Zero Interest Policy & the New Abnormal: A Critique, by Michael Beenstock (Oxford University Press, Oxford, UK, 2022)

Abstract

In his book, Michael Beenstock labels the recent responses to macroeconomic crises that involve setting policy interest rates close to zero and running large budget deficits as ‘the New Abnormal’, with Japan providing an early example of this approach in the 1990s, and many OECD economies following suit during the global financial crisis in 2008 and with the onset of the COVID-19 pandemic in 2020. Beenstock's central thesis is that such policies generate an ‘existential risk’ in the sense of ultimately being unsustainable and setting up future debt crises with high inflation that cannot be tamed by higher interest rates due to Sargent and Wallace's (1981) ‘unpleasant monetarist arithmetic’ under which spiralling debt service costs necessitate implicit default via monetisation of nominal government debt. Beenstock also sees the ‘stay-at-home’ mitigation policy responses adopted in many countries during the COVID-19 pandemic as fitting in with the ‘New Abnormal’ paradigm.

Cutting interest rates and running large budget deficits might sound to some (like me) as being fairly textbook Keynesian responses to deep recessions rather than being all that ‘abnormal’. But it is the persistent maintenance of such policies following the last two global crises, along with the unconventional nature of monetary policy at the zero lower bound, that particularly concerns Beenstock as being highly ‘unorthodox’ and dangerous. Specifically, he argues that modern macroeconomic theory misinterprets Wicksell's (1898) natural rate of interest in a way that has influenced policymakers to keep interest rates too low for far too long and to rely too much on Quantitative Easing (QE) and large budget deficits without realizing the existential risks to inflation that these policies might pose.

The big mistake in Beenstock's view is that New Keynesian theory, as developed by Woodford (2003) and others, defines the natural rate of interest as the level consistent with stable inflation at a target level and a zero output gap, while Beenstock prefers Patinkin's (1966) interpretation of Wicksell that the natural rate is the marginal product of capital (minus depreciation) from neoclassical growth theory. Implicitly, Beenstock believes that zero interest rate policy can keep real interest rates away from this ‘neoclassical’ rate beyond the horizon over which prices are sticky, with central banks apparently somehow doing so under the ‘New Abnormal’. Based on back-of-the-envelope calculations of marginal products of capital for different countries, Beenstock's preferred estimates suggest positive neoclassical rates at the same time when estimates of long-run real rates have been negative, while inflation has remained stable at close to target levels and the output gap has also averaged close to zero.

Why, one might ask, would policymakers not see that they are holding interest rates too low for too long? Beyond Beenstock's view that the existential risk of ‘kinetic inflation’ is hidden during a liquidity trap, his bigger-picture answer is that policymakers have been blinded by a lack of empiricism in modern macroeconomics. In particular, he criticizes DSGE analysis that has been used to justify the policies associated with the ‘New Abnormal’ as effectively being ‘postmodernist’ by taking models that don't fit the data far too literally.

Given Japan's uniquely long experience with the ‘New Abnormal’, Beenstock devotes a fair amount of space in his book to considering its relevance to his broader thesis about existential risks. First, he considers a ‘quality-quantity’ model of fertility to ponder whether persistently low interest rates might be a driver of Japan's demographic trends and concludes that the ‘data are not inconsistent with the hypothesis’, although South Korea had an even larger fall in birth rates despite substantially higher interest rates over the same period of time. Second, he makes some uncontroversial points about how pensioners don't like low interest rates and that low interest rates would tend to boost real house prices. Third, he ruminates on whether Japan provides a ‘crystal ball’ for what will happen to other economies more recently engaged in the ‘New Abnormal’ but reaches a negative and surprisingly sociological conclusion that economic outcomes for Japan can be explained as much by ‘culture’ as by zero interest rate policy and large budget deficits.

The main digression in the book from its central thesis about zero interest rate policy eventually leading to inflation is provided by Beenstock's evaluation of mitigation policies during the COVID-19 pandemic. There is a questionable suggestion that these mitigation policies greatly increased debt-to-GDP ratios by inducing economic recession, thus fuelling ‘kinetic inflation’. However, even if one attributed the full decline in GDP in 2020 to mitigation policies for countries engaged in them, which is a ‘big if’ as I will discuss below, nominal GDP for these countries has generally since recovered to well above pre-COVID levels. The increase in debt-to-GDP ratios, then, was clearly more related to broader fiscal support measures during the pandemic rather than just being about the costs of mitigation and vaccine development.

Beenstock presents an epidemiological SIR (susceptible, infected and recovered) model to make a point that uncertainty about the effectiveness of mitigation reduces the optimal level of mitigation. Then he compares mitigation stringency and outcomes across countries, with the interpretation that the cost per life saved from COVID was generally very high, although he also acknowledges, but does not reconcile, the fact that ‘countries that suffered most economically also suffered most from Covid-19’.

