00:01 Great.
00:02 Why,
00:02 why don't we start?
00:03 Because,
00:04 um,
00:05 You know,
00:06 this morning's uh seminar is incredibly useful
00:10 and,
00:10 um,
00:11 I'll say really hits the core of a lot of what we do here in the macro,
00:16 uh group
00:17 and um,
00:19 we have lots to discuss.
00:20 Mark and Emmanuel wrote this uh great book
00:24 that kind of gives a
00:26 a broad
00:28 Framework
00:30 To think about,
00:31 uh,
00:32 sovereign debt,
00:33 um,
00:34 the cost of it,
00:35 the cost of default,
00:37 um,
00:37 there's interesting discussion on the debt weight,
00:40 um,
00:41 cost
00:42 of defaulting,
00:43 and I think a,
00:44 a lot of,
00:45 uh,
00:45 of this discussion
00:47 really hits the,
00:48 what's,
00:49 what's going on right now and what our teams are doing.
00:52 We are quite involved in discussing,
00:54 say,
00:55 uh,
00:55 the Sri Lanka debt crisis,
00:57 what should they do?
00:58 Uh,
00:59 we are in the middle of discussing the,
01:01 how to better set up the common framework,
01:03 what's the best way of thinking
01:05 about contracts
01:06 in a way that they facilitate,
01:08 uh,
01:09 renegotiations and minimize
01:12 the cost of default.
01:13 Anyway,
01:14 Mark's book
01:15 touches all these points and,
01:17 um,
01:18 Mark's a professor at Princeton now,
01:19 has had a,
01:20 a,
01:21 a kind of a very interesting career,
01:23 uh,
01:24 both in academics and,
01:25 and in policy.
01:26 He spent some time at the Federal Reserve.
01:28 Bank in Boston.
01:30 So,
01:31 um,
01:31 he's a leading thinker on this issue and you're very excited to have him,
01:35 and you're very excited to have Sergio,
01:38 uh,
01:38 discussing his work as Sergio also has done
01:41 work in the area and leads our,
01:43 um,
01:44 macro team in DC,
01:47 so,
01:47 uh,
01:47 in the research department.
01:49 So let me pass to,
01:50 to,
01:51 to Mark.
01:52 Uh,
01:52 Mark has
01:54 Uh,
01:54 we arranged about 45 minutes to speak and Sergio has a bit under 15 minutes to
02:01 discuss,
02:01 and then you go to a Q&A,
02:03 but I think we can be a bit,
02:04 um,
02:05 flexible on this,
02:06 on this,
02:07 uh,
02:08 timing depending on how the discussion goes
02:10 and how the presentation goes.
02:12 So,
02:12 Mark,
02:12 uh,
02:13 many thanks for coming
02:14 and we are very excited to hear from you this morning.
02:20 All right.
02:20 Well,
02:20 thank you very much.
02:22 This is a great opportunity for me and,
02:24 and,
02:24 uh,
02:25 I hope you guys all find it uh useful as well.
02:29 So,
02:29 uh,
02:30 as,
02:30 this is the book,
02:31 uh,
02:32 written,
02:32 uh,
02:33 with Manuel Amador.
02:35 Um,
02:36 And in it,
02:37 we try to distill some of the
02:40 lessons and the way we think about sovereign debt,
02:43 uh,
02:44 models and markets and reality,
02:46 uh,
02:47 um,
02:48 and the link between the two.
02:50 So I'm hoping to,
02:51 to just kind of give you a sense today
02:53 of what we do,
02:54 how we approach this,
02:56 kind of the basic framework we use to,
02:58 to analyze it,
02:59 which we think is useful,
03:01 um,
03:02 and some of the lessons that we can pull out of it.
03:05 Um.
03:07 So
03:08 You know,
03:09 as you know,
03:10 uh,
03:10 better than me,
03:11 even that,
03:11 you know,
03:11 we have these repeated waves of debt buildups and then defaults,
03:16 uh,
03:17 uh,
03:17 kind of
03:18 repeated cycles of this,
03:21 and
03:21 there's a large literature that has developed
03:24 that
03:25 is designed to document these facts and then match them using
03:30 large quantitative models,
03:32 and what we try to do is take
03:35 a step back from all that and try to think.
03:38 It's just basic,
03:39 the basic economics of what's going on,
03:42 and
03:42 isolate key
03:44 lessons,
03:45 uh,
03:46 and then in the book,
03:47 we go back to the quantitative models and point
03:49 out like why this particular lesson shows up at
03:53 at is useful to match this particular moment,
03:55 etc.
03:56 um.
03:58 But some of these,
03:59 some of these large quantitative models,
04:00 as you know,
04:01 if you've read the literature,
04:02 are a bit of black box.
04:03 They're a bit of a black box.
04:04 So,
04:05 so we try to unpack that.
04:07 Um,
04:09 OK,
04:10 so
04:12 Where do we start?
04:13 So we can start thinking about spreads.
04:15 OK.
04:17 Where did I leave off?
04:18 So,
04:19 we're just started talking about these spreads,
04:21 and uh
04:23 yeah,
04:23 obviously,
04:24 we know,
04:24 they're very volatile.
04:26 Uh,
04:26 um.
04:28 And they captured two components,
04:31 the
04:32 default risk
04:33 and the risk premia.
04:35 And,
04:35 and we,
04:36 like most of the literature,
04:36 are going to focus on the default risk,
04:38 and this is going to be crucial kind of in the,
04:41 the way I'm going to approach
04:42 the economics of these markets.
04:45 Uh,
04:45 to think about these spreads as capturing
04:47 something informative about the default risk.
04:49 Obviously,
04:51 without the default risk,
04:52 we wouldn't really need a risk premium,
04:54 uh,
04:55 uh,
04:55 but,
04:55 um,
04:56 that's kind of how I'm gonna
04:58 just kind of focus on that first one.
05:01 OK,
05:01 and then.
05:04 When we think through the equilibrium of these markets.
05:08 We have to think about this interplay.
05:11 How do the
05:12 bond prices influence the government decisions to borrow or not?
05:16 And then how does government borrowing influence pricing?
05:19 And this is kind of the core of a lot of what we discussed.
05:23 So think about the pricing of a long-term bond.
05:25 Um.
05:27 You know,
05:27 Q would be the price,
05:29 um.
05:31 You know,
05:32 went over the yield,
05:33 the implied yield,
05:34 uh,
05:35 M is some just stochastic discount factor.
05:40 This
05:41 fancy one is the indicator function
05:44 that takes 1 if they repay and 0 if you default.
05:47 So you could think of a haircut as well.
05:49 It could take,
05:50 you know,
05:50 you could get a fraction
05:52 of the payments in a default,
05:53 but,
05:53 you know,
05:54 for simplicity,
05:54 let's just go to the extreme and say you get 0 in a default,
05:57 and,
05:58 you know,
05:58 obviously with repay,
05:59 you get the full amount.
06:01 Kappa is some coupon,
06:03 and then Q is the price going forward in the next period.
06:08 And this is like a key object of,
06:10 of what we do and what many of the models in the literature do is to think through
06:14 that
06:15 the next period's price is going to depend on the government's
06:19 future borrowing decisions,
06:21 uh,
06:22 and the government's future borrowing decisions are going
06:24 to depend on what prices they face.
06:26 And so this is the fixed point that
06:29 we spend a lot of time trying to unpack,
06:31 and it's a complicated one,
06:33 but,
06:34 uh,
06:34 I think we think we've made some progress on this.
06:37 Um.
06:39 OK,
06:40 so.
06:42 And then,
06:43 that was kind of from the lender's perspective is that
06:45 they're trying to forecast what the government's gonna do,
06:48 both in the repayment decision and how much to borrow,
06:51 and of course the government's got to make these decisions.
06:54 Now the very early literature,
06:56 going back to Eaton Gershovitz and then Bella Rogos's,
06:59 uh,
06:59 a well-known critique of Eaton Gersowitz,
07:02 uh,
07:02 focused on why the government should pay at all.
07:05 Um,
07:06 and I think,
07:08 uh,
07:08 uh,
07:09 what that,
07:10 um,
07:11 literature concluded is that direct costs of default are key.
07:15 Uh,
07:16 now,
07:16 there's an enormous bureau rograph literature.
07:18 In fact,
07:18 I was just chatting with uh Sergio before this about,
07:21 uh,
07:22 you know,
07:22 Fernando Bronner and Jama Ventura and Alberto Martine of,
07:24 well-known papers on the secondary market,
07:27 um,
07:28 being a useful way of enforcing repayment and,
07:31 and discrimination default.
07:33 But I would say most of the literature is focused on just adding direct costs,
07:36 and there do seem to be direct costs of default.
07:39 But how do you measure them?
07:41 Uh,
07:41 and I'll give you one.
07:43 Uh,
07:43 mentioned this Herbert Schrager paper was a very creative way of doing it,
07:48 looking at the sequence of court cases that Argentina faced,
07:52 uh,
07:52 in New York,
07:53 uh,
07:54 with holdout
07:55 creditors from their default,
07:57 their 2001 default,
07:58 and then their,
07:59 their subsequent restructuring,
08:01 and they kept,
08:01 you know,
08:02 they kept losing these,
08:03 these court battles and the price,
08:05 uh,
08:05 the probability of default kept creeping up,
08:07 and they looked at the implications for um.
