White Paper
Inflation

The Disparate Geographic Impact of Inflation

Pacific states have experienced the highest inflation, while the Plains states have experienced the lowest.
May 19, 2024
Print This Article

Executive Summary

Discussions about inflation almost always focus on how inflation affects the average consumer. But inflation affects different people differently. For example, our previous research shows that inflation disproportionately adversely affects lower-income households, especially over a longer time horizon.

Regulations, income, and transit costs all affect inflation and vary widely across the country, prompting us to look more deeply into inflation between regions and within them by income group.

  • Over the last 40 years, prices are cumulatively about 20 percent higher on the coasts than in the Midwest. This is largely due to increases in housing costs, which are a function of regulatory burden and the cost of building. The discrepancy is about twice as large as that between the top and the bottom income decile over the same time period, meaning that the disparity is economically very large.
  • Real wages have grown the most in the Midwest over the last 40 years. In the southern and coastal regions, nominal wages have not kept up with prices to the same extent as they have in the Midwest. Real wages are 2–6 percent higher in the Midwest than the rest of the country.
  • Since the start of the inflation surge, real wages are highest in the Pacific and Mid-Atlantic regions. Indeed, these are the only two regions with positive real income growth since January 2021 and both are barely above that level.
  • Despite having the lowest average inflation rate, the Midwest has the most inflation inequality. There, the average gap between the bottom and top income quartiles is 26 basis points, compared to only 11 in the Mid-Atlantic.
  • Since the start of the inflationary surge, the wage gap has significantly compressed, but this varies substantially by region. For example, the South Atlantic has not seen positive real wage growth for any income group since January 2021, but the Pacific region has seen positive real wage growth for the bottom three quartiles and the top quartile is about the same as three years ago.

Altogether, our results indicate that accounting for inflation inequality — both within and between regions — is critical for understanding the economic and political consequences of policymakers’ failure to adequately control inflation. Responsibility for this failure belongs not only to the Federal Reserve, but also to states and localities that have increased cost of living through regulation.

Introducing Geographic Inflation Inequality Indices

Two years ago, FREOPP released the first version of the Inflation Inequality Indices, showing that inflation tends to run slightly higher over time for low-income households. Low social mobility has pernicious compounding effects over time, leading to a cumulative price level in 2021 that was significantly higher for households in the bottom income decile (measured from the late 1970s). In FREOPP’s tradition of uncovering consequences of monetary policy across the income distribution, we go beyond distributional effects at the national level to drill down to regional and regional-by-income levels. We use publicly available data from the Bureau of Labor Statistics to create a new database of regional distributional inflation inequality indices. We focus on the nine regional divisions defined by the Census Bureau and replicated in Table 1.

As a preview of our results, we plot cumulative inflation inequality for incomes, geography, and a combination of the two in Figure 1. Income inflation inequality is a more traditional metric and comes straight from FREOPP’s earlier work. It is simply the ratio of price indices between the top and bottom deciles from 1978 onward. Evidently, prices for the bottom decile are cumulatively 10 percent higher than the top income decile. Geographic inflation inequality is the ratio of price indices between the highest inflation region (the Pacific coast) and the lowest (the upper Midwest). Geographically, inflation is cumulatively about twice as bad as for income inflation inequality. However, when we combine them and compare prices at the bottom in the highest inflation region to prices at the top in the lowest inflation region, we see a dramatic magnification of both income and geographic inflation inequality. This, to us, is the key case for examining geographic inflation inequality.

Figure 1. Cumulative inflation inequality is defined as the ratio of price levels normalized to be the same starting in 1978. We define geographic inflation inequality between the Pacific Region and the Upper Midwest. Income inflation inequality is between the top and bottom deciles of income.

