Some more links
The jobless trap--Paul Krugman
A potentially significant change to national income accounting, noted by Antonio Fatas
Destructive creativity--Paul Krugman
Is the "underground economy" resurgent"--James Surowiecki (although I've often thought that the size of the "underground economy" is exaggerated. In this case, all those construction companies that are not reporting employees and not paying payroll taxes? Are they also under-reporting their revenue? Wouldn't they have to be in order to make this work?)
What if the demand for college grauates is really declinding? FromStumbling and Mumbling
Housing bubbles and land fever--Robert Schiller
That's it so far today.
posted by Don Coffin at 11:30 AM
2 Comments:
I have a couple of concerns, as I read through your analyses (which are intelligent, thoughtful and interesting, by the way).
1. how often do you construct a hypothesis (such as goverment income inequality doesn't decrease as income rises, or personal savings is unaffected by personal income) and find no statistically significant data to reject it, and therefore not worth blogging? If you just sit for a moment and think about it, you may not remember it being as often as it really was (or, maybe you do). In any event, if you only report on the investigations that were significant, you (or dear reader) may be in danger of running afoul of confirmation bias. ('Hey, blogger SoC doesn't very often chase down a mental blind alley...')
2. Correlation and causation. You are definitely aware of the distinction, because you mention it, but it seems like you intimate (or even assert a relationship 'suggests') causation, without backing it up with a carefully-reasoned argument. I like where you're going with many of your posts, but I usually end dissatisfied with the half-finished line of argument.
9:59 PM
I'm not really trying to do any formal hypothesis testing here, simply to raise some issues and questions that seem interesting to me. And, in fact, I try to be careful about saying that what I'm showing may or may not be causal, or that it may or may not confirm a hypothesis. Sorry if that doesn't seem enough...
2:25 PM
Post a Comment
<< Home