The Theory of Inflation Expectations on Trial
The concept of expectations has, arguably, been the crown jewel of contemporary macroeconomics. In itself, it represents roughly half of all so-called “micro-foundations” innovations that distinguishes modern macroeconomics from its Keynesian origins. Incorporated into economic modelling beginning with the 1970s, it is the theory of expectations more than anything else that basically lays the foundation for modern-day central bank policy which – until a decade or so ago – seemed to have finally mastered the ups and downs of the economy. Inflation targeting, policy guidance, transparency, commitment, credibility, time consistency and so on are just another way of saying the successful central banks manage the economy by managing expectations. Thoughts are future actions and words can shape reality before it even comes into being, altering 19th century style classical causality – this is in a nutshell the Sci-Fi appeal and semiotic charm of a concept that undergirds current economic theory! Recent economic events has put the theory of expectations to the test like never before, but the outcomes are far from satisfactory.
Expectations are defined by John Muth in a widely recognized article on the matter as “predictions according to theory”. Economists also call this type of expectations “rational expectations”. In short, they assume that economic agents behave like economists and that the economists have the right model regarding how the economy works. Of course, economists are not so silly to think that all economic agents are economists, so in practice they recognize that economic expectations can diverge substantially from the rational norm. Most of them however think that economic agents form expectations on the basis of some kind of rule of thumb, like tomorrow will mostly be like yesterday, with some optimistic and pessimistic outliers averaging out. These are what economists call adaptive expectations. Nevertheless, a shock of some kind (rising unemployment, bank failures, rising prices, a pandemic or a war) can unhinge the public’s expectations and fuel a cycle of more or less self-fulfilling doomsday prophecies. In this setting, the Central Bank’s job resembles less that of a firefighter and more that of a therapist, since a fire on the loose already constitutes a failure of guidance about the future, a failure of expectations management.
The extent to which modern central bank policy and even organisational structure revolves around the theory of economic expectations is often overlooked. From the legal ban to directly acquire government debt title in order to prevent its monetization to the compartmentalisation of the Bank’s decision-making bodies in order to isolate the monetary policy committee from the rest, practically everything a modern Central Bank and its officials should do is the consequence of the importance acquired by expectations in macroeconomic theory, and this includes the frequent press conferences of Bank governors unknown in earlier times as well as the many analyses published in real time by the Bank’s junior staff. The Central Bank must constantly assure and guide the myriad decisions of the economic agents and for this guidance to be effective it must build and preserve its credibility through an outstanding reputational capital.
“The Great Moderation”, which began with the end of the previous inflationary episode around 1982 and lasted until the deflationary episode that followed the 2007 financial crisis, was the heyday of expectations management-based monetary policy. Beginning with the 1970s stagflation, economists derived a measure of expected inflation from econometric analyses of historical time series and economic models based on the Phillips curve, the generalized inverse relationship between the rate of change of the unemployment rate and the rate of change of the price level, which was used to estimate whether there was slack or excess growth in the rate of change of the economy’s aggregate output, as measured by the Gross Domestic Product (GDP). However, these measures were largely backward-looking and therefore constituted an imperfect foundation for a monetary policy called upon to anchor the price level and stabilise the variability of annual GDP. New, forward-looking, measures of expected inflation were needed in order to properly estimate the inflation expectations in the upgraded Phillips Curve, proposed by Edmund Phelps, and thus for the Central Bank to act in a timely and proportional manner, avoiding painful episodes of overshooting as well as undershooting of the economy’s potential output.
A host of surveys began to regularly feature among the most looked at and publicized data. Consumer sentiment surveys first and foremost, but also producers’ and management surveys. These were followed by professional forecasters’ surveys, generally made up of financial industry analysts trying - like the consumers or the corporate acquisition managers - to guess the inflation rate one, two or five years ahead. But the most complex and highly regarded inflation measure devised during this period was the floating of inflation-indexed bonds, first in the UK, then in the US and, since the 1990s, in the major economies of Continental Europe as well. The investors who bought and sold these financial assets were deemed to have more “skin in the game”, as economists say, that is better incentives to forecast the future than the average survey responder because they spoke by risking their savings and, as a group, made more knowledgeable decisions. Unlike the usual government bond, the inflation-indexed bonds were promised a net return, which made it possible to estimate the expected future inflation rate by subtraction from the former. Or so it was thought.
