Is Inflation a Global Phenomenon?

The persistence of inflation at relatively low rates despite years of monetary stimulus has led to wide-ranging investigations into its determinants. Traditionally the rate of inflation has been linked via a Phillips curve relationship to domestic factors, such as slack in the labor market. But is there also a global element?

Maurice Obstfeld, who has returned to UC-Berkeley from his post as chief economist at the International Monetary Field, examines some of the mechanisms by which global factors could affect U.S. inflation in a new National Bureau of Economic Research working paper, “Global Dimensions of U.S. Monetary Policy.” He reviews the evidence on the Phillips curve, and reports that there is little evidence that globalization has had a direct impact on the response of wages to unemployment. An indirect linkage, however, may exist through labor’s lower share of GDP, which could respond to foreign factors such as global supply chains. There may also be a relationship through the correlation of import prices and Consumer Price Index (CPI) inflation.

Another mechanism is based in the linkages of U.S. financial markets with those abroad. If these markets are integrated, then the natural rate of interest (r*) depends on foreign savings and investment as well as the domestic counterparts. An increase in foreign savings will lower the global r* which will stimulate domestic spending. Former Federal Reserve Board Chair Ben Bernanke claimed that this effect the cause of the housing boom in the U.S. that led to the global financial crisis.

Obstfeld points out this financial linkage is intensified by the status of the U.S. dollar as a safe asset and as a reserve currency. He also cites the special role of the U.S. currency as an invoice currency for international trade and a vehicle currency for cross-border lending. Consequently, actions taken to affect domestic spending have significant spillover effects.

Kristin Forbes of MIT also examined the role of global factors in the determination of prices in her Bank for International Settlements working paper, “Has Globalization Changed the Inflation Process?” In this analysis she uses three methodologies: principal components, the Phillips curve and trend-cycle decomposition. In the principal component investigation she looks at inflation in 43 advanced economies and emerging market countries from 1990 through 2017. She reports that 40% of the total variance in CPI inflation is explained by one common principal component. Moreover, this global component of CPI inflation has increased over time. On the other hand, the common component of core inflation (a measurement of inflation without volatile food and energy costs) is smaller and has fallen.

In the Phillips curve analysis, she includes changes in the real exchange rate, the world output gap, changes in oil and other commodity prices, and world producer price dispersion, with the domestic variables. The results for CPI inflation indicate that the foreign variables are significant in explaining inflation. The results for core inflation, however, do not show the same pattern of responses.

When Forbes tests the stability of the coefficients over time, she finds that the global output gap and world commodity prices, which were insignificant in the determination of CPI inflation at the beginning of the sample period, were significant during the period that began in 2007. But these changes are not seen when the measure of inflation is core inflation. As a further test, she compares the predicted changes in CPI and core inflation in regressions using the full set of variables and others with only the domestic variables. The results indicate that the models using the full set of coefficients do better in predicting both inflation rates than the domestic alternatives.

Finally, Forbes utilizes a “trend-cycle” approach that separates inflation into a persistent trend component and a cyclical component. She calculates these components of CPI and core inflation, and then investigates how the trend component and the variables in the Phillips curve analysis affect cyclical inflation. As in the Phillips curve results, she reports that most of the global variables are significant in the regressions for CPI inflation, but not core inflation. She also finds that there was a change in these relationships over time. But when she uses trend inflation as the dependent variable, she finds that the global variables are less significant, even with CPI trend inflation.

Forbes concludes that the evidence she has presented show that global variables should not be considered as ancillary in models of inflation dynamics. Moreover, these dynamics are evolving. Changes in the world output gap and commodity prices now have an impact on CPI inflation that was not evident before the most recent period. Whether or not they will continue to do so is a topic for future research.

Obstfeld’s and Forbes’ results pose a challenge for monetary policymakers. If it is difficult to formulate policies based on domestic economic conditions, it is even more so with foreign factors. This challenge is exacerbated by the constraints on central bank actions due to the current low levels of interest rates. Coordination among central bankers could provide some assistance, but it comes with its own limitations.

Even if the Trump administration is successful in scaling back the trade and financial ties of the U.S. with the rest of the world, inflation will continue to possess a global dimension. The cross-border integration of markets will not be reversed, and domestic prices will respond to foreign shocks. Central bankers are expected to avert another slowdown,  but their ability to maneuver the economy has become more constrained.