col-xs-12
col-sm-12
col-md-12
col-lg-12
col-xs-12
col-sm-12
col-md-12
col-lg-12
videoType
dynamic-media
videoDmUrl
https://delivery-p136806-e1377785.adobeaemcloud.com/adobe/assets/urn:aaid:aem:82233d58-1b63-496d-a608-2c43410f0c85/play?assetname=ABCDE2024-Session-2.mp4
keyFrameImage
videoDescription
A video recording of the second session of The Annual Bank Conference on Development Economics 2024 "The Great Incoherence: Growth and Human Development in An Era of Stagnation." This session discusses "Sovereign Debt and Default."
timestamp

00:06 Welcome back everyone

00:08 to the afternoon session.

00:10 Uh,

00:10 for those joining online,

00:11 please join using the discussion

00:13 hashtag ABCDE 2024.

00:17 Uh,

00:17 I'm

00:18 very excited to welcome you to the second session

00:20 which will focus on sovereign debt and

00:23 default.

00:24 Our moderator is Manuela Francisco,

00:26 and I,

00:27 she,

00:28 it's,

00:28 it's now July,

00:29 so she has a new title.

00:30 So Manuela is the Global Director for Economic

00:33 policies,

00:34 um,

00:34 in the prosperity vertical.

00:36 Um,

00:36 I'll hand over to you,

00:37 Manuela.

00:43 Thank you,

00:43 thank you very much

00:45 and good afternoon everyone here in Washington and

00:48 good evening to everyone that is connected online.

00:51 Um,

00:52 I'm very pleased to see the inclusion of

00:54 a session on sovereign debt at this distinguished

00:58 conference because as we all know it's a very relevant topic.

01:03 Uh,

01:03 to motivate this discussion,

01:06 I would like to make 3 points that I will be very,

01:09 and I will be very brief.

01:11 First,

01:12 Is the following,

01:13 debt levels have stabilized,

01:16 these are the good news,

01:17 but they have stabilized at very high levels,

01:20 and many countries are facing liquidity challenge.

01:24 Just to give you a few numbers,

01:26 8 years ago in 2015.

01:29 25% of the low income countries that prepared DSF

01:34 were at high risk of that distress or

01:37 in distress.

01:38 By 2019,

01:41 this number reached 50%.

01:45 And this was because

01:47 between 2010 and 2019,

01:50 countries borrowed significantly,

01:53 which was facilitated by widely available capital at a very low cost.

01:59 Between 2010 and 2019,

02:01 during this period,

02:03 Public debt.

02:06 In low-income countries,

02:07 and middle-income countries grew by 131% in nominal terms.

02:14 Then,

02:15 as we know,

02:15 came COVID.

02:17 And then in 2022,

02:19 it started the tightening of monetary policy

02:22 and many countries were pushed over a cliff.

02:26 But there are encouraging news.

02:29 Uh,

02:29 some countries have started to tighten fiscal policy.

02:34 Uh,

02:34 some countries have gained access to the capital markets,

02:38 uh,

02:39 but we know that gross financing needs remain still very,

02:42 very high.

02:44 And while credit spreads have declined,

02:47 we know that the underlying interest rates are still very high.

02:52 My second message is that the that landscape

02:56 has changed.

02:58 And the debt structuring process are still adapting to this new world.

03:04 So as public debt levels increase,

03:06 the composition of the creditors also changed.

03:09 There has been a shift

03:11 towards non-traditional creditors

03:13 and bondholders.

03:16 There is also more domestic debt.

03:18 So now over 50 low-income countries have

03:22 domestic debt and total debt is about 25% of GDP.

03:28 There are good news with that.

03:30 Countries will be shielded from exchange rate volatility and currency mismatch.

03:36 But

03:37 when it comes the time of that restructuring,

03:40 this can become very complicated because of

03:43 the wider implications for the economy.

03:47 My last comment is that

03:49 our point is that countries are struggling

03:53 to roll over their debt.

03:56 We are living

03:57 a period

03:58 of very expensive funding costs.

04:02 There is uncertainty regarding the pace of monetary policy easing.

04:08 But I think there is certainty about the fact

04:11 that

04:12 the era

04:13 of very low interest rates will not come back.

04:19 So while major central banks around the world

04:22 are in a path

04:23 of monetary policy easing

04:26 and the Fed has already cut policy rates,

04:30 The,

04:31 the ECB I mean has cut policy rates.

04:34 The Fed is still waiting to see what's happened in the job market and the,

04:37 and prices.

04:39 But even

04:41 When the Fed cuts policy rates,

04:43 the low income countries will still face

04:46 very expensive

04:48 financing costs.

04:50 So that's why this is also relevant today.

04:54 So,

04:54 um,

04:55 it is my honor to moderate this distinguished

04:58 panel and let me introduce the panelists,

05:01 and I will start by the order of their presentation.

05:05 So

05:05 on my left I have Christoph Tribes from Keo University,

05:10 Enrico Mallucci from the Federal Reserve Board,

05:14 Ron Kang

05:16 from the World Bank.

05:17 And following their presentations of these three speakers,

05:21 Marcos Chasmos from the Fund

05:23 and Luca Bandiera from the World Bank will provide discuss and remarks.

05:28 So

05:29 let me pass the floor to

05:31 um

05:32 Christo.

05:40 Thank you so much,

05:40 uh,

05:42 for the moderation.

05:43 Thank you so much for inviting me,

05:44 uh,

05:45 and allowing me to,

05:46 uh,

05:47 present to you our most recent

05:49 paper,

05:50 uh,

05:50 we just joined with,

05:51 uh,

05:51 Lukas Franz uh from the Keen Institute,

05:53 Sebastian Horn from the Bank,

05:55 Bradley Parks from William and Mary,

05:57 and Carmen Reinhardt from Harvard University.

05:59 And this is part of a

06:00 broader research agenda on official lending,

06:04 so state-driven

06:05 lending on official creditors.

06:07 In general,

06:08 and uh

06:09 in particular on China's

06:11 overseas lending.

06:12 So despite the

06:13 growing literature on China's overseas lending,

06:15 we still have very little

06:17 insights on the financial terms,

06:19 especially in,

06:19 in particular,

06:20 the

06:21 returns,

06:22 the financial returns

06:23 of this 1 trillion

06:26 lending program that China has rolled out over the last 15 years,

06:29 and this is the focus.

06:30 Of today.

06:32 Um,

06:33 we see this paper speaking to a bigger phenomenon a long history

06:37 in which great powers have shown

06:39 a,

06:40 uh,

06:40 tendency to invest or to,

06:42 to launch these large scale infrastructure

06:45 investment

06:46 programs,

06:47 uh,

06:47 this just two historical examples.

06:50 Germany in the late 19th century launched the Berlin-Baghdad

06:52 railway or the US with this famous Marshall Plan.

06:55 Both of them investing heavily in infrastructure and faraway lands and facing

07:00 substantial

07:01 risks on this kind of large scale

07:03 uh investments.

07:04 Um,

07:04 so the benefit of looking at China today is that,

07:07 uh,

07:07 thanks to uh consider research by a data and,

07:10 and Uh,

07:10 and this other scholarly work,

07:12 we actually have a pretty

07:13 good idea,

07:14 more granular insight into these,

07:16 into one of the most prominent

07:18 lending programs of this type,

07:20 uh,

07:20 and we can assess

07:21 the dissociated default risk

07:23 and the returns

07:24 to the creditors,

07:25 to the,

07:26 to the great power,

07:27 uh,

07:27 lender,

07:28 uh,

07:28 in a more systematic way than we could do before.

07:31 Um,

07:32 so the key question I'm gonna focus on is,

07:35 is China's project of the century

07:37 a financial success?

07:41 There's two main contributions of the paper.

07:44 Uh the first,

07:45 and that's an essential ingredient to understanding

07:48 the financial returns,

07:49 is to understand defaults and credit losses.

07:52 So,

07:53 uh,

07:53 events of great events and the associated restructurings or,

07:56 um,

07:57 credit loss,

07:58 so haircuts,

07:59 uh it's often called

08:01 on

08:02 uh either the Chinese banks,

08:03 and we put together this comprehensive data set with

08:06 really the details of the sovereign debt restructuring.

08:09 Um,

08:09 deals,

08:10 uh,

08:10 with Chinese creditors

08:12 and then,

08:13 uh,

08:13 compute haircuts,

08:14 um,

08:15 and compare it to the Paris Club and other

08:17 and private external creditors.

08:19 And the second

08:20 contribution is we use the haircuts and the default information.

08:23 Um,

08:24 to compute exposed returns,

08:27 uh,

08:27 on,

08:28 uh,

08:28 on China's,

08:29 uh,

08:30 lending portfolio.

08:31 Uh,

08:31 so for every

08:32 one of these 3500 Chinese loans with the total volume of volume of 1 trillion,

08:39 we compute,

08:39 you know,

08:40 the exact cash flows,

08:41 we consider the,

08:42 uh,

08:43 credit events,

08:44 uh,

08:45 towards each of those loans

08:46 and come to the,

08:48 uh,

08:48 the keys inside is that realized rate of return is about 1.7.

08:53 Uh,

08:53 percent per year,

08:55 uh,

08:55 in real terms,

08:56 inflation adjusted,

08:57 which is neither particularly high

08:59 and not particularly low,

09:00 I'm gonna benchmark that to other,

09:02 uh,

09:02 creditor types.

09:04 So let me turn to the first part,

09:07 haircuts.

09:08 So the losses of Chinese banks on their

09:11 Belt and Road portfolio abroad.

09:14 The central contribution is that uh we we uh we

09:18 compile this granular data set,

09:20 um,

09:21 and importantly,

09:22 uh,

09:23 most of these deals

09:25 are rescheduling

09:26 agreements.

09:26 So they push out the payment

09:28 by a couple of years,

09:29 often not even cutting the interest rate.

09:31 So it's basically just a five year

09:33 kicking down

09:36 uh the can operation with very few instances of face value reduction.

09:39 So nominal debt reduction is rare in these deals.

09:42 Uh,

09:42 which is why you really need to,

09:43 to look at the present value or the,

09:45 the,

09:45 the,

09:46 uh,

09:46 payment streams of these loans

09:48 in detail.

09:49 Um,

09:51 So we use a standard uh to get Fettelmeyer haircuts uh

09:55 to compute uh haircuts.

09:56 So comparing

09:58 like a hypothetical scenario in which you say

10:00 99% of the credit is agreed and 1% does not agree,

10:04 uh,

10:04 and is repaid as if the default would never have occurred.

10:08 Uh,

10:08 what is the loss of the 99% participating vis a vis that 1% holdout is being paid,

10:14 right?

10:14 So,

10:15 so that's kind of at the point of the restructuring.

10:17 How much do creditors lose?

10:18 That's the way haircuts are computed here.

10:21 Um,

10:21 and,

10:22 uh,

10:22 as of course discount rate is crucial,

10:24 we use kind of uh an upper bound here,

10:26 the market yield,

10:27 an exit yield,

10:28 uh,

10:28 based on bond,

10:29 uh,

10:30 uh,

10:30 bond yields at the exit from the restructure but we also

10:33 use a official creditor uh

10:35 uh discount rates of CIRR as a,

10:38 a robustness check.

10:40 This is one of the key graphs of the paper,

10:43 so this is the size of the haircuts on Chinese

10:46 loans and various

10:48 um restructuring operations.

10:50 Um,

10:51 I guess I can't,

10:52 you can't see the,

10:53 uh,

10:53 no here.

10:54 So what you can see,

10:55 the big takeaway is that That

10:58 in the large number of cases,

10:59 haircuts are surprisingly low.

11:00 Uh,

11:01 so from

11:02 0 to 0 to

11:05 20%,

11:05 so the average

11:06 is just 12%,

11:08 and that's in historical comparison,

11:10 low.

11:11 Uh,

11:12 this is just one comparison,

11:13 we also do the comparison with Paris Club haircuts from our other uh work.

11:17 Um,

11:18 and what you can see is that

11:20 if we

11:22 Compare the haircuts on the Chinese restructuring events,

11:25 which are the red dots,

11:27 uh,

11:27 with haircuts,

11:28 uh,

11:29 on private external credits from the Cruz's Trevich data set,

11:32 uh,

11:32 we get a,

11:33 we get substantial differences,

11:35 so the mean

11:36 Chinese haircut is

11:37 10% or close to 0 if you want,

11:40 and the mean,

11:41 uh,

11:41 on private external credit is the mean haircut is 36%,

11:45 uh,

11:45 so considerably higher.

11:47 And if you consider the Paris Club haircuts,

11:50 those are above 50% on average,

11:52 so significantly higher than that even.

11:56 Let's move to the second part,

11:58 creditor returns.

11:59 Um,

12:00 we,

12:01 since these are not market instruments,

12:03 these are not traded instruments,

12:04 these are loans by Chinese banks,

12:06 um,

12:07 we compute internal rates of return.

12:10 Um,

12:10 and therefore,

12:11 for every single of those loans,

12:13 we,

12:13 we,

12:14 you know,

12:14 map the information we have on commitments,

12:17 uh,

12:18 the financial characteristics of the,

12:19 of the term of the terms,

12:21 the grace periods,

12:21 etc.

12:22 uh,

12:22 and then incorporate

12:24 the information on default and missed payments,

12:26 OK,

12:26 and the haircuts,

12:27 uh,

12:27 which gives us

12:29 this,

12:29 uh,

12:29 realized rate of return measure.

12:32 Uh,

12:32 just,

12:33 this is just an example,

12:34 what,

12:34 what kind of a

12:36 Just an illustration of the repayment terms,

12:38 once we do bond by bond and we,

12:40 uh sorry,

12:41 loan by loan and then aggregated

12:43 at a yearly level,

12:44 right?

12:44 What,

12:45 what are the repayments towards Chinese creditors from

12:48 the global south or from the debt Belt and Road countries,

12:51 you can see

12:52 that there is no,

12:53 it's no surprise that we're currently talking a lot about

12:56 uh debt sustainability and about repayment problems

12:59 towards China because this since 2020,

13:01 we really see a peak in repayment.

13:03 And we see that,

13:04 you know,

13:05 there's two big elephants in the room in terms of creditors,

13:07 and that's China Development Bank

13:09 and uh Export Import Bank,

13:12 so Exim Bank of China.

13:15 This graph shows you

13:17 the difference in realized payments.

13:20 Um,

13:21 so what,

13:21 what part of these repayments,

13:23 uh,

13:24 did not occur,

13:25 right?

13:25 So the red

13:26 parts of these bars are the missed payments.

13:28 And what you can see is that,

13:29 of course,

13:30 the Blue bars are much larger.

13:31 Missed payments account for a relatively small share

13:35 of total debt due to China.

13:37 Uh,

13:38 so overall,

13:40 China has not suffered massive arrears,

13:42 massive missed payments,

13:43 defaults on its portfolio.

13:47 We can also look at this by comparing ex ante returns or the promised returns at,

13:52 at,

13:52 at the landing point.

13:54 And exposed returns accounting for

13:56 uh losses and haircuts.

13:58 Um,

13:59 this is

14:00 the nominal return now,

14:02 uh,

14:02 and so the every country is colored according to the average return.

14:07 Um,

14:08 and when we compare these ex ante returns and compare that

14:12 to the exposed ones,

14:13 we see very little change in the coloring.

14:15 Only a few African countries see a substantial shift in the aggregate return.

14:20 Uh,

14:20 on,

14:21 uh,

14:21 Chinese,

14:22 uh,

14:22 loans from a,

14:23 from a Chinese perspective.

14:24 Uh,

14:25 so the

14:26 difference

14:27 in short between ex ante and exposed returns

14:29 is not particularly large

14:31 except for a few outlier countries.

14:33 OK?

14:35 So how does the return on China's

14:38 lending portfolio compare in the bigger picture,

14:41 right?

14:41 So

14:42 would it have

14:44 been a better alternative uh to put this money,

14:47 say,

14:47 in a sovereign wealth fund,

14:49 safe,

14:49 China's uh sovereign wealth fund

14:51 and invest in a portfolio of emerging market bonds instead

14:55 or maybe in a treasury

14:57 account with treasury.

14:58 bonds.

14:58 Uh,

14:59 to,

14:59 to understand that,

15:01 we compare the average real

15:03 uh return uh on a yearly basis of the BRI loans

15:07 with other

15:08 asset types,

15:09 asset classes,

15:10 and we can see that

15:11 the returns are higher

15:12 than the alternative of investing in US treasuries

15:16 and even higher than the alternative investing in Chinese government bonds.

15:20 But they are significantly lower.

15:23 Uh,

15:23 towards the benchmark of investing that same money

15:26 into Chinese equities

15:28 or into the standard,

15:29 you know,

15:30 JP Morgan MB,

15:31 uh,

15:31 portfolio,

15:32 uh,

15:32 or even MSCI

15:34 World Global Equity portfolio.

