I had three real GDP forecasts ranging from -2.2 to +.8 (annual growth rate from Q4). The "experts" consensus was -5.4.
According to market participants at intrade.com, there was only a 10 percent chance that the BEA would report something this morning that was better than -4.0.
This morning we learned that the BEA's "Advance" estimate was -3.8. It will be followed in about a month by its preliminary estimate, and then later its final estimate. I mentioned this morning that they have more work to do on deflators.
I had previously predicted that real GDP "might" go up, although I still said that slightly down was the more likely. I was very precise about that: I said that it was 1/3 likely to go up and 2/3 likely to go down.
According to market participants at intrade.com, there was only a 10 percent chance that the BEA would report something this morning that was better than -4.0.
This morning we learned that the BEA's "Advance" estimate was -3.8. It will be followed in about a month by its preliminary estimate, and then later its final estimate. I mentioned this morning that they have more work to do on deflators.
I had previously predicted that real GDP "might" go up, although I still said that slightly down was the more likely. I was very precise about that: I said that it was 1/3 likely to go up and 2/3 likely to go down.
There are no rules in the blogosphere, but it is factually incorrect to quote me as predicting that real GDP WILL go up. "WILL" and "MIGHT" are different in plain English (not to mention that I also put it terms of probabilities).
I haven't seen anyone comment on it. Considering everyone else's forecasts, you didn't do that bad.
ReplyDeleteCasey,
ReplyDeleteNice job on the GDP numbers. I agree that the deflator was out to lunch and will probably be revised. Your guess is as good as mine as to magnitude.
By the way, my estimate range was -3 to -4. Mine was based on a variety of indicators but the two that were most helpful were the ISM and the spending numbers. And really mine was just a guess based on past data. Anyway, kudos for providing a good framework for evaluating the numbers.