- A new Federal Reserve Bank of Cleveland working paper examines the inflationary and employment responses of different US cities to interest rate rises.
- It finds inflation declines faster in poorer cities than wealthier ones, but at the cost of greater employment losses.
- With inequality 27% higher than during the 70s and 80s, the paper’s results imply that the economic costs of reducing inflation may be relatively higher this time around.
We are only beginning to understand how monetary policy impacts different individuals in society. The models in most pre-2010s textbooks show central bank decisions affecting different regions and their populations equally. But new spending data and a growing archive of consumer surveys is giving researchers the empirical firepower to uncover the wide-ranging household responses to Fed policy.
A new Federal Reserve Bank of Cleveland working paper uses income, inflation and employment data for 28 US city areas to estimate the diverse effects of a change in monetary policy. Their results show that the inflationary impact of interest rate hikes vary systematically with the relative income of the cities. Also, a 1% rise in the interest rate decreases employment by up to 2.5% in the poorest 10% of cities, while it has only a negligible effect on the richest 10%.
Data & Methodology
The authors collect quarterly data on 28 metropolitan areas from 1969 to 2008. The data includes CPI and the various sub-indexes thereof from the BLS; employment data from the QCEW; and personal income per capita data from the BEA.
The idea is to see how unexpected changes in Fed policy influence prices and employment, conditional on an areas real income per capita. The authors obtain these Fed policy changes from seminal papers in the field.
A popular method for estimating the impact of shocks is the local projection method. This is like a VAR. It is handy if you want to include non-linear variables, which the authors do by considering real incomes by area.
How Does CPI Respond to Fed Tightening?
First, the authors start at a national level. What happens to overall CPI inflation when the Fed increases rates by 1%? The effect on prices is initially positive, and close to zero for the first two years, followed by a sharp decline, reaching a value of -6pp after 20 quarters (Chart 1).
This result is unsurprising. And it aligns with popular beliefs – the stickiness of prices means inflation takes time to abate in response to tightening monetary policy.
Do Prices Respond Differently in Different US Regions?
Next, the authors turn to the metropolitan areas. What happens to local inflation when the Fed increases rates by 1%? The inflation response for a city of average income is very much like Chart 1, as you would expect. However, there are smaller price declines in richer cities than poorer ones (Chart 2).
Chart 2 shows a city with an income per capita that is $1,000 higher than the average gets 1pp less cumulative inflation following a monetary policy shock of 1% after 20 quarters. This leads to a large cumulative difference between the richest and poorest three metropolitan areas in the US (Chart 3).
Is It Only Prices That Respond Differently?
No. The paper runs the same exercise above but for employment. Their results indicate that poor cities shape the national profile of employment effects. That is, employment declines when the Fed increases rates but recovers after three years. Meanwhile, they find no significant employment effects for cities as rich as those in the 90th percentile of the geographic income distribution (Chart 4).
Specifically, a city in the bottom 10th percentile of the geographical income distribution faces a peak employment loss of 2.5%. The average city experiences a 1% loss, while the richest three US metropolitan areas suffer no loss.
Why Do We See These Differences?
As a starting point, consumption in rich and poor areas could differ. And it does, but the authors rule out that this is driving the results. Both tradable and non-tradable CPI components move in the same direction to a Fed hike. So, regardless of which goods or services each area consumes, and what proportion of them, the results would stay the same. In other words, price differences do not reflect different economic structures across locations.
If the data cannot explain systematic differences between areas, perhaps a model will. On this front, greater computing power in recent years has opened the possibility of modelling economies with many different households and in many different states. Here, the authors run a variant of a heterogenous agent model (TANK).
They find the proportion of hand-to-mouth (HtM) households, i.e., those with little disposable income, is a key parameter of the model. In other words, the model can reproduce the qualitative regional patterns estimated in the data by varying the number of HtM households.
They reason, then, that we see regional differences because poorer areas hold more HtM households. When the Fed increases rates, these households must reduce consumption ahead of a temporary worsening of economic conditions (richer households do not need to because they have savings). A reduction in local demand in poorer areas hurts the local economy, which affects employment. HtM individuals consume exclusively from their labour income, which further impacts local consumption.
Here we see that the poorer the region, the more severe the local impacts from a Fed policy change. A decline in local demand leads to a larger decline in prices and is associated with a larger decline in employment.
What Does This Mean in Practice?
As the Fed continues to hike to combat inflation, we should expect poorer US regions to suffer greater declines in economic activity. Job losses here are likely to be higher, and disinflation will kick in. Below, I map income per capita by US state, followed by the inflationary pressures by US state (Chart 5).
Many states that have above-average income per capita, such as California, Massachusetts, Colorado, and New Hampshire, also face above-average inflationary costs over the next year. The paper’s results imply inflation in these states is likely to remain higher for longer, with employment remaining relatively constant.
Meanwhile, in below-average income states that are facing above-average inflationary pressures, such as Utah, Idaho, Arizona, Nevada and Montana, we could expect higher job losses but quicker disinflation.
The paper reveals the inflation dynamics of different US areas in response to Fed policy. But we can extract a deeper conclusion. The results imply that an increase in inequality across cities in the US raises the impact of monetary policy on both prices and employment. Since the 1970s, inequality has risen by 27% in the US (Chart 6). This is unlikely to have impacted states uniformly. So, while this may increase the speed of disinflation, it reduces the likelihood of a soft landing relative to the Great Inflation.
Sam van de Schootbrugge is a Macro Research Analyst at Macro Hive, currently completing his PhD in international finance. He has a master’s degree in economic research from the University of Cambridge and has worked in research roles for over 3 years in both the public and private sector.