15:36 So

15:37 in that sense,

15:37 the markets,

15:38 uh,

15:39 returns in standard,

15:40 uh,

15:40 asset classes in emerging markets are considerably higher,

15:43 uh,

15:44 but it is,

15:45 uh,

15:45 more profitable than,

15:47 than investing,

15:48 uh,

15:48 holding more treasury bonds.

15:50 It's uh in the,

15:51 in the mid-range.

15:53 So why are creditor returns so low compared to that kind of market

15:57 benchmark?

15:59 To answer this question,

16:00 we're now gonna

16:01 make use of the large

16:03 loan by loan variation in returns.

16:05 We know the ex ante returns of every single loan.

16:08 And here it's just plotted the variation for every year

16:12 uh in the,

16:12 in the uh contracted return,

16:15 right,

16:16 ranging between 0

16:17 and uh even up to 10% or more,

16:19 right?

16:19 And what you can see that there's considerable variation

16:22 in the lending terms that China gives out.

16:25 So what drives this variation?

16:27 What,

16:27 what explains why some of these loans are really

16:30 low,

16:31 below market rates?

16:32 Um,

16:34 and to make sense of that,

16:35 we,

16:36 uh,

16:36 exploit the fact that China's overseas lending is state-driven.

16:40 So likely to have a,

16:42 uh,

16:43 or at least there is a possibility

16:44 of direct political

16:46 control and strategic lending because this is the state lending,

16:49 right?

16:50 Or state entities lending.

16:51 Um,

16:53 with that perspective,

16:54 we,

16:55 uh,

16:55 look at,

16:56 uh,

16:57 projects that might be in the interest

16:59 of China's,

17:00 you know,

17:01 overall,

17:02 uh,

17:02 um,

17:03 uh,

17:03 so

17:04 overall activities abroad.

17:06 First of all,

17:07 uh,

17:08 projects that might foster

17:10 market access or exports,

17:12 um,

17:12 like

17:13 infrastructure projects,

17:15 uh,

17:15 projects that are implemented by the Chinese entities or construction.

17:19 In Africa,

17:20 for example,

17:21 or loans denominated in renminbi.

17:23 So it's,

17:24 it's China trying to push

17:25 uh the internationalization of its currency

17:28 by subsidizing

17:29 loans that are renminbi dominominated,

17:32 right?

17:32 So

17:33 the question of uh non-financial externalities is what we're after and proxies for

17:38 project with such

17:39 non-financial externalities.

17:41 And the second is

17:42 uh looking at political potential uh project with political um uh

17:47 Uh,

17:48 implications of prestige projects.

17:50 It could be stadiums or a parliamentary buildings,

17:53 uh,

17:53 military security projects in the recipient countries,

17:56 uh,

17:57 loans to public recipients or loans to the region

18:00 from which the leader of the country,

18:03 uh,

18:03 originally comes from.

18:04 OK?

18:04 So this is classic leader favoritism,

18:06 uh,

18:07 literature.

18:08 And what you can See

18:09 in this graph,

18:10 and that's my penultimate slide here,

18:12 is that uh

18:14 projects that are associated or likely to have higher non-financial returns

18:19 beyond the mere

18:20 financial relationship

18:22 uh like prestige projects,

18:24 those denominated in RMB,

18:26 uh,

18:26 military security projects or projects for trade infrastructure,

18:30 um,

18:31 uh,

18:32 see a higher,

18:33 uh,

18:33 see,

18:33 see a significantly lower return,

18:35 so they're subsidized.

18:37 Uh,

18:37 whereas,

18:38 uh,

18:39 other projects that are,

18:40 uh,

18:40 you know,

18:40 for private recipients or for emerging market recipients

18:43 see higher,

18:44 uh,

18:45 uh,

18:45 returns and are not subsidized.

18:47 And we confirmed this in a regression framework,

18:49 but since I promised to be short,

18:51 let me wrap up.

18:52 Uh,

18:53 overall,

18:54 uh,

18:54 the financial returns of the BRI so far,

18:57 uh,

18:58 show that this is a no gain,

18:59 no loss project,

18:59 so the,

19:00 the returns are needed.

19:01 Catastrophic as sometimes

19:03 assumed.

19:04 Um,

19:04 and looking ahead,

19:05 of course,

19:05 the big question is

19:07 whether the haircuts of the China will have to accept higher losses,

19:13 uh,

19:13 and whether,

19:14 uh,

19:14 which would imply lower returns,

19:15 uh,

19:16 and what the current shifts in the way China lends that do,

19:19 especially the focus to,

19:21 uh,

19:21 towards more state-owned commercial bank lending with higher returns ex ante.

19:25 Thank you so much.

19:30 Thank you so much.

19:32 And let's now hear from Eric.

19:36 Enrico Mallucci.

19:38 OK

19:40 All right.

19:41 So thank you very much for including the paper,

19:43 uh,

19:44 in the program.

19:45 So this is joint work uh with AOEE,

19:48 uh,

19:49 University of Navarra and Mattia Piccarelli,

19:51 uh,

19:52 from the European Stability Mechanism,

19:54 uh,

19:54 and,

19:54 uh,

19:55 of course,

19:55 uh,

19:56 the usual disclaimer applies.

19:59 So it's a well-known fact that,

20:01 uh,

20:02 sovereign debt markets have evolved quite a bit in the last 30,

20:05 40 years.

20:06 So

20:07 if you think about it,

20:08 I mean,

20:08 uh,

20:09 bank loans were replaced by bonded debt,

20:12 uh,

20:12 and the size of emerging markets government debt

20:16 has increased,

20:17 uh,

20:17 uh,

20:18 substantially,

20:19 but perhaps,

20:19 uh,

20:20 at least it's our taking the biggest change of all.

20:23 has been the growing role of domestic markets.

20:27 So we have seen,

20:28 uh,

20:28 just to give you an example,

20:30 uh,

20:31 think about Mexico.

20:32 So in Mexico,

20:33 uh,

20:34 domestic,

20:35 uh,

20:35 the domestic share of government debt was about 25% in 1995.

20:41 If you take the same number in 2010,

20:44 you find an 80%.

20:46 So this huge increase in the size of,

20:48 uh,

20:48 of domestic markets.

20:49 And of course,

20:50 the growing role of domestic markets for government debt

20:54 has also translated uh into a greater involvement uh in uh

20:59 Domestic,

20:59 uh,

21:00 uh,

21:00 of domestic debt in defaults and restructuring events.

21:03 So,

21:04 Greece,

21:04 for instance,

21:05 uh,

21:05 is a poster child example.

21:07 So if you think about the,

21:08 the Greek restructuring of 2012,

21:11 well,

21:11 92% of the debt that was restructured uh was actually domestic debt.

21:16 So it was really

21:17 a domestic restructuring.

21:19 And as we speak,

21:20 of course,

21:20 countries,

21:21 uh,

21:22 Both under and and outside the cover of

21:24 the common framework are conducting sovereign debt restructuring involving

21:29 uh bonds issued domestically,

21:32 so it's very much uh in today's topic.

21:35 Um,

21:37 so the problem is,

21:38 uh,

21:38 that we have seen this,

21:39 uh,

21:40 increasing the importance of domestic debt,

21:43 but there is limited systematic understanding

21:46 of how,

21:47 of the differences,

21:48 uh,

21:49 between,

21:49 uh,

21:50 uh,

21:50 defaults,

21:51 uh,

21:52 between sovereign defaults that involve,

21:53 uh,

21:54 domestic debt and external debt.

21:56 And so this is exactly what we do in this paper.

21:59 So we systematically compare sovereign defaults.

22:03 of government debt issued in domestic and

22:06 international debt markets and we compare them.

22:09 And of course,

22:09 then we use the insights to inform the academic literature

22:13 as well uh as uh policymakers.

22:18 Now,

22:19 to,

22:19 to,

22:20 to,

22:20 to really compare what,

22:22 uh,

22:22 what is happening when you default on domestic

22:24 and when you default on external debt,

22:26 of course,

22:27 you need the

22:28 data,

22:29 OK?

22:29 So you need

22:31 uh to have a comprehensive view,

22:33 you need to look at the universe of uh the domestic and external debt restructuring.

22:39 So this is why in this,

22:40 in our paper,

22:41 we leverage on uh two state of the art databases.

22:45 So we take information uh about uh domestic

22:49 defaults and restructuring uh from uh the paper

22:52 by Chris Treves here and uh and Tamon which I also saw in the audience.

22:57 And then we complement this information about external

23:00 restructuring with information about the domestic data restructuring,

23:04 which we take from a paper that I or I,

23:07 and,

23:07 uh,

23:07 and Matthia

23:09 we constructed,

23:10 we constructed over the last 10 years,

23:12 OK?

23:13 And then we harmonize the two databases

23:16 and uh we just uh again analyze how the two

23:21 phenomenon differ.

23:24 So the resulting database has uh about 100 and 116

23:30 uh domestic default restructurings and 177 external default uh uh

23:36 events.

23:36 We span from 1980 until 2018,

23:40 and we have information of really a bunch of things.

23:43 So we have information about the timing of this restructuring,

23:46 so when they start and when they end.

23:49 We have information about the instruments involved,

23:50 so we see if,

23:51 uh,

23:51 if we are talking about bond,

23:53 if we are talking about loans,

23:55 we see the volumes involved.

23:56 So we see how big these defaults actually are.

23:59 And then we also look at the,

24:01 uh,

24:01 we also have information about the restructuring terms and the restructuring,

24:05 uh,

24:05 approach that are taken in,

24:07 uh,

24:07 in,

24:08 uh,

24:08 in this restructuring events.

24:10 And we also have information about the

24:12 NPV losses,

24:13 but this is only a few,

24:14 for a,

24:15 for a,

24:15 for a subset of,

24:16 uh,

24:17 Of the restructuring event.

24:18 And then a key feature of the database is that it also has a global coverage.

24:23 So we,

24:24 uh,

24:24 we have information coming from 84 countries that span,

24:28 uh,

24:28 uh,

24:28 all continents and all income group.

24:32 So this being said,

24:34 uh,

24:35 how the two differ,

24:36 how the,

24:37 uh,

24:37 how these two phenomena,

24:39 uh,

24:39 differ.

24:40 So the first thing we find is that domestic defaults

24:44 have become more frequent than external ones.

24:46 So if you look at the red line,

24:48 that is like uh,

24:50 uh,

24:50 the three-year sum of uh external default episodes,

24:54 and you see that essentially they peaked in the 80s,

24:57 but now they are,

24:59 uh,

24:59 if not rare,

25:00 they are rarer than they used to be.

25:02 The black line instead plots the same,

25:05 uh,

25:05 the same uh information for uh domestic uh restructurings.

25:10 And you see that from uh about like

25:13 the end of the 90s,

25:14 the black line has crossed the,

25:16 the,

25:16 the,

25:16 the red line,

25:17 meaning that nowadays,

25:19 uh

25:20 domestic

25:21 uh restructuring are actually uh more frequent than,

25:24 uh,

25:25 than,

25:25 uh,

25:25 than external ones.

25:28 The second thing that we find that selective defaults are the norm.

25:31 So we define selective defaults as the restructuring episode that only involved

25:35 either domestic or external,

25:37 uh,

25:38 or external debt.

25:39 And,

25:40 uh,

25:41 essentially,

25:41 uh,

25:42 what we find is that 70% of the default episodes are selective.

25:47 So when we think about,

25:48 uh,

25:49 uh,

25:50 sovereign defaults,

25:51 we are really

25:52 normally facing,

25:54 uh,

25:54 uh,

25:54 selective defaults.

25:56 And we find that about 60% of the domestic

25:58 defaults and 77% of the external defaults are selective.

26:02 Now,

26:03 the additional

26:04 piece of information is that

26:06 not only

26:08 Selective defaults are the norm,

26:10 but non-selective defaults,

26:12 so if we look at the last,

26:13 the,

26:13 the remaining 30% of,

26:15 uh,

26:15 of the episode,

26:17 they are also somehow selective in the

26:19 sense that they discriminate between lenders.

26:21 So in this

26:23 Uh,

26:23 scattered plot,

26:24 uh,

26:24 essentially you have on the horizontal axis,

26:27 the domestic,

26:27 uh,

26:28 restructured debt as a percentage of total

26:31 restructure debt and on the vertical axis,

26:33 uh,

26:33 the,

26:33 the,

26:34 the domestic versus external composition of government debt.

26:37 So this is

26:38 what this graph is telling you is that

26:40 even when governments do

26:43 default on both external and domestic debt,

26:46 they tend to discriminate

26:49 against the type of debt

26:51 which has the largest share in the overall composition of,

26:55 uh,

26:55 of total debt.

26:56 So if you have a very high

26:58 external debt,

26:59 you tend to default more

27:02 on,

27:02 uh,

27:02 on,

27:03 uh,

27:03 on the external portion of,

27:05 uh,

27:05 of,

27:05 of the debt.

27:08 Next,

27:10 we look at the size of uh domestic versus external default.

27:14 We find that the domestic defaults are smaller,

27:17 so here in the table,

27:18 we report

27:19 Uh,

27:20 the,

27:20 the,

27:21 the,

27:21 the,

27:21 the,

27:22 the percentage as a fraction of the GDP of the debt in default and you see that,

27:26 uh,

27:27 on average,

27:28 uh,

27:28 domestic,

27:29 uh,

27:29 defaults are,

27:30 uh,

27:30 they,

27:31 they involve uh

27:32 debt which is about 5% points lower as a fraction of GDP of,

27:36 uh,

27:36 of,

27:36 uh,

27:37 external debt restructurings.

27:40 Uh,

27:41 we also find that the domestic restructurings are faster.

27:45 So in the table here,

27:45 you have,

27:46 uh,

27:46 the,

27:46 the duration in months

27:49 of external and domestic,

27:50 uh,

27:50 debt restructurings,

27:51 and what we find is that,

27:53 uh,

27:54 uh,

27:54 domestic restructuring takes on average

27:57 about 9 months less,

27:58 uh,

27:58 to be resolved than,

28:00 uh,

28:00 than,

28:01 uh,

28:01 than,

28:01 uh,

28:01 external one.

28:03 And then interesting thing here is perhaps like,

28:05 uh,

28:06 the evolution of these,

28:07 uh,

28:07 of these numbers.

28:08 So,

28:09 Uh,

28:10 we interpret the fact that,

28:11 uh,

28:12 domestic restructuring are faster to resolve

28:15 as a reflection of the fact,

28:17 uh,

28:17 that when you have domestic debt,

28:19 it's actually easier to restructure it because typically

28:22 domestic debt is governed by the domestic law.

28:24 And so changing the domestic law is much easier than,

28:27 uh,

28:28 than changing the international law.

28:30 But then the interesting aspect is that uh

28:32 there,

28:32 there is a,

28:33 we have seen a convergence uh

28:35 in the speed of uh uh the restructuring of domestic and external restructuring,

28:41 and we

28:42 attribute this to the fact that external restructurings

28:46 are getting faster thanks to to the introduction of cuts in,

28:50 uh,

28:50 in,

28:50 uh,

28:51 in,

28:51 in the contracts.

28:54 Um,

28:55 I'm gonna skip this,

28:57 uh,

28:57 because,

28:58 uh,

28:58 we need to be quicker.

28:59 Uh,

29:00 the next thing I want to say about domestic and external default is that,

29:03 uh,

29:03 what we are seeing is that domestic defaults are smaller,

29:06 quicker,

29:07 but yet,

29:08 uh,

29:08 for the subset

29:10 of,

29:11 uh,

29:11 uh,

29:12 of episodes for which we have the information about NPD losses,

29:16 NPV losses,

29:16 we see that domestic defaults are actually more punitive.

29:19 So here,

29:20 we report NPV losses for domestic and external debt,

29:24 and you see that,

29:25 uh,

29:26 uh,

29:27 that,

29:27 that the NPV losses are actually higher for,

29:30 uh,

29:30 for domestic,

29:31 uh,

29:31 for domestic restructuring.

29:33 And this is for us a bit surprising because there is often

29:37 this belief,

29:38 uh,

29:38 that governments tend to be nicer,

29:40 let's say,

29:41 uh,

29:42 towards,

29:43 uh,

29:43 their,

29:43 uh,

29:44 towards domestic creditors,

29:45 but here we don't,

29:46 we don't see this.

29:50 I'm gonna also jump,

29:51 uh,

29:52 uh,

29:52 this slide,

29:53 and I'm gonna concentrate to the last bit,

29:55 uh,

29:55 of,

29:55 uh,

29:56 the paper.

29:56 So in the last bit of the paper,

29:58 we change a bit the focus and,

30:01 uh,

30:01 we compare essentially the economic and political context,

30:06 uh,

30:07 in which domestic and external debt restructuring happens.

