It matters how you measure things. When economic data construction was in its infancy, Simon Kuznets, the father of National Income Accounting, was unhappy about the decision to include government expenditure in Gross Domestic Product. Kuznets felt government output was an intermediate product that fed into consumption and investment, which are the ultimate output of an economy. Political considerations during the Great Depression of the 1930s dictated the outcome of this debate, so the government is included in GDP. But imagine how different economic policy would be if every time the size of the government increased, our estimates real output would fall?
Something similar is going on with inflation statistics. The adoption of various forms of both explicit and implicit inflation targeting by dozens of central banks around the world has made the measurement of inflation more than just some abstruse subject for economists and government statisticians. Today, decisions about how we compute inflation statistics can have a direct impact on monetary policy decisions.
The big question for inflation measurement – as big as the decision about how to treat government expenditures in calculating GDP – is whether to include owner-occupied housing in aggregate consumer price statistics. And if the answer is that we should, which I think it is, then how should we do it?
Before getting to the details of the issue, let’s step back and recall both what price indices are supposed to measure and why 21st century central bankers are focused so intently on them. Textbooks (including my own) define inflation as a sustained rise in the general price level. It is a situation in which the prices go up on.
There are a variety of ways we can measure inflation. The most common is a consumer price index, or CPI. A CPI is designed to answer the following question: “How much more would it cost today to purchase the same basket of goods and services that was bought at some fixed date in the past?” In order to reduce the systematic upward bias created by using fixed weights, the “date in the past” is changed relatively frequently – every two years in the United States.
Consumer price measures are put to a variety of uses. For example, government tax and expenditure programs are typically indexed to them. In addition, the CPI provides a guide for the adjustment of wages and salaries. The idea is that if you make adjustments based on the CPI you are somehow keeping purchasing power from changing. But the purchasing power of what? Well, that depends on what prices are included in the index, and we’ll come back to it in a moment.
Importantly for public policy, the inflation goals of monetary policymakers are usually evaluated using a CPI. (Granted the Federal Reserve’s public comments tend to refer to a more obscure index based on personal consumption expenditure, but I’m optimistic that will eventually change.) There is wide agreement that even modest amounts of inflation are costly, reducing the efficiency of the economy as a whole and making long-term planning more difficult for individuals.
Returning to the question of measurement, not surprisingly how you measure prices depends on why you are doing it. And to get back to where we started, this is even more true for owner-occupied housing than it is for most goods and services. If, for example, the purpose of measuring housing prices is to convert residential construction expenditure from nominal to real, then the appropriate price is the sale price of new homes. But if you are trying to figure out the price associated with keeping a household running at some benchmark level, then that’s not the right number. But what is depends quite a bit on the way in which you think about housing.1
National income accountants computing consumption – the C in the identity Y=C+I+G+(M-X) – need to include housing and have a method that is designed to treat renters and owners equally. The idea is that if all owners were to become renters, or all renters were to become owners, GDP should not change. To make sure that evolution in the rate of home ownership doesn’t influence measured consumption, government statisticians pretend that owner occupiers rent their homes from themselves. As a result, in the national accounts homeowners receive additional income equal to this rent, and on the expenditure side of their balance sheet, they are given an equal amount of additional consumption. This flow of housing services to owner occupiers is known as imputed rent and its price is based on the rental prices for a sample of homes that roughly matches what is owned.
Construction of the U.S. CPI mirrors the national income accounts treatment of owner-occupied housing. Owners are assumed to rent their homes from themselves, creating a category called Owner-Equivalent Rent (OER). And because more than two-thirds of U.S. households own the house that they live in, OER’s weight in the CPI is substantial; it accounts for 23.8% of the headline CPI and 30.8% of the traditional core CPI that excludes food and energy. Not surprisingly, how you measure something this important is a very big deal.
