James Manzi on Stargazing

Arnold, this strikes me as a deeply misleading post.

Whether the number of stars in our galaxy is odd or even is a question that we can't answer because we can only produce an estimate of this number that we can infer from what we observe. Estimates have errors and as long as we estimate these figures we can't say anything about them that you need zero error to say.

Similarly, we can't say whether the stimulus produced an odd or an even number of jobs. We just can't.

But we can provide an estimate, with error, of the number of stars in the galaxy and we can provide an estimate, with error, of the number of jobs created by the stimulus. Those are the scientifically accessible and the practically relevant questions.

So there's about a fifty-fifty chance that the stimulus helped or did nothing, except we know that either way it cost hundreds of billions.

It reminds me of Lew Pulsipher's comment on why it's a bad idea to drink potions you find in an old-school Dungeons & Dragons game until they're analyzed: "If I gave you a liquid and told you it has a 5/6 chance of making you very strong or fast for an hour, and a 1/6 chance of killing you, would you drink it?"

John,
It's an obvious fallacy to assume that because you can't put an exact number on it that it then follows that the odds are about fifty-fifty.

You can estimate how many jobs were "created" during the time period of ARRA, but I have not seen a really convincing argument that the ARRA made that number bigger or smaller. This was, I think, Manzi's real point.

Both questions are theoretically unanswerable in any exact way----but his point is still valid. We cannot know the number of stars at any point in time because we do not know which ones flamed out after the light reached us----yet we believe there are a specific number of stars at any point in time.

It is harder to make a judgment on economic/job growth because the question is more complex. But, we can state with certainty that the "number of jobs created and saved" by the stimulus is a politically made up number derived from a model presumed, without any evidence, to be true.

We can also make comparisons between different policies made during recessions in various countries during various times and estimate whether recent policies were positive by analogy and hypothesis. We can also count the number of people working today versus, say, 5 years ago.

We can also measure whether this recovery compares favorably with others. So even if there were 6 or 20 million jobs "created or saved", that still may not be good enough.

Fortunately, most people can figure out whether some policy was likely good for them---and vote accordingly.

Daniel Kuehn,

and we can provide an estimate, with error, of the number of jobs created by the stimulus

I dispute this. Yes, I am sure a human could build what he would claim to be a good macroeconomic model which would spit out a number for its estimate of 'jobs created by the stimulus' as well as a number for 'error of the number returned by this model'.

The problem is that both numbers would depend entirely on a huge number of approximations, guesses, and embedded assumptions about how humans behave and coordinate (in this, as well as in counterfactual universes) that they embody what is essentially a guess about something unknowable. You know, like whether the galaxy has an odd or even number of stars.

Do not let the fact that the assumptions and guesswork have been pushed into the inner workings of some macroeconomic model (about, e.g., the shape and parameters of the distribution of the errors in the things you are estimating) fool you into thinking you are not guessing when you run such a model.

Sonic -
Exactly right - that's why they're estimates and that's why we argue about specification.

You don't think these issues come up in counting stars and planets?

Of course it does:

http://blogs.discovermagazine.com/80beats/2010/12/01/the-estimated-number-of-stars-in-the-universe-just-tripled/

People who can't deal with imprecision and uncertainty should go into theology, not science.

Right. So, both are unknowable, for similar reasons, and thus the analogy is a good one, after all.

Does anyone know where Manzi got the stars analogy?

I'd be interested to know if he got it from the ancient Stoics. They were supposed to have used the "Is the number of stars even or odd?" question as a typical example of an unanswerable question.

Source: Outlines of Skepticism (I think that's the one), by Sextus Empiricus.