There are several reasons to care about regional inflation inequality beyond the fact that it is relatively large. First, many welfare programs adjust for inflation based on national inflation rather than regional inflation, let alone demographic-specific inflation within a region. With systematic inflation differences across regions over time, the benefits one receives may be degraded by inflation depending on where they live. Second, a clear understanding of historical and contemporaneous regional inflation disparities is informative about the effects of monetary policy, the usefulness of targeted fiscal assistance, and political debates about the relevance of inflation. Much of this inflation inequality is driven by differences in housing costs which are not fundamental to monetary policy. FREOPP’s own Roger Valdez and a large academic literature show that most of the variation in regional housing costs are driven by poor regulatory policy, which means that the burden for correcting that inequality falls on local authorities. Third, precisely because inflation inequality varies substantially by region, real wages can vary even if nominal wage growth is equal across the country. Indeed, we find that the Midwest leads the country in real wage growth over the past forty years because their inflation was relatively tame.

Inflation is typically calculated by taking an initial basket of goods, holding fixed the share of consumption on each good in the basket, and then estimating how much the price of the basket of goods changes over time. Overall inflation is the weighted average of price changes. Weights are calculated from consumption shares. Building on our earlier Inflation Inequality Indices measure of different consumption baskets by income group, our new dataset contributes to FREOPP’s history of rich data work by estimating consumption baskets for regions as a whole as well as income groups within states. This, to our knowledge, forms the first publicly available set of inflation inequality indices by state from 1978 to the present. The price level data start in 1978 and the inflation data from 1979 because the latter are computed as the 12-month percentage change in price levels. The U.S. Congress Joint Economic Committee has published state-level inflation rates since 2021. The Bureau of Labor Statistics also publishes price indices for select goods for the same geographic divisions, but only going back to 2018 and not for inequality within regions.

Our data largely come from publicly available sources. We use the Bureau of Labor Statistics’ Current Expenditure Survey paired with detailed price information from the Consumer Price Index. Additionally, we use the Current Population Survey to construct estimates of regional nominal wages. We compute nominal wage growth following the methodology of the Federal Reserve Bank of Atlanta. All of our data and programs are available for download here along with an accompanying paper.

Table 1. Regional divisions from the U.S. Census Bureau.

In the following two sections, we cover regional disparities in inflation and then analyze regional disparities interacted with income groups.

Regional inflation disparities

Understanding distributional inflation by state addresses a key criticism of our earlier work on inflation inequality. In particular, with social mobility, compounding inflation rates to income groups may not yield useful results. With strong social mobility, there is little reason to think that an inflation rate specific to an income group will apply for the duration of a person’s life. Consequently, compounding inflation rates may be inappropriate because people accumulate significantly more income over the lifecycle. However, when considering the distribution of state-level inflation inequality, the criticism does not apply because the relevant mobility is geographic rather than economic. Precisely because geographic mobility is comparatively lower than economic mobility, it is reasonable to consider compounding inflation by state. For example, about two percent of Americans move to different states every year whereas 34 percent of Americans see their incomes go up or decline by more than 25 percent annually.

Indeed, it is difficult to overstate how much more significant inflation inequality tends to be at the regional level. In Figure 2, we plot the maximum value of inflation inequality for both geographic inflation inequality and income inflation inequality by decile at the national level for every month dating back to 1979. This does not target any particular inequality between groups. Rather, it simply targets the absolute value of the maximum amount of inequality between any groups within either geographic or income inflation inequality. It is rare for income inflation inequality to exceed geographic inflation inequality in magnitude in any month. This underscores the importance of understanding geographic inflation inequality.

Figure 2. Inflation inequality for each type of inflation. The green line is the twelve month rolling average of the maximum inflation inequality between regions for each month, while the red line is the same for income deciles. That means we are not targeting inequality between any specific groups, but only the maximum value for each month.

To get a sense of the cumulative importance of geographic inflation inequality, we plot the cumulative price level by region since the start of the series in 1978 in Figure 3. The cumulative price level documents how much higher prices are since 1978. We set the base value to 100, so that a cumulative price level of 200 would mean prices doubling since then. Darker values indicate more cumulative inflation, with a high approaching 530 in western regions and lows around 450 in the middle of the country. By comparison, the same number for the United States aggregate CPI is 490, meaning that prices are about five times higher today than in January 1978. Simply looking at aggregate inflation misses a great deal of the spread in inflation between states, about eighty percentage points cumulatively in this case. Notably, the middle of the country has been historically insulated from inflation.

Figure 3. Cumulative price level by region since 1978.

