The cost of borrowing should probably be included in price indices

The construction of a price index is far fuzzier than it would appear on the outside. It seems so simple — just figure out what people buy, and track the changes in price over time. The bundle of goods which people buy is called a consumption basket. If the recorded price gets higher, you have inflation. You don’t track investments, because those investments are ultimately claims on future consumption, and you can track that directly. You want to choose a basket which provides constant utility over time

Yet, this masks considerable difficulties. What if the goods change over time? What if people’s preferred consumption basket changes, independent of changes in price? What if, for that matter, the basket changes due to changes in relative price? How should investments be treated? How about goods which have elements of being investments and goods, and could be entirely one or the other? How about if the quality of goods changes? 

These are not insubstantial questions. To answer questions which occur through time we need to be able to compare relative price, and so our choice of deflator has a massive impact on outcomes. Consider the question: have wages kept up with productivity over time? You may have seen this claim passed around online — in 1971, wages no longer kept up with productivity. As it turns out, this is entirely due to the choice of deflator, using CPI rather than PCE. (CPI keeps its basket of goods more constant, while PCE changes them. The advantage of PCE is that it takes into account people’s ability to substitute between goods. If there’s a bad harvest of apples in the Pacific Northwest and the price rises, people aren’t going to just consume the same combination of apples and oranges — instead they might shift their purchases to include relatively more oranges and fewer apples). Your choice of deflator leads to wildly different views on the true rate of real wage growth.

So back to the question of the cost of borrowing. A recent working paper, from Bolhuis, Cramer, Schulz and Summers, argues that we should return to including the cost of borrowing. The main expense which borrowing affects is buying a house, although it does also matter for cars. The trouble with houses is that it is both a consumption good, and an investment good. We don’t want to include investment goods, because they aren’t consumption at all — rather, they are a claim on future consumption goods. If we included investments, we would be engaging in double-counting. In 1983, the Bureau of Labor Statistics changed their statistical approach, and calculated it using “owner equivalent rent”. The owner of a house is in effect renting it to themselves, so you can take that as a cost. Not including borrowing costs also means that inflation is not spuriously raised every time the Central Bank increases interest rates. 

This approach solves real problems. We no longer get total nonsense, like that inflation is increased every time that the central bank raises rates. Yet, I cannot help but think it is inadequate. Lending is a service, like that of a lawyer or doctor. If there were a reduction in its availability, we would be harmed every bit as much. It is striking, to me, that it simply wouldn’t notice the collapse of banks and credit availability during the Great Depression. Given that Bernanke 1983 links this to the severity of the Great Depression, we should surely want to include it in our price indices! I find myself in agreement with the working paper authors — borrowing is a service people use too.

It may help provide a better explanation of why people are so averse to inflation. Originally, economists wanted to play nice with people. We do not like to assume that people are stupid — rather, that there is something we don’t understand. Inflation is harmful for the normal reasons — it causes price dispersion, distortions, etc. Now the consensus has swung a bit more towards “no, people are actually just retarded”. Stephanie Stantcheva has a recent working paper on this. The reason we dislike inflation is because we don’t like prices going up, ever, and we don’t put it together that our wages rising is also part of inflation. This is a middle ground. Yes, people’s dislike of inflation is a bit weird and excessive and likely influenced by cognitive biases. It is also a bit worse in practical effect than we thought. It is more expensive to buy things when interest is higher.