At the beginning of the Depression, more than a fifth of all Americans worked on farms. Between 1929 and 1932, these people saw their incomes cut by somewhere between one-third and two-thirds, compounding problems that farmers had faced for years. Agriculture had been a victim of its own success. In 1900, it took a large portion of the U.S. population to produce enough food for the country as a whole. Then came a revolution in agriculture that would gain pace throughout the century--better seeds, better fertilizer, better farming practices, along with widespread mechanization. Today, 2 percent of Americans produce more food than we can consume.
What this transition meant, however, is that jobs and livelihoods on the farm were being destroyed. Because of accelerating productivity, output was increasing faster than demand, and prices fell sharply. It was this, more than anything else, that led to rapidly declining incomes. Farmers then (like workers now) borrowed heavily to sustain living standards and production. Because neither the farmers nor their bankers anticipated the steepness of the price declines, a credit crunch quickly ensued. Farmers simply couldn't pay back what they owed. The financial sector was swept into the vortex of declining farm incomes.
You just can't do macro like that. It all goes wrong from the very beginning. The fall in the price of gizmos, for a given quantity of gizmos sold, and for given prices of non-gizmos, reduces the real incomes of gizmo producers, but increases the real incomes of non-gizmo producers by an equal amount. If the price of gizmos falls by $1 each, and one million gizmos get sold each year, gizmo producers have $1 million less income, but non-gizmo producers now have an extra $1 million to spare which they can spend on something else.
Stiglitz is focused on telling an aggregate demand story, and Rowe calls him out on that one. When you change the terms of trade between two sectors, that in itself is not a shock to AD. It can be the basis of a shock, but you need to tell a more complicated story, about relative spending propensities or somesuch. I won't go into detail--it's Stiglitz's job, not mine.
(Scott Sumner calls out Stiglitz on the same issue, and more. But Nick got there first.)
Instead, let me try to tell a PSST story. That won't be easy, either, but here goes.
Let us stipulate that in the long run, resources shift out of sectors where productivity rises faster than demand and into sectors where demand rises faster than productivity. (Also, there may be a shift toward leisure, which is a superior good.)
In the short run, the marginal revenue product of labor in the gizmo industry falls to nearly zero (even as the average product of labor in that industry goes up). Think of a farmer who acquires a tractor. Now he can produce more by himself than he could previously with four farmhands. Higher average productivity for the farmer, but lower marginal productivity for the farmhands. (On the one farm, the fixed amount of land is what keeps marginal productivity low. In agriculture as a whole, inelastic demand is the factor that limits marginal revenue product.)
What about the consumers of gizmos who do not have to spend as much on gizmos? In the short run, their wants are not connected to the skills of the workers freed from gizmo production. (Note that the consumers might want more leisure, and again this may even be true in the long run.) Entrepreneurs have to figure out a way to satisfy consumer wants using the resources freed from the gizmo sector. This is the process that takes time. Nobody knows what the consumers want (or might want if it were invented). Nobody knows the potential uses of unemployed workers. Instead, new products, services, and production methods have to be discovered, by trial and error. I would call this the recalculation problem, while Peter Howitt would call it the coordination problem. Again, see our articles in Capitalism and Society.
In the long run, new patterns of specialization and trade have to be established. But we do not know what those will be. Government does not know, either, so its fiscal policy is a very clumsy instrument.
On the latter score, Stiglitz is quite disingenuous. He says that wartime spending magically cured unemployment by moving people from farms to the military and then to urban factories. I do not think that story will hold up. I think that a lot of what went on is that between 1930 and 1950 quite a few manual workers aged out of the labor force and were replaced by a new vintage more suited to clerical work that was a better fit for the technology available. You could not have slotted the work force of 1930 into the jobs of 1950. Conversely, if you had kept the jobs of 1930, you would have been way inside the production possibility frontier and badly under-utilized the skills of the work force of 1950.
Stiglitz also complains that in the most recent recession 700,000 government workers lost jobs. Well, first of all, 7 million private sector jobs were lost. Second, not a single government job had to be lost due to budgetary restraint--a small cut in pay could have saved all of those jobs.
Anyway, Stiglitz refers to research that he has been doing on the Great Depression, and the role of higher productivity in the farm sector as a causal factor. I look forward to reading more about this, because I think it fits at least as well with a PSST story as with an AD story.