30:10 So in particular,

30:12 we are interested in,

30:13 uh,

30:14 how,

30:14 uh,

30:15 macrofinancial and political variable differ,

30:18 uh,

30:18 when,

30:18 uh,

30:19 when,

30:19 uh,

30:20 Uh,

30:20 governments take the decision to,

30:22 to restructure their debt.

30:24 Um,

30:25 so the first thing that we look at,

30:27 of course,

30:28 uh,

30:28 is,

30:28 uh,

30:29 at,

30:29 uh,

30:31 how GDP evolves around domestic and external,

30:34 uh,

30:34 external debt restructurings,

30:36 uh,

30:36 and,

30:37 uh,

30:37 this is what we show in the first,

30:38 uh,

30:39 in the first column.

30:40 And what you see here is that the coefficients are

30:42 very similar and they are both negative and significant.

30:45 So,

30:45 what we find is that both domestic and external

30:48 defaults tend to happen in periods of low growth,

30:51 OK?

30:51 And so this is actually

30:53 a very standard result.

30:54 And the,

30:55 the second,

30:55 uh,

30:56 the second column tells you that,

30:57 uh,

30:58 both also happen in periods of very high debt.

31:01 So this is nothing surprising.

31:02 So we see defaults when the economy is doing very badly and when,

31:05 uh,

31:06 Uh,

31:07 that is very high.

31:08 But then in the next table,

31:10 we actually

31:11 look into

31:12 essentially what are the different channels that might explain,

31:15 uh,

31:16 this,

31:16 uh,

31:16 uh,

31:17 uh,

31:18 this decline in GDP.

31:19 And we see that domestic and external,

31:21 uh,

31:22 restructuring,

31:23 uh,

31:23 happening very different,

31:24 uh,

31:25 in very different environments.

31:26 So we see

31:28 that the external default happen in periods of substantial

31:31 external adjustment and you can see it in the first

31:34 corner.

31:34 So you can see that

31:35 when you have an external default,

31:38 then you have like

31:39 uh

31:40 uh like an improvement of the current account

31:44 and,

31:44 and,

31:44 and also like a drop in the,

31:46 in the foreign inflow.

31:47 So this is an external adjustment and we don't see

31:50 this happening this as much when you,

31:53 when you,

31:53 when you have like a domestic restructuring.

31:56 Instead,

31:56 when we look at

31:58 Private credit,

31:59 so when we look at one of the variables that

32:02 is telling us what is happening on the credit side,

32:05 we see a negative significant coefficient only for domestic,

32:09 uh,

32:10 for domestic restructuring.

32:11 So this is telling us that domestic default

32:13 happens in period or are associated with the,

32:16 uh,

32:17 credit crunches.

32:19 Um,

32:20 so

32:21 the conclusion here is that even though like perhaps the,

32:24 the,

32:24 the,

32:24 the overall

32:26 economic environment is not to be similar because we see,

32:30 uh,

32:30 we see this contraction of GDP,

32:32 the channel that explains this contraction of GDP can be very different,

32:36 uh,

32:36 in the two,

32:37 in two different,

32:38 uh,

32:39 Episodes

32:41 I'm gonna jump the,

32:42 the,

32:42 the slides and gonna like talk about the last two

32:46 things that we look in the paper which is uh

32:49 the political context in which domestic and external uh uh restructuring happens.

32:55 And then I'm gonna talk about the role of the IMF.

32:57 Um,

32:58 so,

33:00 What we see in the first,

33:02 sorry,

33:02 uh,

33:03 in the first column

33:05 is,

33:05 uh,

33:05 a variable that measures political,

33:07 uh,

33:08 stability.

33:08 So what we find,

33:09 and this is a commonality between domestic and external default,

33:12 is that both of them,

33:14 they are associated with political instability,

33:16 OK?

33:18 But then,

33:18 uh,

33:19 we see that the elections uh are twice as

33:21 likely during domestic defaults than during external default,

33:26 OK?

33:27 So,

33:28 Possibly,

33:28 voters are very displeased when you default on,

33:31 uh,

33:31 on domestic debt.

33:33 And then instead,

33:34 what we see is that the radical governments

33:37 are more often in power

33:39 when,

33:39 uh,

33:40 when you have uh external,

33:41 uh,

33:42 external,

33:42 uh,

33:43 external defaults because perhaps you need a reckless government

33:47 to challenge,

33:47 uh,

33:48 you know,

33:48 the,

33:49 the,

33:49 the international,

33:50 uh,

33:50 rule of law.

33:54 And finally,

33:54 as I anticipated before,

33:56 I,

33:56 I,

33:56 we also look a bit at the role of uh bilateral lenders.

34:00 And so we see,

34:01 we look at what uh

34:02 IMF programs and official debt relief,

34:05 so what is their impact on,

34:06 uh,

34:07 on external and domestic uh restructurings,

34:09 uh,

34:10 and we see that,

34:11 uh,

34:12 In the,

34:13 uh,

34:13 and you can see this in the first column,

34:15 countries that receive support from the IMF,

34:18 uh,

34:18 they tend to prioritize uh repayment

34:22 of external debt

34:24 at the cost of actually increasing,

34:27 uh,

34:27 their,

34:27 uh,

34:28 their domestic,

34:28 uh,

34:29 domestic restructuring.

34:30 And here for,

34:31 for the IMF people in the,

34:33 in the,

34:33 in the room,

34:34 uh,

34:34 I mean,

34:34 we interpret this a reflection of the uh of the lending into arrears policies.

34:38 So the lending into arrears policies adds

34:41 An additional layer

34:43 of uh uh requirements uh to,

34:45 to,

34:45 to actually

34:47 Uh,

34:48 to actually,

34:48 for the IMF to borrow money to,

34:50 to,

34:50 to lend money to countries in troubles.

34:53 And so,

34:54 the,

34:55 and the lending into a real policy

34:57 only covers external debt.

34:59 So,

35:00 our

35:01 intuition is that perhaps uh

35:03 Countries want to avoid this additional layer of

35:07 like requirements and so they prefer to,

35:09 to,

35:09 to default on domestic debt to avoid entering uh into

35:13 the uh lending into arrears programs uh with the IMF.

35:17 And finally,

35:18 when we look at the official uh

35:21 debt relief,

35:22 uh,

35:22 we see that countries that uh receive official

35:24 debt relief are more likely to default,

35:26 uh,

35:27 externally,

35:27 and we,

35:28 we think this,

35:29 uh,

35:29 as a reflection of the comparability of

35:31 treatment policies that are often associated with,

35:33 uh,

35:34 uh,

35:35 with,

35:35 uh,

35:36 with this official debt relief.

35:38 In conclusion,

35:40 uh,

35:40 compared to the faults abroad,

35:42 the faults at home are smaller,

35:43 quicker,

35:44 and deliver larger,

35:46 uh,

35:47 Uh,

35:47 larger losses and the environment surrounding domestic and external defaults,

35:52 uh,

35:52 is very different.

35:54 Domestic defaults are associated with credit crunches.

35:56 External defaults happen at times of fiscal and external adjustments,

36:01 and also what we see

36:03 is that external defaults are less likely with moderate governments,

36:07 while domestic defaults are more likely to trigger elections.

36:17 Many thanks a week.

36:19 Now let's hear from Rong Cheng.

36:21 The floor is yours.

36:29 Thank you for inviting me to present this paper

36:31 which is joint work with Francisco from the IMF.

36:34 Um,

36:34 so,

36:35 the previous two presenters show a little bit of the

36:38 historical,

36:40 uh,

36:40 default history.

36:41 What I'm gonna present

36:42 is uh

36:43 more or less a theoretical model to

36:45 explain why some countries deform more than others

36:49 and the role of institutions.

36:51 So,

36:51 I will go very quickly on the motivation is to

36:56 try to understand

36:57 what explains the difference behavior between high-income

37:01 countries and middle and low-income countries.

37:03 In principle,

37:05 the,

37:05 the level of income doesn't matter for default risk,

37:08 but we do observe from the history.

37:11 And when we look at it among the middle and low-income country,

37:14 we also see that Latin American countries

37:17 tend to devolve more often historically.

37:20 And I'm from Argentina,

37:21 uh,

37:22 so that's my motivation.

37:24 Um

37:25 and then

37:26 I,

37:27 well,

37:27 I have two anecdotal evidence,

37:29 but I will only present the case of Argentina

37:31 in the 80s that had two sovereign default crises.

37:34 So,

37:35 in the 80s,

37:35 the,

37:36 the consolidated fiscal deficits reached to 15% of GDP

37:40 from the federal government,

37:42 local government,

37:43 and uh SOEs.

37:45 And at that time,

37:46 there was a,

37:47 a revenue sharing scheme.

37:49 Uh,

37:49 basically,

37:49 the federal government raised revenue and then transfer

37:53 to the,

37:54 uh,

37:54 local government,

37:55 and they do have an extraordinary uh treasury transfer system

37:59 for um

38:01 unusual events.

38:02 But during the three years.

38:04 In 1985 and 1987,

38:07 this revenue sharing scheme was suspended because the law that governed

38:11 it expired and took them three years to put back.

38:14 So during these three years,

38:16 basically the transfer between the federal government to the local government,

38:20 local government,

38:21 it was a

38:22 bilateral negotiation.

38:24 Uh,

38:25 and that created the uncooperative behavior by the local government,

38:30 tried to get maximum,

38:32 uh,

38:32 they can from the,

38:33 uh,

38:33 central government,

38:35 uh,

38:35 and the federal government couldn't,

38:38 uh,

38:38 push back because there was no legal framework in it.

38:41 So that's just keeping in mind this,

38:43 the,

38:44 the influence of the local government

38:46 on the federal government borrowing decision.

38:50 I will skip,

38:50 uh,

38:51 Brazil.

38:52 So,

38:52 just very quickly,

38:53 the

38:54 related literature,

38:55 look at the sovereign default,

38:57 um,

38:58 and usually in one country like Argentina,

39:01 try to explain what explained it,

39:03 and more recently,

39:05 there's a series of papers that incorporate political.

39:08 Economic consideration

39:09 in to explain that,

39:11 such as

39:12 the degree of impatience from the incumbent government

39:16 or

39:17 there's a left-wing government and

39:19 right-wing government that has different preference

39:22 that can lead to overborrowing.

39:25 And that is um

39:26 something that uh our paper as well as show.

39:29 So let me show you

39:30 quickly,

39:31 uh,

39:32 two-period model

39:33 with a closed form solution so you can see the intuition of the model.

39:38 So this is the government structure.

39:40 Um,

39:40 the government is

39:42 composed by a group of powerful groups

39:45 that want to influence in policy decisions.

39:49 In countries with good institutions,

39:52 there's rules set up

39:54 that limit the influence of these powerful groups in

39:58 central government decisions on spending and on borrowing.

40:02 Uh,

40:03 in weak institutions,

40:04 the central government cannot

40:06 limit the influence,

40:08 so something,

40:09 uh,

40:10 will happen,

40:10 as you will see,

40:12 and these,

40:12 um,

40:13 groups,

40:14 they also have different

40:16 ideas about what to spend,

40:18 how to spend,

40:19 so there's existence of a polarized,

40:22 uh,

40:22 polarization in ideology.

40:25 That,

40:26 that means that they cannot actually coordinate in

40:28 the optimal solutions because they have different ideas about

40:32 how to govern and how to spend.

40:35 And the distribution of power is captured by the,

40:38 the their power in the Congress,

40:40 for example,

40:41 or in terms of the share they get from the central government.

40:46 And these powerful groups,

40:48 they consume,

40:49 um,

40:50 they derive utility from the consumption goods.

40:53 They receive revenue or external borrowing from the central government.

40:58 They cannot do it directly.

40:59 They have to go through the central government.

41:01 And the good institutions,

41:03 because the central government

41:06 uh

41:07 solution is the optimal solution,

41:10 they borrow,

41:11 they make the optimal decision and transfer

41:14 to the local government.

41:16 In a weak institutions set up,

41:18 these groups.

41:20 Kind of uh maximize their own utility function,

41:23 make

41:24 decisions,

41:25 and as a result of that,

41:26 there's a negative externality that will arise because they do not internalize

41:31 the action of their

41:32 maximization function

41:34 into other groups.

41:36 So that's the,

41:37 the,

41:38 how the powerful groups try to influence policy.

41:40 And then we have the foreign investor that this is very typical in the literature,

41:44 that they are risk neutral

41:46 and they,

41:47 they are competitive,

41:48 uh,

41:48 they try to maximize,

41:50 um,

41:50 um,

41:51 the,

41:52 the interest rate that to match the risk-free interest rate.

41:55 They know

41:56 the institutional quality of the government.

41:59 Uh,

41:59 and the degree of polarization in the economy,

42:02 they lend

42:03 to the government,

42:04 uh,

42:05 as a whole,

42:06 not to individual groups.

42:10 And for the foreign investor,

42:11 as long as they don't get the full repayment,

42:14 they constitute a sovereign default.

42:17 So these are the,

42:18 they are all extreme,

42:19 but it's uh the setup for the modeling purpose.

42:23 So in the two-group and two-perio model,

42:26 we have two groups.

42:28 They have,

42:28 uh,

42:28 they are trying to uh uh consume goods through the transfer,

42:34 and in the model,

42:36 I just separate

42:38 countries with good institutions,

42:40 what I call unified government

42:42 or

42:43 countries with weak institutions,

42:45 what I call polarized government.

42:47 And remember that in the polarized government,

42:50 The powerful group tried to influence the central government policy,

42:54 and they succeeded.

42:56 Uh,

42:56 so there's a,

42:57 a negative externality because they cannot coordinate.

43:02 So I won't go into detail.

43:04 The unified government is the typical uh central planner solution.

43:08 They maximize utility of the two periods subject

43:11 to a budget constraint which derived from the endowment

43:15 and the borrowing.

43:17 And in the second period,

43:18 the revenue is the only shock in the economy,

43:22 so they either default or repay depending on the

43:25 realization of the revenue shock in the second period.

43:29 In a paralyzed government,

43:31 each group makes this same maximization problem.

43:34 Uh,

43:35 as you say,

43:36 uh,

43:36 see,

43:36 there's,

43:36 uh,

43:37 the maximization problem is for group one and two.

43:40 And the foreign investor

43:43 know that uh in a weak institution,

43:46 which means in this case,

43:47 the paralyzed government,

43:49 the default risk is the default risk,

43:51 the maximum default risk among the two groups.

43:54 As long as one group defaults,

43:56 it's optimal for the other group to default as well.

44:00 So that's the equilibrium interest rate.

44:02 And,

44:03 OK,

44:03 uh let me skip this.

44:04 Uh,

44:04 so,

44:04 uh,

44:05 the,

44:05 the,

44:06 the theoretical model can derive that

44:09 for all,

44:10 uh,

44:11 degree of polarization

44:12 in a,

44:14 a weak institution,

44:15 which means it's the polarized government,

44:17 the default

44:19 rate is higher,

44:20 the interest rate is higher,

44:21 and the debt level is higher.

44:24 And this is the intensive margin.

44:25 On the extensive margin,

44:27 the more polarized is the government,

44:30 the higher the default and the debt and interest rate.

44:33 And

44:34 let me just quickly

44:36 tell about the intuition.

44:38 It's the marginal cost

44:40 of borrowing

44:41 for one group

44:43 in a weak institution set up

44:46 is lower than the aggregate cost

44:48 of borrowing because they,

44:50 they do not internal.

44:52 The cost of borrowing that affects the whole country as a group while they only

44:57 receive a portion of that debt.

45:00 That's the source of a negative externality,

45:03 not internalized in a paralyzed government,

45:06 which means it's a weak institution set up.

45:10 So let me

45:11 quickly show you,

45:12 uh,

45:13 we try to incorporate this uh in an empirical study

45:17 and use two variables to capture the institution.

45:21 One is the standard CIRG and one is the

45:25 regulation of influence by the different interest groups in the,

45:30 in the economy.

45:31 So,

45:32 and the number 5 is the regulator's

45:35 case,

45:35 the country.

45:35 Has

45:36 set up the clear rule of how these different powerful group can

45:40 influence in the central government decision

45:43 and which policy cannot be influenced.

45:45 In an unregulated situation is,

45:49 remember the case of Argentina during those three years,

45:51 there's no legal framework how this negotiation happened.

45:55 And then we tested,

45:56 uh,

45:57 let me just show this one.

45:59 so we separate the countries,

46:01 uh,

46:01 in

46:03 good institution

46:04 versus,

46:05 uh,

46:06 weak institution.

46:07 So the variable para is the high,

46:10 the higher is the better.

46:12 So for countries below the medium,

46:15 uh,

46:15 the first column,

46:16 and above the medium,

46:18 and the 3 and 4 is the countries with para

46:22 above a medium,

46:23 uh,

46:24 below medium plus 1 standard deviation and above.