08:11 Uh,
08:11 equity market prices in Argentina,
08:13 they found very big
08:14 implied costs of default.
08:17 Um,
08:19 But
08:20 From our perspective is how do these costs
08:23 affect.
08:25 The government's uh borrowing decisions.
08:29 So this is what we refer to as the,
08:31 the deadweight costs of default.
08:33 Uh,
08:34 and the way I think about them is,
08:36 think about what,
08:38 think about
08:38 the lender gets zero
08:40 in a default.
08:40 Let's go to that extreme again.
08:42 The lender gets 0 in that default.
08:44 Um,
08:45 what does the,
08:46 what does the government get in return?
08:48 OK?
08:48 And it might be less than the full amount,
08:50 uh,
08:51 of the,
08:51 of the value of the bonds that are being defaulted on.
08:54 And that gap is the deadweight loss.
08:56 I'm gonna come back to that in a second.
08:59 And then
09:01 This object uh becomes an important equilibrium uh force
09:06 for borrowing and pricing,
09:08 and,
09:08 and that's kind of what we trace out.
09:11 OK.
09:13 All right,
09:13 so,
09:14 some lessons that we gained.
09:17 I'll summarize the lessons and then I'll talk a little bit about the
09:20 framework just to give you a sense of how we do it.
09:24 So the crucial thing is that the,
09:26 the,
09:27 the,
09:27 the relationship between the lender and the borrower is not zero-sum.
09:32 The
09:33 lenders lose,
09:34 let's again,
09:35 let's take the extreme of the lenders lose everything.
09:37 You could think about a haircut of 40%,
09:39 50%,
09:39 whatever you want,
09:40 but the lenders lose a certain amount.
09:43 The government doesn't gain that much.
09:45 So if the government promised to pay
09:47 $1 in a particular state and defaults,
09:50 the lender loses that $1.
09:53 The government doesn't capture the full dollar
09:55 because of these disruptions to the economy,
09:57 disruptions to the trading relationships,
09:59 spillovers to other relationships,
10:02 all these things that have built up in the literature and we try to measure,
10:06 all these things
10:07 factor in,
10:08 so the government doesn't capture the full.
10:11 Uh,
10:11 amount that the lenders lose,
10:12 so it's not a zero sum.
10:15 These deadweight costs are the,
10:16 the loss to the surplus between the lender and the borrower.
10:20 And the way I,
10:21 the,
10:21 the
10:22 thought experiment I use is what if the government,
10:24 I mean,
10:24 what if the lenders had just forgiven this amount of debt?
10:27 What if they had just said,
10:28 OK,
10:28 we're going to write it off?
10:29 And we saw a bit of that in um
10:31 Europe.
10:32 Uh,
10:33 what if they had just written it off and then we go from there.
10:35 And we,
10:36 we see it uh uh uh in restructuring sometimes,
10:39 but
10:39 That's not
10:41 the equivalent of a default,
10:42 right?
10:43 In the data,
10:44 a default typically
10:46 involves.
10:48 Months,
10:48 if not years
10:49 of
10:50 renegotiation,
10:51 you know,
10:51 the median
10:52 uh uh
10:54 struct restructuring negotiation is multiple years.
10:59 Complicated negotiations
11:03 during which the governments are excluded
11:06 from from borrowing and lending typically,
11:09 and they have disruptions to the domestic financial system.
11:13 So all these things just show up as costs.
11:17 OK,
11:18 so then if you think about that relationship.
11:21 And you think about kind of an efficient
11:23 contract between lenders and borrowers,
11:26 the idea would be,
11:27 let's avoid those costs,
11:29 right?
11:29 There are a loss to the joint surplus
11:31 of this partnership,
11:32 let's avoid it.
11:34 How does the competitive equilibrium
11:37 uh lead the government,
11:39 yeah.
11:42 Oh.
11:45 Take care.
11:48 Um,
11:50 can you please,
11:50 everybody,
11:50 can you please mute.
11:52 Thank you.
11:53 Um.
11:54 OK,
11:55 so in a competitive equilibrium,
11:56 only the government can minimize the chance of default,
11:59 uh,
11:59 and the way they can do that is by restricting how much they borrow,
12:04 or,
12:04 or imposing some additional kind of institutional controls perhaps,
12:07 but really it's kind of avoiding,
12:09 uh,
12:10 Getting a lot of debt in the first place.
12:14 And so what is it about a competitive equilibrium that
12:18 provides the correct incentives,
12:20 correct,
12:20 from the carbon efficiency standpoint,
12:22 for the government to minimize these deadweight costs?
12:26 Now,
12:27 outside of competitive equilibrium,
12:28 there may be paral improvements,
12:30 very,
12:30 uh,
12:31 you know,
12:31 there could be borrowing,
12:32 uh,
12:32 bargaining,
12:33 restructuring,
12:34 thirty-party interventions,
12:35 but
12:36 it's useful to think through what happens in a competitive equilibrium
12:40 before we get to those,
12:41 uh,
12:42 uh,
12:42 non-market interventions.
12:45 OK,
12:46 so,
12:47 going to this idea that only the
12:49 government can minimize uh the probability of default
12:53 in a competitive equilibrium,
12:54 what are some of the
12:55 implications?
12:57 Security design is critical for aligning incentives.
13:01 All right.
13:01 And the lesson that we,
13:03 we emphasize in the models and then,
13:05 you know,
13:05 kind of the discussion for policy or or or.
13:09 Practitioners,
13:10 is that the security design
13:12 needs to align the incentives.
13:14 Now what do I mean by that?
13:16 If the government
13:18 increases the probability of default
13:20 by a certain percentage,
13:23 it's important that the government bears that cost
13:26 at the margin,
13:26 the full cost.
13:27 And if they reduce the probability of default
13:30 by a certain percentage through their actions,
13:32 they need to reap the entire gains.
13:34 Otherwise,
13:35 there's going to be an inefficient um.
13:38 Amount of borrowing or lending.
13:40 And we show that in
13:43 kind of in our simplified model
13:45 that short-term debt does this.
13:48 And what does short-term debt do?
13:49 It forces the government
13:52 to return to the market after every decision and say,
13:56 let's roll over all of our debt at new prices.
14:00 And so if the government has done something
14:02 to reduce the probability of default,
14:04 they capture that full benefit.
14:06 If the government has done something to enhance the probability of default,
14:08 they,
14:09 they pay the full cost.
14:11 And that's gonna lead to a uh what we call the constrained efficient allocation,
14:15 and we show that that's the case.
14:16 Obviously,
14:17 short-term
14:18 debt has other risks,
14:20 mainly rollover risk.
14:22 Uh,
14:23 uh,
14:23 that's,
14:24 I would say the primary shortcoming.
14:25 There's an additional
14:27 hedging,
14:27 uh,
14:28 um,
14:29 argument as well,
14:30 but I think that's secondary.
14:31 The primary one is rollover risk.
14:34 Having long,
14:35 we show having long maturity bonds with the variable coupon does
14:38 exactly what
14:39 the short-term debt does
14:41 in terms of the incentives,
14:43 uh,
14:44 but minimizes the rollover risk cause the principal doesn't have to be rollover,
14:49 rolled over.
14:49 And so that's another thing that we've,
14:51 we argue in the book and we've done some other work on quantifying this,
14:55 that really
14:56 kind of the ideal instrument,
14:58 the,
14:58 the
15:00 Happy medium
15:01 would be a long-term bond
15:04 with a
15:05 coupon that's tied towards
15:08 the government's default probability.
15:10 So maybe the spread between the government's uh um
15:14 Uh,
15:15 short term bond price and
15:17 a comparable emerging market,
15:19 let's say basket,
15:21 uh,
15:21 so they don't have to pay the fluctuations in the risk premium,
15:24 just the,
15:25 the,
15:25 the marginal net
15:27 change between its default probability and its and its
15:31 benchmark peers.
15:34 The other thing we show is that the yield curve
15:36 is not really a useful measure for this cost.
15:38 The yield curve is the average cost,
15:41 uh,
15:41 and it's not a useful
15:43 in the model,
15:44 at least,
15:45 and I think,
15:45 you know,
15:46 we believe that that's kind of an implication for the data.
15:48 It's not a good guide for maturity choice,
15:52 right?
15:52 So,
15:53 I mean,
15:53 that's one of the things that we talk about,
15:55 and I'll I'll come back to that.
15:58 OK.
15:59 Now,
16:00 some alternative non-market mechanisms for doing this,
16:04 and really it's kind of aligning the fiscal
16:08 authorities'
16:09 incentives to minimize this deadweight cost
16:12 occurring in reality.
16:14 Conditionality is exactly what this is trying to do.
16:17 For these governments to pursue a fiscal reform
16:21 uh
16:22 through a non-market mechanism.
16:24 It's an imperfect proxy.
16:26 And the reason it's imperfect,
16:27 it's very hard to monitor and very hard to enforce.
