I'd like to add a paper I published in 1988. There I presented a model in which unemployment arises from a drop in the demand for the output of a particular sector. The unemployed workers could consider trying to retrain or relocate, or might instead decide to wait it out in hopes that the demand for their specialized skills will come back.
If instead of a drop in the demand for sector A there was a boom in the demand for sector B, it is true that some workers in sector A might choose to retrain or relocate, and be temporarily unemployed as a result. But the key kind of unemployment that I think this sort of model describes-- waiting for an opening in the particular area in which you've specialized-- is caused by drops in demand, not increases.
Asset bubbles create value out of thin air. Price trends develop that deviate sharply from historical norms, and eventually a new, albeit temporary, norm is established. As a result, asset bubbles make the bubble-asset look like a much better investment than it will eventually turn out to be in the long term.
That story helps answer the question of why booms create jobs, but it does so in what amounts to a one-sector model. That is, over-optimism leads to excess capital, which in turn is going to increase the demand for labor.
The Recalculation story is a multi-sector model, in which unemployment results from workers being stuck in between sectors when shifts are required. The asymmetry question is why a boom in one sector is any less of a problem than a bust in that sector. If you have one steady sector and one cyclical sector, why is shifting workers out of the cyclical sector when it is slumping any harder than shifting them in when it is booming?
I have suggested that one asymmetry is that booms are long and slow, while busts are short and sharp, so that busts create more difficult adjustment problems.
Another asymmetry that has been suggested is that a boom takes place in a single, known sector, whereas the bust releases workers into a diffuse economy. In 1849, everyone knows to rush to California. When the Gold Rush ends, to which other state should workers retreat? However, I'm not sure that, in principle, it should be harder to be absorbed back into a diffuse economy than it should to be pulled out of it to pursue a bubble.
As an aside, John Haltiwanger reminded us at lunch today that his findings are that the most dramatic restructuring takes place within an industry, rather than between industries. In the 1980-1982 recession, integrated steel mills in the United States disappeared, while smaller mills emerged instead. This sort of recalculation goes on all the time, but it seems to happen faster during deep recessions. This suggests that the challenge for the market is not to figure out which new industry will grow but to figure out which firms in existing industries need to die and which firms in existing industries are poised to expand.
One way to think of this is that there are always firms in all sectors of the economy that are somewhat inefficient and barely profitable. A recession starts when a bust in one sector knocks out all the marginal firms in that sector, but the impact reverberates onto other sectors, where marginal firms get knocked out as well. The recovery takes place in the firms that were most promising and most efficient to begin with.
I think that the most dangerous habit that macroeconomists have have been drawn into is the habit of thinking that we have a cure for unemployment. It is always easy to say, "If the government followed my policies, unemployment would be a lot lower." And it is easy to rationalize afterward why those policies fail to achieve their promised results. It is hard to come forward and say that we may not have an answer, and harder still to get anyone to listen if that is your message.