46:27 What I see,

46:28 what I try to show here is that the degree of polarization

46:32 only matters

46:34 in the setting of weak institutions.

46:37 So you can have,

46:38 you know,

46:39 the ideological difference among the government as

46:42 long as the institution is strong enough

46:44 to limit the influence of these groups,

46:47 the default risk is still,

46:48 uh,

46:49 manageable or the,

46:50 the variable doesn't even matter for default risk.

46:53 So let me conclude,

46:55 uh,

46:56 how this is,

46:58 um,

46:58 research is relevant today as Mueller already said,

47:02 many countries have higher debt to GDP ratio even before the pandemic.

47:06 It's even higher now,

47:08 and we also know there's a rising inequality

47:11 which could translate into ideological polarizations.

47:15 So what do we do to

47:18 limit these two factors that are already there,

47:22 is to set up institutions

47:24 to limit the spending pressure that will arise from the different groups.

47:29 And now,

47:29 currently I'm,

47:30 I'm,

47:30 I'm in Indonesia,

47:32 the only country in Asia that had a

47:35 a default crisis

47:37 after the Asian financial crisis,

47:38 and they do have a fiscal rule since then.

47:41 So I see that this political influence still exists in the context of Indonesia,

47:46 not maybe from the local government,

47:48 but from SOEs,

47:50 but they have the 3% of GDP as a fiscal rule

47:54 that is hard since 2003 and still remain very relevant.

47:58 As a result,

47:59 their debt to GDP ratios.

48:01 Remain the lowest in among the lowest in Asia Pacific,

48:04 but at the same time,

48:06 um,

48:06 this budgetary

48:08 process,

48:09 the PFM that the bank also uh uh advised a lot,

48:13 is also very important to make it transparent

48:15 so

48:16 this political negotiation doesn't happen.

48:18 And finally,

48:19 the debt transparency and.

48:21 Part of the other speakers um

48:24 show,

48:24 it's very important to know actually how much is the total debt

48:28 that we are talking about

48:29 and try to

48:31 conduct fiscal consolidation to lower the risk

48:34 because all defaults happen for only 3

48:37 reasons

48:38 either

48:39 capital flow reversal.

48:42 Interest increased by the um developed country or commodity shock.

48:47 And to lower the

48:48 risk for default is only to limit the debt level

48:51 for many of the developing countries.

48:53 Thank you.

48:58 Many times um

49:00 wrong and with this we concluded the presentations

49:03 and now we will hear from the discussions.

49:06 Let's start with Mark Chamon from the fund.

49:14 Thank you.

49:14 Uh,

49:14 it's a pleasure to be here to discuss the 3 papers.

49:17 Uh,

49:17 we have 3 papers to cover,

49:19 and I think,

49:19 uh,

49:19 we are behind schedule,

49:20 so let me just jump right in.

49:23 So,

49:24 uh,

49:24 first paper,

49:25 uh,

49:25 the one on the returns to China's Belt and Road Initiative,

49:29 um,

49:30 I wanna focus the comments on the issue of the haircuts

49:33 because I think they're also tied very closely to the,

49:36 uh,

49:37 financial returns on those loans.

49:39 Um,

49:39 the paper goes through,

49:40 uh,

49:40 systematically estimating the haircuts.

49:42 I just want to highlight a few,

49:44 uh,

49:44 points on this issue of haircuts.

49:47 First,

49:47 that

49:48 it,

49:48 it is somewhat constrained the choice of haircuts.

49:50 There are many reasons why,

49:52 uh,

49:53 even a large creditor may not have that much wiggle room to decide on its own haircuts

49:58 because we tend to have fairly robust,

50:01 uh,

50:02 creditor coordinating mechanisms when it

50:04 comes to official sector restructuring.

50:05 So notably the price club process.

50:08 And even if a,

50:09 uh,

50:09 uh,

50:09 a creditor is not a member of the Paris Club,

50:11 it will still end up being bound by the process because of the,

50:15 uh,

50:15 need to respect comparability of treatment.

50:17 So that could,

50:18 uh,

50:18 constrain the choices.

50:20 Uh,

50:20 it can also be constrained by

50:22 some hard restructuring targets.

50:23 For example,

50:24 if the country has an IMF supported program,

50:26 uh,

50:27 there might be a requirement to restore that sustainability,

50:30 which

50:30 will also,

50:31 uh,

50:31 in a sense constrain the,

50:32 the choice of haircuts of the different creditors,

50:34 uh,

50:35 could propose.

50:36 And

50:37 last but not least,

50:38 in the particular case of HPE,

50:40 there were like very clear rules for how each,

50:44 like,

50:44 you could go low one by low one and apply those rules that would then determine

50:48 what type of treatment would be expected on those low ones.

50:50 Um.

50:51 So if,

50:52 maybe if one wants to focus on this sort of bargaining dimension,

50:55 it might be useful to just focus on,

50:57 on episodes where there was only a restructuring,

50:59 a bilateral restructuring between the debtor and China,

51:03 which was not part of,

51:04 say,

51:04 a broader uh debt restructuring.

51:07 On,

51:08 on the issue of haircuts,

51:09 I also want to mention the discount rates.

51:11 Uh,

51:12 they,

51:12 they consider a range of discount rates.

51:14 I would suggest focusing

51:16 on the sample on the HIPIC discount rates and maybe on the more recent ones,

51:19 on the LDSF 5% discount rates,

51:22 which

51:22 is the,

51:23 uh,

51:24 yardstick that the other official creditors would use,

51:26 uh,

51:27 when looking at comparability of treatment.

51:29 Um,

51:30 on this,

51:31 uh,

51:32 and also encourage maybe a bit more comparison between the compare the haircuts,

51:35 but also looking at the rate of returns,

51:38 how it would compare to the other official creditors to also get a sense of how

51:42 usual or unusual the,

51:43 the lending behavior is.

51:46 But on,

51:46 on this issue of HP,

51:47 going back to this issue of HPI,

51:48 this is one of the charts that was used in the presentation.

51:52 And,

51:52 and just visually,

51:53 the point I want to highlight is that

51:55 A lot of the very episodes with very high haircuts tend to occur

51:59 early in the sample where the circle where the circles are quite small.

52:02 Remember,

52:03 the circles show the size of China's exposure to the country.

52:07 So in,

52:07 in a sense,

52:08 since China is a relatively new,

52:10 uh,

52:11 bilateral creditor,

52:12 it sort of sit out hit peak.

52:13 So when HI was taking place and all these very large haircuts were

52:18 Taking place,

52:18 China had very small exposures,

52:21 and to the extent they had small exposure,

52:22 it still

52:23 did participate in the HIPIC initiative,

52:25 uh,

52:26 but the exposures were small,

52:27 and I think that's

52:29 maybe that factor alone

52:30 may be driving the low average haircut and

52:34 taking account HIPIC maybe would bring it closer to the average of,

52:37 uh,

52:37 the other official creditors.

52:41 Turning to the second paper looking at domestic defaults,

52:44 this is actually something quite topical.

52:45 I mean,

52:46 as the paper highlights,

52:47 uh,

52:48 domestic debt is playing,

52:49 um,

52:50 a relatively more important role,

52:52 uh,

52:52 in recent episodes.

52:53 Uh,

52:54 I think

52:54 they may have too broad a definition of a domestic default.

52:58 Uh,

52:58 I think the,

53:00 the,

53:00 it may be including a lot of episodes that may be 1 may not necessarily,

53:04 uh,

53:04 considered a default,

53:05 but I think

53:06 Either way,

53:07 11 will cut it,

53:07 the,

53:08 the pattern will remain the same,

53:09 that this has become uh more important uh in recent years.

53:13 And I think this time dimension is something that maybe

53:15 the paper could go uh in a bit more detail.

53:17 They do have different breakdowns by time,

53:20 but I think this is quite an important dimension to explore.

53:23 On this issue of domestic restructuring,

53:25 there is this perception that maybe restructuring domestic debt is,

53:29 is easier because you're operating in your own jurisdiction,

53:31 but I think that's quite a

53:33 simplistic reading of the situation because any domestic restructuring would

53:37 bring a lot of financial stability considerations into play.

53:41 Uh,

53:41 and however bad the debt crisis is,

53:43 the financial crisis can be,

53:45 uh,

53:45 a lot worse.

53:46 So

53:47 you really need to be careful,

53:49 uh,

53:49 and,

53:50 and I think countries do tend to be quite conservative when approaching this issue.

53:53 Uh,

53:54 and

53:55 there was a reference to the IMF and we also tend to be quite,

53:57 quite cautious about it as well.

53:59 There's a recent,

54:00 uh,

54:00 board paper on domestic debt restructuring about 2 years ago where

54:04 this financial stability,

54:05 uh,

54:06 considerations to the future quite importantly.

54:09 Uh,

54:10 the paper finds larger losses for domestic debt,

54:12 uh,

54:13 which is,

54:13 uh,

54:14 uh,

54:14 a bit of a surprising result,

54:15 uh,

54:16 again,

54:16 because this financial stability considerations are,

54:18 are,

54:19 are so prominent and you have to be very careful about pushing losses

54:22 beyond

54:23 what the domestic bank can take.

54:26 Uh,

54:27 and,

54:27 uh,

54:27 it's also surprising because in the,

54:29 some recent cases,

54:30 there is indeed the,

54:32 the view that the domestic debt was restructured more likely,

54:35 and there's no expectation of,

54:37 uh,

54:37 comparability of treatment,

54:38 uh,

54:39 between domestic and external debts,

54:41 including the limiting case,

54:42 which are many cases where

54:44 external debt is restructured and domestic debt is not touched at all.

54:48 Uh,

54:49 so I,

54:49 I would,

54:50 uh,

54:50 I think part of the challenge here would be finding

54:53 the right discount rate for the domestic debt restructuring.

54:56 I think this is something that's

54:57 hard to do even for external debt restructurings.

55:00 For domestic debt,

55:01 it would be even more complicated because chances are,

55:04 you are in a very volatile situation,

55:06 high inflationary environment,

55:07 so what discount rate should one plug in?

55:10 Uh,

55:10 maybe one way around it could be to,

55:13 to complement the final analysis of the

55:16 comparing the expost and the ex ante market prices of the restructured claims.

55:19 I think

55:20 maybe that should be

55:22 doable,

55:22 at least for a few recent cases where this data may not be so hard to get.

55:26 And,

55:26 and from there,

55:27 you can compute different measures of either market or MPV haircuts

55:31 and use it as a cross-check.

55:34 And last but not least,

55:36 uh,

55:36 I'll turn to the,

55:38 to the third paper on the role of institutions.

55:40 Uh,

55:41 it presents a

55:43 very interesting model where you have two

55:44 different groups in the economy making separate,

55:47 uh,

55:47 borrowing and spending decisions,

55:49 and they're default decisions,

55:51 uh,

55:51 will affect the other groups.

55:52 So it's a very,

55:53 uh,

55:54 Easy way to get this overborrowing problem,

55:57 which is an alternative way to get over borrowing problem,

56:00 like

56:00 more traditional political economy channels in the literature like related to uh

56:05 short horizons or,

56:06 or high discounting.

56:08 And they,

56:08 they motivate the paper by the role of the subnationals,

56:11 uh,

56:12 giving examples of,

56:12 of states in Brazil and provinces in Argentina.

56:16 And the channel highlighted is quite strong as,

56:18 as the model calibration shows.

56:20 I mean,

56:21 it,

56:21 it is so strong that there has been a lot of effort,

56:23 uh,

56:24 to,

56:24 to try to address this institutional shortcomings.

56:27 So,

56:28 I,

56:28 I,

56:28 in the case of Brazil,

56:29 I know there was a,

56:30 a,

56:30 a fiscal responsibility law that was passed

56:32 in 2000 that really constrained how much,

56:35 uh,

56:35 The,

56:35 the fiscal policy of the states and local governments,

56:38 and I think uh probably something similar may have taken place in Argentina.

56:42 But

56:43 so this initial uh size and indeed the

56:46 anecdotal evidence all dates back from the 1980s.

56:49 So when I first read the paper,

56:51 my first reaction was that,

56:52 OK,

56:53 maybe this was a,

56:53 a big problem back then,

56:55 but we have seen uh graduated and the problem has been solved.

56:58 But then on second thought,

56:59 I,

57:00 I think this problem is still quite relevant because there are many,

57:03 many cases where we have state-owned enterprises

57:06 which do make a lot of borrowing decisions and they

57:09 are outside the direct control of the Ministry of Finance.

57:12 They may not necessarily be reporting what they're doing to the debt offices.

57:15 So I think this

57:17 Sort of channel of the centralized borrowings,

57:19 make borrowers making these decisions and then having spillovers of the country,

57:23 uh,

57:24 can,

57:24 is still quite relevant,

57:25 maybe not in the initial in the in the way that they,

57:29 uh,

57:30 related to the states or provinces,

57:32 but I think through the role of SOEs,

57:34 it's still,

57:35 it can be quite relevant.

57:36 So I think it's quite,

57:37 quite an important paper

57:39 in that regard.

57:40 And that's it.

57:41 I hope I,

57:41 I'm OK on time.

57:43 It's very well.

57:44 Thanks.

57:47 And now the last speaker

57:49 of this session,

57:49 Luca Bandera.

57:55 Thank you very much.

57:56 So I'll try to be brief.

57:57 I took,

57:58 I,

57:58 I kept time for everybody now,

57:59 I keep time for myself.

58:01 OK.

58:02 All right.

58:03 So,

58:04 uh,

58:05 yeah,

58:05 uh,

58:05 so comments on these three papers,

58:07 I will focus my comment on,

58:08 on what's,

58:09 uh,

58:09 basically

58:10 the,

58:11 the,

58:11 the policy implications since we are the World Bank and I work in the World Bank.

58:15 Uh,

58:15 then,

58:15 uh,

58:16 the date and definitions,

58:17 uh,

58:17 the regression and the theoretical model.

58:20 So let me go very fast.

58:22 So the key contribution,

58:23 I would say,

58:24 You know,

58:25 I,

58:25 I'm working with a lot of data on defaults and,

58:28 and these papers,

58:29 these two papers,

58:30 no,

58:30 the,

58:30 the,

58:31 the papers that,

58:31 that,

58:31 uh Christoph presented and,

58:33 and,

58:33 uh,

58:33 and uh Enrico presented,

58:35 um,

58:36 were,

58:36 were really,

58:37 uh,

58:37 really,

58:38 uh,

58:39 you know,

58:39 uh,

58:40 gave key contributions,

58:41 right?

58:41 Because the granularity of this information is extremely important,

58:44 uh,

58:44 knowing when a country defaults,

58:46 how long the default lasts,

58:47 it's,

58:48 it's extremely important,

58:49 and

58:49 there is,

58:50 uh,

58:50 uh,

58:51 say paity of data.

58:53 On the domestic debt default episodes and,

58:56 uh,

58:56 and,

58:57 you know,

58:57 this,

58:57 this analysis that was presented today really feels,

59:00 feels,

59:00 uh,

59:00 uh,

59:01 part of the gap,

59:02 um,

59:03 um,

59:03 with,

59:04 with some,

59:05 some,

59:05 uh,

59:06 empirical uh statistics that it's,

59:08 uh,

59:08 it's important to notice,

59:09 uh,

59:10 so far,

59:10 no,

59:10 the,

59:11 the,

59:12 uh,

59:13 you know,

59:13 conventional wisdom is that the default episode basically overlap,

59:16 but they said no,

59:16 they don't,

59:17 right?

59:17 And it will be important to understand exactly in which countries in,

59:20 in,

59:21 in what circumstances.

59:22 Also,

59:23 the importance of political-related variables

59:25 and um

59:27 the fact that,

59:27 that uh different defaults have different outcomes in terms of uh

59:31 haircuts,

59:32 uh uh they are preventive uh which creators have lost,

59:35 so it's really rich,

59:37 uh,

59:37 the,

59:37 the contribution of this paper is really rich.

59:40 Uh,

59:40 and there are policy implications,

59:41 no,

59:42 the,

59:42 if I can really summarize,

59:43 the defaults are very expensive,

59:45 so better to avoid them,

59:46 even though

59:46 the contribution that,

59:47 for example,

59:47 an institution like the World Bank could give to this default is very different.

59:50 Uh,

59:50 external defaults

59:52 are linked to,

59:52 you know,

59:53 balance of payment problems,

59:54 uh,

59:54 domestic default connected with,

59:56 you know,

59:57 uh,

59:57 financial,

59:58 uh,

59:58 stress,

59:59 domestic financial stress.