16:29 You,
16:30 you know,
16:30 the,
16:31 uh,
16:31 the World Bank and the IMF know this,
16:33 you know,
16:33 better than anybody.
16:34 It's very difficult,
16:36 uh,
16:37 to set these targets and then,
16:39 um,
16:40 enforce them.
16:42 Forcing the,
16:43 the government to go back to the market
16:46 with at least the coupon payments and pay the costs if they fail and,
16:50 and,
16:50 and reap the benefits uh
16:52 if they succeed,
16:54 uh.
16:55 Is a credible
16:57 enforcement mechanism.
17:00 Uh,
17:01 So this is what I mean by the role for for better security design.
17:05 The other thing that we show is that the lender of last resort functions.
17:09 So think about a rollover crisis,
17:11 uh,
17:12 and the IMF can step in,
17:14 like,
17:14 like they did in Mexico in '94,
17:16 95,
17:17 like the ECB or the troika did in,
17:20 in Europe
17:21 in the 2010s.
17:23 Uh,
17:24 these are very crucial
17:26 uh
17:27 to resolve
17:28 runs.
17:29 Uh,
17:29 what we show is that with long-term bonds,
17:32 uh,
17:33 having this role,
17:34 this lender of last resort can complicate things.
17:37 And,
17:38 uh,
17:39 I'll,
17:39 I'll mention that again.
17:40 I'll come back to these once I showed this kind of the basic framework.
17:45 Then another thing that we show is that political economy frictions.
17:50 Get amplified.
17:51 So small
17:53 gaps in the discount rate
17:55 between the political incumbents and the population get amplified.
17:59 Uh,
18:00 and we show that
18:03 these frictions
18:04 can lead to very big welfare costs because of these deadweight costs of default.
18:08 The government is willing to take on a
18:10 little too much risk relative to the population.
18:13 The other thing that we've shown in other papers
18:15 and we talk about in the book is that uh
18:18 when you just look at
18:20 um outcomes in the world,
18:23 uh,
18:24 Borrowing from it,
18:26 from,
18:26 from international agencies or from uh uh private debt bond markets
18:32 are not
18:32 uh
18:34 um
18:34 common in the countries that have really had long-run growth.
18:37 And you think,
18:37 you know,
18:38 think about the Asian economies
18:40 that have really kind of succeeded in this,
18:42 uh,
18:43 versus kind of some of the Latin America versus some of the uh African countries.
18:47 Obviously,
18:48 they are very,
18:48 these are very different experiences,
18:50 but certainly when you look at kind of what's happening with the
18:54 Uh,
18:54 government
18:56 debt
18:57 process in these economies,
18:58 we don't see a strong argument for government
19:00 borrowing as a as a mechanism for growth.
19:04 And then the final lesson for IFIs
19:06 is that these debt sustainability formulas really need to focus on maturity,
19:10 coupon type,
19:12 bond type,
19:13 flexible versus uh fixed rate,
19:15 uh,
19:16 and then obviously the history
19:17 um
19:18 of debt and default,
19:20 cause that seems to play a role.
19:22 OK.
19:24 So,
19:25 let me just give you an overview of the,
19:26 of the basic uh uh arguments in the book.
19:29 Uh,
19:30 I think I,
19:30 yeah,
19:30 I have
19:31 20 minutes left.
19:32 So,
19:32 um,
19:34 Here's kind of the partial list of frictions that we,
19:36 we eventually put into the model.
19:38 So the lack of commitment obviously is crucial.
19:41 Uh,
19:42 limited state contingency,
19:44 so the bond markets don't
19:46 have a lot of contingent,
19:48 uh,
19:49 to them.
19:50 The GDP warrants,
19:51 you see them,
19:51 but they're very,
19:52 um,
19:54 Very rare.
19:55 Uh,
19:56 even things like in the Caribbean,
19:58 you know,
19:58 you'd make,
19:58 you'd think like hurricane,
20:00 uh,
20:00 um,
20:01 covenants would be a common feature,
20:03 and those are only
20:05 more recent and they're very difficult.
20:07 Uh,
20:08 uh,
20:08 they have had a difficult kind of history.
20:11 There's an incentive to dilute legacy creditors,
20:13 and that's gonna be crucial when we talk about long-term bonds.
20:16 Large deadweight costs of default,
20:18 it seems to be the case
20:19 that it's not a zero-sum game,
20:22 self-fulfilling runs,
20:23 currency mismatch,
20:24 and then these political economy distortions that I've talked about.
20:27 All right,
20:28 so this is kind of just a partial list.
20:31 Uh.
20:34 So we tried to
20:35 Embed those one by one
20:38 in a simple framework.
20:40 Most of the literature,
20:40 as I said,
20:41 uses these medium scale quantitative models,
20:44 uh,
20:44 and we try to unpack that,
20:46 and we,
20:47 we have a way of doing it,
20:48 which,
20:48 which kind of captures these core mechanisms,
20:51 but leaves out
20:52 certain things.
20:54 All right,
20:54 so let me give you the basic overview.
20:58 For those interested in the modeling,
21:00 yes,
21:00 this is kind of how we approach it.
21:03 Um,
21:04 So think about a small open economy.
21:06 They face a large pool of risk neutral lenders,
21:09 atomistic lenders.
21:10 So this is kind of gives rise to the inability to coordinate on better outcomes,
21:14 rollover risk,
21:15 uh,
21:16 Pareto
21:17 suboptimal outcomes.
21:19 There's just a lack of this coordination that goes on.
21:22 The lenders discounted some risk-free rate R,
21:25 and the government discounts at some rate row,
21:28 which is probably
21:29 relatively impatient compared to the world interest rate.
21:35 That
21:36 one of the tricks we use is that the government faces a constant endowment.
21:40 Why?
21:41 Where does that leave out?
21:41 That leaves our business cycle at risk.
21:45 I think
21:46 having studied these models both quantitatively in the data,
21:49 kind of normal business cycle fluctuations are the least of the concern in these,
21:54 these economies.
21:55 In terms of their debt default dynamics,
21:59 you know,
22:00 uh,
22:00 you know,
22:01 think about a normal business cycle,
22:02 1% movement in output.
22:04 It's just not,
22:05 um,
22:06 what we're talking about when we think of these major crises.
22:09 Um.
22:10 Concave utility,
22:11 this gives kind of an incentive to smooth things
22:14 over time so you don't get these lump sum payments and,
22:17 and,
22:18 uh,
22:18 and there's an incentive to default if the payments become owners.
22:22 Uh,
22:22 they trade this non-contingent bond,
22:24 that's what I said.
22:25 This seems to be a defining feature of sovereign debt markets,
22:28 the lack of contingency,
22:29 uh,
22:30 and they cannot commit to repayment.
22:32 All right,
22:33 so those are the core ingredients.
22:35 Then,
22:36 where do we bake in the risk?
22:38 We bake in the risk of payoffs to default.
22:40 So there's some
22:42 outside option the government can grab
22:44 uh the superscript D,
22:46 which we think of kind of like.
22:49 Pay zero debt or some fraction of debt,
22:52 you can,
22:52 we go to a full haircut,
22:54 but it wouldn't be hard to conceptually think about something in between.
22:58 They don't pay anything,
23:00 and they capture their endowment,
23:01 but perhaps with some loss of the endowment because of the dislocation.
23:05 So that's the deadweight cost of default,
23:06 right?
23:07 And you can just think about
23:08 losing some fraction of the endowment and
23:11 The,
23:11 the,
23:12 the
23:13 lenders don't get it.
23:14 So,
23:15 you know,
23:15 think about
23:16 debt forgiveness would be
23:18 the government captures this Y,
23:20 whichever,
23:21 however you want to think about that,
23:22 the output of the economy or the revenues of the government,
23:26 uh,
23:26 without any payments,
23:28 uh,
23:28 if the debt were forgiven and they get 1 minus tau y if they default,
23:33 and so this tau y
23:35 is the
23:36 deadweight cost of default.
23:38 And that varies over time.
23:41 Uh,
23:41 in our model,
23:42 and this idea here is we can think about
23:45 the political consequences of default vary over time,
23:48 the economic consequences.
23:50 We saw in Europe
23:52 that um,
23:55 Uh,
23:55 you know,
23:55 there was some reluctance at the beginning to allow default,
23:59 but then they said,
23:59 OK,
24:00 we're going to have this private sector involvement
24:03 and minimize these costs of default.
24:05 This is the kind of idea that could be bailouts.
24:08 This is the idea that we want to bake into this,
24:10 and we do it in an extremely simple way
24:12 that the normal state
24:14 is this harsh punishment,
24:16 and so you're just not going to see defaults
24:19 in the normal state and then with some probability lambda.
24:23 There's almost like a,
24:24 a,
24:24 a weak enforcement regime,
24:27 right?
24:27 There's almost like the stigma or however you wanna think about these costs
24:31 become
24:32 less,
24:32 and then that's the risk for the creditors.
24:34 It's a,
24:34 it's actually a windfall for the,
24:36 for the lender,
24:37 for the borrower
24:38 that they get to kind of write off this debt at,
24:40 at,
24:41 at less cost,
24:41 but it's a,
24:42 it's a risk to the creditors,
24:44 OK?
