Donnerstag, 25. Juni 2009

GDP Imputations

Following up on my previous post, I started wondering about the logic of behind imputations as part of GDP calculation.

In case you don't know what an "imputation" is:

Parts of GDP are not based on market transactions for goods and services. Instead, they are "imputed", i.e. valued "as if" there was a transaction.

The most important category is "imputed rent" for owner-occupied housing: If you live in your own house, you are assumed to pay rent to yourself, which is part of GDP. No kidding!

In the US, "rent for housing" is part of "services consumed", and makes up 10 % of GDP. As Americans are into home ownership, I assume that 2/3 or more of this are imputed payments as opposed to actual rent payments.

The interesting thing: In spite of the housing price crash, "rent for housing" has increased 3 % in nominal terms from Q1 2008 to Q1 2009 (it increased 0.5 % in real terms, i.e. the statisticians apparently calculated average rent increases of 2.5 %).

As most of this is imputed as opposed to "real rent", the question beckons: Does this make sense, considering the situation of the US housing market?

(Roughly 15 % of US GDP is due to imputations. My guess is that at least half of this is imputed rent. German GDP also includes imputations. However, I was unable to find any details as to the extent. Imputed rent should be less important than in the US, as Germans are more likely to be renters than owners)

(Another side note: For some reason, there is no imputed rent on cars. So if I understand correctly, leasing a car as opposed to buying it increases GDP, because the leasing payments are counted, whereas the use of an owned car isn't. A tad bizarre, isn't it?)


  1. Nice to know how you can cheat the statistics. Just "impute" a value you like. A different view can be found here:

  2. Hey there!
    I am so happy google showed your blog in my search! I am an economics student and my professor just went over GDP imputation today. I had the exact same question! If we want to double count rent, why don't we include car lease too? because for sure the dealers will include the lease payment under income if you calculate GDP using the income approach!

    Similarly, what about other things that we buy right out instead of doing the monthly installments? Example, TV from best buy, furnitures, etc.

    I find this rent imputation very baffling... I cannot believe nobody looked into this yet. As indicated in your numbers, it makes up a substantial part of a nation's GDP.