The book concludes with a discussion of cryptocurrencies. Beenstock suggests that their (sometimes) high values are driven by having, in the case of Bitcoin at least, stronger limits on supply than fiat currency, with investors in crypto evidently more aware than policymakers of the existential risks associated with the ‘New Abnormal’. In this sense, Beenstock sees cryptocurrencies as something akin to a new Gold. However, having made the analogy, he wisely argues against trying to establish a ‘Cryptocurrency Standard’, given that a relatively fixed supply in the presence of economic growth could produce deflation in the prices of goods and services if there were a fixed money price of a cryptocurrency. Of course, as I will also discuss below, Beenstock's earlier argument against New Keynesian theory that interest rates are somehow too low even when inflation is at target and the output gap is zero might be interpreted as leading down a similarly deflationary path.

My main wish when reading this book was that there had been more engagement with the related literature on the various topics addressed. There were a number of assertions made that struck me as having been made in too definitive of a manner that could have been tempered if considered within the context of the broader literature. In the remainder of my review, I make a few observations related to the literature that particularly struck me when reading the manuscript.

First, on the question of whether QE is ultimately inflationary, the related literature is admittedly vast and beyond my remit to summarize. However, in a helpful review of the effects of unconventional monetary policies, Kuttner (2018) highlights a point made by many others that interest paid on reserves by the Federal Reserve (‘the Fed’ hereafter) helps explain why there has not been a massive money multiplier process that some economists predicted would lead to high inflation with the Fed's large-scale asset purchases. Meanwhile, the high inflation after the pandemic seems easier to tie to both traditional demand and global supply drivers corresponding to tight labour markets and cost push shocks than to QE, with inflation coming down recently as those pressures have eased (see Blanchard & Bernanke, 2023, for a comprehensive analysis of the sources of inflation since the pandemic). Debt-to-GDP ratios are certainly higher than they were pre-pandemic, but any unpleasant monetarist arithmetic has thus far failed to stop central banks from raising interest rates to address inflation. Meanwhile, the idea that high government debt might have driven the recent inflation has certainly been hinted at by proponents of the ‘fiscal theory of the price level’ (Cochrane, 2023). So it would have been useful to have some discussion of this theory too, even if it also has its controversies (Buiter, 2023).

In terms of the natural rate of interest, the literature on it is also vast. As just one recent example, we find in Morley et al. (2023) that the trend in real interest rates has been driven by a number of factors, including the marginal product of capital, demographics and safe asset supply/demand. Our estimates of trend are significantly negative at times during the 2010s, which could reflect a low marginal product of capital and/or high levels of market power (see the neoclassical growth model with market power in Ball & Mankiw, 2023). Meanwhile, it is also notable that updated estimates related to Abel et al. (1989) in Geerolf (2018) imply the possibility of dynamic inefficiency and low marginal products of capital, especially for Japan. Beenstock reports high estimates of the marginal products of capital for a number of countries but does not engage with the possibility that low and negative estimates of the natural rate of interest might actually be valid. Interest rates might have been low for long periods of time in Japan and, more recently, in other countries because the neoclassical rate has actually been low, not necessarily because policy is holding real rates down persistently beyond the horizon of sticky prices. Indeed, one might even wonder on theoretical ground whether central banks can actually control real interest rates at longer horizons (see Nelson, 2008, on this point).

Beenstock's criticism of modern macroeconomics as being ‘postmodernist’ was certainly the most fun part of the book to read, including an amusing speculation about how Foucault might have framed what has often been referred to colloquially in the blogosphere as the ‘macro wars’. I'm personally sympathetic to some aspects of Beenstock's critique of modern macroeconomics, having made related points, especially about the Lucas critique, in Morley (2010). But, again, it would have been better to engage more with the literature. For example, reference to a comprehensive review of different epistemological approaches in macroeconomics by Glandon et al. (2023) would have bolstered Beenstock's arguments about the rise of the DSGE methodology and a decline of a positivist hypothesis testing using time series methods. At the same time, it should be noted that Glandon et al. (2023) highlight that the DSGE approach has largely embraced model fit in recent years, contrary to Beenstock's claims about it.

More generally, Beenstock's criticism of modern macroeconomics suffers from a number of bold claims that aren't at all convincing. For example, he suggests ‘VAR models [do] not seek to identify the causal effect of policy on the macroeconomy’ despite the fact that this is exactly what structural VARs are intended to do. One can be critical of the identification assumptions for exogenous monetary policy shocks often employed, but the aim of using structural VARs is very clearly to identify causal effects. Perhaps this quest for causal inference has become clearer with the rise in consideration of ‘external instruments’ in VARs (e.g., Stock & Watson, 2018). Meanwhile, in terms of Beenstock's suggestion that ‘SVAR enthusiasts impose untestable restrictions’, I would note that structural change and identification through heteroskedasticity allow for some testing of restrictions used when identifying structural shocks (see, for example, Lanne & Lütkepohl, 2008).

