): The discounted Euler equation: a note, in: Economica, 84(336), 820–831. Oh, H., Reis, R. (2012): Targeted transfers and the fiscal response to the great recession, in: Journal of Monetary Economics, 59, S50–S64. Palley, T. (2019): The economics of the super‐multiplier: a comprehensive treatment with labor markets, in: Metro, 70(2), 325–340. Palley, T.I. (2013): Horizontalists, verticalists, and structuralists: the theory of endogenous money reassessed, in: Review of Keynesian Economics, 1(4), 406–424. Parker, J.A., Souleles, N.S. (2019): Reported effects versus revealed-preference estimates: evidence from the propensity to spend tax rebates, in: American Economic Review: Insights, 1(3), 273–290. Prante, F., Bramucci, A., Hein, E., Truger, A. (2020): Pluralist macroeconomics - an interactive simulator, in: International Journal of Pluralism and Economics Education, 11(1), Artikel 10031790, 55–78. Ramey, V.A. (2016): Macroeconomic shocks and their propagation, in: Taylor, J.B., Uhlig, H. (eds), Handbook of Macroeconomics, Volume 2, 1st edn, Amsterdam: Elsevier, 71–162. Ramey, V.A. (2019): Ten years after the financial crisis: what have we learned from the renaissance in fiscal research?, in: Journal of Economic Perspectives, 33(2), 89–114. Rannenberg, A. (2021): State-dependent fiscal multipliers with preferences over safe assets, in: Journal of Monetary Economics, 117(4), 1023–1040. Ravn, M.O., Sterk, V. (2017): Job uncertainty and deep recessions, in: Journal of Monetary Economics, 90, 125–141. Reis, R. (2018): Is something really wrong with macroeconomics?, in: Oxford Review of Economic Policy, 34(1–2), 132–155. Rendahl, P. (2016): Fiscal policy in an unemployment crisis, in: Review of Economic Studies, 83(3), 1189–1224. Romer, P.M. (2016): The trouble with macroeconomics, mimeo, URL: https://paulromer.net/troublewith-macroeconomics-update/WP-Trouble.pdf (accessed 30 May 2023). Sahm, C., Shapiro, M.D., Slemrod, J. (2010): Household response to the 2008 tax rebate: survey evidence and aggregate implications, in: Brown, J.R. (ed.), Tax Policy and the Economy,Cambridge MA: National Bureau of Economic Research; University of Chicago Press, 69–110. Schoder, C. (2015): Methodological, internal and ontological inconsistencies in the conventional micro-foundation of post-Keynesian theory, New School for Social Research Working Papers, No 1518. 354 European Journal of Economics and Economic Policies: Intervention, Vol. 20 No. 2 © 2023 The Author Journal compilation © 2023 Edward Elgar Publishing Ltd Schoder, C. (2017): Are dynamic stochastic disequilibrium models Keynesian or neoclassical?, in: Structural Change and Economic Dynamics, 40, 46–63. Setterfield, M. (2019): Time variation in the size of the multiplier: a Kalecki–Harrod approach, in: Review of Keynesian Economics, 7(1), 28–42. Shefrin, H.M., Thaler, R.H. (1988): The behavioral life-cycle hypothesis, in: Economic Inquiry, 26(4), 609–643. Spilimbergo, A., Symansky, S., Schindler, M. (2009): Fiscal multipliers, IMF Staff Position Note, SPN/09/11, Washington, DC. Woodford, M. (2011): Simple analytics of the government expenditure multiplier, in: American Economic Journal: Macroeconomics, 3(1), 1–35. Fiscal policy: post- or New Keynesian? 355 © 2023 The Author Journal compilation © 2023 Edward Elgar Publishing Ltd