People shopping at a food market in New York. Picture taken January 30, 2024.

Blog Name

Supply shocks were the most important source of inflation in 2021-23, but raising rates to curb demand was still appropriate

Date

Photo Credit: Sipa USA/Richard B. Levine

Body

Inflation dropped sharply in the United States and around the world in late 2023 after surging in 2021 and 2022, spurring some retrospective second-guessing over recent monetary policies. Some analysts have argued, for example, that inflation was almost exclusively driven by supply shocks that receded without any help from tighter monetary policy. That line of thinking prompted some to conclude that central banks did not need to raise interest rates at all.

But that contention oversimplifies what happened. Though the evidence strongly supports the powerful role of negative supply shocks in driving inflation up, it seems likely that positive demand shocks also played a supporting role, most notably in the United States. It would therefore be a mistake to argue that central banks did not need to raise interest rates.

To the extent that demand contributed to the surge in inflation, higher policy rates were necessary. Moreover, central banks needed to be seen responding to inflation in order to prevent expectations of higher inflation from becoming entrenched in wage and price setting behavior.

This post reexamines the inflationary shocks in the major advanced economies over the past few years, expanding on a previous post that decomposed the surge of inflation into demand-side and supply-side forces as of mid-2022.

Inflation and growth since 2019

The table below displays outcomes for inflation and economic growth in the major advanced economies over the past four years. The first two columns display outcomes over the eight quarters starting with the spread of COVID-19. The United States stands out with the strongest economic growth and highest inflation. The United States is the only economy to have maintained a normal growth rate on balance over the period, but it came at the cost of a much higher inflation rate. The euro area and the United Kingdom had weak growth and inflation close to target. Japan had weak growth and inflation far below target, though it should be noted that annual inflation in Japan was only 0.5 percent over the eight quarters before 2019Q4.

The last two columns display outcomes over the eight quarters starting with the Russian invasion of Ukraine. All of these economies saw a pickup in inflation, though in Japan inflation was only modestly above target. Again, the United States stands out with the highest growth rate, but the euro area and Japan grew reasonably closely to estimates of their long-run potential.1The United Kingdom had the worst growth performance and the highest inflation over this period.

Deviations from forecast as demand and supply shocks occur

Figure 1 displays a stylized demand and supply diagram for the entire economy. The demand curve slopes downward because lower prices encourage spending while higher prices discourage spending. The supply curve slopes upward because higher prices elicit more production while lower prices reduce production.

The curves intersect at the levels of prices and output (normalized at 0) that are expected to prevail in the absence of unexpected shocks. Deviations in prices and output from their expected levels can be decomposed into demand and supply forces using the diagram. For example, a leftward shift of the supply curve from the solid to the dotted line moves the economy up the demand curve to an outcome with higher prices and lower output, labeled point A. The exact location of point A depends on the slope of the demand curve, but the outcome of a negative supply shock is always in the upper left quadrant. Similarly, a positive supply shock lands the economy in the lower right quadrant, while positive and negative demand shocks move the economy to the upper right and lower left quadrants, respectively.2

Outcomes that are not on either initial curve require both demand and supply shocks. For example, an increase in the price level without any change in output (point B) can happen only if there is a positive demand shock and a negative supply shock. The magnitudes of the demand and supply shocks needed to explain any outcome depend on the slopes of the curves. In addition to moving the curves left or right, it is possible that shocks can change the slopes of the curves. For example, supply chain bottlenecks may make supply curves steeper.

Demand and supply shocks differ across countries

Panel a of figure 2 displays percent deviations of GDP prices and real GDP from their expected levels in 2021Q4 for the major advanced economies. The expected levels are based on projected growth rates from 2019Q4 in the Organization for Economic Cooperation and Development's (OECD) Economic Outlook November 2019.3 The major surprises over this period are the COVID-19 pandemic and its effects on workers and firms as well as the policy responses, including fiscal packages and monetary easing.

All of these economies except Japan had a sharp recession in 2020 followed by a strong recovery in 2021. The United States is the only economy to exceed its forecasted real GDP, albeit by a small amount. The US location close to the vertical axis strongly suggests that both a positive demand shock (from massive fiscal stimulus) and a negative supply shock (COVID-related labor market disruptions) were at work. Negative supply shocks were also at work in the euro area and the United Kingdom, with relatively little role for any net demand shock.4

Japan's outcome with a significant shortfall in prices and a modest shortfall in output suggests a negative demand shock, though the magnitude may be overstated owing to a faulty initial projection. The OECD projected a 2 percent increase in the GDP price level, whereas private forecasters were projecting less than a 1 percent cumulative increase in core consumer prices over 2020 and 2021 (Consensus Forecasts January 2020). In the end, GDP prices changed little over the period.

The apparent positive demand shock in the United States, minimal demand shocks in the euro area and the United Kingdom, and negative demand shock in Japan are consistent with the fact that general government expenditures increased the most in the United States and the least in Japan over 2020–21.5

Panel b displays percent deviations of GDP prices and real GDP from their expected levels in 2023Q4. The expected levels are based on projected growth rates from 2021Q4 in the OECD Economic Outlook November 2021.6 The major surprises over this period are the Russian invasion of Ukraine with subsequent reduction of oil and gas supplies and the supply chain bottlenecks and semiconductor shortages arising from COVID and the shift of consumer demand away from services and into durable goods.7

All of these economies were pushed into the negative supply shock quadrant after 2021. Although inflation continued to surprise on the high side in the United States, the inflation shock in the euro area and the United Kingdom jumped from considerably less than that in the United States to considerably more.8 Both the euro area and the United Kingdom are more affected by energy shocks than the United States because they are net energy importers, whereas the United States is now a net exporter of oil and gas. Also, they relied heavily on cheap piped Russian gas and had to switch to expensive liquefied natural gas.

