tag:blogger.com,1999:blog-1740670447258719504.post6740893633906883897..comments2024-03-27T03:00:27.024-04:00Comments on Facts & other stubborn things: Can someone tell me why my labor demand subsidy is giving me numbers that look like a labor supply subsidy?Evanhttp://www.blogger.com/profile/12259004160963531720noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-1740670447258719504.post-19528191083343276552012-11-25T10:58:40.575-05:002012-11-25T10:58:40.575-05:00Sorry - that's not a downward sloping supply s...Sorry - that's not a downward sloping supply story... I was reading up on downward sloping supply too (that of course would also explain this) - but I'm not getting my head around how that could be the case here. This explanation - which I think is more plausible - is not a downward sloping supply argument. It's a "your data is fooling you" argument.Daniel Kuehnhttp://www.factsandotherstubbornthings.blogspot.comnoreply@blogger.comtag:blogger.com,1999:blog-1740670447258719504.post-26472773654545091892012-11-25T10:57:06.971-05:002012-11-25T10:57:06.971-05:00I have an interesting downward sloping supply stor...I have an interesting downward sloping supply story associated with the fact that:<br /><br />1. My wages aren't really wages... they're monthly earnings (it's derived from UI administrative data so there's really no choice) - so it's got the intensive quantity margin mixed in as well, and<br /><br />2. The program requires the creation of full time jobs.<br /><br />It's possible that the firms want the tax credit but they're only willing to hire low-wage workers for full-time positions. So average monthly earnings actually falls as they shift towards that type of worker.<br /><br />I'm not sure how plausible that is, but it seems possible. I doubt they're claiming this credit for any but the very lowest wage worker, and they don't make much in a month. They could probably pull down an average monthly wage figure.<br /><br />I don't know if I believe this... my identification strategy for this one is very strong, I think, but the effect size is still probably pretty small. It's quite possible this is being driven by something odd in my execution of it all.<br /><br />But I think that's a plausible enough story to tell if the specification is still right and these results don't go away.Daniel Kuehnhttp://www.factsandotherstubbornthings.blogspot.comnoreply@blogger.com