- 05:30 You don’t ask is a model good or bad. You ask is it good for a particular question. (For example, What effect will school vouchers have on inequality? or Why do politicians “race to the center” in a two-party system but not in a three-party system? or Why was a small-but-predictable voting bloc of Evangelical voters so influential in the election of George W. Bush?). At the same time, it’s useless and possibly misleading for other questions.
- 04:30 Exogeneity and endogeneity. There’s always a tension in modelling, between deriving something and assuming it. There are values in both ways of going. If you start with a non-basic assumption, you’re closer to where you want to get; you can focus on the parts that you want. On the other hand, there is the risk that the underlying reasons might make what you do with the model unattractive.
- 03:30 When you’re thoroughly embedded in a theory [like Arrow-Debreu general equilibrium], you’re as aware of the shortcomings as you are of the successes.
- 06:15 The government may have an advantage or a disadvantage vis-à-vis private markets. For example governments have difficulty making commitments. (Private enterprise, having the government to enforce contracts, has less difficulty.) On the othe rhand, the government has the ability to do things like collect taxes, and so impact large numbers of people with very low administrative costs. So that goes either way.
- 06:45 A full-blown theorem by Geneakoplos & Polemarchakis (1 2 3 4 5) stood the welfare theorem in the Arrow-Debreu model on its head. In incomplete markets, the equilibrium is never Pareto optimal.
(Diamond says “the government can do better” but I think he might mean an omnipotent social organiser could do better.)
- 15:00+ Talking about why he built his search model the way he did.
- If you took out all of the typical “Keynesian” elements (sticky prices, …), could you still get a picture of aggregate demand that looks Keynesian? The answer is yes.
- 17:30 Uncertainty in economics is unlike quantum-mechanical uncertainty.
- 19:00 Increasing returns to scale throughout the circle of production. (Does this have anything to do with growth?)
- 24:00 Changes in unemployment are more complicated than just a two-way unemployed to employed and back. In fact there are 3² paths, including:
- e → i. employed to out-of-the-labour-force. Long vacations, retirement, I’ve-saved-up-enough-and-don’t-want-to-work-for-a-while, I’m-going-to-quit-and-write-a-book.
- i → u. inactive to unemployed. The unemployment rate of American men during the Great Recession was 2.3 percentage points higher than the unemployment rate of American women during the same period. The New York Fed says that’s partly because a lot of men who were comfortably out-of-the-labour-force lost wealth (in the stock market crash) and started looking for work again — but couldn’t find work. In other words, new men who hadn’t been looking for jobs before, started looking for them — and that raised the unemployment rate.
- e→e. A lot of people change jobs without ever being unemployed. They may take a different job at the same company, or switch companies. What does that do to the unemployment rate?
- i→e. ”I wasn’t actively looking for work, but then I got an offer and took it.” So the number of unemployed people stays the same, and one of the open positions has now been filled by someone who wasn’t even counted in the unemployment statistics.
- 25:45 “Job availability is the small difference between two large flows.” (job destruction and job creation each dwarf the job availability numbers)
- Did you know that Harvard & MIT professors are ideological opponents of the Minneapolis Fed?
- 24:24 Picture of employment inflows and outflows across countries. Macroeconomists think that higher turnover leads to a more efficient economy.
- 25:45 Picture of the breakdown of the job separation numbers. (Quits + Layoffs + Retirements + Other) Job separation != job destruction because about half of separating employees are soon replaced with other workers.
- Beverage Curve. The jokes write themselves.
- 31:00 Sometimes a hiring occurs without a vacancy posted first. (This is like at 24:00 where he said that unemployment and employment numbers are more complicated than you might assume.) In fact, maybe 20% of jobs are filled without the statisticians noticing a vacancy ever existed — and another 20% of jobs are filled without a vacancy ever actually existing at all.
- 34:00 Maybe the high U.S. unemployment rate (unemployment above Okun’s Law) doesn’t signal a dysfunction of the economy. Maybe the sectoral pieces of the economy are behaving just as usual, but different sectors have different Okun’s Law parameters — and the sectors with higher Okun’s constants are experiencing the greater shocks. (Okun’s constant = 30% historically has related changes in GDP to changes in employment)
- 35:00 Aggregate numbers didn’t show anything was wrong with the capital markets after the collapse of Lehman Brothers. And the Minneapolis Fed got it wrong for that reason. But if you disaggregated, you would have noticed a problem.
Posted on Sunday, 6 November 2011