24:44 And so that's the underlying risk.
24:46 And so,
24:47 um,
24:50 What this captures is this lack of commitment.
24:53 The deadweight cost of default,
24:55 and again,
24:55 this is,
24:56 you know,
24:56 the payoff to default,
24:58 this is the payoff to just consuming your endowment forever,
25:02 uh,
25:03 having zero debt,
25:04 default gives the government less value than zero debt,
25:07 uh,
25:07 despite the full write-off,
25:09 and that's the gap.
25:10 Lack of state contingency,
25:12 and then we can have this impatience,
25:15 uh,
25:15 which I'm going to ignore for,
25:16 for now and come back to at the very end.
25:19 You know,
25:19 one of the things that we missed,
25:21 risk premium,
25:21 consumption hedging,
25:23 the default renegotiating haircut process,
25:26 all these things are important.
25:27 We just put them to the side to capture what we really want to focus on.
25:32 All right.
25:33 Then with the simple model.
25:35 What would be the first best?
25:37 Well,
25:38 you,
25:38 we show that you keep consumption constant,
25:41 uh,
25:42 then
25:43 when
25:44 this high outside option arrives,
25:46 you don't want the lender,
25:48 the lenders don't want the government to walk away,
25:51 so they adjust
25:52 and just keep them in the,
25:54 they basically write down the debt
25:56 and keep them in the relationship,
25:58 and this makes consumption contingent.
25:59 This would be the ideal thing.
26:01 We see,
26:02 you know,
26:03 something like the COVID pandemic,
26:05 and the private sector,
26:06 we had the ability to just say,
26:08 OK,
26:08 we're gonna write down the debt because of this.
26:10 This is obviously,
26:11 as you know better than I,
26:12 very difficult to do in practice,
26:14 and we rule it out in the,
26:15 in the model.
26:17 So what does it do
26:18 with non-contingent lending?
26:20 Well,
26:20 we
26:21 solve a planning problem.
26:23 And it involves backloading consumption.
26:26 The government is forced to save to avoid default.
26:29 They get to what we call a safe region,
26:32 and then at some point,
26:34 they're
26:35 immune from this outside option.
26:37 They have low enough debt that they're not going to default,
26:40 and that's kind of the goal.
26:42 And so it's a race between how quickly they can get there before the arrival.
26:46 That's what we'd like to see in the competitive equilibrium.
26:50 We show that
26:51 that can be implemented with short-term bonds,
26:54 all right?
26:54 And that's
26:55 the,
26:56 the key
26:57 uh
26:58 role in the competitive equilibrium of prices.
27:01 But let me just kind of walk through this,
27:03 the equilibrium in one slide.
27:07 This is the government's value
27:09 function as a function of debt.
27:11 If they have too much debt.
27:14 Uh,
27:14 this is this the upper bar.
27:16 If they have too much debt,
27:17 when that arrives.
27:20 If they repay,
27:21 they would get this.
27:23 If they default,
27:25 they get this job.
27:27 And they're going to default.
27:28 If they have too much debt,
27:29 they're gonna default.
27:31 Now,
27:31 the interesting thing is,
27:33 why do they save out of this crisis zone?
27:36 As I said,
27:37 why don't they just wait for this.
27:39 Quasi-debt forgiveness or weak enforcement regime,
27:42 or,
27:42 or however you want to interpret this high outside option.
27:46 Why don't they just roll over the debt and wait for that?
27:49 And the insight is that.
27:52 There's a
27:53 price they have to pay
27:55 in the safe region,
27:57 right,
27:57 where they're not going to default,
27:58 they pay the risk-free price,
27:59 R.
28:00 In the crisis region where there's this probability
28:03 hazard of default lambda,
28:06 every period,
28:07 they pay our
28:08 plus lambda.
28:11 The lenders are competitive.
28:12 They're going to get completely compensated for this risk,
28:15 so the government has to pay the full lambda times
28:18 the amount of debt.
28:22 To compensate them for this risk of default
28:24 with a deadweight cost of default,
28:26 the government doesn't capture,
28:28 this is not equivalent to giving them lambda B.
28:31 The government doesn't capture the full loss,
28:33 so they're paying the lenders to compensate them for this,
28:36 but they're not capturing it in the default state.
28:39 We show that that gives them enough incentive
28:42 to save their way out of the crisis region and get to the safe region,
28:46 OK?
28:47 But only
28:48 if
28:49 they're rolling over,
28:50 this is just basically what the slide says,
28:52 only if they're rolling over the full
28:54 amount of debt.
28:56 They fully internalize the deadweight costs of default.
29:00 It's exactly what the planning problem would implement.
29:02 It captures all of the government captures all
29:05 the benefit of reducing the probability of default.
29:07 They go from R plus lambda to our star.
29:10 Everything is aligned,
29:11 OK?
29:16 What if we had
29:17 long term bonds?
29:21 Well,
29:22 We can do a planning problem again.
29:25 There are these legacy bonds.
29:27 They have to pay some coupon.
29:29 We just set it at the R star.
29:31 We look at the new planning problem.
29:34 The key is to the planning problem is that
29:38 like
29:38 the the new lenders get Y,
29:41 the government gets C,
29:42 and then the legacy bondholders get this R times BL.
29:46 And then we assume this paripasu condition that default happens to everybody,
29:50 which is reality.
29:53 Basically it's like
29:55 there's a smaller endowment is split between the new lenders and the government,
29:59 and there's
30:00 a smaller endowment only in the repayment states,
30:03 so
30:04 there's a smaller deadweight cost
30:06 of default,
30:07 ignoring the legacy lender because they're they're not
30:10 the ones who are pricing the new bonds.
30:14 So this is what gives rise to delusion.
30:18 All right,
30:19 we can decentralize this modified planning problem.
30:22 We show that there are these legacy bondholders.
30:26 The government would only issue short-term debt and then going forward,
30:30 they have to service this legacy bondholders.
30:33 We saddled them that at the beginning.
30:35 They never trade the long-term bond,
30:37 they service the coupon.
30:39 But they face this weaker incentive to save,
30:42 right?
30:43 And the weaker incentive comes because they're only rolling over this new debt.
30:48 They're not rolling over the legacy debt,
30:50 and they don't have to then internalize the cost to the legacy lenders of a default.
30:55 They don't reap the benefit of reducing that probability.
30:59 All right.
31:02 The government avoids issuing long-term bonds.
31:04 Actually,
31:05 Sergio has a beautiful paper.
31:06 It started a lot of this literature
31:08 showing that in these crises,
31:10 the government stopped issuing long-term bonds,
31:14 and that was a bit
31:15 puzzling.
31:15 They have an argument about risk premium,
31:17 but it shows up in our model too,
31:19 and it really shows up through this incentive.
31:22 The government doesn't want to be.
31:25 Uh,
31:26 and an inefficient allocation going forward,
31:28 so they rely on short-term debt to align the incentives with the new lenders,
31:33 and it's the,
31:34 it's the efficient thing for the government to do,
31:35 and it's going to be the equilibrium outcome.
31:38 It turns out the yield curve is irrelevant for this.
31:41 The yield curve is kind of the average price,
31:44 uh.
31:45 What we want is the government to have the marginal
31:48 change in the default probability.
31:51 Interestingly,
31:52 we can show in this model.
31:53 What about debt buybacks?
31:56 Debt buybacks
31:57 are.
31:59 A bad idea for the government in equilibrium.
32:02 And the idea is that after the exchange,
32:05 the government is going to have a greater incentive to
32:08 reduce their debt because they have the shorter term bonds.
32:12 The market
32:13 priced in
32:14 this new efficient allocation.
32:17 The,
32:17 the old legacy debt is worth more.
32:20 So if they're going to
32:22 repay these,
32:23 repurchase these at market prices,
32:25 they're going to be very expensive.
32:28 This is never profitable for the government.
32:30 We show this.
32:30 It's very similar to the old Bella Rogoff buyback boondoggle.
32:35 Result,
32:36 but it's a very different mechanism.
32:37 The buyback boondoggle involve the liquidation value and diluting
32:42 uh
32:42 uh the liquidation value.
32:44 Uh,
32:45 this is all about
32:46 incentives going forward.
32:49 Uh,
32:49 you could get a Preto improvement
32:51 through a swap
32:52 at non-market prices,
32:54 but requires a collective,
32:55 uh,
32:56 uh,
32:56 bargaining because there's always this incentive to hold out.
33:00 All right.
33:01 How else to kind of
33:03 mimic
33:04 these incentives,
33:05 fiscal rules,
33:06 hard to enforce.
33:07 I already talked about that.
33:08 We saw in the Eurozone,
33:09 they had built into the Eurozone,
33:11 they had.
33:13 Deficit and actually debt
33:15 to GDP targets that were violated,
33:18 it's just difficult to enforce.
33:20 Conditionality.
33:23 It's very difficult as well.
33:24 What do I mean by conditionality?
33:26 Look at the US Liberty bonds and after World War One.
33:30 It's kind of a rare
33:32 example.
