tag:blogger.com,1999:blog-2499715909956774229.post7119323918356333883..comments2024-03-22T22:37:02.639-07:00Comments on Stephen Williamson: New Monetarist Economics: Government Debt and Intergenerational DistributionStephen Williamsonhttp://www.blogger.com/profile/01434465858419028592noreply@blogger.comBlogger41125tag:blogger.com,1999:blog-2499715909956774229.post-14004863978444654952012-10-18T07:28:28.154-07:002012-10-18T07:28:28.154-07:00I see John D is back. Never believe I won't b...I see John D is back. Never believe I won't be here to mock you John, wherever you manifest.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-57797153707900439602012-10-17T19:41:26.607-07:002012-10-17T19:41:26.607-07:001. You write, "This is unrelated. We typicall...1. You write, "This is unrelated. We typically don't think of private debt as being a cause of macroeconomic phenomena. It's an outcome." I do not agree but think it unlikely that if Minsky cannot teach you where you are wrong it will not be productive for me to attempt such a teaching moment here. I will say that if private debt is not a cause, then neither is public debt for an economy cannot distinguish between the two, so why are you worrying about the Ricardian Equivalence at all?<br /><br />2. You falsely charge, "There are no obvious answers to those questions, but apparently you think so." To the contrary, I don't think there are "right answers" because of what Soros calls irrationality.<br /><br />The reason can be simply stated. To be truly effective, gov't spending must meet two tests.<br /><br />The first question is whether the expenditure increases our productive potential more than it costs.<br /><br />The second is whether the amount being spent meets the public expectation of what is required or necessary. <br /><br />If it is widely perceived that the gov't is not making sufficient investment expenditures, the effect will be the duel horsemen of declining private investment and declining private spending.<br /><br />You and I live in such a State (Missouri). Our public spending, etc. is so bad (lack of support for education, roads, etc., that smart people are running for the exits.). Angry Bear has had several recent posts on the key local driver being the % of college educated adults. A locality that cannot drive that number up faces a bleak future.<br /><br />My two cents is that the perceived lack of investment by government is why Bush II lead to the Lesser Depression. Look at the rational expectations created by Bush's tax cuts: govn't investment would be insufficient.<br /><br />Said differently, adequate gov't spending creates the expectation that business can be done safely, efficiently, and profitably. Where are these expectations in Ricardian Equivalence?<br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-71679529279115825452012-10-17T18:32:14.204-07:002012-10-17T18:32:14.204-07:00Sorry to keep harping on this, but I’ve taken an i...Sorry to keep harping on this, but I’ve taken an interest in how different people have been expressing the same idea in this discussion. Considering it further, it looks to me that there are a couple of errors of algebra in the post, unless you correct my understanding of it. I’m not sure this makes any difference in terms of the directional thrust of your conclusion, so you may wish to leave this. It’s probably not important to the thrust of your post. In fact, I liked the way you presented the algebra as a model.<br /><br />But, if you are interested, this is my analysis:<br /><br />I think the problem I see is that b seems to be defined initially and explicitly as the debt per capita of those buying it (the T young), but in another place (implicitly by equation) as debt in aggregate (also in your comment above).<br /><br />So, using your original per capita T young notation and meaning for b in the post, I would say:<br /><br />In period T:<br /><br />Number of young<br />1 + n<br />Number of old<br />1<br />Total population<br />2 + n<br /><br />Bonds bought by each young person<br />b<br /><br />Total bonds bought<br />b (1 + n)<br /><br />Transfer to each person (young and old)<br />x<br /><br />Total transfer<br />x (2 + n)<br /><br />Budget constraint:<br /><br />Total T transfer = total bonds bought<br />x (2 + n) = b (1 + n)<br /><br />Transfer to each person<br />x = b [(1 + n)/(2 + n)]<br /><br />(That agrees with the post, although not with your comment above)<br /><br />In period T + 1:<br /><br />Total taxes required to retire the debt = Aggregate bonds accrued with interest<br />b(1 + n)(1 + r)<br /><br />IF:<br /><br />The T + 1 old pay all the taxes then:<br /><br />Each T + 1 old pays:<br />b(1 + r)<br /><br />(Both the post and your comment seem to identify this as:<br />b (1 + r)/(1 + n)<br /><br />IF:<br /><br />All T + 1 alive pay the taxes, the result for one thing depends on the assumption for the size of the T + 1 young population:<br /><br />IF:<br /><br />You assume the T + 1 young population is 1 + n (which the post apparently does, which I don’t understand)<br /><br />and,<br /><br />All T + 1 alive pay the taxes then,<br /><br />The tax per T + 1 person is:<br /><br />-x(T+1) = (b(1 + n)(1 + r))/(2 + n)<br /><br />Both the post and your comment say:<br /><br />-x(T+1) = [b(1+r)]/(2+n)]<br /><br />IF:<br /><br />You assume the T + 1 young population is 1 + 2n (which I do)<br /><br />then,<br /><br />If the young and the old pay all the taxes then,<br /><br />The tax per person at T1 is:<br /><br />-x(T+1) = (b(1+n)(1+r))/(2+2n)<br /><br />Again, both the post and your comment say:<br /><br />-x(T+1) = [b(1+r)]/(2+n)]<br />JKHhttps://www.blogger.com/profile/06322177539880818556noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-74843558243096600762012-10-17T09:50:31.622-07:002012-10-17T09:50:31.622-07:00"...private debt could not cause growth."..."...private debt could not cause growth."<br /><br />This is unrelated. We typically don't think of private debt as being a cause of macroeconomic phenomena. It's an outcome.<br /><br />"...whether the expenditure increases our productive potential more than it costs."<br /><br />Exactly. That's a critical issue. The government's expenditure on goods and services that matters. And the deficit can matter. That doesn't make Ricardian equivalence "misleading." I told you in the post why it's useful.<br /><br />"GDP falls on tax cuts"<br /><br />You're stating this as if it's a fact. It's not.<br /><br />"If Gov't fails to spend..."<br /><br />Fails to spend how much? On what? How big should the government be and what should it do? There are no obvious answers to those questions, but apparently you think so.Stephen Williamsonhttps://www.blogger.com/profile/01434465858419028592noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-36939177729077739372012-10-16T23:23:23.729-07:002012-10-16T23:23:23.729-07:00If your explanation of the Ricardian equivalence i...If your explanation of the Ricardian equivalence is accurate, then private debt could not cause growth, either, and we should not permit such, given the risks.<br /><br />This would be true for example 3, if the student borrows, for this would cause a reduction in disposable income in the future by the amount necessary to pay off the debt.<br /><br />Of course we know that students borrow exactly because the education makes them more productive and thus able to both pay off the debt and make a decent living for themselves.<br /><br />The point of the lesson is that the Ricardian equivalence is both false and misleading, for it wholly fails to focus on the real key and that is whether the expenditure increases our productive potential more than it costs. That in turn seems to me to argue that we should be finding the tools to inform our students of the best ways to determine our most productive expenditures, as opposed to worrying about something that is not true.<br /><br />This might also inform us why GDP falls on tax cuts. Such would be evidence that the tax cuts have hurt the Government's ability to make productive expenditures and that, in reaction to a known future reduction in the standard of living either or both investors or consumers have cut back on spending. This is, in effect, an extension of the arguments of Robert Dugger. <br /><br />Or said a little differently. There is a true corollary to the false Ricardian equivalence, which would be:<br /><br />If Gov't fails to spend such will result in lower present investment and consumption to the extent that future expectations are lowered due to a belief that lower gov't spending will result in foregone opportunities to increase productivity.<br /><br />Doesn't this pretty much explain the last 5 years? We have an immense amount of cash sitting on the sidelines because we have all known that the Bullards and Republicans of this World will keep gov't spending below what we need for future growth.<br /><br />Look at Obama's stimulus, which was mostly tax cuts. No wonder such was so ineffective. Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-53281261374032051462012-10-16T18:14:29.261-07:002012-10-16T18:14:29.261-07:00Addendum to the above:
