> So when I say that there is not an absolute point at which
> 'profitable' turns into 'unprofitable', and that "profitability is a
> complex thing, and always relative to other returns", I think it's an
> answer to your point here:
>>> Okay, but it doesn't follow automatically that disappointing rates of
>>> return to money-capital will result in a greater (system-wide) appetite for
>>> risk. I mean, if you think of a classic case of a deeply depressed economy,
>>> the rates of return are disappointing but people *flee* from risk.
> This is a difference between a period where a sector or sectors are
> expected to outperform generally disappointing other returns, and a
> period in which nothing is expected to outperform, or in which people
> value capital safety and liquidity higher.
[...]
> I don't see that the two arguments are incompatible. Clearly a lot of
> people had to believe real estate returns would outperform those of
> other potential assets for them to put their financial resources into
> it. But the gap can begin to open up either because expectations for
> real estate spontaneously rise, or because expectations for the
> alternatives fall. And the reason why the gap is sustained might be
> different from why it opens up in the first place - because asset
> market movements tend to develop momentum.
Aha - now I see what you were getting at! I was missing this point.
I guess my response is that this is correct as a *general principle*: All bubbles, throughout history, have been focused on some particular sector. So yes, by definition that means a gap has opened up between returns in this sector (whichever one it is) and returns in the other sectors - and therefore, it becomes possible *in theory* to argue that the first cause of the divergence was "too low" returns in other sectors, rather than "too high" returns in the bubble sector.
In response, I'd say three things.
First, in practice it seems very unlikely to me that a bubble in a particular sector can start simply because returns in other sectors have fallen. It just doesn't seem to match up with the typical bubble narrative. For one thing, if there was a decline in returns to other sectors, why was the bubble sector spared from this decline in the first place? Take the housing sector, for instance. In a given local real estate market, when returns to other industries fall, returns to housing almost always fall in response. Think about the local property market in rust-belt-type areas like Michigan. In the US, those areas were never touched by the housing bubble - precisely because returns in other local sectors (e.g., mfg) were too low to support bubble-like expectations for the housing market. And this type of story is even less plausible for other types of bubbles, which usually focus on some *new* fast-growing sector. Could anyone really argue that the reason people poured insane money into internet companies in the 90's was simply that other sectors were disappointing? No - it was because the internet was a genuinely promising new sector. (Although not quite so promising as to fully justify valuations.)
Second, given the point I make above, it seems that to make your argument really convincing one would have to make a case that returns in the other sectors were "too low" in some *absolute* sense. You'd have to show, for instance, that returns in the non-housing sectors in the 2000's were exceptionally disappointing. But I'm sure one could easily point to other historical periods when non-housing returns were just as low or lower than in the 2000s without producing any bubble in housing.
Third, although I still haven't gotten around to reading the whole paper, it appears to give an explanation that contradicts the scenario you evoke. It says the best predictor of the onset of a local housing bubble in the 90's was an unusually fast growth in local *incomes*. I strongly suspect that these early bubble areas were ones where late-90's technology prosperity was bringing an influx of rich people. (E.g., Silicon Valley.) In other words, the process went like this:
(1) returns to other sectors (e.g., tech) *rose* (2) this generated local enclaves where average incomes were rising very fast (3) as a result, the value of land rapidly appreciated (4) incumbent landowners reaped enormous capital gains (5) seeing this, people in other areas started to believe that the same thing would happen (or was happening) in their own area (6) this led to the creation of bubbles outside of the local wealth enclaves (7) as massive house price appreciation spread around the country, capital markets refused to question the rationality of it (8) this attracted more capital into the mortgage market, thus feeding the later, crazier stages of the bubble (e.g., in exurban California or Nevada)
So although I acknowledge that your scenario is theoretically consistent, I think the opposite scenario seems much more plausible: It was *rising* returns in the real economy that triggered the initial episodes of house price appreciation that then formed the psychological basis of the bubble. The capital markets absorbed the psychology and *responded* to price growth, propagating the bubble further, even after the initial shock had dissipated.
SA