Once the turbulence started, some fifteen years ago, these new macroeconomic policy tools failed to provide policymakers and the public alike with an adequate compass to navigate the storm. The observation has already been made during the preceding, quiet, decade that the inflation expectations measures varied widely, but this was seen as just another argument in favour of multiplying the measures of expected inflation. Consumers, for instance, tend to consistently overestimate future inflation compared to actual inflation. This upward consumer bias estimate of inflation was to a certain extent in line with the theory of adaptive or insufficiently anchored expectations and the measure was consequently saved by new rationalisations of the inflation data, most notably by eliminating the most volatile prices - usually energy and food – from the computation. This gave birth to a so-called core measure of inflation and expected inflation, which is deemed more precise and less confusing, although it often allows various central banks to choose ex post facto the numbers behind the realized inflation target, as is convenient for them. On the other hand, professional forecasters tend to underestimate future inflation compared to actual inflation. This is due to a sort of expert bias, less studied or acknowledged than the upward consumer bias, which manifests as overconfidence in the central bank’s ability to hit its announced inflation target and unwillingness to stray too far from its forecasts. But the inflation-indexed bonds produced the greatest perplexity. Ever since these financial instruments were created, economists hypothesized that their low quantity, uncertainty over continual issuance and very safety incorporates a risk premium in their return that hinders the proper estimation of the inflation premium incorporated in the return of non-inflation indexed bonds. The disagreement over the size of the risk premium in the inflation-indexed bonds had major implications regarding the fiscal stimulus packages and the monetary policy stance adopted in the aftermath of the 2007/2008 Financial Crisis to restart growth and has major implications today, in the aftermath of the 2019/2020 Covid-19 pandemic and the Russian war in Ukraine, regarding the course of economic policy we need to adopt in order to contain the inflation surge.
The noise in the inflation expectations measures is so great that, more often than not, during the last decade policy observers were left with the feeling that no economic theory could account for the facts of reality. First, quantitative easing and direct asset purchases by the world’s major central banks failed to raise inflation and even to move inflation expectations, although practically every monetary policy textbook written before their implementation suggested that monetary expansions on this scale should have unleashed an inflationary tsunami. A new central banking policy paradigm basically emerged on the go, one based not on its contemporary price stability mandate but on its older and subsidiary bank-clearing role. Then the stock market – a topical branch of economics previously banned from serious macroeconomic discussions – became the leading measure of the health of the economy. Only the wealth effect, it seemed, could overcome the zero-bound of the liquidity trap. The whole debate regarding the flattening or disappearance of the Phillips Curve simply meant that nobody could quite explain how the economy works anymore, whether it made any sense to connect variables such as interest rates, monetary aggregates or government spending to inflation and to connect inflation, or the lack of it, to unemployment or GDP growth. And then, out of the fear of a relapse into deflation, suddenly came runaway inflation...
The fact that inflation expectation measures vary substantially in their estimates of future inflation, but generally converge with regard to the direction of future inflation allows proponents of the theory to save face and continue to gloss over an impractical set of crucial policy indicators. When the US 10-Years Breakeven Inflation Rate, for instance (which calculates the difference between the 10-year US nominal bond and the 10-year inflation-adjusted bond), indicates an expected inflation rate of only 2.55% on August 30, 2022, it is really hard to square the theory of rational expectations with an 8.52% actual yearly inflation rate in the US for the end of July 2022, no matter how many auxiliary assumptions one makes, let alone formulate monetary policy on this basis. Nevertheless, the concept is elegant and the intuition behind it – an overlooked extension of the very basic tenets of economics as a science – is too strong to be easily discarded. Economists, the worldly philosophers as Robert Heilbroner called them, are no better in this regard than the philosophers proper: like Hegel, they will sacrifice the world for the sake of the theory, at least until a better theory comes along.
References
Robert L. Heilbroner, The Worldly Philosophers, Penguin Books, London, 2000.
John F. Muth, Expectations and the Theory of Price Movements, Econometrica, Vol. 29, No. 3, (Jul. 1961), pp. 315-335.
Edmund S. Phelps, Phillips Curves, Expectations of Inflation and Optimal Unemployment over Time, Economica, New Series, Vol. 34, No. 135 (Aug. 1967), pp. 254-281.