1:00:00 Uh,

1:00:01 debt relief is often not deep that we knew,

1:00:03 and,

1:00:03 uh,

1:00:04 the fact that many countries today require deep debt restructure implies

1:00:08 That,

1:00:08 that debt restructuring are really prolonged and very

1:00:10 costly for the countries and the creditors likewise,

1:00:13 and loans from Chinese banks uh seems not to follow,

1:00:16 you know,

1:00:17 uh,

1:00:17 economic and financial risk assessment,

1:00:19 but they are motivated in a different way.

1:00:21 That implies that maybe one of the largest creditors do not really follow

1:00:25 what we try to preach,

1:00:26 right?

1:00:27 That

1:00:28 the terms of the loans is to be

1:00:30 uh adequated to the,

1:00:31 to the,

1:00:32 to the capacity repayment of the country.

1:00:34 Other,

1:00:35 other implications are less clear,

1:00:36 uh,

1:00:37 like,

1:00:37 for example,

1:00:38 we have very little uh leverage on the political system of the country.

1:00:42 So I wonder if

1:00:43 the analysis,

1:00:44 especially the analysis on the,

1:00:46 uh,

1:00:46 defaulted institutions can look at

1:00:48 You know,

1:00:49 we can use in the regression and the empirical analysis,

1:00:51 the variables that are more

1:00:53 closer to what we can impact upon like countries that have fiscal rules,

1:00:57 countries that have a good understanding of how to,

1:01:00 you know,

1:01:00 regulate SOEs,

1:01:02 uh,

1:01:03 or institutional variables like

1:01:05 federal governments,

1:01:06 right,

1:01:07 uh,

1:01:07 compared to centralized governments.

1:01:10 Um,

1:01:11 also,

1:01:12 another point is,

1:01:14 uh,

1:01:14 Chinese debt restructuring seems

1:01:17 Uh,

1:01:18 you know,

1:01:18 have different

1:01:19 haircuts and maybe a temporal dimension of it,

1:01:22 as also Marco,

1:01:23 uh,

1:01:23 noticed,

1:01:24 uh,

1:01:24 may be important.

1:01:25 Uh,

1:01:25 during HEPIC time,

1:01:26 it seems to us,

1:01:27 uh,

1:01:27 we,

1:01:28 we draft,

1:01:28 uh,

1:01:29 together with the IMF progress report on the HEPIC initiative.

1:01:33 Chinese creditors provided debt relief that was in line

1:01:36 with the debt relief required in the HEPIC initiative.

1:01:38 Uh,

1:01:39 therefore,

1:01:39 maybe it's worth,

1:01:40 uh,

1:01:41 uh,

1:01:41 stressing the fact that maybe there is

1:01:43 The Chi and Chinese banks or,

1:01:45 or different Chinese policy banks may have followed

1:01:49 uh other creditors,

1:01:50 comparability of treatment uh implications apply.

1:01:53 Um,

1:01:54 also,

1:01:54 uh,

1:01:55 in the paper on the domestic around default at the home and away.

1:01:59 Um,

1:02:01 There is a,

1:02:01 a bit of a contradiction.

1:02:03 On one hand,

1:02:03 it seems that

1:02:05 Uh

1:02:07 IMF lending.

1:02:09 Implies that there is not,

1:02:11 there are not external defaults,

1:02:12 but on the other hand,

1:02:14 external defaults are associated with the

1:02:16 debt relief from official creditors which in turn are also

1:02:19 take place always in the context of an IMF program or very,

1:02:23 very often,

1:02:24 I would say.

1:02:25 So there is a kind of internal contradiction

1:02:28 in this,

1:02:28 in these results.

1:02:31 Uh,

1:02:31 moving to data and definitions,

1:02:33 um,

1:02:34 I think that,

1:02:35 that both papers on the Belt and Road and the,

1:02:37 uh,

1:02:38 default,

1:02:38 uh,

1:02:39 home and,

1:02:39 uh,

1:02:39 and abroad,

1:02:40 uh,

1:02:42 could be more precise on the definition of defaults and restructuring periods.

1:02:46 This is really,

1:02:46 really,

1:02:47 uh,

1:02:47 relevant for the empirical analysis

1:02:49 to know exactly when a default happened,

1:02:51 right?

1:02:51 Default is different from restructuring,

1:02:53 so it's very important to be very clear on

1:02:57 what,

1:02:58 what is the,

1:02:59 what they measure.

1:03:00 Um,

1:03:01 also,

1:03:02 surprisingly,

1:03:03 the paper on the Belt and Road implies that many loans,

1:03:05 this is what is in slide 5,

1:03:08 are,

1:03:08 are not from China Exim Bank or China Development Bank.

1:03:11 And maybe,

1:03:13 you know,

1:03:13 it's worth

1:03:14 digging through,

1:03:15 right?

1:03:15 Provide more detail.

1:03:16 Would these

1:03:17 banks somehow or these creditors behave differently

1:03:20 and provide more or,

1:03:22 or less

1:03:23 uh PV reduction compared to the

1:03:26 two big banks,

1:03:27 not the China Exim Bank and the China Development Bank.

1:03:30 Uh,

1:03:31 finally,

1:03:32 the,

1:03:32 the,

1:03:32 the,

1:03:32 the default and institution paper uses probabilities that are from,

1:03:36 you know,

1:03:37 the,

1:03:37 the initial papers on the default like the

1:03:39 Reiner Rogoff paper in 2009 and other papers,

1:03:42 but there are,

1:03:43 you know,

1:03:43 sources of,

1:03:44 uh,

1:03:45 of,

1:03:45 uh,

1:03:45 of,

1:03:46 um,

1:03:47 you know,

1:03:47 of distress,

1:03:48 including these two papers now presented that could

1:03:50 be used to update and use much longer time period compared to the,

1:03:55 to the data presented.

1:03:57 Uh,

1:03:57 so let me skip the last one because I already made the point that,

1:04:00 that maybe we need,

1:04:01 we need,

1:04:01 uh,

1:04:02 it,

1:04:02 it could

1:04:03 be helpful to use indicators that are

1:04:06 closer to

1:04:07 Uh,

1:04:07 you know,

1:04:08 causes that,

1:04:09 that the World Bank or IFIs in general can act upon,

1:04:12 uh,

1:04:13 to,

1:04:13 to,

1:04:14 to help countries.

1:04:16 Finally,

1:04:17 moving to regressions,

1:04:18 um,

1:04:20 The Belt and Road paper focuses on the regressions basically on

1:04:23 the financial returns and put emphasis on loan purposes and characteristics.

1:04:27 But what about the characteristic of the country?

1:04:29 Would

1:04:29 the,

1:04:30 would

1:04:31 PV reduction be higher in low in low-income

1:04:34 countries or um compared to middle-income countries,

1:04:37 uh,

1:04:38 would,

1:04:39 you know,

1:04:39 Chinese

1:04:40 lenders

1:04:42 provide more concessional terms to,

1:04:44 uh,

1:04:45 countries that are also recipient of concessional financing from IFI.

1:04:49 So is there a

1:04:50 You know,

1:04:50 similar behavior,

1:04:51 uh,

1:04:52 that could be considered since you have all the terms,

1:04:54 um,

1:04:55 and

1:04:56 also

1:04:57 there is,

1:04:58 there was not similar analysis for the PB reduction,

1:05:01 right?

1:05:01 There are countries where China provide more or less debt reduction,

1:05:05 the original loan terms matter for the amount of PB reduction,

1:05:08 or,

1:05:09 you know,

1:05:09 different banks provide different debt relief.

1:05:12 Um,

1:05:13 the same,

1:05:14 same thing on the,

1:05:15 on the,

1:05:15 uh,

1:05:16 paper,

1:05:16 uh,

1:05:17 on the

1:05:19 default at home and,

1:05:20 uh,

1:05:20 and abroad,

1:05:21 where

1:05:22 basically regressions measure the association of certain

1:05:25 characteristics in the case of default,

1:05:27 however,

1:05:27 if the focus would be on more on a profit model,

1:05:30 Uh,

1:05:31 then,

1:05:32 you know,

1:05:32 probability of

1:05:34 default could have been

1:05:36 estimated,

1:05:37 uh,

1:05:37 that would have been very helpful to researchers and,

1:05:40 uh,

1:05:40 you know,

1:05:41 users like ourselves,

1:05:43 uh,

1:05:43 to understand the causes of,

1:05:44 of,

1:05:45 of defaults.

1:05:46 Finally,

1:05:46 the default and institutions paper uses

1:05:49 regressors that are are highly correlated,

1:05:52 uh,

1:05:52 richer countries have better institutions,

1:05:54 uh,

1:05:55 and therefore some robustness checks and uh

1:05:58 Uh,

1:05:58 for example,

1:05:59 use of instrumental variable would have been,

1:06:01 uh,

1:06:01 possible,

1:06:02 then theoretical model are a few,

1:06:04 um,

1:06:05 so just to the bare minimum,

1:06:07 is that in the theoretical model of the,

1:06:10 uh,

1:06:10 defaulted institutions,

1:06:11 basically,

1:06:13 the,

1:06:14 it is important to understand that,

1:06:15 that,

1:06:16 so a country with high

1:06:17 level of institutions basically does not suffer polarization.

1:06:21 This is what comes out from the empirical analysis.

1:06:24 But the model

1:06:25 in a sense implies that there is always impact on polarization,

1:06:28 and I wonder if

1:06:29 this,

1:06:29 uh,

1:06:30 this,

1:06:30 uh,

1:06:31 you know,

1:06:31 an additional variable could be introduced

1:06:33 in order to maintain the same

1:06:35 regularity that we see in the,

1:06:37 uh,

1:06:38 in the empirical data.

1:06:40 I'll stop here.

1:06:41 Thank you very much.

1:06:44 Yeah,

1:06:44 thank you.

1:06:47 And I believe we still have some time according to the revised agenda.

1:06:52 So,

1:06:52 um,

1:06:53 we'll now take questions from the floor,

1:06:56 5 minutes and so I think there are mics in the room.

1:07:00 So if you could go to a mic

1:07:02 and uh if you could briefly introduce yourself.

1:07:08 Thank you.

1:07:16 Hi,

1:07:17 thank you very much.

1:07:18 My name is Mohindra Gulati,

1:07:19 and my question is on the presentation on the BRI.

1:07:24 Where he showed that the rate of return,

1:07:26 financial rate of return was

1:07:28 Less than 2%.

1:07:30 I assume that this is on the basis of the

1:07:33 project cost,

1:07:34 structuring,

1:07:35 and the financing plan

1:07:37 as presented.

1:07:39 Did you take into account

1:07:41 the fact

1:07:43 that the goods and services

1:07:46 in the BRI projects

1:07:47 are generally supplied by the Chinese firms?

1:07:51 The clawback that you can get

1:07:54 from the inflated prices of those contracts,

1:07:58 which can be 1415,

1:07:59 16%,

1:08:00 and believe me,

1:08:01 that's a conservative estimate.

1:08:03 So if you have a clawback of 15% from inflated prices of those contracts,

1:08:09 Your rate of return would be

1:08:11 14 to 16%,

1:08:12 even if you deflate it

1:08:14 over a period of 3 or 4 years of implementation.

1:08:17 So the financial rate of return

1:08:19 is not 2%,

1:08:20 it could be

1:08:21 14 to 16%.

1:08:24 And one more variable,

1:08:25 if you add to it,

1:08:26 if these goods and services

1:08:29 are being supplied by firms

1:08:31 which have overcapacity,

1:08:34 then just imagine

1:08:36 the multiplier effect

1:08:38 of the utilization of that idle capacity,

1:08:41 uh,

1:08:42 which,

1:08:42 which would otherwise be a,

1:08:43 be a very expensive

1:08:45 financial and economic cost to the,

1:08:47 to the,

1:08:47 to the Chinese economy.

1:08:49 So the haircut that you see of 2%,

1:08:52 compare that with the

1:08:54 cost of ideal capacity

1:08:56 of the goods and services which are being supplied

1:08:58 in the BRI projects.

1:09:00 So,

1:09:00 were you able to take any of these things into account

1:09:03 in your calculations?

1:09:06 Let's take 2 more questions.

1:09:14 Hello,

1:09:17 I'm sorry.

1:09:17 Hello,

1:09:17 my name is Benny Lara,

1:09:19 um,

1:09:19 previously a member of parliament from Nigeria.

1:09:23 I'm here with the members of the Foundation and our heads,

1:09:27 Solomon and Mary La Foundation,

1:09:28 my mother,

1:09:29 former ambassador,

1:09:30 and Doctor Lara.

1:09:31 So the question

1:09:32 I have is,

1:09:34 you did say that

1:09:36 um there's

1:09:40 Domestic borrowing,

1:09:41 there's usually higher returns of,

1:09:43 um,

1:09:44 you know,

1:09:44 repayment on domestic loans than external loans.

1:09:48 What are you doing to ensure that because my country,

1:09:51 usually Nigeria,

1:09:52 we have,

1:09:52 you know,

1:09:53 huge volumes,

1:09:53 we borrow so much money and it becomes a political issue.

1:09:57 And it is going to be an issue in the next election.

1:09:59 What are you doing to try to ensure that some countries such as ours,

1:10:03 um,

1:10:04 we do more of

1:10:06 domestic borrowing rather than external loans,

1:10:09 um,

1:10:10 because that is always a,

1:10:11 a big political issue when it comes to elections.

1:10:15 Um,

1:10:15 and then also,

1:10:16 are we,

1:10:17 is Nigeria considered a low income,

1:10:19 medium,

1:10:19 middle income,

1:10:20 or

1:10:21 What,

1:10:22 you know,

1:10:22 because I hear in different parlance,

1:10:24 some say we're lower income,

1:10:25 some say we're middle income.

1:10:26 What exactly are we,

1:10:27 is our income level,

1:10:29 according to the World Bank.

1:10:30 Thank you.

1:10:32 Thank you.

1:10:34 Do we have uh another question?

1:10:38 OK.

1:10:39 If not,

1:10:39 let's start with this.

1:10:44 Yes,

1:10:45 mhm,

1:10:46 yeah,

1:10:46 it was,

1:10:46 I guess a question for,

1:10:47 uh,

1:10:48 for me.

1:10:48 I'm not quite sure honestly how this is classified,

1:10:50 uh,

1:10:51 by the World Bank.

1:10:52 Uh,

1:10:53 the,

1:10:53 the,

1:10:53 the income level for,

1:10:54 uh,

1:10:55 for your country,

1:10:55 I think it's a low-income country,

1:10:57 but honestly,

1:10:57 I would need to,

1:10:58 low middle income country.

1:11:00 OK,

1:11:00 thanks for,

1:11:01 uh,

1:11:01 for the,

1:11:02 I don't work for the World Bank,

1:11:03 so I'm not

1:11:04 supposed to know this.

1:11:05 So,

1:11:06 um.

1:11:07 In terms of domestic versus external,

1:11:10 uh,

1:11:10 uh,

1:11:10 we,

1:11:11 I mean,

1:11:11 uh,

1:11:12 we are not so sure that it's so straightforward to say that,

1:11:16 uh,

1:11:17 domestic debt is better than external one when it comes to default.

1:11:20 I mean,

1:11:21 we have a more nuanced,

1:11:22 uh,

1:11:23 actually take,

1:11:23 uh,

1:11:24 on,

1:11:24 uh,

1:11:24 on our result.

1:11:25 Uh,

1:11:26 for sure,

1:11:27 uh,

1:11:28 We see,

1:11:29 uh,

1:11:30 that,

1:11:30 uh,

1:11:31 widening,

1:11:32 uh,

1:11:33 the investor base,

1:11:34 uh,

1:11:35 is helpful for countries and is also

1:11:37 helpful for government because government can,

1:11:39 can actually default selectively

1:11:42 and therefore,

1:11:43 uh,

1:11:44 uh,

1:11:45 essentially,

1:11:46 uh,

1:11:46 you,

1:11:47 you go from a very discrete choice of either default or I don't default to a more like,

1:11:51 uh,

1:11:52 Uh,

1:11:53 continuous choice of uh a default on

1:11:56 a fraction of the debt.

1:11:58 So you,

1:11:58 you essentially,

1:11:59 you increase,

1:12:00 uh,

1:12:00 your,

1:12:00 uh,

1:12:01 your,

1:12:01 uh,

1:12:01 your,

1:12:02 your,

1:12:02 your possibilities as a government when you have,

1:12:05 uh,

1:12:05 uh,

1:12:06 domestic and external debt market.

1:12:07 I think a key issue

1:12:09 to develop domestic,

1:12:10 uh,

1:12:10 debt market is,

1:12:11 uh,

1:12:11 the debt uh of financial,

1:12:13 of domestic financial markets.

1:12:14 You need,

1:12:14 uh,

1:12:15 to have the ability to absorb the

1:12:17 debt domestically.

1:12:19 Thank you.

1:12:20 Do you want to take the question,

1:12:21 right?

1:12:21 Yes,

1:12:22 sure.

1:12:22 Uh,

1:12:22 thank you very much for the question.

1:12:24 These are very good points.

