Following mounting evidence, central banks now recognise that their actions may affect various dimensions of household inequality (e.g. Draghi 2016, Schnabel 2021, Carstens 2021). Yet, whether this observation should inform monetary policy decisions has remained notoriously controversial (Chang 2023).
Until recently, there was little guidance that macroeconomic theory could provide on this normative question. When Clarida et al. (1999) summarised the principles of optimal monetary policy according to the then-emerging New Keynesian synthesis, their framework shared with real business cycle theory the critical assumption of perfect consumption insurance, implying the existence of a representative agent. Under this assumption, if households are symmetric ex-ante, they stay so throughout their lives even if repeatedly hit by idiosyncratic income shocks. Ultimately, there is never any inequality among households, and aggregate consumption is the only consumption metric relevant to policymaking.
While the representative agent assumption has long been subject to serious critique (Kirman 1992), until recently it remained the dominant assumption in mainstream macroeconomic models used to study monetary policy. Recently though, various authors (e.g. McKay et al. 2016, Kaplan et al. 2018, Auclert 2019) have laid the groundwork for a generalisation of the New Keynesian model that assumes away full insurance and makes room for persistent inequality among households. This heterogenous agent New Keynesian model – ‘HANK’, as opposed to ‘RANK’ – implies that consumption dispersion and not only aggregate consumption may be relevant to policy. So far, much of the HANK literature has focused on how inequality alters monetary policy transmission. For example, it stresses that aggregate demand responds differently to shocks and policies when households are borrowing-constrained (Kaplan et al. 2018, Auclert 2019) or engage in precautionary savings in response to idiosyncratic income risk (e.g. Challe et al. 2017, Acharya and Dogra 2020). However, whether and how household inequality calls for revisiting the New Keynesian monetary policy guidelines – the “science of monetary policy” in Clarida et al.’s words – has received much less attention.
In Acharya et al. (2023), we answer this question by deriving optimal monetary policy in a HANK model that incorporates essential dimensions of household heterogeneity, such as income and wealth inequality and limited participation in the stock market. In doing so, we provide a tractable framework allowing us to characterise simple optimal policy rules that directly generalise those implied by the RANK benchmark – thereby making the contribution of inequality to optimal policy fully transparent.
Policy mix under imperfect consumption insurance
Before we discuss the properties of these rules, a few words about how we think of the macroeconomic environment and the role of policymakers therein are in order. First, our analysis draws a sharp distinction between the average level of inequality versus fluctuations in inequality over the business cycle – and, correspondingly, a clear division of tasks between the fiscal and monetary authorities. Consistent with central bankers’ view that long-run trends in inequality are beyond their control and should stay out of their objectives (e.g. Schnabel, 2021), we assume that fiscal policy, not monetary policy, addresses inequality on average. But we also recognise that even the best-intentioned fiscal authority cannot fully eliminate inequality on average. Given that, the best policymakers can hope for is stabilising inequality around its (positive) average level.
On the other hand, fiscal instruments (such as tax rates) are typically inertial and cannot easily adjust to changing economic conditions – a feature of the real world that we capture in our model by assuming that distortionary taxes are time-invariant. This specification gives the central bank the driver’s seat in managing business cycles – including fluctuations in inequality.
Monetary policy trade-offs in RANK versus HANK
How are monetary policy trade-offs altered by the concern for inequality? The traditional trade-off faced by the central bank in RANK is between stabilising prices versus the welfare-relevant output gap (i.e. the distance between actual output and efficient output). The optimal solution to this trade-off is embodied in a simple targeting rule – a form of ‘flexible price level targeting’ which also puts some weight on stabilising the output gap – where the respective weights on the price level and output gap reflect the relative importance of the two goals (Clarida et al. 1999). For example, the optimal monetary policy response to cost-push inflation (as recently experienced in Europe and elsewhere) consists of mitigating the upward pressure on prices by contracting the economy – i.e. by generating a negative output gap. The target criterion pins down the sweet spot where the two conflicting goals (stabilising prices and closing the output gap) are efficiently balanced.
This simple trade-off breaks down in HANK because imperfect insurance and the ensuing persistent consumption inequality across households reduce social welfare. Thus, while in RANK the central bank only seeks to stabilise prices and the output gap, in HANK it also seeks to stabilise consumption inequality. To see why and how this additional goal affects the optimal policy rule, we must describe how inequality fluctuates along the business cycle and what monetary policy can do about it.
Inequality channels and the optimal policy rule in HANK
There are essentially three ways the central bank can stabilise fluctuations in consumption inequality. First, to the extent that individual income risk is countercyclical (i.e. larger in recessions than in expansions), as the evidence suggests (Storesletten et al. 2004, Guvenen et al. 2014), recessions tend to increase consumption inequality, prompting the central bank to mitigate that increase by being more accommodative. Because such fluctuations in inequality are tied to the level of output, that variable shows up as an additional argument in the HANK policy rule – next to the output gap and the price level (the latter two being inherited from the RANK policy rule). In our calibration, the coefficient on output stabilisation turns out to be similar in size to the coefficient on output-gap stabilisation (Acharya et al. 2023, Section IV).
