tag:blogger.com,1999:blog-1331441403058020963.post8743146086801057941..comments2024-03-28T06:49:24.930-04:00Comments on International Political Economy at the University of North Carolina: What to Think About Debt, Deficits, and Interest RatesThomas Oatleyhttp://www.blogger.com/profile/14092437150746625670noreply@blogger.comBlogger4125tag:blogger.com,1999:blog-1331441403058020963.post-73612869947088719222010-08-14T02:16:35.529-04:002010-08-14T02:16:35.529-04:00Thanks for your reply Kindred.
But please ignore...Thanks for your reply Kindred. <br /><br />But please ignore my previous post. There were many things I was confusing in there, and this may very well be the right thing to do in such a model. <br /><br />That's what you get for writing before thinking...Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-1331441403058020963.post-63594817569352062542010-08-13T23:58:16.959-04:002010-08-13T23:58:16.959-04:00Emmanuel, I have looked the IMF projections, as we...Emmanuel, I have looked the IMF projections, as well as the OECD projections. I don't agree that they say what you think they say, and in any case they revise them pretty significantly every few months.<br /><br />Anon, if you have questions about Mosley's methodological choices I'd suggest you e-mail her, since she'd be able to give you more info than I. My guess is that since she modeled an additive relationship between the variables, she thought that adding the coefficients for a "total effect" was appropriate. She did report them separately, as you noticed, and they retain significance in almost all cases. She also modeled it as an AR(1) process, which includes an exponential decay function.<br /><br />Anyway, I don't want to comment too strongly because I'm not exactly sure why she made those choices. I'm sure she'd be happy to answer any questions via e-mail.Kindred Winecoffhttps://www.blogger.com/profile/14330671232391851377noreply@blogger.comtag:blogger.com,1999:blog-1331441403058020963.post-6781750313331096402010-08-13T21:33:50.929-04:002010-08-13T21:33:50.929-04:00So, I glanced at the Mosley piece you link to and ...So, I glanced at the Mosley piece you link to and I have a question/comment. <br /><br />She uses lags for most of her independent variables (t-1 & t-2) and she writes that "to determine the cumulative effects of each (lagged and current) independent variable, I combine the coefficients."<br /><br />I wondered what this meant and looked at the "disaggregated coefficients" in appendix. Basically, the author takes all coefficients of the lags and of the main terms, and simply adds them together. <br /><br />I'm not too familiar with work with multiple time lags, but I know that this approach would be wrong in the simple case of a single t-1 lag. To get the dynamic effect, you need to multiply both coefficients at t, and trail it in an exponentially decreasing function for all t+n until the effect vanishes. You can't just add them together and interpret that sum as your effect. <br /><br />Similarly, you can't just add standard errors and treat this as your estimate of uncertainty. An analytical solution is actually hard to get, but it's easy to work with simulations in order to provide a confidence band based on the standard deviation of simulated values (e.g. using CLARIFY). <br /><br />I want to believe the author, but I'd really need someone to explain to me why this could be right. <br /><br />This makes me think of another Mosley (2007) paper on which I did some basic replication work. If I remember right, she misinterprets an interaction term, and it turns out that one of the main findings of the paper is essentially a null result.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-1331441403058020963.post-32031450376751352672010-08-12T22:27:24.562-04:002010-08-12T22:27:24.562-04:00The arguments of those who call for fiscal consoli...The arguments of those who call for fiscal consolidation do not rely exclusively on fears of future spikes in interest rates. Even at historically low interest rates, the cost of debt service will be exceptionally high and, more importantly, unsustainable for the US. <br /><br />For instance, a guy called Erskine Bowles, I think he has something to do with the UNC ;-), is wary of interest payments of the US government amounting to <a href="http://ipezone.blogspot.com/2010/07/we-have-just-received-word-from-us.html" rel="nofollow">$2 trillion annually</a> by 2020. <br /><br />The best way for anyone to learn this stuff is to work through estimates. Study how IMF, OMB, David Walker or Laurence Kotlikoff come up with their fiscal projections. If you disagree with them and think "more deficit fiscal expansion" is possible, explain why using figures. Perhaps you can tell us why their discount rates are wrong, why projections of non-discretionary spending are overstated and so forth. <br /><br />Do the math. Otherwise, you end up like just another Drezner (ideologically biased) or Pettis (numerically challenged) with no numbers to show for it.<br /><br />More likely than not, you too will come to the conclusion that the US is bankrupt.Emmanuelhttps://www.blogger.com/profile/04615366847433704476noreply@blogger.com