1:12:25 Uh,

1:12:25 so

1:12:26 the way we see the paper is that it's a first step towards moving away from anecdotes

1:12:31 and having a more systematic overview on the returns.

1:12:34 Now,

1:12:34 this said,

1:12:35 this is the first step in the sense that we look at

1:12:37 the perspective of Chinese creditor banks,

1:12:40 right?

1:12:40 So these are just the loans,

1:12:42 the pure financial returns or the

1:12:43 loans accounting for defaults and restructurings.

1:12:46 Um.

1:12:47 We do not

1:12:48 compute that this would be another project,

1:12:50 uh,

1:12:51 very ambitious one,

1:12:52 the returns for Chinese firms at large,

1:12:54 for the construction firms,

1:12:56 for the railway companies engaged there,

1:12:57 etc.

1:12:58 etc.

1:12:58 So that's,

1:12:59 you know,

1:12:59 the equity returns or the construction returns,

1:13:01 they are not in our paper,

1:13:03 but it would be fascinating to move that step and,

1:13:06 uh,

1:13:06 the,

1:13:07 the,

1:13:07 the points you raised,

1:13:08 uh,

1:13:09 you know,

1:13:09 raise further my interest in going in that direction,

1:13:11 but it's hard to get systematic.

1:13:13 Data

1:13:14 on these projects and to compute uh return on equity or return on investments,

1:13:20 both in the construction and the management of

1:13:21 these projects and what it all involves.

1:13:23 So I think

1:13:25 in a way,

1:13:26 this is a bigger agenda.

1:13:27 We are doing an,

1:13:28 I,

1:13:29 I think an important first step to understand the lending dimension,

1:13:32 the financial uh returns,

1:13:34 uh,

1:13:34 but there's much more of course to understand uh the,

1:13:37 the,

1:13:38 uh,

1:13:38 additional returns involved.

1:13:40 First,

1:13:41 the equity,

1:13:41 but also the non-financial returns,

1:13:42 the geopolitical returns that I was also uh mentioning in,

1:13:46 in my,

1:13:46 in my talk.

1:13:47 Thank you.

1:13:48 I believe we have time for one

1:13:51 or two more questions.

1:13:58 Please

1:14:04 I just wanted a clarification,

1:14:06 uh,

1:14:07 when you were presenting,

1:14:08 I think,

1:14:08 uh,

1:14:09 wrong.

1:14:09 I think it was your presentation about inequality,

1:14:12 increasing polarization.

1:14:15 Can you please clarify that?

1:14:17 Because

1:14:17 in terms of Gini coefficient,

1:14:19 which is what came to mind as you were talking about inequality.

1:14:23 I was wondering why you would say it increases.

1:14:26 Doesn't polarization

1:14:28 also worsen

1:14:29 inequality?

1:14:30 Because you have sort of people who

1:14:33 control the system in terms of governance.

1:14:36 And so when they control the system,

1:14:38 they manipulate

1:14:39 whatever resources comes in.

1:14:41 To me,

1:14:42 that seems to actually increase

1:14:44 inequality in terms of the Gini coefficient as opposed to the reverse.

1:14:48 Or maybe it's just a

1:14:50 You know,

1:14:51 maybe I got it wrong,

1:14:52 so please clarify that.

1:14:53 Thanks.

1:14:54 Yeah,

1:14:54 very good question.

1:14:55 Uh,

1:14:56 I was just waiting to see if there is,

1:14:58 uh,

1:14:59 one more question from the audience,

1:15:00 and will that,

1:15:01 with that will conclude the questions.

1:15:04 OK.

1:15:05 I think uh we are good from the audience,

1:15:07 so over to you.

1:15:09 Yeah,

1:15:09 uh,

1:15:09 this,

1:15:10 uh,

1:15:10 from,

1:15:10 so,

1:15:11 uh let me separate the,

1:15:13 the,

1:15:13 uh,

1:15:13 policy implication versus in the model.

1:15:15 In the model,

1:15:16 uh,

1:15:16 in the empirics,

1:15:17 the polarization is measured as ideological difference.

1:15:22 So in a country,

1:15:23 the largest 3 groups

1:15:26 was their ideology,

1:15:27 was their seat in the government to capture this ideological or political

1:15:32 polarization.

1:15:34 What I mean by increasing income inequality,

1:15:37 there's actually a recent

1:15:40 growing literature to try to understand

1:15:42 whether the increasing income.

1:15:44 Equality

1:15:45 also lead to ideological polarization.

1:15:48 And I think I see this in working in the countries

1:15:51 is

1:15:52 since

1:15:52 income inequality is so wide,

1:15:55 only the,

1:15:56 uh,

1:15:57 the elites are in the government

1:15:59 and they have a very different preference in spending decision

1:16:02 than other groups.

1:16:04 So,

1:16:05 if they are more.

1:16:06 Likely to be in making decision,

1:16:10 they might prefer different spending or borrowing decision

1:16:13 than if the inequality is less.

1:16:15 So,

1:16:16 so that's

1:16:17 where the policy implication come from.

1:16:19 But this,

1:16:19 this is not in the paper per se.

1:16:22 It's more on the reflection of how this paper,

1:16:25 the theoretical model can be applied in the current context.

1:16:29 Thank you.

1:16:29 And with this,

1:16:30 let's conclude.

1:16:31 Thank you so much for your participation

1:16:35 and many thanks to.

1:16:41 Thank you so much,

1:16:42 Manuela.

1:16:43 Uh,

1:16:43 we're taking a 5-minute break,

1:16:46 right.

1:16:47 Greater than the incoherence is the challenge of keeping us on time.

1:16:49 So please be back in 5 minutes.

1:16:50 There is coffee outside.