33:32 They issue these bonds and they promise
33:35 to compensate the bondholders if they go back to the market and issue new bonds
33:40 at higher spreads.
33:42 Uh,
33:43 and they basically promised to exchange the old bonds
33:45 for the new bonds if they did that,
33:47 and it did happen.
33:48 But that aligns the incentives
33:50 perfectly in our model.
33:52 Limits on future fiscal policy are hard to enforce.
33:55 You could always have
33:56 the banking system,
33:57 for example,
33:58 accumulating debt,
33:59 which are really implicit liabilities of the of the fiscal authority.
34:04 Again,
34:05 I'm just gonna echo it again.
34:06 Floating rate that does it.
34:07 Uh,
34:08 it,
34:08 uh,
34:09 a perpetuity with a coupon tied to some market rate,
34:12 uh,
34:12 does the efficient outcome,
34:14 and it provides safety from rollover crises.
34:16 I'm not 100% sure why we don't see more of these in the data.
34:20 Um.
34:22 Uh,
34:22 they would have to
34:24 be immune from kind of movements and risk premia.
34:27 But that's not hard to do with these liquid,
34:30 uh,
34:30 and well-known EMBI indices.
34:32 You could just benchmark off of them.
34:35 All right.
34:37 Last,
34:37 uh,
34:38 2 points.
34:41 With long-term bonds,
34:43 we show that there's multiplicity.
34:45 There's well-known uh uh
34:48 benefits of long-term bonds to avoid rollover crisis.
34:53 What is the multiplicity that arises with long-term bonds?
34:57 With long term bonds,
34:58 future fiscal policy determine prices.
35:01 Prices influence the government's willingness to implement a fiscal reform.
35:05 If the government can't commit to these fiscal reforms,
35:08 uh,
35:09 there may be
35:10 multiple equilibrium,
35:11 one where the government.
35:14 Implements the fiscal reform and is rewarded with a very good prices,
35:18 very low spreads
35:20 going forward.
35:21 There's also another equilibrium
35:23 where the government fails to implement the fiscal
35:26 reform and that's priced into the long-term bonds,
35:29 um,
35:30 what we call self-fulfilling delusion,
35:32 OK?
35:33 Interestingly,
35:34 in that model,
35:35 if there is a lender of last resort.
35:39 Which we
35:40 interpret as a floor
35:41 on the price of government bonds.
35:44 And it's.
35:45 Too loose of a floor in terms of uh without conditionality,
35:49 it will select the bad equilibrium.
35:51 It will induce this dilution,
35:53 right?
35:53 And it's because
35:54 with the floor and the prices,
35:55 the government
35:57 doesn't get punished for
35:59 failure to implement a fiscal reform,
36:01 and the reward is not as big
36:03 for implementing it.
36:05 And that's exactly this incentive I've been talking about,
36:07 and it will select the bad equilibrium.
36:10 Really,
36:10 what you need is a lender of last resort plus a hard conditionality.
36:14 If you get to a certain level of debt to
36:15 GDP we'll provide the lender of last resort function.
36:19 That will work in our model,
36:20 but,
36:20 but I've already talked about the um
36:23 Weakness of that in terms of implementability and credibility of these things.
36:28 All right.
36:29 Finally,
36:30 the other things we talk about in the book,
36:31 political economy frictions,
36:33 and we
36:34 mimic them
36:35 by looking uh at
36:37 uh
36:38 impatient governments
36:40 due to political turnover.
36:42 And we showed that just a small.
36:47 gap between the private agents and the government.
36:51 Plus,
36:52 long term bonds,
36:53 plus deadweight costs of default.
36:55 If you put these things in there,
36:57 you get very big welfare losses.
37:00 Now,
37:00 you know,
37:01 and I hope I've convinced you that with short-term debt,
37:04 uh,
37:05 uh,
37:05 these small movements,
37:06 these small gaps don't have a big implication because
37:10 welfare implication because the government has enough.
37:14 Market incentives not to,
37:15 to,
37:16 to,
37:16 to dilute or borrow or to have a high probability of default.
37:20 But
37:21 given the long maturity bonds,
37:23 the non-contingencies,
37:25 and a small
37:26 gap in the discount factors between private agents and,
37:29 and the government,
37:30 you'll see big welfare consequences.
37:33 To the point when we've done the quantitative implementation,
37:36 I'm skeptical now
37:38 of even the value.
37:40 Of
37:41 allowing these governments to borrow.
37:43 Obviously they have debt outstanding,
37:44 so it's hard
37:46 to unwind that.
37:47 But starting from a state in which there's very low debt to GDP,
37:52 facilitating
37:53 their access to international loan markets,
37:56 to me it's not a clear welfare win.
37:59 The
38:00 the debt overhang,
38:01 and then we have an extension with investment.
38:04 The debt overhang not only brings these,
38:06 these deadweight costs to default,
38:08 there's a debt overhang to investment because once
38:11 the government has this high debt to GDP,
38:13 there's an incentive for the government to tax.
38:16 Or expropriate,
38:17 however you wanna
38:18 term it,
38:19 uh,
38:20 capital that's installed in the economy,
38:22 OK,
38:23 and,
38:23 and,
38:23 uh,
38:24 and that.
38:26 That overhang then becomes a,
38:27 a
38:29 disincentive to invest in these economies.
38:31 Uh,
38:32 this is what I referred to at the very beginning.
38:34 There's this allocation puzzle that was,
38:36 I think,
38:36 first documented,
38:37 at least I first saw it by Gerrius
38:39 and Jean,
38:40 uh,
38:41 and we
38:42 kind of
38:43 built on that empirically
38:45 to,
38:45 to look at government borrowing versus private sector.
38:48 And so let me just say what the allocation puzzle is.
38:51 And the way we document it with the government borrowing.
38:54 If you look at government borrowing,
38:57 uh,
38:57 uh,
38:57 net borrowing versus net saving,
39:00 uh,
39:00 as a function of long-run growth,
39:02 you get a relationship that the countries that grow,
39:05 the governments were net lenders.
39:07 So think of China as the extreme example.
39:09 They're net lenders to the rest of the world.
39:12 But,
39:12 you know,
39:12 there are many of these.
39:13 Singapore,
39:14 even Botswana will show up,
39:16 um,
39:17 in the,
39:17 in the scatter plot as a
39:19 Uh,
39:20 as a high growth,
39:21 uh,
39:21 low,
39:22 low debt country.
39:23 And then you have on the other side,
39:25 countries that have a lot of debt to GDP just have very low
39:29 growth.
39:30 And,
39:30 you know,
39:30 we do a lot of controls for this,
39:32 uh,
39:32 uh,
39:33 initial debt and,
39:34 and things like that,
39:35 um,
39:37 but it seems to be a,
39:37 a fairly robust pattern.
39:40 Uh,
39:40 and we have,
39:41 you know,
39:41 we show in the models that that actually,
39:44 uh,
39:44 with debt overhang to capital taxation,
39:47 it makes perfect sense in the model.
39:49 So,
39:50 one of the things after all these,
39:51 uh,
39:52 all this work,
39:52 both empirical and theoretical,
39:54 is I'm a skeptic
39:56 about the value of access to international loans and,
39:59 you know,
39:59 the World Bank,
40:00 you might have a different view or you might have
40:02 come around to this,
40:03 but,
40:04 uh,
40:04 um,
40:05 given the political economy frictions,
40:07 the deadweight cost of default,
40:09 and the debt overhang,
40:10 I think it's a high bar.
40:12 Uh,
40:14 before we start saying that this is actually a good thing.
40:17 So let me
40:18 stop there,
40:19 and I apologize for the technical difficulties.
40:23 Thank you so much,
40:24 Mark.
40:25 Um,
40:26 I basically,
40:26 I have a million questions,
40:27 but let,
40:27 let me ask,
40:28 let me ask,
40:29 uh,
40:29 um,
40:31 The discussion to come in session.
40:47 Can you see this?
40:49 Yup.
40:50 OK.
40:50 Thank you very much.
40:51 Um,
40:51 thank you very much,
40:52 uh,
40:52 Marcelo,
40:53 for the invitation to be here.
40:54 It's a pleasure to discuss,
40:57 uh,
40:57 Mark's,
40:58 uh,
40:58 work with,
40:59 um,
41:00 Manuel.
41:03 So,
41:03 the
41:04 type of um analysis that they are making in the book is super interesting
41:08 and it's done by two leading economists with a lot of brainpower.
41:13 So I think that we should listen to them carefully.
41:16 Uh,
41:16 when you talk to both Mark and,
41:18 and Manuel,
41:19 I have interacted with them over the years,
41:21 um,
41:22 they,
41:22 they think hard about these issues.
41:24 So I think it's,
41:25 it's a privilege for us to have him here.
41:28 And,
41:29 and what I try to do in my comments is to
41:32 try to take a look at the book and some of the topics that they cover
41:36 and,
41:37 and that Mark covered in his presentation
41:40 and put them from a
41:41 different perspective here
41:43 and try to pick,
41:44 um,
41:44 use this opportunity that we have,
41:46 Mark,
41:46 to pick his brain on some of these issues that he mentioned and some other,
41:51 uh,
41:51 issues that might be missing.