Second, in HANK, monetary policy affects households’ ability to self-insure against uninsured idiosyncratic shocks. For example, low interest rates make it easier for a household to borrow against future income when facing temporarily low earnings, thereby mitigating the pass-through from the income shock to consumption. A lower pass-through, in turn, implies less consumption inequality across households for a given dispersion of individual earnings. Conversely, high interest rates raise pass-through and increase consumption inequality. To see why this channel affects the target criterion, take again the example of a cost-push shock, to which the central bank responds in RANK by hiking the interest rate to fend off inflation. In HANK, hiking interest rates also raises inequality above its average level, which is socially costly. This additional effect leads the central bank to mute the interest-rate hike and thus tolerate more inflation than in RANK. In terms of the targeting rule, this translates into a lower coefficient on price stability – about half of that in RANK in our calibration.
Third, and as stressed by Bhandari et al. (2020), under imperfect insurance macroeconomic shocks typically redistribute aggregate income between different groups of households, ultimately leading to excess fluctuations in consumption inequality across groups. For example, in most countries, only a subset of the population holds stocks. In this context, an increase in firms’ market power (aka a ‘markup shock’) putting upward pressure on prices also redistributes aggregate income away from wages and towards dividends, eventually magnifying consumption inequality in favour of stockholders and at the expense of non-stockholders. To the extent that aggregate dividends and wage payments both depend on the level of output, the concern for between-group inequality provides an additional motive to avoid large output fluctuations – and not only fluctuations in the output gap and the price level. In fact, in our calibration this motive can be so strong that the central bank does not permit any decline in output following a markup shock (Acharya et al. 2023, Section V).
To conclude
In summary, our analysis shows that the central bank should care about stabilising consumption inequality, but also that this concern does not require incorporating an explicit measure of inequality in the policy rule. Stabilising the level of output (not only the output gap), and reducing the weight on price stability, can help reduce inefficient fluctuations in inequality arising from idiosyncratic earnings risk or macroeconomic shocks.
Authors’ note: The views expressed here are the authors’ and not necessarily those of the Bank of Canada, the Federal Reserve Bank of New York, or the Federal Reserve System.
References
Acharya, S and K Dogra (2020), “Understanding HANK: Insights from a PRANK”, Econometrica 88(3): 1113-1158.
Acharya, S, E Challe and K Dogra (2023), “Optimal Monetary Policy According to HANK”, American Economic Review 113(7): 1741-1782.
Auclert, A (2019), “Monetary Policy and the Redistribution Channel”, American Economic Review 109(6): 2333-2367.
Bhandari, A, D Evans, M Golosov and T J Sargent (2021), “Inequality, Business Cycles and Monetary-Fiscal Policy”, Econometrica 89(6): 2559-2599.
Carstens, A (2021), “Central Banks and Inequality”, Remarks at the Markus’ Academy, Princeton University’s Bendheim Center for Finance.
Challe, E, J Matheron, J Rubio-Ramirez and X. Ragot (1997), “Precautionary Saving and Aggregate Demand”, Quantitative Economics 8: 435-478.
Chang, R (2023), “Equality, monetary policy, and the central bank mandate”, VoxEU.org, 25 January.
Clarida, R, J Gali and M Gertler (1999), “The Science of Monetary Policy: A New Keynesian Perspective”, Journal of Economic Literature 37: 1661-1707
Draghi, M. (2016), “Stability, Equity and Monetary Policy”, 2nd DIW Europe Lecture, German Institute for Economic Research (DIW), Berlin, 25 October 2016.
Guvenen, F, S Ozkan and J Song (2014), “The Nature of Countercyclical Income Risk”, Journal of Political Economy 122(3): 621-660.
Kaplan, G, B Moll and G J Violante (2018), “Monetary Policy According to HANK”, American Economic Review 108(3): 697-743.
Kirman, A (1992), “Whom or What Does the Representative Individual Represent?”, Journal of Economic Perspectives 6 (2): 117-136.
Le Grand, F, A Martin-Baillon and X Ragot (2022), “Should Monetary Policy Care About Redistribution? Optimal Fiscal and Monetary Policy with Heterogenous Agents”, Working Paper.
McKay, A, E Nakamura and J Steinsson (2016), “The Power of Forward Guidance Revisited”, American Economic Review 106(10): 3133-33158
Schnabel, I (2021), “Monetary Policy and Inequality”, Speech at the conference on Diversity and Inclusion in Economics, Finance, and Central Banking, 9 November 2021.
Storesletten, K, C Telmer and A Yaron (2004), “Cyclical Dynamics of Idiosyncratic Labor Market Risk”, Journal of Political Economy 112(3): 695-717.
