New York Times, 10/5/2012, "Taming Volatile Raw Data for Jobs Reports", By CATHERINE RAMPELL
Obama 在接下來一個月最喜歡的數字將會是 7.8。美國九月份的失業率降到了自從他就任總統以來的最低點，而這個數據是大選之前最後一個數字了。在前幾天以經濟議題為主題的第一場總統選舉電視辯論表現不佳之下，這個數字給了 Obama 以及他的競選團隊一劑強心針。
GOP 的第一個反應是這個數字有造假可能，因為這個失業率 "too good to be true"。New York Times 這篇文章簡單的說明了這個數字是怎麼出來的。Conspiracy theory? 大概不至於，不過仔細拆解之後我們也可以看出 Obama 其實沒什麼好高興的。這個數字出乎意料的原因有二：失業率數字來自於規模有限的問卷調查，本身的誤差就可能不低 (文中說 margin of error 達到 400,000，不過沒有說清楚這是幾倍的標準差）；更嚴重的問題在於季節性調整的模型在今年可能不適用，因為往年 20-24 的年輕人在九月離開工作，作者猜測可能是結束暑假的實習工作回學校唸書，不過今年這個時間點提前到七、八月，使得經過季節性調整後的 20-24 就業人數特別漂亮。
這些數據未來幾個月會持續修正。等到塵埃落定後，我們很可能會發現失業率仍然穩穩的在 8% 以上，不過到那時候總統大選早已結束了。
1 comment:
Its really interesting how last month, when the numbers came out, all the news mentioned people stopped looking for jobs as the main reason for the decrease in unemployment. But this time around, nobody even talked about why numbers dropped so dramatically with the exception of the 100K+ new jobs created and the "revision" that was done by DOL. But the numbers don't quite add up...
Numbers were cooked? Probably not. but survey bias? Pretty likely I think...
As one NPR commentator put it, if one look back far enough on the "revised numbers" the unemployment rate has in fact remained stagnant instead of the so called "significant" drop.
If I was Romney's advisors, I'd look up all the revised unemployment rate for the past couple of months and use the trend it as a zinger.
That ought to burst Obama's fairy tale...
Post a Comment