41:53 From,
41:53 from the book,
41:54 but they tend to come out when we think about
41:57 uh debt,
41:58 uh,
41:59 particularly in developing countries.
42:02 So the book provides a very,
42:04 as Marx mentioned,
42:05 is a very simple,
42:06 unified and canonical framework,
42:08 theoretical framework
42:09 on sovereign debt
42:11 and the incentives to default,
42:13 the cost of default,
42:14 etc.
42:15 The book covers several aspects like frictions in debt market,
42:19 commitment,
42:20 enforcement,
42:21 self-fulfilling crisis,
42:22 government choices,
42:24 uh,
42:24 about debt and default,
42:26 uh,
42:26 the issue of long-term bonds
42:28 that is,
42:28 uh,
42:28 important,
42:29 and he characterizes different equilibrium outcomes
42:33 and how the model behaves under different scenarios.
42:36 I think that for grad students or economists working on debt,
42:41 It's a must read to,
42:42 uh,
42:43 to try to understand some of the issues
42:45 when government start issuing debt.
42:48 There's plenty of uh food for thought in for one talk and for um this discussion here.
42:54 So let me pick up on some of the issues that I,
42:57 I think to me
42:58 were in my mind before
43:00 and they're still there when,
43:01 when reading the book.
43:03 So we'll um make different points.
43:06 One is to compare first,
43:07 uh,
43:07 sovereign debt with private debt so we can see the differences and why
43:11 uh sovereign debt might behave in this way that uh Mark was mentioning.
43:16 Then talk a bit about this issue of multiple equilibrium,
43:20 um,
43:20 and the choice of uh instruments.
43:22 Talk about defaults,
43:24 talk about,
43:24 about the political economy and the role of
43:27 institutions in trying to uh limit uh debt,
43:31 the role of IFIs
43:32 and heterogeneous debtors,
43:34 uh,
43:35 and also,
43:36 um,
43:38 the other questions.
43:42 So in terms of the,
43:43 the corporate finance literature and the finance literature look at,
43:47 typically looks at private debt,
43:49 uh,
43:49 and there is a difference,
43:50 big difference between private debt and sovereign debt.
43:53 Uh,
43:53 in
43:54 one is that the contracts are for private debt are enforceable.
43:58 Uh,
43:58 if a company
44:00 doesn't pay,
44:01 usually you have the courts that come in,
44:03 you have the,
44:04 the investors,
44:05 uh,
44:05 the debt,
44:06 the creditors,
44:07 they,
44:08 they are,
44:08 or they start acting.
44:10 With sovereign debt,
44:11 um,
44:12 the,
44:12 the courts have limited power.
44:14 They have limited power to grab assets of countries,
44:17 uh,
44:18 sovereign countries.
44:20 So the,
44:21 the countries,
44:21 um,
44:23 it's impossible for them to commit given that they are not going to be published,
44:26 uh,
44:26 punished.
44:28 In terms of the resolution,
44:29 uh,
44:30 it's a very different,
44:31 um,
44:31 uh,
44:31 resolution mechanism.
44:33 In,
44:33 in the case of private debt,
44:34 there is a bankruptcy,
44:35 and liquid an eventual liquidation of the company
44:38 after Chapter 11.
44:40 In the case of sovereign debt,
44:41 there is a lack of liquidation capacity,
44:44 and countries continue to exist.
44:46 That doesn't matter what happens.
44:48 The,
44:48 the country might not pay,
44:50 it might be in default for many years.
44:51 The country is still there,
44:52 and eventually it might pay.
44:55 And I think that that may
44:56 play a role in terms of the,
44:57 the amount of debt that is accumulated by governments versus the private sector.
45:04 There is the other issue which are the principals.
45:07 If the shareholders,
45:08 uh,
45:09 uh,
45:09 of the companies,
45:11 they are at risk from default.
45:12 If the government,
45:13 if the company defaults,
45:15 the share equity holders lose everything,
45:18 uh,
45:18 so they try,
45:19 uh,
45:20 the company not to default.
45:21 In the case of voters,
45:23 households,
45:24 the,
45:25 the,
45:25 the incentives are not that aligned in the sense that,
45:29 uh,
45:29 when they don't pay,
45:30 um,
45:32 They don't have to go through these adjustments.
45:35 They have to balance the cost of default,
45:37 but there is also incentives to renege on the debt.
45:41 So the,
45:42 the incentives from the principals are different.
45:44 And,
45:45 and also um the agents behave differently.
45:48 Managers,
45:49 uh,
45:49 they are having a short-term monitoring,
45:51 but they might have a long-term view in many cases,
45:54 not always,
45:55 but in some cases.
45:57 And the government,
45:57 they have a political cycle that they need to get re-elected.
46:01 And these short term incentives to kind of win
46:04 the next election and do well in within that cycle
46:08 is important and I think it's a central piece on what
46:12 Mark mentioned
46:13 in his book.
46:15 Uh,
46:15 so the lenders in the private sector are at risk of losing everything.
46:19 In the sovereign side,
46:21 they can lose everything,
46:22 as Mark mentioned,
46:23 but I think in many cases they're don't lose everything.
46:25 It's,
46:26 it's kind of
46:27 tends to be rare,
46:28 the cases in which lenders
46:31 lose everything,
46:32 so they usually recuperate part of the,
46:34 of the,
46:35 of the,
46:36 of the debt later on with future
46:38 debt renegotiations.
46:39 That might take long,
46:40 but eventually they get paid off.
46:42 And that might uh
46:44 add into the willingness of investors to lend aside from the
46:48 high risk premium that they can capture to begin with.
46:51 So I think that that might lead to large
46:53 uh debt accumulation sometimes on the government side,
46:56 uh,
46:56 plus this
46:58 um incentives sometimes to,
47:00 to go into default.
47:01 That is very different dynamics from the private sector.
47:05 So what happens,
47:06 um,
47:08 In
47:09 with the government is that usually during normal times,
47:12 the government pays low interest rates,
47:14 the debt is sustainable,
47:15 and then you have a feedback loop there,
47:18 everything looks good.
47:19 Normal times you don't have a problem,
47:21 but many times we enter into crisis times.
47:24 And crisis times,
47:24 you can enter because you either shock,
47:27 um,
47:27 sometimes you
47:28 accumulate a bit more debt that becomes unsustainable,
47:32 and then you start
47:33 triggering high interest rates,
47:35 or sometimes there can be a shock to high interest rates to
47:38 marking his top kind of discounts with the business cycle.
47:42 But during the business cycle,
47:43 you might get
47:44 a shock to,
47:45 let's say,
47:46 high interest rates
47:47 that might make an otherwise sustainable debt unsustainable very fast.
47:54 And this happens especially when you have short-term debt.
47:57 You have short-term debt is necessary.
47:59 Uh,
48:00 it could be a disciplining mechanism,
48:02 as Mark mentioned,
48:03 to,
48:04 for the government to reap the benefits from behaving well.
48:08 Um,
48:08 but that puts you,
48:10 uh,
48:10 in a situation in which you have multiple equilibrium very
48:15 easily.
48:16 Uh,
48:17 so,
48:17 so therefore,
48:19 the current
48:21 Because
48:21 of this,
48:21 myopath tries to maybe uh get the
48:25 uh the short-term debt
48:26 that usually entails low,
48:28 low cost,
48:29 but the,
48:30 at the risk of running into this
48:32 multiple equilibrium and crisis.
48:35 The long term debt
48:37 Can avoid this type of self-fulfilling crisis or multiple equilibrium,
48:40 but not the fundamental ones.
48:42 You can see how a crisis with long-term debt.
48:45 So I think that the,
48:45 the short-term debt
48:47 that um Mark was mentioning that can align incentives can put you in these
48:52 uh type of situations very easily and that's where countries end up being.
48:57 Uh,
48:57 the short term debt could be also,
48:59 uh,
49:00 thought about some type of,
49:01 um,
49:02 foreign currency debt that you,
49:03 you might be hit when you borrowing short in foreign currency.
49:06 You,
49:07 uh,
49:07 there is a,
49:08 um,
49:08 an exchange rate shock.
49:09 That exchange rate shock might trigger again,
49:12 um,
49:12 a kind of a multiple equilibrium type of situation.
49:18 So
49:19 that leads me a little bit to default when,
49:21 when you run into these,
49:22 uh,
49:23 debt problems,
49:23 a sustainable debt,
49:25 or the,
49:26 the,
49:27 the,
49:27 the,
49:27 the
49:28 cause of default that that uh Mark mentioned is a big debt
49:32 weight loss that the government doesn't capture when there are these default.
49:36 There are different ways to,
49:37 to measure it.
49:38 Uh,
49:39 one
49:39 way to capture the cost of default
49:42 is,
49:42 uh,
49:43 that the average default,
49:44 uh,
49:44 spell is 8 years,
49:45 and this is work that,
49:46 um,
49:46 people,
49:47 um,
49:49 at Carmen's,
49:49 uh,
49:50 uh,
49:50 Reinhard's Group,
49:51 uh,
49:51 did for the WDR,
49:53 Clemens,
49:53 and Sebastian.