In example 3, you could al...Addendum to the above:<br /><br />In example 3, you could also imagine at different scenario. Suppose there is someone who has zero wealth. He or she has been admitted to Harvard, and he or she claims that if the government does not give him or her an outright grant, then he or she cannot go to college. There are two decisions to be made: (i) Should the government give him or her the grant? (ii) How should the government finance the grant if they make it? The second decision is independent of the first. The government has two options: tax someone today to finance the grant, or issue debt to finance the grant, and tax someone in the future to pay off the debt. If there is Ricardian equivalence, then it doesn't matter. But, if you think that generations are not linked together by bequests, then the choice is between making someone worse off today to finance the grant, or making someone worse off in the future.Stephen Williamsonhttps://www.blogger.com/profile/01434465858419028592noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-40141154597273838592012-10-16T15:48:36.222-07:002012-10-16T15:48:36.222-07:00"A tax cut that increases the government debt..."A tax cut that increases the government debt today has no effect because everyone understands that government debt is just deferred taxation."<br /><br />You're taking my words out of context. I explain above exactly what that means, and how you use Ricardian equivalence to think about the real world.<br /><br />In any case:<br /><br />1. So GDP fell after the Kennedy tax cut. How do you know the latter caused the former?<br /><br />2. The Ricardian experiment, and issues about the long-run effects of changes in the tax structure are two different issues. The Ricardian experiment is: hold government spending constant, change taxes today, with future taxes increasing to pay of the debt. What happens? If we change the income tax code by, say, increasing marginal tax rates, possibly with other changes in the government budget, that's entirely different.<br /><br />3. This is confused. Someone wants to accumulate human capital, and we all agree this would be a good idea. Suppose that person has zero wealth. They have to borrow to get an education. Maybe a private financial institution will make the loan. Maybe we think there is some market inefficiency, and the loan won't be made unless the government does it. Different issue entirely.Stephen Williamsonhttps://www.blogger.com/profile/01434465858419028592noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-78662269546690955112012-10-16T15:36:11.208-07:002012-10-16T15:36:11.208-07:00Hopefully I did the algebra correctly. Everything ...Hopefully I did the algebra correctly. Everything comes from the government budget constraint, written in per-young-person terms. Government debt is one-period bonds, so if b(t) is the government debt issued in period t, and x(t) is the transfer the government makes to each person alive in period t(assuming everyone gets the same transfer), then<br /><br />b(t+1)=<br />[(1+r)b(t)]/(1+n) + x(t)[(2+n)/(1+n)]<br /><br />Everything follows from that.Stephen Williamsonhttps://www.blogger.com/profile/01434465858419028592noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-29001704473600639382012-10-16T15:25:59.505-07:002012-10-16T15:25:59.505-07:00You write, "A tax cut that increases the gove...You write, "A tax cut that increases the government debt today has no effect because everyone understands that government debt is just deferred taxation."<br /><br />Please apply such to the two following real data examples.<br /><br />1. GDP fell after the so called Kennedy tax cut.<br /><br />The 1964 Tax Cuts and Economic Growth - Paul Ryan Edition<br /><br />http://www.angrybearblog.com/2012/10/the-1964-tax-cuts-and-economic-growth.html<br /><br />2. GDP Growth Caused By Tax Cuts Has Never Happened<br /><br />http://www.angrybearblog.com/2012/10/gdp-growth-caused-by-tax-cuts-has-never.html<br /><br />3. Please explain how private borrowing can increase GDP, etc., but not public borrowing. How does the economy or Harvard, for that matter, know the money borrowed to educate the young person who will save the World, tomorrow, came from a public or private borrowing?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-49681578402922971742012-10-16T05:23:01.547-07:002012-10-16T05:23:01.547-07:00Andy
Exactly!
And when social security or compar...Andy<br /><br />Exactly!<br /><br />And when social security or comparable schemes are themselves fully funded through bonds, efficient asset insurance converts into efficient social insurance.<br /><br />And we're in Tobin's world. The point about dynamic efficiency, risk-free and risky rates of capital was also made succintly by him.<br /><br />Perry Mehrling has an excellent paper on this, on the role of the state as a risk manager, as a social mutual fund.Ritwikhttps://www.blogger.com/profile/00616694597577112758noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-18685933872283998862012-10-16T03:59:14.339-07:002012-10-16T03:59:14.339-07:00Very helpful post, thanks.
