3 ms·
Yeah. I'd take a lot of convincing that they hadn't just failed to specify adequate identification tests, since local newspaper closures and raised municipal bo
by notahacker 8y ago
Yeah. I'd take a lot of convincing that they hadn't just failed to specify adequate identification tests, since local newspaper closures and raised municipal borrowing costs are both clearly and obviously linked to local economic conditions, and the potential linkage between municipal bond rates and local newspaper scrutiny far more questionable.
- Eridrus 8y agoThe article links to the paper, where this is the number one concern they address. They seem to have a bunch of different ways of slicing the data to prove their effect, e.g. pairing demographically similar districts, looking at newspaper shutdowns likely induced by Craigslist, looking at number of newspapers in a district, looking at distance of political center to economic center. I don't know if this is a complete list of things that economists would look at, so I don't know if they're cherry-picking good ways to slice the data, but they have 47 pages of trying to convince you it's right if you want to dig that deep :) There could obviously be other explanations, e.g. markets thinking that less information rich environments are just riskier because they won't know what's going on, but they seem to have been fairly thorough in establishing the link between newspapers closing and bond rates increasing.
- mirimir 8y agoYes, they do lots of checking. But still, I don't see anything clear beyond the association of newspaper shutdowns and increasing bond rates. Nothing about mechanism except speculative argument. This is very much like epidemiology. And the literature is loaded with retrospective studies that found correlations which turned out to reflect selection bias and/or confounding. You really need prospective studies with carefully matched controls.