I think it is worthwhile to spell out explicitly that attacking GDPism is a much more limited goal than attacking productivism. A productivist can easily share many or most of our criticisms of GDP as The One True Proxy without sharing our view that Growth^TM is not unequivocally good.

GDP is a useful indicator - it would be difficult to do economic planning without that indicator. As the diary points out, there are several caveats in interpreting it: Rebuilding a house knocked down by a storm increases GDP, because the house didn't count when it was standing up; raw materials are presumed to be worthless; and those informal favours and pleasures of other people's company which are judged to be too trivial to count as income for tax purposes are discounted.

The two former could actually be resolved by using the net domestic product instead of the gross. The net domestic product includes amortisation of assets, which includes depletion of non-renewable raw materials. However, amortisation of assets is for the overwhelming part a purely internal process on the balance sheet of an economic actor, so getting reality-based data for it is a real PITA.

But the fundamental problem with GDP is not that it's a bad indicator. The fundamental problem with GDP is actually not a problem with GDP at all. It is the universal problem inherent in elevating any single indicator to The One True Proxy. There is no law in econometrics more ironclad than the observation that the value of an indicator as a proxy is inversely proportional to the amount of political attention being paid to it.

Not because politicians are stupid and pick the wrong indicators, but because once you become fixated on The One True Proxy, you are liable to use it as a proxy for variables that it really cannot reliably reproduce - such as using GDP as a proxy for overall societal achievement. If you become fixated on The One True Proxy, you are also much more likely to enact policies that target the proxy, rather than the causal mechanism that made the proxy useful in the first place. Lowering the fever is much easier than curing the patient - and breaking the thermometer is even easier.

A second technical point is that many common uses of GDP implicitly assume that price can be used as a proxy for value. The theoretical justification for this only really makes sense under conditions that do not obtain in any modern economy - namely, if consumer wants are independently established and well-behaved, and producers maximise discounted profits (and there is a well-behaved financial sector which arrives at a uniform discount rate). In a planned economy GDP is a different sort of beast (still useful, just for different things). And all modern economies are substantially planned economies, although the sovereign may not be the entity doing the planning.

Moving beyond the technical problems with GDP, there is another technical problem with national accounts, in that financial assets are counted as capital. That capital and financial assets are not equivalent is actually another point that might confuse new readers: Capital is the physical means of production - financial asssets are promises of command over the physical means of production, or their products, at some future date. The two are only interchangeable if the financial sector is well-behaved. Which it stops being almost immediately when policymakers start assuming that it is, because keeping the financial sector well-behaved requires constant political vigilance.

Further, and staying in the national accounts, production of goods and services that are free at the point of delivery is, by definition, assumed to make no profit. There is a technical reason for this: Goods and services that are provided free of charge are assigned a virtual price, which is then booked as production income for the producer of the free good or service, as a subsidy (expense) for the producer and as a gift (income) for the consumer of the service. This virtual price is, by definition, equal to the cost of production.

While this does sort of make sense, it means that there is a bias against free stuff in the GDP calculation. Imagine that, instead of having a public hospital system to treat patients, we outsource hospital services to a wholly owned subsidiary of the national retirement fund, pay them out of income taxes (which are secondary allocation, and thus do not enter into the computation of GDP), and use the profits to defray the outlays on retirement benefits that would ordinarily accrue from general revenue. The hospitals are the same. The staff is the same. The patients are the same. The cash that actually changes hands is the same, except for some accounting technicalities purely within the consolidated sovereign balance sheet and cash flow. But this game of three-card monte has magically increased GDP by the full profits that the national retirement fund accrues from owning hospitals...

All of the above exposition is a purely technical one that any student of economics, regardless of ideology, should be able to follow. So, as should be obvious, GDPism has far more problems than productivism: GDPism is a subset of productivism, which means that all the diary's excellent criticisms against productivism apply with full force to GDPism, on top of the technical problems outlined above.

- Jake

Friends come and go. Enemies accumulate.