Figure 3 suggests that even if nominal wages grow at the same rate around the country, real wages may differ substantially because of regional inflation inequality. Indeed, a potential rejoinder to Figure 3 is that nominal wage demands have been strong enough along the coasts that they make up for the inflation deficit. That does not appear to be true. In Figure 4, we plot real wages for each region normalized to the lowest inflation region, the West North Central. Real wages come from applying the Current Population Survey applying the methodology of the Federal Reserve Bank of Atlanta. The East North Central and West North Central region — both in the Midwest — have the highest real wage growth in the country. The Northeast is not far behind. However, the West Coast and the South lag. The latter is largely because of lower nominal wage growth, while the former struggles because of historically high inflation relative to other regions. Altogether, this breakdown indicates that correcting for regional inflation is vital to discuss real wages across the country.

Figure 4. Real wages by region from 1983 onward computed as the difference between regional nominal wage growth and regional inflation and then converted to an index of 100 at the beginning of the sample. Regional wage data come from the Current Population Survey using the methodology of the Federal Reserve Bank of Atlanta. Normalized relative to the West North Central region and smoothed with a twelve month rolling average.

In Figure 5, we repeat the same cumulative inflation exercise in Figure 3, but this time with a start date of February 2021. This way, we get a sense of cumulative exposure to the pandemic-era inflation which continues to this day. Here, the figure is substantially different from the long-run cumulative inflation plot in Figure 3. Today, a typical consumption basket in the Middle Atlantic, New England, and Pacific regions costs about 17 percent more than it did in February 2021, while that same figure is about 21 percent for the rest of the country. The geographic disparity combined with the relatively high incomes of the low-inflation regions may even help explain some of the political disagreement over the relevance of inflation.

Figure 5. Cumulative price level by region since February 2021.

Geographic disparities in real wages tell an interesting story. In Figure 6, we plot the evolution of real wages by region since the start of the pandemic inflationary episode. Across the country, real wages are either typically lower than they were in 2021 or barely higher. Indeed, the Pacific and the Mid-Atlantic regions — strongholds for Democrats, aside from Pennsylvania — are the only ones with real wages exceeding 2021. Every other region has lower wages than they did over three years ago, an astounding outcome for the strongest post-pandemic economy in the world and a damning indictment of the destructive consequences of inflation.

Figure 6. Real wages by region since the beginning of the inflation surge.

The heat maps in Figures 3 and 5 indicate that aggregate inflation misses significant underlying regional heterogeneity. In Figure 7, we chart the average contribution of each major expenditure type to inflation for each region. Loosely, a major expenditure type’s contribution to inflation is its expenditure share multiplied by its own inflation as a share of total inflation within the region. Clearly, there are large differences between regions driven by housing, transportation, and to a lesser extent, food. These components are almost exactly those excluded in analyses of “supercore” inflation favored by economists like Paul Krugman.

Figure 7. Average contribution of major expenditure groups to inflation within each group during the pandemic inflation episode. Loosely, the average contribution for each group is inflation for each group multiplied by the corresponding expenditure share.

Traditionally, economists prefer looking at “core” or “supercore” inflation rather than headline inflation to monitor the performance of monetary policy. Core inflation excludes energy and food prices and supercore inflation excludes energy, food, and housing. These more volatile components of the Consumer Price Index are largely outside the control of the Federal Reserve, which makes them better metrics for evaluating Fed performance. However, they fail to capture significant regional heterogeneity. In Figure 8, we plot the difference between headline and core inflation for each region during the pandemic era inflation. A larger gap for a particular region indicates that emphasizing core inflation is less relevant for that region. As the gap grows, wonkish discussions of core inflation become more likely to fall on the deaf ears of households who see their cost of living continually increasing and persistently expensive contrasted with the claim that because core inflation is low, the Fed is doing a good job. The headline-core gap is substantial between states, with a difference of nearly two percentage points in July 2022 between the West South Central region and the neighboring Mountain region. Such differences become even larger when using the supercore measure instead.

Figure 8. Difference between headline and core inflation from February 2021 to March 2024. Core inflation excludes energy and food.