49:55 Uh,
49:56 so,
49:56 and it,
49:57 it takes,
49:57 it entails usually to restructurings.
50:00 And,
50:00 and the times that they have these big,
50:02 um,
50:03 um,
50:05 These defaults,
50:06 usually you have big um GDP collapses for a long time.
50:10 So in the future,
50:11 you pay the cost eventually of these defaults,
50:14 they are there for sure.
50:16 But one thing that is important is that these high default
50:20 uh costs make probably uh financing possible.
50:23 If you don't have a
50:25 high cost of default,
50:26 uh,
50:27 governments will tend
50:28 to default very easily because they don't face any cost.
50:32 And given that the government will tend to default very easily,
50:36 um,
50:37 lenders might not be willing to lend to
50:39 the government because the incentive to default is,
50:41 is too great.
50:43 So there are a lot of efforts as um
50:46 Marcelo was mentioning to try to reduce the,
50:49 the,
50:49 the,
50:50 the,
50:50 the cost of default.
50:51 The question is to what extent
50:53 these efforts
50:55 are,
50:55 can lead to some type of resolution
50:58 um
50:59 because you need some type
51:01 of high cost of default
51:03 to avoid the government from defaulting.
51:05 Otherwise,
51:05 uh,
51:06 governments will default.
51:07 So,
51:08 so the question is whether we can have a happy meddling with
51:13 But at the same time,
51:14 align the incentives of government to default,
51:17 um,
51:18 to not default.
51:19 Uh,
51:20 and that's,
51:20 I don't,
51:21 I don't know the answer to that,
51:22 but that's something that maybe,
51:23 um,
51:24 Mark can elaborate on.
51:28 Uh,
51:28 but it could be also that defaults are hard
51:31 to resolve because of other inefficiencies like market microstructure.
51:35 You have many bond holders,
51:36 you all have to agree,
51:37 etc.
51:38 You have a core process that is lengthy,
51:40 etc.
51:41 So I don't know to what extent the,
51:43 the cost of the high cost of default comes from these microstructural,
51:47 um,
51:48 issues in the market,
51:49 or they might
51:50 be intrinsic to,
51:52 uh,
51:52 the nature of the debt,
51:53 and,
51:54 and,
51:54 and what type of penalty you need to impose.
51:57 On,
51:57 on,
51:57 on the borrowers
51:59 to not default.
52:03 Um,
52:03 the other part that Mark mentioned,
52:05 I think,
52:06 is this issue of um impatient politicians that uh
52:10 plays a big role there.
52:12 I think that's an essential part,
52:15 and I think that's key in the case of sovereign debt.
52:19 And it,
52:20 it tends to lead to very risky bets,
52:22 both,
52:23 both ex ante and exposed.
52:25 Ex ante in terms of
52:26 governments trying to borrow short term
52:29 because it is kind of a lower,
52:31 uh lower cost.
52:32 It allows also governments to borrow a lot.
52:36 And when there is a crisis,
52:38 uh,
52:38 governments tend to kind of bet for survival
52:43 and take even riskier bets
52:45 to be able to make it to the next election and survive,
52:48 otherwise they wouldn't be in power.
52:50 So,
52:51 so this kind of
52:52 impatient politicians or short term horizonal political cycle
52:56 might induce to kind of bad situations
52:59 for governments.
53:00 So the big question here is,
53:04 If we assume that cutting debt is beneficial in terms of
53:09 market is very negative about debt
53:12 financing growth opportunities,
53:15 and it could be also the case that debt is not good for consumption smoothing,
53:19 but a big chunk of the literature argues that
53:22 there might be some room for consumption smoothing.
53:25 There might be some room for
53:28 financing.
53:30 Grow
53:30 through debt
53:31 instruments.
53:33 So to the extent that we want
53:35 some type of debt,
53:36 uh,
53:36 the question is to what extent we have,
53:38 can have some type of uh ins fiscal institution or institution
53:42 that might help,
53:43 um,
53:45 uh,
53:46 alleviate these problems that,
53:48 um,
53:48 Mark was mentioning.
53:50 I know that the fiscal stability pact might be hard to enforce,
53:53 uh,
53:54 but at least it imposes some constraints on,
53:57 on the executive.
53:59 To take some risk.
54:00 I mean,
54:01 I have experienced the case of Chile when they have this stability pact.
54:04 It's very different when you have some type of
54:06 agreement with Congress that you have to go through to get more,
54:09 more debt.
54:10 It's cumbersome,
54:11 but at the same time it constrains a little bit
54:14 the executive from maybe taking more risky behaviors
54:19 than otherwise.
54:21 Um,
54:23 the other,
54:23 other type of solution that can be imposed is,
54:25 is to have,
54:26 uh,
54:27 uh,
54:27 limits on certain type of debt that puts you
54:30 in,
54:30 in certain situations.
54:32 Uh,
54:32 the question is,
54:33 once you impose limits on,
54:35 on certain type of debt,
54:36 then the markets are less complete,
54:38 you,
54:38 the incentives might not be aligned in,
54:40 in other senses,
54:41 etc.
54:42 So
54:43 I don't know,
54:44 they,
54:44 I think they,
54:46 Mark mentioned is
54:47 negative on debt
54:48 and maybe the solution is set on debt.
54:50 But I think we might need to start to think creatively about other
54:55 solutions that is not zero debt because governments are going to borrow.
54:58 There might be a benefit from that.
55:01 To what extent we can implement debt arrangements
55:04 that are incentive compatible.
55:06 Another issue that comes
55:08 a lot in the discussions
55:11 in the political economy is when our governments have shocks,
55:14 and this happens not only with debt but with asset accumulation
55:17 and let's say the sovereign wealth funds,
55:20 to what extent shocks are transitory or permanent.
55:23 Uh,
55:24 so there is a lot of discussion there.
55:27 Uh,
55:27 government faces a shock.
55:28 They don't know,
55:30 the,
55:30 the,
55:30 the argument usually is that the shock is transitory,
55:34 saying you need to borrow or you need to
55:36 use the accumulated assets to smooth the shock,
55:39 but,
55:40 uh,
55:40 the shocks might be more permanent.
55:42 To what extent we can distinguish the more permanent
55:44 versus the transitory shocks might play a big role there
55:48 and the political economy.
55:50 Uh,
55:50 plays a big part
55:51 on that issue as well.
55:54 There is another issue that I think is central
55:56 to both the bank and the IMF and other institutions
56:00 is that the debt sustainability
56:02 um
56:03 analysis,
56:04 um,
56:04 agreements that they have.
56:06 And usually,
56:07 a lot is focused on the level of debt
56:10 over time and whether debt is sustainable over the long term,
56:13 long term under different scenarios.
56:16 And a lot of discussion
56:17 goes there into looking at the rate of
56:20 growth of the economy.
56:22 And the cause of death over the long term.
56:27 Um,
56:28 this is not the only indicator.
56:29 Mark mentioned that we,
56:30 we should be considering
56:32 other issues like maturity and seniority,
56:35 which I think are already considered
56:37 in the debt sustainability analysis that the bank and the fund make on an
56:42 almost daily basis.
56:43 I don't know how they are incorporated
56:45 and,
56:46 and how they should be modified in practice.
56:50 to make this debt sustainability analysis
56:54 more robust and less influenced by political decisions.
56:58 And what I mean by the political economic frictions that Mark mentioned is not only
57:03 the political economic frictions that we have at
57:05 the domestic level with the political cycles,
57:08 but also at the international levels.
57:10 When there are decisions about whether debt is sustainable or not,
57:14 there are also,
57:14 also pressures at the institutions from different stakeholders
57:19 about whether at the end it's a judgment call whether debt is sustainable or not
57:24 to some degree,
57:25 because you have to make assumptions about growth rates.
57:28 Uh,
57:28 and,
57:29 and cost of borrowing,
57:30 etc.
57:31 in the future.
57:32 So to,
57:33 to what extent,
57:33 we can make that type of framework a little bit more robust to uh
57:38 political and econo economy frictions,
57:39 I think uh might help
57:41 a lot.
57:43 So
57:44 Mark might have some thoughts about it
57:46 without going too negative.
57:48 On that.
57:49 The other issue that I think is um they are on the,
57:52 when
57:53 IFIs are called
57:55 is usually
57:56 um
57:57 when you have this uh a situation of crisis,
58:00 you have high interest rates,
58:01 unsustainable debt,
58:03 then you call
58:04 in many cases,
58:04 the IMF.
58:07 So the idea of this IMF interventions or IFI's intervention more generally,
58:12 is that um
58:13 you anchor the interest rate.
58:15 The interest rates are very um
58:18 um high.
58:19 So the idea is that you start lending at low-interest rates
58:22 that kind of stabilize will calm the markets
58:25 and then will become sustainable.
58:27 Again,
58:28 and the idea is that
58:29 you lower,
58:30 you lend at low interest rates and the assumption basically from the IFIs
58:35 is that the,
58:35 the debt that um
58:38 the IFIs are lending is senior debt.
58:40 So it's going to be paid first.
58:43 But one thing that I'm not sure that is taken into account so often.
