The FOMC will meet on September 16 and 17 to once again review incoming data and decide whether the time is ripe to begin normalizing its accommodative monetary policy.
That decision, as Vice Chairman Stanley Fischer clearly stated in Jackson Hole, is still on the table, but is now complicated by three factors: First, uncertainty about China and the market turmoil in equities across the world have obviously created a situation where responsible policy makers will seek to “do no harm.”
Second, the faltering in Friday’s labor market report, coupled with possible other negative incoming data, may also give the FOMC pause.
Finally, the absence of any movement in the inflation rate towards the FOMC’s two percent objective, whether due to the drop in energy prices or some other factor, has caused some committee members to argue that it is still too soon to begin a liftoff.
To be sure, the FOMC has seen significant progress in achieving its employment objectives but has made essentially no progress on inflation. This inflation problem suggests that the Kansas City Fed’s Jackson Hole symposium focusing on inflation dynamics could not have been timelier.
“Consumer Markets and Financial Frictions: Implications for Inflation Dynamics”
So what did we learn from the papers presented on inflation dynamics? The short answer is essentially nothing. Let’s look briefly at some of the key papers and their principal conclusions.
The symposium opened with an empirical paper, “Consumer Markets and Financial Frictions: Implications for Inflation Dynamics,” by Simon Gilchrist and Egon Zakrajsek. Its authors sought to explore why standard Phillips curve equations over-predicted the extent to which disinflation would occur during and after the Great Recession. Their story focused on empirical evidence regarding the pricing behavior of firms and its impact on the Producer Price Index. In this sense, the paper is mistitled, since its focus is entirely on the behavior of producer prices and not on consumer prices.
Gilchrist and Zakrajsek argue that firms’ pricing decisions were affected by market frictions, and in particular by differences in financial conditions as measured by excess bond premiums and firms’ dependence on and ease of access to external finance. More constrained firms were less likely to cut prices in times of stress. However, given the fact that none of their estimated equations had R-squares exceeding 9% (and most were much lower), the authors tend to oversell the importance of their findings because so much of the variation in prices remains unexplained. Furthermore, if frictions were important during the Great Recession, why were they not important previously? This would seem to be an obvious question, which they do not address.
So what do their results mean for the measures of inflation that constitute the focus of central bank policy decisions? One important conclusion is that the financial frictions they have explored may play a role in undermining the Phillips curve and other real business cycle models. However, also absent from their discussion was any attempt to link movements in producer prices to the FOMC’s preferred measure of consumer prices (the Personal Consumption Expenditure Index) or what the shape of the leads and/or lags might be. So in the end, while the paper may have provided some clues about firm behavior, it was silent in terms of what happens consumer prices or to wages – the key factor in current equations driving inflation dynamics in the Phillips curve. Thus we learned essentially nothing from the paper that would help policy makers better understand the inflation dynamics they are trying to model and influence.
“The Impact of International Spillovers on Inflation Dynamics and Independent Monetary Policy: the Swiss Experience”
The second paper, presented in a panel by Thomas J. Jordan, chairman of the Governing Board of the Swiss National Bank, was titled “The Impact of International Spillovers on Inflation Dynamics and Independent Monetary Policy: the Swiss Experience.” The purpose was to discuss whether and how a central bank in a small, open economy could pursue independent policies to control inflation, given the globalization of financial markets that has taken place. His discussion reviewed three time periods: the aftermath of Bretton Woods, the high-inflation period from 1974–1999, and the recent Great Recession period.
Because of two unique characteristics of the Swiss economy – the large role of trade and internationalism and the disproportionate role played by the Swiss banking system and the Swiss franc in international capital markets – inflation and deflation within the country were significantly impacted by international events and fund flows. Virtually no attention was paid by the author to the Swiss policy of fixing its exchange rate to the euro, except to explain that in 2015 that policy became impossible to maintain. Nor did he discuss the international implications of the way the devaluation took place or the lack of credible communications surrounding that policy move.
He really reached two conclusions in his paper. The Swiss experience suggests that monetary policy could be effectively implemented in a small, open economy. But sometimes it can be hard to implement and hampered by international developments. This is an interesting set of observations but hardly relevant to either other small, open economies – since the Swiss economy and financial system are not mirrored elsewhere – or to the US, which is not particularly open or small.
“Reinflation Challenges and Inflation: Targeting Paradigm”
The third paper, by Jon Faust and Eric Leeper, titled “Reinflation Challenges and Inflation: Targeting Paradigm,” was an interesting and wide-ranging thought piece. It stands conventional approaches to inflation dynamics on their heads – mainly the Phillips curve and its derivatives – and argues that there are no really normal periods to which policy might return. For this reason, the models employed and that capture those supposedly normal periods fail to capture the confounding effects that impact economies and that cannot easily be accounted for or anticipated ex ante. They argue that the existing models are too simplistic and the underlying dynamics and behavior of economies in the past can only imperfectly provide data with which to anticipate future events because the underlying structure generating those data is constantly changing.