But I’m embarrassed to...Very helpful post, thanks.<br /><br />But I’m embarrassed to say I found your algebraic notation too subtle for my older brain.<br /><br />Not only that, I keep changing my question regarding the very same point.<br /><br />But here it is at this time:<br /><br />x(T) = [b(1+n)]/(2+n)<br /><br />I assume there that in period T the “normalized” number of old people is 1 and the number of young people is 1 + n (given a growth rate of n), for a total of 2 + n.<br /><br />... in period T+1 requires total taxes per young person alive equal to [b(1+r)]/(1+n)<br /><br />Does “young person alive” mean young person in T + 1 or young person in T (and old person in T +1)?<br /><br />And: <br /><br />-x(T+1) = [b(1+r)]/(2+n)<br /><br />If the population grows at rate n, would it not be 2 + 2n at that point?<br /><br />With the number of old people 1 + n and the number of young people 1 + n?<br /><br />I know I've erred disastrously - but where?<br /><br />Everywhere?<br /><br />thx<br />JKHhttps://www.blogger.com/profile/06322177539880818556noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-49371957661305980182012-10-15T18:59:08.140-07:002012-10-15T18:59:08.140-07:001. Social insurance? I was talking about asset in...1. Social insurance? I was talking about asset insurance, like a CDS, insurance for the people who own the bonds, not for society as a whole.<br /><br />2. For purposes of determining whether the gov't can run a stable Ponzi scheme, the relevant n includes productivity growth as well as population growth (call it "n+g" if you want), so it is at least .02 p.a. (possibly less in the future, but considerably more on average over the past 50 years). Compare actual growth rates for the US since WWII to actual yields, the growth rates have almost always been higher (80's is an exception).Andy Harlesshttps://www.blogger.com/profile/17582263872850949568noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-90495385458353528502012-10-15T17:49:43.079-07:002012-10-15T17:49:43.079-07:00OK. I'm now fairly sure that as long as r <...OK. I'm now fairly sure that as long as r < n+g (permanently) the debt is too small for Pareto Optimality. Because the government can make a lump sum transfer to the current cohort, and then roll over the debt+interest forever, with the debt/GDP ratio falling over time. The current cohort is better off, and future cohorts cannot be worse off, since they don't have to pay higher taxes. In fact, the higher debt will raise r, and make future cohorts better off, since they get a higher rate of return on their savings.Nick Rowehttps://www.blogger.com/profile/04982579343160429422noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-9597034252124594352012-10-15T14:24:52.974-07:002012-10-15T14:24:52.974-07:00The only OG models I have seen are ones that are s...The only OG models I have seen are ones that are stationary - population can grow, but the technology, endowments, and preferences are the same across generations. Maybe it's too hard otherwise.Stephen Williamsonhttps://www.blogger.com/profile/01434465858419028592noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-59754179471866516932012-10-15T13:17:53.090-07:002012-10-15T13:17:53.090-07:00Now I'm not sure if I was right either. I need...Now I'm not sure if I was right either. I need to think about it.Nick Rowehttps://www.blogger.com/profile/04982579343160429422noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-88891753109541335812012-10-15T12:08:44.612-07:002012-10-15T12:08:44.612-07:00Mysteries of the OG model. What I said in the post...Mysteries of the OG model. What I said in the post is correct. Now I'm not so sure about my 6:35 am comment (too early maybe). Once you allow the endowment to grow in the endowment model, or allow TFP to grow in Diamond's model, I'm not even sure the Pareto optimum is easy to characterize. Maybe someone knows about this.Stephen Williamsonhttps://www.blogger.com/profile/01434465858419028592noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-42874516477007014702012-10-15T12:01:55.120-07:002012-10-15T12:01:55.120-07:00Had to check my graduate copy of Romer. Government...Had to check my graduate copy of Romer. Government expenditures are expressed in units of effective labor, so once this is taken into account I think debt can grow at n+g. CAnoreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-82946136123714621532012-10-15T09:00:25.824-07:002012-10-15T09:00:25.824-07:001. Social insurance is another issue entirely.
2. ...1. Social insurance is another issue entirely.<br />2. "...empirically r < n has typically been the case historically for government debt." n is at most .01 per annum. For r, it depends what maturity you are looking at. For 30-year bonds, it looks like r > .03 prior to the recent recession.Stephen Williamsonhttps://www.blogger.com/profile/01434465858419028592noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-35460245372605240382012-10-15T08:53:09.166-07:002012-10-15T08:53:09.166-07:00You and Andy are both confused.You and Andy are both confused.Stephen Williamsonhttps://www.blogger.com/profile/01434465858419028592noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-32718596572409547932012-10-15T08:51:05.849-07:002012-10-15T08:51:05.849-07:00Steve: are you sure that's right?