58:47 Is that,
58:48 that by doing that,
58:50 um,
58:50 you might have,
58:51 you might be
58:52 diluting the existing bond holders.
58:55 Many times there is a,
58:57 in a situation in which you have
58:59 um
59:00 high debt,
59:02 um once uh debt that looks unsustainable given the interest rates.
59:06 Uh,
59:07 you lend
59:09 from,
59:09 let's say,
59:09 the IMF,
59:10 and that debt is usually in many cases used to repay existing bondholders.
59:15 Part of them and the other bondholders remain,
59:18 but you're replacing kind of junior debt for
59:22 senior debt,
59:23 debt that cannot be defaulted,
59:25 so you won't be able to write down that debt.
59:28 So that puts the country,
59:30 even the,
59:30 when the interest rates are lower,
59:32 you put the country
59:33 in a
59:34 difficult position
59:35 in which you cannot
59:37 just take a write down on,
59:39 on that debt
59:40 because um you're going to pay in full back
59:43 to the IMF.
59:44 That might make
59:45 the remaining debt more unsustainable given that it's junior to the senior debt
59:51 that the IFIs are lending.
59:53 So,
59:54 so you,
59:55 when you have these IFI interventions,
59:57 um,
59:57 you,
59:58 you,
59:59 you can have two scenarios.
1:00:00 Uh,
1:00:00 you have the scenarios in the,
1:00:01 the good equilibrium in which things
1:00:04 improve,
1:00:04 but it could be very easily the bad equilibrium
1:00:07 in which you end up with a lot of
1:00:09 um
1:00:10 Uh,
1:00:11 senior debt,
1:00:12 uh,
1:00:12 government indebted with the IFIs,
1:00:15 and at the same time,
1:00:16 some type of private debt that becomes,
1:00:18 uh,
1:00:18 that has,
1:00:19 even though it's lower than,
1:00:20 might be lower than before,
1:00:21 it has,
1:00:22 um,
1:00:22 high spreads,
1:00:23 uh,
1:00:24 and the high spreads
1:00:26 might
1:00:27 be a drag on the economy,
1:00:28 um,
1:00:28 moving forward.
1:00:31 Um,
1:00:32 I have other questions.
1:00:33 Um,
1:00:33 it,
1:00:34 it,
1:00:34 um,
1:00:35 some of them are related to the,
1:00:36 uh,
1:00:36 previous ones that are,
1:00:37 are,
1:00:37 are related,
1:00:38 but I think also
1:00:39 if Mark can elaborate on some of these issues as well,
1:00:42 that would,
1:00:42 that'd be great.
1:00:44 Um,
1:00:44 Mark mentioned this issue of,
1:00:46 um,
1:00:47 the,
1:00:47 the work that,
1:00:48 um,
1:00:49 Fernando Bruner,
1:00:50 Jae Ventura,
1:00:51 Alberto Martin have done
1:00:53 on between having
1:00:55 um
1:00:55 heterogeneous agents.
1:00:57 It,
1:00:57 the model that in the framework that Mark has is a representative agent,
1:01:02 but you can have secondary market trading between domestic
1:01:04 and foreign investors that might be playing a role.
1:01:07 You also,
1:01:08 aside from IFIs,
1:01:09 um,
1:01:10 you also have bilateral lenders now.
1:01:12 You have China
1:01:13 lending in debt markets
1:01:15 with
1:01:15 maybe a different seniority relative to others,
1:01:18 and that might be playing a role
1:01:20 in terms of the debt dynamics that and
1:01:22 the incentives that governments might have to default.
1:01:26 And it could,
1:01:27 you see it also in this pattern of reserve accumulation,
1:01:31 sometimes that happens with debt accumulation
1:01:34 uh that you see that uh governments have debt,
1:01:37 but they also want to prop up the reserves
1:01:39 to kind of
1:01:41 The balance sheet of the government that don't look very uh bad.
1:01:45 Um,
1:01:46 and even um China with the debt swaps is doing that.
1:01:50 They,
1:01:50 you reinforce reserves at the same time that you have debt.
1:01:53 And in fact,
1:01:54 um,
1:01:54 some work that um
1:01:56 our colleague here at the bank now,
1:01:58 Austin Samano,
1:01:59 that was a previous,
1:02:00 uh,
1:02:00 a student of uh Manuel at Minnesota.
1:02:04 Yeah,
1:02:04 on his job market paper,
1:02:05 he was documenting
1:02:07 the increase in both
1:02:09 uh reserves and debt.
1:02:11 So it's not that
1:02:13 government has assets and liabilities that the
1:02:15 two things accumulate at the same time.
1:02:17 Uh,
1:02:17 it's not that the,
1:02:19 the government accumulates.
1:02:20 I mean,
1:02:21 when they have,
1:02:22 uh,
1:02:22 reserves,
1:02:23 they don't use it to pay,
1:02:24 and write down the debt.
1:02:25 They just
1:02:26 keep the two things on the balance sheet
1:02:29 and they are kind of happy with that,
1:02:30 uh,
1:02:31 type of situation.
1:02:32 So we observed those types of cases.
1:02:35 So I was wondering whether
1:02:36 Mark has any thought about why
1:02:38 you might have incentive to accumulate both uh assets and liabilities
1:02:42 at the same time.
1:02:45 Um,
1:02:45 the other issue that that Mark was mentioning,
1:02:48 um,
1:02:49 and this is something that he mentioned in passing,
1:02:50 but I would like to,
1:02:51 to,
1:02:52 to see
1:02:53 if he can elaborate more,
1:02:55 is this issue of floating a rate debt that he,
1:02:57 he mentioned is might be a solution.
1:02:59 Is equity-like
1:03:01 in the sense that
1:03:02 um it's variable,
1:03:04 it's different from,
1:03:05 from equity in the sense that inequity during good
1:03:08 uh states,
1:03:10 you pay more than shareholders.
1:03:12 Here the good states,
1:03:13 you pay less,
1:03:14 the government reach the reach the benefit.
1:03:17 That's kind of the beauty of um this
1:03:19 floating rate debt.
1:03:21 In some sense,
1:03:22 it's similar to index debt.
1:03:24 Um,
1:03:26 and Mark mentioned that this type of debts,
1:03:28 uh,
1:03:29 instruments have not,
1:03:30 um,
1:03:30 um,
1:03:31 taken off.
1:03:33 So I was wondering why
1:03:34 that might have been the case.
1:03:36 I mean,
1:03:36 this issue of index debt had been proposed by many people like,
1:03:39 I mean,
1:03:39 Jeff Frankel
1:03:41 many years ago and others,
1:03:43 but,
1:03:43 um,
1:03:44 for some reason,
1:03:45 he,
1:03:45 I mean,
1:03:45 uh,
1:03:46 Eduardo Boristen also has papers on that,
1:03:47 and he,
1:03:48 he hasn't picked up.
1:03:50 Uh,
1:03:50 so I was wondering whether he has any,
1:03:52 any,
1:03:52 uh,
1:03:53 thoughts about that.
1:03:54 The same thing about domestic currency debt.
1:03:56 Now we have more domestic currency debt than before,
1:03:59 but probably not as much as before.
1:04:03 Usually,
1:04:03 in domestic currency debt,
1:04:05 you don't have incentive to default,
1:04:07 even though they might default on domestic currency debt as well.
1:04:10 But to some extent,
1:04:11 it isolates more,
1:04:12 you at least from the um
1:04:14 from the exchange rate shocks.
1:04:16 Uh,
1:04:16 so,
1:04:17 so the question is to what extent these type of debts that are
1:04:20 more kind of variable payments
1:04:22 are
1:04:23 not used more often.
1:04:26 Um,
1:04:27 the framework that Mark mentioned,
1:04:29 I think,
1:04:30 it seems to fully account for infrequent,
1:04:32 uh,
1:04:33 uh,
1:04:34 and cultural repurchases,
1:04:35 that might be the case.
1:04:37 Um,
1:04:37 that,
1:04:38 that I think is a useful,
1:04:40 um,
1:04:41 aspect of the book because when we were doing
1:04:43 And they were,
1:04:44 for example,
1:04:45 with Fernando,
1:04:45 we are wondering to what extent are repurchases are frequent or not,
1:04:49 and we found that they are not very frequent.
1:04:52 So I think in that sense,
1:04:53 the book provides a good
1:04:54 understanding of the incentives of why
1:04:56 uh those things might not happen.
1:04:59 And there is this issue of risk premium that um Mark,
1:05:03 he kind of um
1:05:04 sets aside in the book to do an analysis and to focus on default probability,
1:05:08 which is a good start,
1:05:09 but I think risk premium
1:05:11 uh is very important.
1:05:13 It might be difficult to incorporate in the analysis
1:05:15 that I would like to hear more from Mark,
1:05:17 but I think
1:05:18 might be a next step in terms of some of the work that they do.
1:05:21 But overall,
1:05:22 I found
1:05:22 the book and his work
1:05:24 super enlightening.
1:05:25 And I'd like to engage Mark in uh some
1:05:28 of these topics and congratulations on the book.
1:05:31 Thank you very much.
- add-style
- lp-body-content