In recent years, as the sale prices of existing homes rose sharply, rents languished. This brought close scrutiny to the decision (originally made around 1981) to use OER in the U.S. CPI. In the short-run, the inverse relationship between sale prices and rents makes sense. Individuals, not wanting to miss out on the housing boom, leave the comfort of their rentals and rush to purchase their own homes.2 In the long run, however, the user cost of owning has to equal the price of renting, but the deviations from this equilibrium can be large and long-lived. Looking at the U.S. since 2000, we see that OER has risen roughly 26% (cumulatively), while the government’s index of the price of existing homes increased 80%. Taking these numbers at annual rates, the difference is 3.17% versus 8.48% on average per year for seven and one-quarter years. To be clear, sale prices of new and existing homes rose more than 5% per year faster than rents since the turn of the millennium.
Doesn’t this suggest that OER is understating inflation? There is an argument that, rather than including observed rents, the existing price of a home should be in the consumer price index. The reasoning goes like this. A homeowner’s foregone income is based on the current sale price of the house. This means that the implicit rent, which is essentially a measure of the opportunity cost of owning rather than renting, should be based not on the rental market but on the price of the house.
Making this change in the consumer price index would make an enormous difference. To see how big, start with the fact that since 2000, the U.S. headline CPI has risen at an average annual rate of 2.75%, while the traditional core CPI has gone up 2.20% per year on average. If government statisticians had been using the price of homes sold rather than rents, consumer price inflation would have registered an annual increase of something like 4% per year – roughly one and one-quarter percentage points higher. And core CPI inflation would have been something like 3.8%; that’s more than one and one-half percentage points above the official reading. Had these been the inflation readings, it’s hard to imagine the Fed keeping their federal funds rate target below 2% for three years.
There are other methods for treating owner-occupied housing in price measures. For example, the Harmonized Index of Consumer Prices (HICP) that forms the basis for the ECB Governing Council’s definition of price stability and effectively serves as their target, ignores owner-occupied housing entirely. (Interestingly, when the Bank of England’s official inflation target changed from their retail price index, similar to the U.S. CPI, to an HICP, they reduced their target from 2.5% to 2%. It matters how you measure inflation!)
Some countries base their measure of owner-occupied housing prices on acquisition costs, usually without land. Australia is an example.3 This method treats housing in the same way as durable goods like televisions and automobiles. The procedure is to track the price of the new ones and weight them by proportion of current expenditure that goes to the purchase. While the sale price tends to be more volatile than rent, the weight is roughly half as much.4
Maybe I’m just defending the home team, but after thinking long and hard about this question, I’ve concluded that the U.S. has this one right and that the rental-equivalence measure balances all of these things just about right. OER really is the current opportunity cost of the house. It measures the income I could receive if I were to vacate my current living space and rent it to someone else.
I find all of the arguments against the use of imputed rent unpersuasive. Housing may be an investment, but the dividend I receive is the housing service that I use. The long-term capital gain is miniscule, averaging roughly 20 basis points per year in the United States during the 100 years of the 20th century. So if it’s an investment, it’s a bad one.
The second primary argument against the inclusion of owner-occupied housing in the price index is that owners are hedged against these price increases. This is surely true, but why single out housing as the one place where we will take account of this hedging? We don’t remove asparagus because asparagus farmers are hedged against changes in the price of asparagus. More seriously, we don’t account for the use of financial assets to hedge other consumption risks, so why single out housing? I could, for example, hedge changes in my consumption of energy or medical services by purchasing equity in the suppliers. Does that mean that the aggregate price index should exclude these? I think not.
Price stability is about helping people make their long-term plans. Keeping inflation low means that individuals can do their retirement saving without having to worry about the difference between real and nominal interest rates, something the vast majority of people don’t understand. And that means including housing to the extent that it is actually used.
1 For a general discussion of the treatment of housing in consumer price indices see Erwin Diewert “The Treatment of Owner Occupied Housing and Other Durables in a Consumer Price Index,” Department of Economics, University of British Columbia, Discussion Paper No. 03-08, November 2003.
2 For a discussion of these dynamics see Richard Peach and Jonathan McCarthy “Recent Housing Trends: Their Effects on Rent Inflation and Its Measurement of the CPI,” Federal Reserve Bank of New York, forthcoming.
3 For a survey of the treatment of owner-occupied housing in consumer price measurement around the world see the Proceedings of the Irving Fisher Committee Workshop on CPI Measures, April 2006.
4 See Diewert op. cit. for a discussion of why this is inevitably the case.