Conceptually, it's useful to think of a house as an investment good that produces housing services (shelter, living space, etc.). These housing services are what you get when you rent a house. If you're a home owner, you don't pay rent explicitly, but it really is as if you had rented out your house to yourself and you were paying rent to yourself. Therefore the cost of housing services is best estimated through rent prices.
The other thing is the cost of the house itself, not the housing services. This includes the purchasing cost and the cost of repairs that you make periodically. The repairs are not a problem but the purchasing cost would need to be spread over the useful life of the house. If the house is bought on credit then the buyer will also incur funding costs (i.e. in the form of interest payments), but these can't be considered costs that are related to housing. They are separate. It's like if you buy a car on credit you'll pay more than if you paid for it with your savings, but that doesn't mean the car is more expensive.
CPI includes housing services and repairs. As far as I know, it doesn't include houses, probably because they're considered an investment good, while CPI is concerned with consumer goods. At any rate, mortgage payments are not a good estimate of how much houses cost. The reasons are interest, which is a different cost, and the fact that principal payments are arbitrary as they vary depending on things like mortgage length, and so on.
Anyway, I understand that many people feel strongly about CPI and inflation measurements, but while cost accounting is not rocket science it's not completely straightforward either, and most of the time criticism of CPI by layman people turns out to be misguided.
Thanks for your time. Renting looks like simpler to estimate, but it confuses me as well.
> Every month, the rental prices paid by the households in the sample are compared with the rental prices they paid the previous month
When rent control caps rent increases this becomes a lagging indicator. Why don't they try to estimate market rent?
This feels similar to my problem with the owner cost estimate. Consider a hypothetical about cars - in a world where everyone buys new cars with 8 years of financing, suddenly there is a car shortage (chips or whatever). The price of a new car doubles. As very few owners need to buy a new car every year, on average their costs are about the same as last year despite a 100% increase in the cost of a new car.
The other thing is the cost of the house itself, not the housing services. This includes the purchasing cost and the cost of repairs that you make periodically. The repairs are not a problem but the purchasing cost would need to be spread over the useful life of the house. If the house is bought on credit then the buyer will also incur funding costs (i.e. in the form of interest payments), but these can't be considered costs that are related to housing. They are separate. It's like if you buy a car on credit you'll pay more than if you paid for it with your savings, but that doesn't mean the car is more expensive.
CPI includes housing services and repairs. As far as I know, it doesn't include houses, probably because they're considered an investment good, while CPI is concerned with consumer goods. At any rate, mortgage payments are not a good estimate of how much houses cost. The reasons are interest, which is a different cost, and the fact that principal payments are arbitrary as they vary depending on things like mortgage length, and so on.
Anyway, I understand that many people feel strongly about CPI and inflation measurements, but while cost accounting is not rocket science it's not completely straightforward either, and most of the time criticism of CPI by layman people turns out to be misguided.