1:16:55 That

showAllTimestamps
no
transcript
Welcome back everyone to the afternoon session. Uh, for those joining online, please join using the discussion hashtag ABCDE 2024. Uh, I'm very excited to welcome you to the second session which will focus on sovereign debt and default. Our moderator is Manuela Francisco, and I, she, it's, it's now July, so she has a new title. So Manuela is the Global Director for Economic policies, um, in the prosperity vertical. Um, I'll hand over to you, Manuela. Thank you, thank you very much and good afternoon everyone here in Washington and good evening to everyone that is connected online. Um, I'm very pleased to see the inclusion of a session on sovereign debt at this distinguished conference because as we all know it's a very relevant topic. Uh, to motivate this discussion, I would like to make 3 points that I will be very, and I will be very brief. First, Is the following, debt levels have stabilized, these are the good news, but they have stabilized at very high levels, and many countries are facing liquidity challenge. Just to give you a few numbers, 8 years ago in 2015. 25% of the low income countries that prepared DSF were at high risk of that distress or in distress. By 2019, this number reached 50%. And this was because between 2010 and 2019, countries borrowed significantly, which was facilitated by widely available capital at a very low cost. Between 2010 and 2019, during this period, Public debt. In low-income countries, and middle-income countries grew by 131% in nominal terms. Then, as we know, came COVID. And then in 2022, it started the tightening of monetary policy and many countries were pushed over a cliff. But there are encouraging news. Uh, some countries have started to tighten fiscal policy. Uh, some countries have gained access to the capital markets, uh, but we know that gross financing needs remain still very, very high. And while credit spreads have declined, we know that the underlying interest rates are still very high. My second message is that the that landscape has changed. And the debt structuring process are still adapting to this new world. So as public debt levels increase, the composition of the creditors also changed. There has been a shift towards non-traditional creditors and bondholders. There is also more domestic debt. So now over 50 low-income countries have domestic debt and total debt is about 25% of GDP. There are good news with that. Countries will be shielded from exchange rate volatility and currency mismatch. But when it comes the time of that restructuring, this can become very complicated because of the wider implications for the economy. My last comment is that our point is that countries are struggling to roll over their debt. We are living a period of very expensive funding costs. There is uncertainty regarding the pace of monetary policy easing. But I think there is certainty about the fact that the era of very low interest rates will not come back. So while major central banks around the world are in a path of monetary policy easing and the Fed has already cut policy rates, The, the ECB I mean has cut policy rates. The Fed is still waiting to see what's happened in the job market and the, and prices. But even When the Fed cuts policy rates, the low income countries will still face very expensive financing costs. So that's why this is also relevant today. So, um, it is my honor to moderate this distinguished panel and let me introduce the panelists, and I will start by the order of their presentation. So on my left I have Christoph Tribes from Keo University, Enrico Mallucci from the Federal Reserve Board, Ron Kang from the World Bank. And following their presentations of these three speakers, Marcos Chasmos from the Fund and Luca Bandiera from the World Bank will provide discuss and remarks. So let me pass the floor to um Christo. Thank you so much, uh, for the moderation. Thank you so much for inviting me, uh, and allowing me to, uh, present to you our most recent paper, uh, we just joined with, uh, Lukas Franz uh from the Keen Institute, Sebastian Horn from the Bank, Bradley Parks from William and Mary, and Carmen Reinhardt from Harvard University. And this is part of a broader research agenda on official lending, so state-driven lending on official creditors. In general, and uh in particular on China's overseas lending. So despite the growing literature on China's overseas lending, we still have very little insights on the financial terms, especially in, in particular, the returns, the financial returns of this 1 trillion lending program that China has rolled out over the last 15 years, and this is the focus. Of today. Um, we see this paper speaking to a bigger phenomenon a long history in which great powers have shown a, uh, tendency to invest or to, to launch these large scale infrastructure investment programs, uh, this just two historical examples. Germany in the late 19th century launched the Berlin-Baghdad railway or the US with this famous Marshall Plan. Both of them investing heavily in infrastructure and faraway lands and facing substantial risks on this kind of large scale uh investments. Um, so the benefit of looking at China today is that, uh, thanks to uh consider research by a data and, and Uh, and this other scholarly work, we actually have a pretty good idea, more granular insight into these, into one of the most prominent lending programs of this type, uh, and we can assess the dissociated default risk and the returns to the creditors, to the, to the great power, uh, lender, uh, in a more systematic way than we could do before. Um, so the key question I'm gonna focus on is, is China's project of the century a financial success? There's two main contributions of the paper. Uh the first, and that's an essential ingredient to understanding the financial returns, is to understand defaults and credit losses. So, uh, events of great events and the associated restructurings or, um, credit loss, so haircuts, uh it's often called on uh either the Chinese banks, and we put together this comprehensive data set with really the details of the sovereign debt restructuring. Um, deals, uh, with Chinese creditors and then, uh, compute haircuts, um, and compare it to the Paris Club and other and private external creditors. And the second contribution is we use the haircuts and the default information. Um, to compute exposed returns, uh, on, uh, on China's, uh, lending portfolio. Uh, so for every one of these 3500 Chinese loans with the total volume of volume of 1 trillion, we compute, you know, the exact cash flows, we consider the, uh, credit events, uh, towards each of those loans and come to the, uh, the keys inside is that realized rate of return is about 1.7. Uh, percent per year, uh, in real terms, inflation adjusted, which is neither particularly high and not particularly low, I'm gonna benchmark that to other, uh, creditor types. So let me turn to the first part, haircuts. So the losses of Chinese banks on their Belt and Road portfolio abroad. The central contribution is that uh we we uh we compile this granular data set, um, and importantly, uh, most of these deals are rescheduling agreements. So they push out the payment by a couple of years, often not even cutting the interest rate. So it's basically just a five year kicking down uh the can operation with very few instances of face value reduction. So nominal debt reduction is rare in these deals. Uh, which is why you really need to, to look at the present value or the, the, the, uh, payment streams of these loans in detail. Um, So we use a standard uh to get Fettelmeyer haircuts uh to compute uh haircuts. So comparing like a hypothetical scenario in which you say 99% of the credit is agreed and 1% does not agree, uh, and is repaid as if the default would never have occurred. Uh, what is the loss of the 99% participating vis a vis that 1% holdout is being paid, right? So, so that's kind of at the point of the restructuring. How much do creditors lose? That's the way haircuts are computed here. Um, and, uh, as of course discount rate is crucial, we use kind of uh an upper bound here, the market yield, an exit yield, uh, based on bond, uh, uh, bond yields at the exit from the restructure but we also use a official creditor uh uh discount rates of CIRR as a, a robustness check. This is one of the key graphs of the paper, so this is the size of the haircuts on Chinese loans and various um restructuring operations. Um, I guess I can't, you can't see the, uh, no here. So what you can see, the big takeaway is that That in the large number of cases, haircuts are surprisingly low. Uh, so from 0 to 0 to 20%, so the average is just 12%, and that's in historical comparison, low. Uh, this is just one comparison, we also do the comparison with Paris Club haircuts from our other uh work. Um, and what you can see is that if we Compare the haircuts on the Chinese restructuring events, which are the red dots, uh, with haircuts, uh, on private external credits from the Cruz's Trevich data set, uh, we get a, we get substantial differences, so the mean Chinese haircut is 10% or close to 0 if you want, and the mean, uh, on private external credit is the mean haircut is 36%, uh, so considerably higher. And if you consider the Paris Club haircuts, those are above 50% on average, so significantly higher than that even. Let's move to the second part, creditor returns. Um, we, since these are not market instruments, these are not traded instruments, these are loans by Chinese banks, um, we compute internal rates of return. Um, and therefore, for every single of those loans, we, we, you know, map the information we have on commitments, uh, the financial characteristics of the, of the term of the terms, the grace periods, etc. uh, and then incorporate the information on default and missed payments, OK, and the haircuts, uh, which gives us this, uh, realized rate of return measure. Uh, just, this is just an example, what, what kind of a Just an illustration of the repayment terms, once we do bond by bond and we, uh sorry, loan by loan and then aggregated at a yearly level, right? What, what are the repayments towards Chinese creditors from the global south or from the debt Belt and Road countries, you can see that there is no, it's no surprise that we're currently talking a lot about uh debt sustainability and about repayment problems towards China because this since 2020, we really see a peak in repayment. And we see that, you know, there's two big elephants in the room in terms of creditors, and that's China Development Bank and uh Export Import Bank, so Exim Bank of China. This graph shows you the difference in realized payments. Um, so what, what part of these repayments, uh, did not occur, right? So the red parts of these bars are the missed payments. And what you can see is that, of course, the Blue bars are much larger. Missed payments account for a relatively small share of total debt due to China. Uh, so overall, China has not suffered massive arrears, massive missed payments, defaults on its portfolio. We can also look at this by comparing ex ante returns or the promised returns at, at, at the landing point. And exposed returns accounting for uh losses and haircuts. Um, this is the nominal return now, uh, and so the every country is colored according to the average return. Um, and when we compare these ex ante returns and compare that to the exposed ones, we see very little change in the coloring. Only a few African countries see a substantial shift in the aggregate return. Uh, on, uh, Chinese, uh, loans from a, from a Chinese perspective. Uh, so the difference in short between ex ante and exposed returns is not particularly large except for a few outlier countries. OK? So how does the return on China's lending portfolio compare in the bigger picture, right? So would it have been a better alternative uh to put this money, say, in a sovereign wealth fund, safe, China's uh sovereign wealth fund and invest in a portfolio of emerging market bonds instead or maybe in a treasury account with treasury. bonds. Uh, to, to understand that, we compare the average real uh return uh on a yearly basis of the BRI loans with other asset types, asset classes, and we can see that the returns are higher than the alternative of investing in US treasuries and even higher than the alternative investing in Chinese government bonds. But they are significantly lower. Uh, towards the benchmark of investing that same money into Chinese equities or into the standard, you know, JP Morgan MB, uh, portfolio, uh, or even MSCI World Global Equity portfolio. So in that sense, the markets, uh, returns in standard, uh, asset classes in emerging markets are considerably higher, uh, but it is, uh, more profitable than, than investing, uh, holding more treasury bonds. It's uh in the, in the mid-range. So why are creditor returns so low compared to that kind of market benchmark? To answer this question, we're now gonna make use of the large loan by loan variation in returns. We know the ex ante returns of every single loan. And here it's just plotted the variation for every year uh in the, in the uh contracted return, right, ranging between 0 and uh even up to 10% or more, right? And what you can see that there's considerable variation in the lending terms that China gives out. So what drives this variation? What, what explains why some of these loans are really low, below market rates? Um, and to make sense of that, we, uh, exploit the fact that China's overseas lending is state-driven. So likely to have a, uh, or at least there is a possibility of direct political control and strategic lending because this is the state lending, right? Or state entities lending. Um, with that perspective, we, uh, look at, uh, projects that might be in the interest of China's, you know, overall, uh, um, uh, so overall activities abroad. First of all, uh, projects that might foster market access or exports, um, like infrastructure projects, uh, projects that are implemented by the Chinese entities or construction. In Africa, for example, or loans denominated in renminbi. So it's, it's China trying to push uh the internationalization of its currency by subsidizing loans that are renminbi dominominated, right? So the question of uh non-financial externalities is what we're after and proxies for project with such non-financial externalities. And the second is uh looking at political potential uh project with political um uh Uh, implications of prestige projects. It could be stadiums or a parliamentary buildings, uh, military security projects in the recipient countries, uh, loans to public recipients or loans to the region from which the leader of the country, uh, originally comes from. OK? So this is classic leader favoritism, uh, literature. And what you can See in this graph, and that's my penultimate slide here, is that uh projects that are associated or likely to have higher non-financial returns beyond the mere financial relationship uh like prestige projects, those denominated in RMB, uh, military security projects or projects for trade infrastructure, um, uh, see a higher, uh, see, see a significantly lower return, so they're subsidized. Uh, whereas, uh, other projects that are, uh, you know, for private recipients or for emerging market recipients see higher, uh, uh, returns and are not subsidized. And we confirmed this in a regression framework, but since I promised to be short, let me wrap up. Uh, overall, uh, the financial returns of the BRI so far, uh, show that this is a no gain, no loss project, so the, the returns are needed. Catastrophic as sometimes assumed. Um, and looking ahead, of course, the big question is whether the haircuts of the China will have to accept higher losses, uh, and whether, uh, which would imply lower returns, uh, and what the current shifts in the way China lends that do, especially the focus to, uh, towards more state-owned commercial bank lending with higher returns ex ante. Thank you so much. Thank you so much. And let's now hear from Eric. Enrico Mallucci. OK All right. So thank you very much for including the paper, uh, in the program. So this is joint work uh with AOEE, uh, University of Navarra and Mattia Piccarelli, uh, from the European Stability Mechanism, uh, and, uh, of course, uh, the usual disclaimer applies. So it's a well-known fact that, uh, sovereign debt markets have evolved quite a bit in the last 30, 40 years. So if you think about it, I mean, uh, bank loans were replaced by bonded debt, uh, and the size of emerging markets government debt has increased, uh, uh, substantially, but perhaps, uh, at least it's our taking the biggest change of all. has been the growing role of domestic markets. So we have seen, uh, just to give you an example, uh, think about Mexico. So in Mexico, uh, domestic, uh, the domestic share of government debt was about 25% in 1995. If you take the same number in 2010, you find an 80%. So this huge increase in the size of, uh, of domestic markets. And of course, the growing role of domestic markets for government debt has also translated uh into a greater involvement uh in uh Domestic, uh, uh, of domestic debt in defaults and restructuring events. So, Greece, for instance, uh, is a poster child example. So if you think about the, the Greek restructuring of 2012, well, 92% of the debt that was restructured uh was actually domestic debt. So it was really a domestic restructuring. And as we speak, of course, countries, uh, Both under and and outside the cover of the common framework are conducting sovereign debt restructuring involving uh bonds issued domestically, so it's very much uh in today's topic. Um, so the problem is, uh, that we have seen this, uh, increasing the importance of domestic debt, but there is limited systematic understanding of how, of the differences, uh, between, uh, uh, defaults, uh, between sovereign defaults that involve, uh, domestic debt and external debt. And so this is exactly what we do in this paper. So we systematically compare sovereign defaults. of government debt issued in domestic and international debt markets and we compare them. And of course, then we use the insights to inform the academic literature as well uh as uh policymakers. Now, to, to, to, to really compare what, uh, what is happening when you default on domestic and when you default on external debt, of course, you need the data, OK? So you need uh to have a comprehensive view, you need to look at the universe of uh the domestic and external debt restructuring. So this is why in this, in our paper, we leverage on uh two state of the art databases. So we take information uh about uh domestic defaults and restructuring uh from uh the paper by Chris Treves here and uh and Tamon which I also saw in the audience. And then we complement this information about external restructuring with information about the domestic data restructuring, which we take from a paper that I or I, and, uh, and Matthia we constructed, we constructed over the last 10 years, OK? And then we harmonize the two databases and uh we just uh again analyze how the two phenomenon differ. So the resulting database has uh about 100 and 116 uh domestic default restructurings and 177 external default uh uh events. We span from 1980 until 2018, and we have information of really a bunch of things. So we have information about the timing of this restructuring, so when they start and when they end. We have information about the instruments involved, so we see if, uh, if we are talking about bond, if we are talking about loans, we see the volumes involved. So we see how big these defaults actually are. And then we also look at the, uh, we also have information about the restructuring terms and the restructuring, uh, approach that are taken in, uh, in, uh, in this restructuring events. And we also have information about the NPV losses, but this is only a few, for a, for a, for a subset of, uh, Of the restructuring event. And then a key feature of the database is that it also has a global coverage. So we, uh, we have information coming from 84 countries that span, uh, uh, all continents and all income group. So this being said, uh, how the two differ, how the, uh, how these two phenomena, uh, differ. So the first thing we find is that domestic defaults have become more frequent than external ones. So if you look at the red line, that is like uh, uh, the three-year sum of uh external default episodes, and you see that essentially they peaked in the 80s, but now they are, uh, if not rare, they are rarer than they used to be. The black line instead plots the same, uh, the same uh information for uh domestic uh restructurings. And you see that from uh about like the end of the 90s, the black line has crossed the, the, the, the red line, meaning that nowadays, uh domestic uh restructuring are actually uh more frequent than, uh, than, uh, than external ones. The second thing that we find that selective defaults are the norm. So we define selective defaults as the restructuring episode that only involved either domestic or external, uh, or external debt. And, uh, essentially, uh, what we find is that 70% of the default episodes are selective. So when we think about, uh, uh, sovereign defaults, we are really normally facing, uh, uh, selective defaults. And we find that about 60% of the domestic defaults and 77% of the external defaults are selective. Now, the additional piece of information is that not only Selective defaults are the norm, but non-selective defaults, so if we look at the last, the, the remaining 30% of, uh, of the episode, they are also somehow selective in the sense that they discriminate between lenders. So in this Uh, scattered plot, uh, essentially you have on the horizontal axis, the domestic, uh, restructured debt as a percentage of total restructure debt and on the vertical axis, uh, the, the, the domestic versus external composition of government debt. So this is what this graph is telling you is that even when governments do default on both external and domestic debt, they tend to discriminate against the type of debt which has the largest share in the overall composition of, uh, of total debt. So if you have a very high external debt, you tend to default more on, uh, on, uh, on the external portion of, uh, of, of the debt. Next, we look at the size of uh domestic versus external default. We find that the domestic defaults are smaller, so here in the table, we report Uh, the, the, the, the, the, the percentage as a fraction of the GDP of the debt in default and you see that, uh, on average, uh, domestic, uh, defaults are, uh, they, they involve uh debt which is about 5% points lower as a fraction of GDP of, uh, of, uh, external debt restructurings. Uh, we also find that the domestic restructurings are faster. So in the table here, you have, uh, the, the duration in months of external and domestic, uh, debt restructurings, and what we find is that, uh, uh, domestic restructuring takes on average about 9 months less, uh, to be resolved than, uh, than, uh, than, uh, external one. And then interesting thing here is perhaps like, uh, the evolution of these, uh, of these numbers. So, Uh, we interpret the fact that, uh, domestic restructuring are faster to resolve as a reflection of the fact, uh, that when you have domestic debt, it's actually easier to restructure it because typically domestic debt is governed by the domestic law. And so changing the domestic law is much easier than, uh, than changing the international law. But then the interesting aspect is that uh there, there is a, we have seen a convergence uh in the speed of uh uh the restructuring of domestic and external restructuring, and we attribute this to the fact that external restructurings are getting faster thanks to to the introduction of cuts in, uh, in, uh, in, in the contracts. Um, I'm gonna skip this, uh, because, uh, we need to be quicker. Uh, the next thing I want to say about domestic and external default is that, uh, what we are seeing is that domestic defaults are smaller, quicker, but yet, uh, for the subset of, uh, uh, of episodes for which we have the information about NPD losses, NPV losses, we see that domestic defaults are actually more punitive. So here, we report NPV losses for domestic and external debt, and you see that, uh, uh, that, that the NPV losses are actually higher for, uh, for domestic, uh, for domestic restructuring. And this is for us a bit surprising because there is often this belief, uh, that governments tend to be nicer, let's say, uh, towards, uh, their, uh, towards domestic creditors, but here we don't, we don't see this. I'm gonna also jump, uh, uh, this slide, and I'm gonna concentrate to the last bit, uh, of, uh, the paper. So in the last bit of the paper, we change a bit the focus and, uh, we compare essentially the economic and political context, uh, in which domestic and external debt restructuring happens. So in particular, we are interested in, uh, how, uh, macrofinancial and political variable differ, uh, when, uh, when, uh, Uh, governments take the decision to, to restructure their debt. Um, so the first thing that we look at, of course, uh, is, uh, at, uh, how GDP evolves around domestic and external, uh, external debt restructurings, uh, and, uh, this is what we show in the first, uh, in the first column. And what you see here is that the coefficients are very similar and they are both negative and significant. So, what we find is that both domestic and external defaults tend to happen in periods of low growth, OK? And so this is actually a very standard result. And the, the second, uh, the second column tells you that, uh, both also happen in periods of very high debt. So this is nothing surprising. So we see defaults when the economy is doing very badly and when, uh, Uh, that is very high. But then in the next table, we actually look into essentially what are the different channels that might explain, uh, this, uh, uh, uh, this decline in GDP. And we see that domestic and external, uh, restructuring, uh, happening very different, uh, in very different environments. So we see that the external default happen in periods of substantial external adjustment and you can see it in the first corner. So you can see that when you have an external default, then you have like uh uh like an improvement of the current account and, and, and also like a drop in the, in the foreign inflow. So this is an external adjustment and we don't see this happening this as much when you, when you, when you have like a domestic restructuring. Instead, when we look at Private credit, so when we look at one of the variables that is telling us what is happening on the credit side, we see a negative significant coefficient only for domestic, uh, for domestic restructuring. So this is telling us that domestic default happens in period or are associated with the, uh, credit crunches. Um, so the conclusion here is that even though like perhaps the, the, the, the overall economic environment is not to be similar because we see, uh, we see this contraction of GDP, the channel that explains this contraction of GDP can be very different, uh, in the two, in two different, uh, Episodes I'm gonna jump the, the, the slides and gonna like talk about the last two things that we look in the paper which is uh the political context in which domestic and external uh uh restructuring happens. And then I'm gonna talk about the role of the IMF. Um, so, What we see in the first, sorry, uh, in the first column is, uh, a variable that measures political, uh, stability. So what we find, and this is a commonality between domestic and external default, is that both of them, they are associated with political instability, OK? But then, uh, we see that the elections uh are twice as likely during domestic defaults than during external default, OK? So, Possibly, voters are very displeased when you default on, uh, on domestic debt. And then instead, what we see is that the radical governments are more often in power when, uh, when you have uh external, uh, external, uh, external defaults because perhaps you need a reckless government to challenge, uh, you know, the, the, the international, uh, rule of law. And finally, as I anticipated before, I, I, we also look a bit at the role of uh bilateral lenders. And so we see, we look at what uh IMF programs and official debt relief, so what is their impact on, uh, on external and domestic uh restructurings, uh, and we see that, uh, In the, uh, and you can see this in the first column, countries that receive support from the IMF, uh, they tend to prioritize uh repayment of external debt at the cost of actually increasing, uh, their, uh, their domestic, uh, domestic restructuring. And here for, for the IMF people in the, in the, in the room, uh, I mean, we interpret this a reflection of the uh of the lending into arrears policies. So the lending into arrears policies adds An additional layer of uh uh requirements uh to, to, to actually Uh, to actually, for the IMF to borrow money to, to, to lend money to countries in troubles. And so, the, and the lending into a real policy only covers external debt. So, our intuition is that perhaps uh Countries want to avoid this additional layer of like requirements and so they prefer