Marijn Bolhuis, Judd Cramer, Karl Schulz, and Larry Summers have a very nice new NBER working paper on inflation. Here is the current measure of inflation along with what inflation would be if the BLS computed the cost of housing as it did before 1983:
Just some unsophisticated thoughts. Rents have gradually gone up due to higher borrowing cost for multifamily and higher taxes and maintenance fees for investor-owned. Using the current measure, core CPI will stay high for a while. If Fed started to cut rates, while it’ll ease up some borrowing cost pressure, it could bring up housing cost, which could cause the old measure even higher. Not to mention using the future consumption measurement. Some argued that we’re slowly absorbing the one time price shock due to government stimulate policies. But I’m afraid the absorbing speed is too slow (given that Fed printed more money to save SVB and Signature) that only a hard landing could quickly solve this inflation problem…
Shouldn't interest rates show up in prices via costs? If the interest rate rises from 2025 to 2026 but nothing else changes, shouldn't care lease payments go up, because the leasing company needs higher prices to break even?
To be sure, if you use lease prices instead of car prices to measure car inflation, you don't need to do any further adjustment-- the interest rate is captured in the lease price. But inflation will have caused, indirectly, an increase in the consumer price index thus measured. And so with any good that takes capital as an input.
Are there measures of the inflation rate for different demographics? Different groups such as 20 somethings starting families and 60 somethings starting retirement clearly have different baskets of goods...
For example, I hear from young folks that buying a first home in most of the country is now essentially impossible without available damily wealth or 0.1% income. Their life satisfaction and trajectory is much more dependent on housing, including interest, costs than for the majority of rising retirees who typically already own their own home...
The average person owns 3 houses in their lifetime. As long as they are able to transfer their equity from one house to another, the inflation component is lower.
That's a very interesting post. If I may add a few comments to the discussion:
When considering whether asset prices should be included in the CPI, there are some theoretical arguments in favor of doing so (e.g., Alchian and Klein, 1973). Therefore, incorporating asset prices into the CPI might better capture the true change in the purchasing power of money over time.
However, there are significant practical challenges to implementing this approach. First, it would require including a wide range of assets beyond just stocks and real estate, such as small businesses, art, and even human capital, which are difficult to accurately value. Second, determining the appropriate weights for these assets in the index would be far more complex than the straightforward process of using expenditure surveys for a standard CPI.
Empirical studies, such as the one conducted by Filardo (2000), have found little evidence that including asset prices in the CPI would lead to better policy outcomes in the United States. While some researchers argue differently (e.g., Goodhart, 2001), I think that the overall literature does not provide strong support for the idea that incorporating asset prices into the CPI would significantly improve economic performance.
Interestingly, in the years that predate the Great Depression, the Fed used the Generalized Price Index, which somewhat included asset prices (Orphanides, 2003). This index pointed to elevated inflation and presumably contributed to the tightening that predated the crisis.
Clearly, if the primary concern is accounting for future inflation, there are already several established measures of inflation expectations, such as surveys of professional forecasters and consumers, which can provide valuable insights without the need to directly include asset prices in the CPI.
The paper by Summers et al., is an attempt to explain the contemporary level of a heuristic consumer confidence measure in terms of two 'state variables', namely, the BLS CPI and the difference between U-3 (unemployment rate) and an unobservable statistic referred to as the 'neutral rate of unemployment'. The fit to the data is not particularly noteworthy after 2019-Q4 for either the BLS CPI measure or the modified CPI measure proposed by the authors. Is the authors' modified CPI measure valid? John seems to think so, but the BLS apparently doesn't concur.
Consumer sentiment declines when the FOMC acts to counter low unemployment rates. This has been its historical pattern in monetary policy implementation. The effect is seen in the charts that Summers et al. include in their paper.
The paper by Fama and French does not explicitly address residential housing (real estate) and it does not give any indication that the hypothesized price model, which is based on the discounted value of an infinite stream of 'net of cost' rents, represents the asset value based on the implied owners' implicit rent a la the BLS CPI owner-occupied housing index construction. 'Net of cost' rent, in the Fama-French model, almost surely refers to economic rent from investment in real estate--typically commercial or industrial or productive agricultural real estate. Their paper is heavy on statistical curve-fitting, but for the purposes of this blog's discussion Figure 1 of the Fama-Farmer paper provides the important takeaway -- namely, that the asset price, P_{t}, reflects the formation and dissipation of rational bubbles, particularly pre-2006 (bubble formed) and post-2006 and pre-2010 (bubble collapsed), and then again post-2020 (bubble formed). The time variation in 'net of cost' rents, {R_{t}}, rises over time, as may be expected in a growing economy, but shows no signs of the presence of rational bubbles. What we can conclude from this is that an asset price component in the consumer price index would result in wide variation in U.S. monetary policy as that policy attempts to deter rational bubbles or recover from the collapse of rational bubbles that policy successfully deterred. The Volcker FOMC's performance in 1981-83 is indicative of the damage to the private economy that would inevitably arise from a consumer price index which included an asset price index component subject to rational bubble episodes.