Over a longer time horizon, it is clear that there are typically large differences between regions in how much core inflation deviates from headline inflation. In Figure 9, we plot the annual average difference between core and headline inflation for each region from 1978–2024. The East South Central typically has the largest gap, averaging around 0.8 percentage points, while it is smallest in the Pacific region at 0.5 points.

Figure 9. Annual difference between headline and core inflation from February 2021 to March 2024. Core inflation excludes energy and food.

Geographic inflation inequality interacts with inflation type in a significantly different way than income inflation inequality does. In Figure 10, we plot cumulative inflation inequality between the West North Central and Pacific regions for each of core, headline, and supercore inflation. Cumulative inflation inequality at the regional level is the Pacific price level divided by the West North Central price level. A value of 1.2 corresponds to cumulative inflation being 20 percent higher in the Pacific region. We do the same for inflation inequality between the top and bottom deciles of the national income distribution, where cumulative inflation inequality is the bottom decile divided by the top decile. There are two interesting patterns. First, the cumulative effect of inflation inequality is stable over time and across inflation types for income inflation inequality, the opposite is true for geographic inflation inequality. Second, whereas geographic cumulative inflation inequality is about twice as large as income inflation inequality for both headline and core inflation, that pattern reverses when we look at supercore inflation. Indeed, geographic inflation inequality goes the opposite direction. This underscores the dramatic importance of housing for determining geographic inflation inequality.

Figure 10. Cumulative inflation inequality is defined as the ratio of price indices. For geographic inflation inequality, we use the ratio of the Pacific region to the West North Central region. For income inflation inequality, we use the ratio of the bottom to top income deciles.

Figure 10 implies that most of the regional inflation variation is driven by housing. Because this is largely a result of poor regulatory policy at the local level, the burden of correcting this falls on local authorities, not on federal fiscal and monetary policy.

Observed significant regional heterogeneity in inflation is important for two reasons. The first reason is economic: higher inflation rates correspond to lower personal well-being. This means that regional inflation inequality exacerbated by monetary policy shocks leads to disparate outcomes for people in different states. Everyone loses during bouts of inflation, but some states simply lose less than others. For states and the federal government, this heterogeneity may present an opportunity to target inflationary treatment and relief to a larger degree geographically than it currently does. The second reason is political and follows from the first: when there is significant regional heterogeneity in inflation rates, geographic perceptions of inflation will differ substantially. Because inflation is especially salient for many Americans, these regional variations cause varying degrees of geographically based frustration with government efforts to reduce inflation. Policymakers too often overlook these differences because they focus on traditional measures of inflation, like the national Consumer Price Index.

Inflation inequality within states

Perhaps the key feature of this dataset is the ability to examine inflation inequality within states over time. Because the inequality estimates come from the Consumer Expenditure Survey designed for analysis of national inflation, it is not possible to be quite as granular at the state level as at the national level. Consequently, we focus on income quartiles rather than income deciles for each region.

Grouping together all national data, the lowest income quartile on average faces an inflation rate 19 basis points higher than the highest income quartile, which is consistent with our earlier research on inflation inequality. Over 45 years, this annual difference compounds into a difference of about nine percentage points of cumulative inflation. For comparison, the difference between the lowest and highest inflation region in cumulative inflation is approximately 80 percentage points, nearly an order of magnitude larger.

To illustrate where inflation inequality has been most prominent historically, we plot the average inflation inequality for each state in Figure 11 over the period 1979–2024. We define inflation inequality as the difference between the bottom income quartile’s inflation rate and the top quartile’s inflation rate, so that positive inflation inequality corresponds to higher inflation for low-income households. Each region has positive average inflation inequality over time, but there are larger spreads between regions. Interestingly, the West North Central region has the most inflation inequality despite having the least inflation. By contrast, the Northeast — led by New York, Pennsylvania, and New Jersey — has almost negligible inflation inequality.

Figure 11. Average inflation inequality by region. Inflation inequality is defined as the lowest income quartile’s inflation rate minus the top quartile’s.