As a sometimes ‘Bayesian’ in my own empirical work, I was also quite surprised to read ‘Bayesians believe that their model is correct’ given that was never the impression I've gotten when discussing inference with other Bayesians and would also seem to suggest, contrary to a large literature, that Bayesian model averaging and model comparison don't exist.

My biggest concern with Beenstock's critique of modern macroeconomics is his suggestion that DSGE modellers don't care about the empirical fit of their models. Specifically, my impression in reading the recent literature is that the ‘macro wars’ are largely over and the empiricists have won some key battles. With the rise of estimated DSGE models, there has been a strong recognition of Beenstock's suggestion that ‘macroeconometric models must fit the data well’. Beenstock argues that ‘the Smets-Wouters model generates VAR coefficients, which are very different to their unrestricted counterparts’. But this misses the point of Sims (1980), who also argues that macroeconometric models must fit the data well but points out that it is better to compare fit in terms of the implied moving average form (i.e., impulse response coefficients) given that a wide range of very different VAR coefficients can fit the data very similarly due to multicollinearity, but impulse responses will tend to be more stable across different estimates of the VAR coefficients. To be clear, the good fit of the Smets-Wouters model is as much about the assumed process of shocks than about endogenous dynamics due to theoretical structure. But the combined effects of both on the dynamics of observables is that the model fits about as well as a reduced-form VAR, at least enough to avoid distorting the estimated quantitative effects of policy.

Perhaps not surprisingly given the pace at which recent events have moved, the material about COVID-19 in the book seems a bit dated (e.g., the discussion of low vaccination rates in Australia and New Zealand). And, again, Beenstock takes too narrow a view of the related literature. A particularly frustrating omission given his consideration of a SIR model is the literature augmenting DSGE models with SIR dynamics to consider optimal mitigation policies when allowing for endogenous economic responses to the pandemic, as exemplified by Jones et al. (2021) (for which there was a working paper version available in April 2020). I worry this literature was missed because of Beenstock's concerns about methodological issues with DSGE analysis, even though a global pandemic provides a particularly useful application of such analysis to inform policy.

A key point of DSGE models of the pandemic is they are able to capture the fact that a pandemic can lead to economic recession even if a country employs no mitigation strategies at all, while Beenstock assumes the economic recession is the cost of the mitigation strategy. This clearly overstates the cost of mitigation strategies if the counterfactual of no mitigation actually still involves an economic recession. Crucially, it is inconsistent with the fact Beenstock acknowledges that ‘countries that suffered most economically also suffered most from Covid-19’.

Jones et al. (2021) show with their quantitative model that, in a decentralized equilibrium (i.e., no mitigation policy) with endogenous labour supply and consumption choices that affect the probability of exposure to a contagious disease, output will still contract in a pandemic as people choose to stay at home to reduce their own infection risk. However, optimal policy taking into account externalities in terms of reducing others' infection risk and healthcare congestion implies that front-loaded stay-at-home orders should be used to ‘flatten the curve’ more than private agents would choose to do on their own, with the benefits of doing so increasing with how soon a vaccine can be rolled out to reduce infection rates.

Another related literature that I wished there had been more engagement with relates to debates over a Neo-Fisherian view of liquidity traps (e.g., García-Schmidt & Woodford, 2019). Beenstock's implicit prescription for Japan and other countries with zero interest policy is simply to raise rates before it is too late and unpleasant monetarist arithmetic takes hold. However, the Bank of Japan arguably tried this approach a couple of times in the 2000 and 2006, and I think it is safe to say that those Neo-Fisherian experiments didn't work out well given deflation worsened afterwards both times (see Kole & Martin, 2008; Ueda, 2012). Another more recent Neo-Fisherian experiment has been Turkey where the central bank was cajoled by President Recep Erdoğan into lowering, not raising, interest rates to tame inflation, and it is hard to see the subsequent massive exchange rate depreciation and higher inflation as a ringing endorsement of the Neo-Fisherian theory (see Gürkaynak et al., 2023).

In the end, I was unsure about whom this book was written for. A reasonable guess might be that it was meant to help convince central bankers not to set interest rates close to zero even when faced with recession and other policymakers not to use mitigation strategies during a global pandemic. If so, I am doubtful the book will succeed in its aim given competing evidence that many of the policies in the ‘New Abnormal’ worked better than alternatives. Perhaps, instead, the book was meant more for academic macroeconomists like me to get us to be a bit more self-aware and interrogate our methodological proclivities. It is, of course, always worthwhile being conscious of and appropriately critical of prevailing methodologies, although a sole focus on doing so at the expense of actually attempting more useful applied analysis strikes me as one of the main sins of the postmodernism Beenstock is so clearly sceptical of. To Beenstock's credit, he doesn't succumb to such navel gazing and does actually ‘get on with it’ throughout the book by providing a number of different analyses of important topics. I just wish he had engaged more with the related literature when doing so.

Open access publishing facilitated by The University of Sydney, as part of the Wiley - The University of Sydney agreement via the Council of Australian University Librarians.