C. Romer's paper on the unemployment index comparison between the pre-1939 index-construction and the post-1948 index-construction serves to demonstrate the necessity of having strongly-founded statistical indices to represent the state of the private economy.
This blog post is an excellent example of commentary which produces thought-provoking examination of concepts that step outside the reader's comfort zone. Good on you, John!
"But “asset price inflation” isn’t necessarily wrong, it’s just an answer to a different question. If you want to know the cost of providing for a lifetime of consumption, higher asset prices and lower real interest rates mean that cost has risen."
The cost of a lifetime consumption stream has risen for someone who doesn't own the assets, for the asset owner the same consumption stream has gotten cheaper. It's a transfer not inflation.
That's for existing assets, for creation of new assets the lower required return makes that cheaper: you sacrifice less of the consumption stream that the asset will yield. That is the point of monetary policy encouraging high asset prices, it encourages real investment by making that investment relatively cheaper.
I wonder if the actual annual cost for owner-occupied housing is the rental rate that includes the eventual appreciation of housing. Therefore, if house prices increase, there is no net increment to the rental cost of housing, and this is true even for people with mortgages.
Question: How much lower do you think inflation would have been if one assumes that fiscal expenditure had been the same 2020-2023 but there had been NO Quantitative Easing.
Recently Sen. Bernie Sanders raised the discussion about lowering the working hours to 32 instead of 40. How should we analyze the impact of this possible policy in the grumpy economist view?
Just some unsophisticated thoughts. Rents have gradually gone up due to higher borrowing cost for multifamily and higher taxes and maintenance fees for investor-owned. Using the current measure, core CPI will stay high for a while. If Fed started to cut rates, while it’ll ease up some borrowing cost pressure, it could bring up housing cost, which could cause the old measure even higher. Not to mention using the future consumption measurement. Some argued that we’re slowly absorbing the one time price shock due to government stimulate policies. But I’m afraid the absorbing speed is too slow (given that Fed printed more money to save SVB and Signature) that only a hard landing could quickly solve this inflation problem…
Makes sense, well said.
Shouldn't interest rates show up in prices via costs? If the interest rate rises from 2025 to 2026 but nothing else changes, shouldn't care lease payments go up, because the leasing company needs higher prices to break even?
To be sure, if you use lease prices instead of car prices to measure car inflation, you don't need to do any further adjustment-- the interest rate is captured in the lease price. But inflation will have caused, indirectly, an increase in the consumer price index thus measured. And so with any good that takes capital as an input.
This was really good. Lots to think about.
Are there measures of the inflation rate for different demographics? Different groups such as 20 somethings starting families and 60 somethings starting retirement clearly have different baskets of goods...
For example, I hear from young folks that buying a first home in most of the country is now essentially impossible without available damily wealth or 0.1% income. Their life satisfaction and trajectory is much more dependent on housing, including interest, costs than for the majority of rising retirees who typically already own their own home...
The average person owns 3 houses in their lifetime. As long as they are able to transfer their equity from one house to another, the inflation component is lower.
That's a very interesting post. If I may add a few comments to the discussion:
When considering whether asset prices should be included in the CPI, there are some theoretical arguments in favor of doing so (e.g., Alchian and Klein, 1973). Therefore, incorporating asset prices into the CPI might better capture the true change in the purchasing power of money over time.
However, there are significant practical challenges to implementing this approach. First, it would require including a wide range of assets beyond just stocks and real estate, such as small businesses, art, and even human capital, which are difficult to accurately value. Second, determining the appropriate weights for these assets in the index would be far more complex than the straightforward process of using expenditure surveys for a standard CPI.
Empirical studies, such as the one conducted by Filardo (2000), have found little evidence that including asset prices in the CPI would lead to better policy outcomes in the United States. While some researchers argue differently (e.g., Goodhart, 2001), I think that the overall literature does not provide strong support for the idea that incorporating asset prices into the CPI would significantly improve economic performance.
Interestingly, in the years that predate the Great Depression, the Fed used the Generalized Price Index, which somewhat included asset prices (Orphanides, 2003). This index pointed to elevated inflation and presumably contributed to the tightening that predated the crisis.
Clearly, if the primary concern is accounting for future inflation, there are already several established measures of inflation expectations, such as surveys of professional forecasters and consumers, which can provide valuable insights without the need to directly include asset prices in the CPI.
