No … for panes of glass or anything else, only the market-clearing price is the market clearing price. That tautology is both true and important: any stimulus policy that changes that price by destruction interferes with the processes of economnic health. Bastiat and Hazlitt are right.

1) There are some generally very smart and mathematically adept people who seem to forget that averaging a bunch of measurements, all made with the same biases, will not give you an endlessly improving estimate of the thing to be measured. Some people (largely optimistic Romney supporters) have argued that the polls are skewed because the samples are unrepresentative, and have adduced the party breakdowns of poll respondents as an argument. If you believe there are shared methodological problems with the 25 polls a day (or whatever it is at this point) that go into the averages, you shouldn’t imagine that those will average out.

2) The closest thing to a good response Silver has given has been that the relative fraction of voters who are Republican or Democrat are not constant, and that that is part of what the polls should capture. The problem that I think most of the other side has with that is that it seems easier to believe right now that there’s some methodological flaw in the polls than to believe that Obama will do better in the turnout battle this year than he did against McCain. My response to them is that I think a significant number of the people identifying themselves to pollsters as “independent” are tea party voters who consider themselves fed up with the GOP, but will continue to generally vote for Republicans over Democrats, and in particular Romney over Obama. I think this is why Romney is polling so well against Obama among “independents”, and is why the likely voter party breakdown is skewing toward Democrats.