The rest of the paper is devoted to analysis and discussion of confounding factors and how they have affected models, policy rules, and conduct of policy. The discussion concentrates on the Phillips curve, Taylor rules, Neo-Keynesian frameworks, and related policy. There is virtually no discussion of money or its role in inflation. On the other hand, there is a fairly harsh criticism of the economics profession and its focus on making the Phillips curve work by tweaking the inputs and framework to fit the data rather than developing better models that actually do explain both normal cyclical behavior of the economy and outlier events.
The authors could have more sharply criticized policy makers for relying upon structural equations like the Phillips curve and Taylor rules, which are statistical descriptions of how the economy and policy have evolved rather than estimated structural equations derived from actual models of the economy itself. The bulk of the paper is devoted to detailed discussions of why each supposed period that might be considered normal is not and how factors such as demographics can change the underlying character of an economy.
All in all, this paper, while interesting, delivers what you hope your children or your employees never bring to you: a problem that doesn’t have a solution.
“Inflation During and After the Zero Lower Bound”
The fourth paper, “Inflation During and After the Zero Lower Bound,” by S. Boragan Aruoba and Frank Schorfheide, undertakes the difficult task of trying to describe inflation dynamics during and after an economy has experienced a zero lower bound. They find that their models predict a return to a moderate inflation environment for the US and Europe but continued deflation for Japan.
The problem with this paper is that it is extrapolating using equations estimated from environments that have never exited from the zero lower bound. None of the three economies examined has actually successfully exited from a zero-lower-bound environment, so one can’t verify the conclusions drawn from the empirical work since we don’t know if the estimated equations are stable when the zero bound environment occurs. This problem is compounded by the fact that the authors use both a backward-looking Phillips curve as well as a New-Keynesian model. The former relies not only upon history, where no zero bound exists in the data, but also uses estimates of unobservables such as the output gap, which typically have wide confidence intervals, not to mention that the estimates can never be verified.
To their credit, the authors recognize many of these limitations, but they proceed anyway. Again, like the empirical work in the previous studies, their effort is best viewed as an exercise in descriptive statistics designed to capture the movement in key economic variables rather than estimation of a functional model of the economy designed to reliably capture inflation dynamics.
Having said that, the authors do treat their equations as if they were a model of the economy; and when they solve them they conveniently obtain multiple equilibria, enabling them to assert that one of those equilibria captures what happened in Japan while another better reflects developments in the US. In the end, they admit that “…multiplicity is a blessing and a curse. It allows us to rationalize disparate cross-country experiences but it also generates a lot of uncertainty about the effect of economic policies.”
“Understanding Inflation Dynamics and Monetary Policy: Panel Remarks”
The final conference paper, “Understanding Inflation Dynamics and Monetary Policy: Panel Remarks,” by Vitor Constancio, vice-president of the ECB, is a refreshing discussion of the current state of the art when it comes to inflation dynamics. The author examines alternative attempts to capture inflation dynamics, especially during the Great Recession. He argues that models developed after an event occurs can capture that experience fairly well, but of course those models were not in place to predict the events. His primary focus is on why forecasters didn’t predict the deflation in the Eurozone or the lack of deflation in the US and on possible changes in the relationship between estimates of economic slack and inflation across countries.
He goes on to address weaknesses in current approaches to both the Phillips curve and New-Keynesian equations and to highlight concerns about the stability of the underlying parameters. With regard to measurement of the output gap, a key component of the real-business-cycle models, Constancio notes the wide confidence intervals implied by different models of economic slack. The intervals have important implications for model forecasts, which typically use point estimates to make their predictions, thereby masking the error rates associated with those estimates.
Given these issues, however, it is surprising that he concludes with such an optimistic view of the usefulness and relevance of these approaches. Perhaps that is because the profession has not developed better alternatives. Put another way, sometimes incremental improvements in a mousetrap don’t necessarily increase the capture rate.
Given the focus of the papers and the number of questions about how much light has been shed on inflation dynamics, what would we have liked to see come out of the conference? The first question we would like to answer is, “how do we identify and separate underlying trend inflation from transitory elements?”
Supposedly, inflation should manifest itself in a broad increase in prices, but we know of course that not all prices move in lock step. Sometimes a movement in a relative price, like the recent decline in the oil price, shows up as a drop in measured inflation – but should that drop trigger a policy response? To answer that question, we need to focus on an important dimension of inflation dynamics that was not touched upon in the conference.
In a world in which underlying inflation is accelerating, what components of measured inflation move first, and what components follow? If it is possible to identify which prices tend to lead an upward movement in the overall price level and which ones tend to lag, then we are partway towards understanding inflation dynamics.
The second remaining question concerns the mechanism that generates that upward movement in the overall price level. The current paradigms imply that inflation is a real side phenomenon driven by a tight labor market and the tendency of companies to pass on the associated increases in wages in the form of higher prices in order to preserve profit margins. The old paradigm, which now seems out of favor, suggests that increases in the money supply (and thus too much money chasing too few goods) generates inflation.
In that world government printing presses are the fuel for excess demand, whereas in the Phillips curve world the source of the excess demand is simply assumed. We have few historical examples where demand has increased spontaneously and, without any increase in the money supply, generated unwanted inflation. On the other hand, we have many examples of an unfettered increase in the money supply and fiscal excesses that did generate unwanted inflation. The implied inflation dynamics in these two worlds are quite different and seem to have been totally missing from the recent Jackson Hole debate.