John, what do think of the view that Owners' Equivalent Rent should be excluded from the consumer price index as other countries do? Alan Reynolds has made this argument (see, for example: https://www.cato.org/blog/we-are-still-measuring-inflation-all-wrong-2?utm_campaign=Cato%20Today&utm_medium=email&_hsmi=295932181&_hsenc=p2ANqtz-9PJQsh2f29CBm_Bn1w2RzEUxu6tvnkdKSAZtfxLA2ZhrvXnTQeoq-xuQ6-CDsz6InjK3TxVarcD_6QHfUlG-k2Kioiww&utm_content=295932181&utm_source=hs_email). He says it is a dubious guestimate, accounts for too large a share of the index, and that without it the inflation reading would be much lower.
The paper by Summers et al., is an attempt to explain the contemporary level of a heuristic consumer confidence measure in terms of two 'state variables', namely, the BLS CPI and the difference between U-3 (unemployment rate) and an unobservable statistic referred to as the 'neutral rate of unemployment'. The fit to the data is not particularly noteworthy after 2019-Q4 for either the BLS CPI measure or the modified CPI measure proposed by the authors. Is the authors' modified CPI measure valid? John seems to think so, but the BLS apparently doesn't concur.
Consumer sentiment declines when the FOMC acts to counter low unemployment rates. This has been its historical pattern in monetary policy implementation. The effect is seen in the charts that Summers et al. include in their paper.
The paper by Fama and French does not explicitly address residential housing (real estate) and it does not give any indication that the hypothesized price model, which is based on the discounted value of an infinite stream of 'net of cost' rents, represents the asset value based on the implied owners' implicit rent a la the BLS CPI owner-occupied housing index construction. 'Net of cost' rent, in the Fama-French model, almost surely refers to economic rent from investment in real estate--typically commercial or industrial or productive agricultural real estate. Their paper is heavy on statistical curve-fitting, but for the purposes of this blog's discussion Figure 1 of the Fama-Farmer paper provides the important takeaway -- namely, that the asset price, P_{t}, reflects the formation and dissipation of rational bubbles, particularly pre-2006 (bubble formed) and post-2006 and pre-2010 (bubble collapsed), and then again post-2020 (bubble formed). The time variation in 'net of cost' rents, {R_{t}}, rises over time, as may be expected in a growing economy, but shows no signs of the presence of rational bubbles. What we can conclude from this is that an asset price component in the consumer price index would result in wide variation in U.S. monetary policy as that policy attempts to deter rational bubbles or recover from the collapse of rational bubbles that policy successfully deterred. The Volcker FOMC's performance in 1981-83 is indicative of the damage to the private economy that would inevitably arise from a consumer price index which included an asset price index component subject to rational bubble episodes.
C. Romer's paper on the unemployment index comparison between the pre-1939 index-construction and the post-1948 index-construction serves to demonstrate the necessity of having strongly-founded statistical indices to represent the state of the private economy.
This blog post is an excellent example of commentary which produces thought-provoking examination of concepts that step outside the reader's comfort zone. Good on you, John!
Such an important and misunderstood point that a rise in a price is a change in relative prices. But do ALL prices really rise when we have inflation?
"But “asset price inflation” isn’t necessarily wrong, it’s just an answer to a different question. If you want to know the cost of providing for a lifetime of consumption, higher asset prices and lower real interest rates mean that cost has risen."
The cost of a lifetime consumption stream has risen for someone who doesn't own the assets, for the asset owner the same consumption stream has gotten cheaper. It's a transfer not inflation.
That's for existing assets, for creation of new assets the lower required return makes that cheaper: you sacrifice less of the consumption stream that the asset will yield. That is the point of monetary policy encouraging high asset prices, it encourages real investment by making that investment relatively cheaper.
John:
Excellent post.
I wonder if the actual annual cost for owner-occupied housing is the rental rate that includes the eventual appreciation of housing. Therefore, if house prices increase, there is no net increment to the rental cost of housing, and this is true even for people with mortgages.
One question and two comments:
Question: How much lower do you think inflation would have been if one assumes that fiscal expenditure had been the same 2020-2023 but there had been NO Quantitative Easing.
You might be interested in my March 2021 Substack post which forecast a serious, if not existential inflation problem. I believe I was alone in making such a prediction. https://charles72f.substack.com/p/aint-nothin-but-a-party
I think it is unfair to blame all of our problems on "Bidenomics." Most of the money printing took place under Trump.
Recently Sen. Bernie Sanders raised the discussion about lowering the working hours to 32 instead of 40. How should we analyze the impact of this possible policy in the grumpy economist view?