Despite being an energy importer, Japan seems to have had a relatively small supply shock because it uses more long-term energy contracts and the government subsidized some energy prices. Japan had a much better public health response to COVID, enabling shorter school closures and fewer mobility restrictions. And its consumers did not splurge on durable goods and create bottlenecks as much as consumers elsewhere.

Unless demand curves are very steep, it is likely that all of these economies had modest positive demand shocks in 2022–23, reflecting the lagged effects of past fiscal and monetary easing.9

Monetary policy and the recent decline of inflation

Inflation dropped sharply over the course of 2023 in the United States, the euro area, and the United Kingdom. (Japan never really had an inflation surge and seems on track to achieve its inflation target over the next couple of years.) Central banks are not going to try to reverse the 2021–23 runup in the price level, which would require negative inflation. Instead, they merely want to keep future inflation at its 2 percent target.

Some economists have argued that the decline in inflation can be attributed entirely to a reversal of the supply shocks and that monetary policy played no role. Many readers may have drawn the erroneous conclusion that central banks did not need to raise interest rates at all.

The results presented here suggest that supply shocks were the main culprit but that demand probably played a supporting role, especially in the United States. To the extent that demand contributed to the surge in inflation, higher policy rates were necessary. In addition, higher inflation directly reduces real interest rates, so that higher policy rates are needed just to keep real interest rates from falling. Moreover, central banks needed to be seen responding to inflation in order to prevent expectations of higher inflation from becoming entrenched in wage and price setting behavior.10

Inflation and growth since 2019 (percent change at annual rate)
  2019Q4 – 2021Q4 2021Q4 – 2023Q4
Country Real GDP GDP price Real GDP GDP price
United States 2.1 3.9 1.9 4.5
Euro area 0.5 2.3 1.0 5.6*
Japan 0.2 -0.1 1.1 2.8
United Kingdom 0.3 2.4 0.2 6.3
*2023Q4 GDP price in the euro area is based on eight member economies with a large majority of area-wide GDP. 
Sources: OECD Economic Outlook November 2023, US Bureau of Economic Analysis, Eurostat, Japanese Cabinet Office, UK Office for National Statistics, and author's calculations.

Figure 1. Demand and supply in the overall economyFigure 2. Forecast surprises in 2021Q4 and 2023Q4

Notes

1. The October 2021 IMF World Economic Outlook projected long-term (5 year ahead) growth rates of 1.7 percent (United States), 1.4 percent (euro area), 1.5 percent (United Kingdom), and 0.5 percent (Japan).

2. Researchers at the Federal Reserve Bank of San Francisco use changes in output and price at the industry level to identify demand and supply contributions to inflation in the United States. Their method uses monthly deviations from statistical models that define "normal" output and price, with deviations characterized as demand- or supply-driven based on their quadrant. They find that both demand and supply surprises were important during the 2021–23 surge in inflation.

3. The projected growth rates are applied to the actual levels of prices and output in 2019Q4, not the estimated values as of November 2019, reflecting an assumption that the forecasts are focused on growth rates and not on levels of prices and output.

4. It seems that fiscal support in these economies roughly offset the negative demand effects on household and business spending caused by the pandemic and subsequent mobility restrictions.

5. See table 3 in Joseph E. Gagnon and Asher Rose, How did Korea's fiscal accounts fare during the COVID-19 pandemic? PIIE Policy Brief 23-8.

6. Outcomes for 2023Q4 are  taken from initial GDP estimates from the US Bureau of Economic Analysis, Eurostat, the Japanese Cabinet Office, and the UK Office of National Statistics.

7. Although an increase in consumer spending is typically evidence of a demand shock, a shift in the pattern of spending across sectors for a given level of total spending is an aggregate supply shock to the extent that sectoral supply curves are nonlinear. That is because prices rise more in the sectors experiencing higher demand than they fall in the sectors experiencing lower demand.

8. The excess of the UK negative growth surprise over that in the euro area likely reflects the impact of Brexit on UK supply chains. For a summary of the effects of Brexit on the UK economy, see this article by Jonathan Portes.

9. A demand slope of around -2.5 would be needed for the outcomes in figure 3 to be consistent with a constant demand curve. For comparison, the demand slope over an eight-quarter horizon in the Federal Reserve's FRB/US model is between -0.25 and -0.40 (depending on assumptions about expectations formation), requiring a significant increase in demand to yield the outcomes displayed. (Demand slopes are calculated as the ratio of price to output responses to a pure productivity (supply) shock.)

10. The big inflation shocks of the 1970s occurred initially as reductions in energy supply, but over time as central banks did not communicate and act aggressively to rein inflation in, expectations of future inflation rose and inflation became self-perpetuating. Higher inflation expectations act simultaneously to increase demand by lowering the real interest rate and to reduce supply by making producers less willing to sell at the existing price level because they expect to face higher costs.

Data Disclosure

The data underlying this analysis are available here [zip].

More From