It is useful to think about how we arrived at Figure 9 over a longer time horizon. In Figure 10, we plot inflation inequality by region for each year in the sample. Interestingly, the relatively high average inflation inequality for the Pacific Region is driven by a high burst of inflation in the early 1980s. Removing this outlier results in a near-total inversion of Figure 1. That is, inflation is lower on average in the middle of the country, but inflation inequality is only a significant phenomenon in the middle of the country. In the context of welfare programs, that means adjusting for regional inflation is insufficient. Moreover, simply observing regional inflation is far from a sufficient indicator of how people in the middle of the country tend to be affected by inflation, even if it might be for more coastal areas.

Figure 12. Inflation inequality for each region from 1978–2024. Inflation inequality is the difference in inflation rates between the bottom and top quartile of income.

Figure 12 also reveals that inflation inequality within regions tends to move in tandem, with the exception of the recent inflationary surge. In 2021 and 2022, inflation was higher for the top income quartile in all regions except the West South Central.

Figure 13. Average inflation inequality by state (monthly), February 2021-March 2024. Inflation inequality is inflation for the lowest income quartile minus inflation for the highest income quartile.

Figures 10 and 12 highlight the difficulty with doing policy around inflation inequality. It is difficult to predict when inflation inequality will emerge, and there are not clear systematic patterns relating to monetary policy in the same way that there are with inequality between regions.

In a pair of impressive papers, Autor, Dube, and McGrew (2023) and Gregory and Harding (2024) document significant wage compression since the beginning of the pandemic. In Figure 14, we complement their analyses by plotting real wages relative to February 2021 by income quartile and region. Two things stand out. First, there is significant heterogeneity in how far wages fell for each income quartile. Indeed, both of the bottom quartiles saw significant real wage gains, while others still have wages below what they were three years ago. Second, the top of the distribution still has lower wages, on average, than they did three years ago.

Figure 14. Real wages by income quartile and region relative to their January 2021 level.

Just as with overall geographic inflation, differences across regions and between income groups boil down to differences in demand for different goods. Indeed, just as the difference between headline and core inflation differs between regions, it differs between income groups as well. Since 2021, the absolute value of the difference between headline and core inflation is about 0.32 percentage points higher for the lowest income quartile than it is for the top income quartile. Consequently, abstract discussions about core inflation will tend to be less reflective of price inflation for the bottom income quartile than for the top income quartile.

Policymakers and data transparency

When inflation is low, policymakers pay less attention to it. It inflicts less pain across the board, leading politicians and media outlets to focus on other issues. While that is sensible in the moment, policymakers then lack the tools to adequately assess who inflation hurts the most when it happens. Congress should adequately fund data collection agencies to study inflation in greater detail, particularly at the regional level.

Extant indices are riddled with measurement error. The Consumer Expenditure Survey does not adequately estimate consumption for each state, let alone state-level consumption inequality. At the same time, the Bureau of Labor Statistics does not publish state-level price indices regularly, so it is not possible to estimate state-level inflation without significant measurement error. These inflation indices rely on state-level consumption paired with national price indices, which is far from ideal.

Granular inflation indices are vitally important for two reasons. First, monetary policy does not affect everyone equally and often makes particular regions or income groups significantly worse off. That inequality cannot be addressed with monetary policy because national monetary policy is necessarily a blunt instrument. However, if policymakers know who monetary policy hurts the most and where the inflationary pressure is, then fiscal policy may be well-suited to alleviate some of that pain without aggravating inflation further. That may mean targeted policy toward certain regions without having to use national monetary policy. Precisely because attention remains fixated on inflation, policymakers should act now to give us the tools to more precisely measure granular inflation going forward.

Conclusion

Overall, inflation varies widely geographically and the degree of inflation inequality within regions varies widely by geography. While there appears to be no systematic relationship between aggregate inflation and income inflation inequality, there is a strong relationship between national inflation and geographic inflation inequality. This revelation should prompt further study of how to use fiscal policy to precisely target those hurt the most by inflation. Such studies will require better and more extensive data from the government.

Data and code are available at this Dropbox folder. Monthly updates available on Dropbox and on Medium.

ABOUT THE AUTHORS
">
Visiting Fellow, Macroeconomics
">
Research Fellow, Financial Services