Take a simp...Steve: are you sure that's right?<br /><br />Take a simple 2 period model with no capital or storage. Zero population growth. Assume an endowment of 100 when young, and 0 when old. Assume zero time preference proper (B=1). The national debt should be 50. But if the endowment is growing at 10% per cohort (so the next cohort gets 110), we would want the national debt to be growing at 10% per cohort too (so it's 55 for the next cohort). And we do that by paying 10% interest on the debt.<br /><br />As long as we can roll over the debt forever, and the debt/GDP ratio stays constant, we are OK.<br /><br />Maybe I'm misunderstanding you.Nick Rowehttps://www.blogger.com/profile/04982579343160429422noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-75517629030305663202012-10-15T08:46:20.144-07:002012-10-15T08:46:20.144-07:00Nick: I would quibble with the word "always&...Nick: I would quibble with the word "always" (since the r < n condition doesn't have to hold continuously), but I think I see what you mean: the expectation that it will hold, in whatever sense it has to hold, is only an estimate, not a certainty. Presumably, even if we were certain that r < n for government debt, it would still be the case that r > n for the aggregate capital stock, because the capital stock involves different, more expensive risks. The uncertainty associated with future r vs n for government debt doesn't explain why r is so much higher for the aggregate capital stock. So I think you are really getting at a different issue.<br /><br />Steve: OK, you're going in a direction similar to something I suggested on Twitter yesterday. I was thinking of the government as selling insurance, as if it first borrowed at the MPK and then sold CDS on its own debt. In that case the government isn't running a Ponzi scheme; it's just running a profitable asset insurance business. The case might be more clear cut if, as you do, one thinks of the government as selling liquidity rather than selling insurance. Either way, the government is some sort of super-profitable financial institution that is actually adding value and not running a Ponzi scheme. Note that, while this may be true now to a greater extent than usual, it is not a new phenomenon, since empirically r < n has typically been the case historically for government debt. It's just that maybe we should think of the actual "r" as being much higher than the empirical yield and regard the difference as some kind of premium that the government is earning, separate from the cost of borrowing.Andy Harlesshttps://www.blogger.com/profile/17582263872850949568noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-56546571952943703112012-10-15T08:44:03.554-07:002012-10-15T08:44:03.554-07:00Bravo to Andy.
I should also point out that this...Bravo to Andy. <br /><br />I should also point out that this whole discussion is a bit skewed by the fact that the median retiree has non-housing wealth of $30,000, and housing wealth of $150,000. <br /><br />Bonds are primarily held by the wealthy, not by the old. If you want to talk about intergenerational transfers, then in the U.S. at least, this is always and everywhere as discussion about housing being bought by one generation and re-sold to another, not about bonds.<br /><br />So, for example, this downturn caused a large hit to the wealth of the median retiree, larger than to other cohorts.<br /><br />For government bonds, or financial assets more generally, it is more of a transfer between the poor and the rich than between the young and the old. rsjhttps://www.blogger.com/profile/05489955485750918419noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-64489725719629964642012-10-15T08:29:42.916-07:002012-10-15T08:29:42.916-07:00When I teach my students Ricardian equivalence, I ...When I teach my students Ricardian equivalence, I note that if you believe the debt is going to be a burden on your children, you must not like them as otherwise you would increase their bequest by exactly enough to compensate for their added taxes. Maybe we should use that line of argument in public -- think the debt is a problem? Then stop farting around and save, dagnabbit!Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-55631432988071480652012-10-15T08:10:38.092-07:002012-10-15T08:10:38.092-07:00"...to cut programs which primarily help the ..."...to cut programs which primarily help the young (they are promising not to cut the benefits of retirees and near retirees, justifiably)."<br /><br />If the programs involve government spending on goods and services, that's a different issue. I'm just dealing with the timing of taxation.<br /><br />"Also, borrowing to finance public investment can have an impact on growth/ future income..."<br /><br />Again, another issue altogether. For both of these things, you have to consider what the special role of government is. Is there some externality the government is correcting? In any case, all government policy actions involve distributional issues, whether across generations or within the population currently alive. You can't avoid these things.Stephen Williamsonhttps://www.blogger.com/profile/01434465858419028592noreply@blogger.comtag:blogger.com,1999:blog-2499715909956774229.post-2205991091871179372012-10-15T07:55:53.835-07:002012-10-15T07:55:53.835-07:00To focus on the political debate here though, thes...To focus on the political debate here though, these results are informative but incomplete. They abstract from a number of relevant factors it seems from a practical policy standpoint. <br /><br />The first is that the Republican response to this "burden of debt" issue is to cut programs which primarily help the young (they are promising not to cut the benefits of retirees and near retirees, justifiably). If we do not increase the debt, future generations (either the next or thereafter depending on whether the debt is paid down) will be better off in that regard of lower taxes all else equal, but with less aid to college students, reduced benefits of social security or medicaid/medicare, and other social safety net programs, they will be worse off in those regards. These are inter-generational effects with an intra-generational nature (primarily redistributing to the top). This seems to be the true Republican agenda (if deficits were the priority the only two specifics in the plan wouldn't be 20% tax rate cuts and increased defense spending with loopholes and spending cuts to be named later). Whether they believe this is good for growth is relevant but separate (and they are then being disingenuous about their concern over deficits). <br /><br />Also, borrowing to finance public investment can have an impact on growth/ future income (contrary to the Barro "Are gov't bonds net wealth" argument where he assumes only transfers), Republican spending cuts which reduce funding on education, infrastructure, and R&D etc are like "eating your seed corn". Anonymousnoreply@blogger.com