Lopez-Salido, David and Loria, Francesca, Inflation at Risk (January 6, 2022). Available at SSRN: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4002673 [Previous version: FEDS Working Paper No. 2020-013]
Abstract:
- David López-Salido and Francesca Loria (Federal Reserve) measure inflation risk based on the predictive distribution of inflation in the United States and the euro area.
- The results show how economic conditions have continued to affect the extreme values of the inflation distribution, despite the « disappearance » of the Phillips curve effects.
- The results also highlight the role of financial conditions in the evolution of inflation risk in the United States and the euro area.
How do macroeconomic and financial factors influence inflation dynamics? Why is the distribution of inflation important for measuring inflation risk? How has inflation risk (and, incidentally, deflation risk) evolved over the last few decades in the United States and the euro area? David López-Salido and Francesca Loria (Federal Reserve) address these questions in this working paper. While these issues were crucial before the health crisis, they are even more so in the post-COVID world and in a context of high inflation, where some central banks (the Federal Reserve in the United States and the European Central Bank in the eurozone, among others) seem to be hinting at a rate hike. In terms of economic policy, the stakes are therefore high, and the main message of this paper echoes this: measuring inflation risk rather than inflation itself provides a better understanding of its dynamics and therefore a more coherent explanation of its evolution over time, enabling more appropriate monetary policy decisions in the event of an inflationary shock.
Inflation risk and the Phillips curve
To answer these questions, López-Salido and Loria estimate the distribution of the core inflation rate (which excludes food and commodity prices) using a quantile regression based on an « augmented » Phillips curve model.[2] More specifically, the estimation consists of analyzing how the macroeconomic factors commonly considered to explain inflation—inflation expectations, unemployment rates, relative price changes—determine inflation quantiles, and more specifically the first quantile (10% of the lowest values) and the last quantile (10% of the highest values). Using a sample of quarterly U.S. data from 1973 to 2019, they show how the effect of these variables on the distribution of the inflation rate evolved from one sub-period to another, between 1973-1999 and between 2000-2019 (Figure 1).
The authors demonstrate that, while the estimated effect of these macroeconomic variables was virtually equivalent across the entire inflation distribution over the period 1973-1999, their impact tended to shift toward the extreme values of the distribution over the period 2000-2019. Furthermore, the results clearly indicate that the effect of the unemployment rate[3] on inflation (top left of Figure 1), commonly interpreted as the slope of the Phillips curve, has dissipated, suggesting that the effects of the Phillips curve have disappeared in the recent period.[4]
The role of financial conditions
In addition to macroeconomic variables, the authors « augment » the traditional Phillips curve by including a variable that measures financial conditions in their quantile regression of the augmented Phillips curve. Following the ideaof Adrian et al. (2019), the idea is to establish a link between the level of financial conditions and the risk of inflation (low or high). The bottom graph in Figure 1 shows that financial conditions have tended to have a relatively stronger effect on the tail of the inflation distribution (i.e., risk of deflation) in the recent period compared to the previous period, when the impact was roughly equivalent across the quantiles of the distribution. The change in the coefficient associated with financial conditions is therefore more in line with the main message of the paper: the impact of macroeconomic and financial factors on the inflation distribution has shifted in favor of the first quantile (i.e., risk of deflation), and not across the entire distribution.
The risk of inflation during a recession
To illustrate this point, the authors cite the consequences of the 2008 financial crisis on inflation trends in the United States and the euro area. By breaking down estimated inflation by distribution quantile, they show how the risk of low inflation, or even deflation, has changed over the past two decades. Figure 9 of the paper presents these results: while the gap between the first quantile and the fifth quantile—the median—widened during the Great Recession of 2009 in the United States, this does not seem to have been the case in the euro area, where this gap tended to narrow during periods of recession (the 2007-08 crisis and the 2010-12 sovereign debt crisis). In other words, the change in the distribution of inflation observed following the 2007-08 crisis stems from a fall in the estimated values at the tail end of the distribution in the United States, and rather from a marked decline in the median and the estimated values at the top end of the distribution in the euro area.[5] The authors thus emphasize how the risk of deflation is returning to pre-crisis levels in the United States, while it appears to be increasing in the euro area, due to the downward trend observed in the first quantile of the distribution.
Focusing on the more recent episode of the health crisis, López-Salido and Loria highlight the importance of studying inflation risk in times of uncertainty. Their results reveal a risk of deflation at the start of the pandemic and a risk of high inflation during the recovery.
Conclusion
In this working paper, David López-Salido and Francesca Loria suggest analyzing the distribution of inflation, rather than its average value, in order to understand and thus appreciate the risk of inflation.
More specifically, they show why it is important to study the extreme values of the distribution by considering a quantile separation (10% of the lowest values, 10% of the highest values). The methodology developed in this working paper highlights two main results:
(i) it is necessary to look at the distribution of inflation to understand that macroeconomic factors still explain the dynamics of inflation over the last two decades, and
(ii) financial conditions are an important component of inflation trends, particularly through their impact on the risk of low inflation or deflation.
Finally, the comparison between the United States and the euro area presented in the paper and its implications in terms of economic policy should lead central bankers and researchers to question the relevance of taking inflation risk into account in monetary policy decisions. Taking into account the full range of possible inflation values, via its distribution rather than its mean or median value, could be a key issue in the conduct of monetary policy and the strategy of central banks in the years to come.
[1]See the recent BSI Minute » Are central banks immune to inflation risk? » dated January 20, 2022.
[2]The Phillips curve in its initial formulation describes a decreasing relationship between the unemployment rate and the inflation rate (or the rate of nominal wage growth). The BSI note » The puzzle of low inflation in advanced countries » dated January 23, 2018, illustrates this.
[3]A measure of the unemployment gap (the difference between the observed unemployment rate and its structural level) is used in the paper rather than the unemployment rate.
[4]The results presented in this graph tend to show a decrease in the impact of macroeconomic factors rather than a significant change in the distribution of these effects on inflation. The absorption of these effects is probably explained by the high coefficient on long-term inflation expectations estimated over the recent period, which could capture a high degree of persistence in core inflation.
[5]This result may seem surprising, as it would mean that the risk of deflation was significantly greater during the Great Recession in the United States than in the euro area, whereas the European Central Bank (ECB) justified the implementation of unconventional policies in response to downward pressure on inflation, including in the wake of the sovereign debt crisis (see Mario Draghi’s press conference on January 22, 2015).
