"It's true that there was strong growth after WWII, this doesn't mean that there couldn't have been stronger growth without WWII.
is the right measure depends on what were interested in. I remember
having this discussion with Anthony Evans. He mentioned a colleague of
his who said that GNP shouldn't interest us and we should use consumer
spending instead because that's a measure of the wealth of the
population. Anthony said well in that case we should use consumer
surplus. That's true for measuring the result of long run changes, for
measuring "material progress".
Output measures, like GNP and GDP
are about estimating how things are going by looking outside of consumer
goods. By looking at investment goods too as a proxy for future
consumer goods. There is a lot to be said for doing this, but it's
necessarily inaccurate. The prices of capital goods are principally
tools of entrepreneurial calculation, using them to estimate the future
is dicey. Mises discusses this problem in Human Action p.520-522, as
far as I can remember. I'm sure Post Keynesians have reams to say on
the matter. Statistics bear this out, if you look at the investment
series in the Penn World Tables, they don't predict the future of the
consumer goods series well, or the growth series well.
prices of investment is likely to be particular dubious in a situation
where the government are spending a lot of money which pushes up the
value of investment goods (existing and new). So, the output we get
isn't likely to be the same as sustainable output.
So, how is output useful in this case?"
The answer, I think, is relatively straightforward: we care about output because we care about employment. If we didn't care about employment we probably would only be interested in a consumption aggregate. Investment aggregates might be produced (or more likely capital stock estimates) as a predictor of consumption production. But it wouldn't be as important. We care about total output because we care about employment.
I would agree investment would be a funny thing to use to predict consumption - I'm not sure why anyone would do such a thing. You'd want to think about capital, not investment.