to, to, to default on domestic debt to avoid entering uh into the uh lending into arrears programs uh with the IMF. And finally, when we look at the official uh debt relief, uh, we see that countries that uh receive official debt relief are more likely to default, uh, externally, and we, we think this, uh, as a reflection of the comparability of treatment policies that are often associated with, uh, uh, with, uh, with this official debt relief. In conclusion, uh, compared to the faults abroad, the faults at home are smaller, quicker, and deliver larger, uh, Uh, larger losses and the environment surrounding domestic and external defaults, uh, is very different. Domestic defaults are associated with credit crunches. External defaults happen at times of fiscal and external adjustments, and also what we see is that external defaults are less likely with moderate governments, while domestic defaults are more likely to trigger elections. Many thanks a week. Now let's hear from Rong Cheng. The floor is yours. Thank you for inviting me to present this paper which is joint work with Francisco from the IMF. Um, so, the previous two presenters show a little bit of the historical, uh, default history. What I'm gonna present is uh more or less a theoretical model to explain why some countries deform more than others and the role of institutions. So, I will go very quickly on the motivation is to try to understand what explains the difference behavior between high-income countries and middle and low-income countries. In principle, the, the level of income doesn't matter for default risk, but we do observe from the history. And when we look at it among the middle and low-income country, we also see that Latin American countries tend to devolve more often historically. And I'm from Argentina, uh, so that's my motivation. Um and then I, well, I have two anecdotal evidence, but I will only present the case of Argentina in the 80s that had two sovereign default crises. So, in the 80s, the, the consolidated fiscal deficits reached to 15% of GDP from the federal government, local government, and uh SOEs. And at that time, there was a, a revenue sharing scheme. Uh, basically, the federal government raised revenue and then transfer to the, uh, local government, and they do have an extraordinary uh treasury transfer system for um unusual events. But during the three years. In 1985 and 1987, this revenue sharing scheme was suspended because the law that governed it expired and took them three years to put back. So during these three years, basically the transfer between the federal government to the local government, local government, it was a bilateral negotiation. Uh, and that created the uncooperative behavior by the local government, tried to get maximum, uh, they can from the, uh, central government, uh, and the federal government couldn't, uh, push back because there was no legal framework in it. So that's just keeping in mind this, the, the influence of the local government on the federal government borrowing decision. I will skip, uh, Brazil. So, just very quickly, the related literature, look at the sovereign default, um, and usually in one country like Argentina, try to explain what explained it, and more recently, there's a series of papers that incorporate political. Economic consideration in to explain that, such as the degree of impatience from the incumbent government or there's a left-wing government and right-wing government that has different preference that can lead to overborrowing. And that is um something that uh our paper as well as show. So let me show you quickly, uh, two-period model with a closed form solution so you can see the intuition of the model. So this is the government structure. Um, the government is composed by a group of powerful groups that want to influence in policy decisions. In countries with good institutions, there's rules set up that limit the influence of these powerful groups in central government decisions on spending and on borrowing. Uh, in weak institutions, the central government cannot limit the influence, so something, uh, will happen, as you will see, and these, um, groups, they also have different ideas about what to spend, how to spend, so there's existence of a polarized, uh, polarization in ideology. That, that means that they cannot actually coordinate in the optimal solutions because they have different ideas about how to govern and how to spend. And the distribution of power is captured by the, the their power in the Congress, for example, or in terms of the share they get from the central government. And these powerful groups, they consume, um, they derive utility from the consumption goods. They receive revenue or external borrowing from the central government. They cannot do it directly. They have to go through the central government. And the good institutions, because the central government uh solution is the optimal solution, they borrow, they make the optimal decision and transfer to the local government. In a weak institutions set up, these groups. Kind of uh maximize their own utility function, make decisions, and as a result of that, there's a negative externality that will arise because they do not internalize the action of their maximization function into other groups. So that's the, the, how the powerful groups try to influence policy. And then we have the foreign investor that this is very typical in the literature, that they are risk neutral and they, they are competitive, uh, they try to maximize, um, um, the, the interest rate that to match the risk-free interest rate. They know the institutional quality of the government. Uh, and the degree of polarization in the economy, they lend to the government, uh, as a whole, not to individual groups. And for the foreign investor, as long as they don't get the full repayment, they constitute a sovereign default. So these are the, they are all extreme, but it's uh the setup for the modeling purpose. So in the two-group and two-perio model, we have two groups. They have, uh, they are trying to uh uh consume goods through the transfer, and in the model, I just separate countries with good institutions, what I call unified government or countries with weak institutions, what I call polarized government. And remember that in the polarized government, The powerful group tried to influence the central government policy, and they succeeded. Uh, so there's a, a negative externality because they cannot coordinate. So I won't go into detail. The unified government is the typical uh central planner solution. They maximize utility of the two periods subject to a budget constraint which derived from the endowment and the borrowing. And in the second period, the revenue is the only shock in the economy, so they either default or repay depending on the realization of the revenue shock in the second period. In a paralyzed government, each group makes this same maximization problem. Uh, as you say, uh, see, there's, uh, the maximization problem is for group one and two. And the foreign investor know that uh in a weak institution, which means in this case, the paralyzed government, the default risk is the default risk, the maximum default risk among the two groups. As long as one group defaults, it's optimal for the other group to default as well. So that's the equilibrium interest rate. And, OK, uh let me skip this. Uh, so, uh, the, the, the theoretical model can derive that for all, uh, degree of polarization in a, a weak institution, which means it's the polarized government, the default rate is higher, the interest rate is higher, and the debt level is higher. And this is the intensive margin. On the extensive margin, the more polarized is the government, the higher the default and the debt and interest rate. And let me just quickly tell about the intuition. It's the marginal cost of borrowing for one group in a weak institution set up is lower than the aggregate cost of borrowing because they, they do not internal. The cost of borrowing that affects the whole country as a group while they only receive a portion of that debt. That's the source of a negative externality, not internalized in a paralyzed government, which means it's a weak institution set up. So let me quickly show you, uh, we try to incorporate this uh in an empirical study and use two variables to capture the institution. One is the standard CIRG and one is the regulation of influence by the different interest groups in the, in the economy. So, and the number 5 is the regulator's case, the country. Has set up the clear rule of how these different powerful group can influence in the central government decision and which policy cannot be influenced. In an unregulated situation is, remember the case of Argentina during those three years, there's no legal framework how this negotiation happened. And then we tested, uh, let me just show this one. so we separate the countries, uh, in good institution versus, uh, weak institution. So the variable para is the high, the higher is the better. So for countries below the medium, uh, the first column, and above the medium, and the 3 and 4 is the countries with para above a medium, uh, below medium plus 1 standard deviation and above. What I see, what I try to show here is that the degree of polarization only matters in the setting of weak institutions. So you can have, you know, the ideological difference among the government as long as the institution is strong enough to limit the influence of these groups, the default risk is still, uh, manageable or the, the variable doesn't even matter for default risk. So let me conclude, uh, how this is, um, research is relevant today as Mueller already said, many countries have higher debt to GDP ratio even before the pandemic. It's even higher now, and we also know there's a rising inequality which could translate into ideological polarizations. So what do we do to limit these two factors that are already there, is to set up institutions to limit the spending pressure that will arise from the different groups. And now, currently I'm, I'm, I'm in Indonesia, the only country in Asia that had a a default crisis after the Asian financial crisis, and they do have a fiscal rule since then. So I see that this political influence still exists in the context of Indonesia, not maybe from the local government, but from SOEs, but they have the 3% of GDP as a fiscal rule that is hard since 2003 and still remain very relevant. As a result, their debt to GDP ratios. Remain the lowest in among the lowest in Asia Pacific, but at the same time, um, this budgetary process, the PFM that the bank also uh uh advised a lot, is also very important to make it transparent so this political negotiation doesn't happen. And finally, the debt transparency and. Part of the other speakers um show, it's very important to know actually how much is the total debt that we are talking about and try to conduct fiscal consolidation to lower the risk because all defaults happen for only 3 reasons either capital flow reversal. Interest increased by the um developed country or commodity shock. And to lower the risk for default is only to limit the debt level for many of the developing countries. Thank you. Many times um wrong and with this we concluded the presentations and now we will hear from the discussions. Let's start with Mark Chamon from the fund. Thank you. Uh, it's a pleasure to be here to discuss the 3 papers. Uh, we have 3 papers to cover, and I think, uh, we are behind schedule, so let me just jump right in. So, uh, first paper, uh, the one on the returns to China's Belt and Road Initiative, um, I wanna focus the comments on the issue of the haircuts because I think they're also tied very closely to the, uh, financial returns on those loans. Um, the paper goes through, uh, systematically estimating the haircuts. I just want to highlight a few, uh, points on this issue of haircuts. First, that it, it is somewhat constrained the choice of haircuts. There are many reasons why, uh, even a large creditor may not have that much wiggle room to decide on its own haircuts because we tend to have fairly robust, uh, creditor coordinating mechanisms when it comes to official sector restructuring. So notably the price club process. And even if a, uh, uh, a creditor is not a member of the Paris Club, it will still end up being bound by the process because of the, uh, need to respect comparability of treatment. So that could, uh, constrain the choices. Uh, it can also be constrained by some hard restructuring targets. For example, if the country has an IMF supported program, uh, there might be a requirement to restore that sustainability, which will also, uh, in a sense constrain the, the choice of haircuts of the different creditors, uh, could propose. And last but not least, in the particular case of HPE, there were like very clear rules for how each, like, you could go low one by low one and apply those rules that would then determine what type of treatment would be expected on those low ones. Um. So if, maybe if one wants to focus on this sort of bargaining dimension, it might be useful to just focus on, on episodes where there was only a restructuring, a bilateral restructuring between the debtor and China, which was not part of, say, a broader uh debt restructuring. On, on the issue of haircuts, I also want to mention the discount rates. Uh, they, they consider a range of discount rates. I would suggest focusing on the sample on the HIPIC discount rates and maybe on the more recent ones, on the LDSF 5% discount rates, which is the, uh, yardstick that the other official creditors would use, uh, when looking at comparability of treatment. Um, on this, uh, and also encourage maybe a bit more comparison between the compare the haircuts, but also looking at the rate of returns, how it would compare to the other official creditors to also get a sense of how usual or unusual the, the lending behavior is. But on, on this issue of HP, going back to this issue of HPI, this is one of the charts that was used in the presentation. And, and just visually, the point I want to highlight is that A lot of the very episodes with very high haircuts tend to occur early in the sample where the circle where the circles are quite small. Remember, the circles show the size of China's exposure to the country. So in, in a sense, since China is a relatively new, uh, bilateral creditor, it sort of sit out hit peak. So when HI was taking place and all these very large haircuts were Taking place, China had very small exposures, and to the extent they had small exposure, it still did participate in the HIPIC initiative, uh, but the exposures were small, and I think that's maybe that factor alone may be driving the low average haircut and taking account HIPIC maybe would bring it closer to the average of, uh, the other official creditors. Turning to the second paper looking at domestic defaults, this is actually something quite topical. I mean, as the paper highlights, uh, domestic debt is playing, um, a relatively more important role, uh, in recent episodes. Uh, I think they may have too broad a definition of a domestic default. Uh, I think the, the, it may be including a lot of episodes that may be 1 may not necessarily, uh, considered a default, but I think Either way, 11 will cut it, the, the pattern will remain the same, that this has become uh more important uh in recent years. And I think this time dimension is something that maybe the paper could go uh in a bit more detail. They do have different breakdowns by time, but I think this is quite an important dimension to explore. On this issue of domestic restructuring, there is this perception that maybe restructuring domestic debt is, is easier because you're operating in your own jurisdiction, but I think that's quite a simplistic reading of the situation because any domestic restructuring would bring a lot of financial stability considerations into play. Uh, and however bad the debt crisis is, the financial crisis can be, uh, a lot worse. So you really need to be careful, uh, and, and I think countries do tend to be quite conservative when approaching this issue. Uh, and there was a reference to the IMF and we also tend to be quite, quite cautious about it as well. There's a recent, uh, board paper on domestic debt restructuring about 2 years ago where this financial stability, uh, considerations to the future quite importantly. Uh, the paper finds larger losses for domestic debt, uh, which is, uh, uh, a bit of a surprising result, uh, again, because this financial stability considerations are, are, are so prominent and you have to be very careful about pushing losses beyond what the domestic bank can take. Uh, and, uh, it's also surprising because in the, some recent cases, there is indeed the, the view that the domestic debt was restructured more likely, and there's no expectation of, uh, comparability of treatment, uh, between domestic and external debts, including the limiting case, which are many cases where external debt is restructured and domestic debt is not touched at all. Uh, so I, I would, uh, I think part of the challenge here would be finding the right discount rate for the domestic debt restructuring. I think this is something that's hard to do even for external debt restructurings. For domestic debt, it would be even more complicated because chances are, you are in a very volatile situation, high inflationary environment, so what discount rate should one plug in? Uh, maybe one way around it could be to, to complement the final analysis of the comparing the expost and the ex ante market prices of the restructured claims. I think maybe that should be doable, at least for a few recent cases where this data may not be so hard to get. And, and from there, you can compute different measures of either market or MPV haircuts and use it as a cross-check. And last but not least, uh, I'll turn to the, to the third paper on the role of institutions. Uh, it presents a very interesting model where you have two different groups in the economy making separate, uh, borrowing and spending decisions, and they're default decisions, uh, will affect the other groups. So it's a very, uh, Easy way to get this overborrowing problem, which is an alternative way to get over borrowing problem, like more traditional political economy channels in the literature like related to uh short horizons or, or high discounting. And they, they motivate the paper by the role of the subnationals, uh, giving examples of, of states in Brazil and provinces in Argentina. And the channel highlighted is quite strong as, as the model calibration shows. I mean, it, it is so strong that there has been a lot of effort, uh, to, to try to address this institutional shortcomings. So, I, I, in the case of Brazil, I know there was a, a, a fiscal responsibility law that was passed in 2000 that really constrained how much, uh, The, the fiscal policy of the states and local governments, and I think uh probably something similar may have taken place in Argentina. But so this initial uh size and indeed the anecdotal evidence all dates back from the 1980s. So when I first read the paper, my first reaction was that, OK, maybe this was a, a big problem back then, but we have seen uh graduated and the problem has been solved. But then on second thought, I, I think this problem is still quite relevant because there are many, many cases where we have state-owned enterprises which do make a lot of borrowing decisions and they are outside the direct control of the Ministry of Finance. They may not necessarily be reporting what they're doing to the debt offices. So I think this Sort of channel of the centralized borrowings, make borrowers making these decisions and then having spillovers of the country, uh, can, is still quite relevant, maybe not in the initial in the in the way that they, uh, related to the states or provinces, but I think through the role of SOEs, it's still, it can be quite relevant. So I think it's quite, quite an important paper in that regard. And that's it. I hope I, I'm OK on time. It's very well. Thanks. And now the last speaker of this session, Luca Bandera. Thank you very much. So I'll try to be brief. I took, I, I kept time for everybody now, I keep time for myself. OK. All right. So, uh, yeah, uh, so comments on these three papers, I will focus my comment on, on what's, uh, basically the, the, the policy implications since we are the World Bank and I work in the World Bank. Uh, then, uh, the date and definitions, uh, the regression and the theoretical model. So let me go very fast. So the key contribution, I would say, You know, I, I'm working with a lot of data on defaults and, and these papers, these two papers, no, the, the, the papers that, that, uh Christoph presented and, and, uh, and uh Enrico presented, um, were, were really, uh, really, uh, you know, uh, gave key contributions, right? Because the granularity of this information is extremely important, uh, knowing when a country defaults, how long the default lasts, it's, it's extremely important, and there is, uh, uh, say paity of data. On the domestic debt default episodes and, uh, and, you know, this, this analysis that was presented today really feels, feels, uh, uh, part of the gap, um, um, with, with some, some, uh, empirical uh statistics that it's, uh, it's important to notice, uh, so far, no, the, the, uh, you know, conventional wisdom is that the default episode basically overlap, but they said no, they don't, right? And it will be important to understand exactly in which countries in, in, in what circumstances. Also, the importance of political-related variables and um the fact that, that uh different defaults have different outcomes in terms of uh haircuts, uh uh they are preventive uh which creators have lost, so it's really rich, uh, the, the contribution of this paper is really rich. Uh, and there are policy implications, no, the, if I can really summarize, the defaults are very expensive, so better to avoid them, even though the contribution that, for example, an institution like the World Bank could give to this default is very different. Uh, external defaults are linked to, you know, balance of payment problems, uh, domestic default connected with, you know, uh, financial, uh, stress, domestic financial stress. Uh, debt relief is often not deep that we knew, and, uh, the fact that many countries today require deep debt restructure implies That, that debt restructuring are really prolonged and very costly for the countries and the creditors likewise, and loans from Chinese banks uh seems not to follow, you know, uh, economic and financial risk assessment, but they are motivated in a different way. That implies that maybe one of the largest creditors do not really follow what we try to preach, right? That the terms of the loans is to be uh adequated to the, to the, to the capacity repayment of the country. Other, other implications are less clear, uh, like, for example, we have very little uh leverage on the political system of the country. So I wonder if the analysis, especially the analysis on the, uh, defaulted institutions can look at You know, we can use in the regression and the empirical analysis, the variables that are more closer to what we can impact upon like countries that have fiscal rules, countries that have a good understanding of how to, you know, regulate SOEs, uh, or institutional variables like federal governments, right, uh, compared to centralized governments. Um, also, another point is, uh, Chinese debt restructuring seems Uh, you know, have different haircuts and maybe a temporal dimension of it, as also Marco, uh, noticed, uh, may be important. Uh, during HEPIC time, it seems to us, uh, we, we draft, uh, together with the IMF progress report on the HEPIC initiative. Chinese creditors provided debt relief that was in line with the debt relief required in the HEPIC initiative. Uh, therefore, maybe it's worth, uh, uh, stressing the fact that maybe there is The Chi and Chinese banks or, or different Chinese policy banks may have followed uh other creditors, comparability of treatment uh implications apply. Um, also, uh, in the paper on the domestic around default at the home and away. Um, There is a, a bit of a contradiction. On one hand, it seems that Uh IMF lending. Implies that there is not, there are not external defaults, but on the other hand, external defaults are associated with the debt relief from official creditors which in turn are also take place always in the context of an IMF program or very, very often, I would say. So there is a kind of internal contradiction in this, in these results. Uh, moving to data and definitions, um, I think that, that both papers on the Belt and Road and the, uh, default, uh, home and, uh, and abroad, uh, could be more precise on the definition of defaults and restructuring periods. This is really, really, uh, relevant for the empirical analysis to know exactly when a default happened, right? Default is different from restructuring, so it's very important to be very clear on what, what is the, what they measure. Um, also, surprisingly, the paper on the Belt and Road implies that many loans, this is what is in slide 5, are, are not from China Exim Bank or China Development Bank. And maybe, you know, it's worth digging through, right? Provide more detail. Would these banks somehow or these creditors behave differently and provide more or, or less uh PV reduction compared to the two big banks, not the China Exim Bank and the China Development Bank. Uh, finally, the, the, the, the default and institution paper uses probabilities that are from, you know, the, the initial papers on the default like the Reiner Rogoff paper in 2009 and other papers, but there are, you know, sources of, uh, of, uh, of, um, you know, of distress, including these two papers now presented that could be used to update and use much longer time period compared to the, to the data presented. Uh, so let me skip the last one because I already made the point that, that maybe we need, we need, uh, it, it could be helpful to use indicators that are closer to Uh, you know, causes that, that the World Bank or IFIs in general can act upon, uh, to, to, to help countries. Finally, moving to regressions, um, The Belt and Road paper focuses on the regressions basically on the financial returns and put emphasis on loan purposes and characteristics. But what about the characteristic of the country? Would the, would PV reduction be higher in low in low-income countries or um compared to middle-income countries, uh, would, you know, Chinese lenders provide more concessional terms to, uh, countries that are also recipient of concessional financing from IFI. So is there a You know, similar behavior, uh, that could be considered since you have all the terms, um, and also there is, there was not similar analysis for the PB reduction, right? There are countries where China provide more or less debt reduction, the original loan terms matter for the amount of PB reduction, or, you know, different banks provide different debt relief. Um, the same, same thing on the, on the, uh, paper, uh, on the default at home and, uh, and abroad, where basically regressions measure the association of certain characteristics in the case of default, however, if the focus would be on more on a profit model, Uh, then, you know, probability of default could have been estimated, uh, that would have been very helpful to researchers and, uh, you know, users like ourselves, uh, to understand the causes of, of, of defaults. Finally, the default and institutions paper uses regressors that are are highly correlated, uh, richer countries have better institutions, uh, and therefore some robustness checks and uh Uh, for example, use of instrumental variable would have been, uh, possible, then theoretical model are a few, um, so just to the bare minimum, is that in the theoretical model of the, uh, defaulted institutions, basically, the, it is important to understand that, that, so a country with high level of institutions basically does not suffer polarization. This is what comes out from the empirical analysis. But the model in a sense implies that there is always impact on polarization, and I wonder if this, uh, this, uh, you know, an additional variable could be introduced in order to maintain the same regularity that we see in the, uh, in the empirical data. I'll stop here. Thank you very much. Yeah, thank you. And I believe we still have some time according to the revised agenda. So, um, we'll now take questions from the floor, 5 minutes and so I think there are mics in the room. So if you could go to a mic and uh if you could briefly introduce yourself. Thank you. Hi, thank you very much. My name is Mohindra Gulati, and my question is on the presentation on the BRI. Where he showed that the rate of return, financial rate of return was Less than 2%. I assume that this is on the basis of the project cost, structuring, and the financing plan as presented. Did you take into account the fact that the goods and services in the BRI projects are generally supplied by the Chinese firms? The clawback that you can get from the inflated prices of those contracts, which can be 1415, 16%, and believe me, that's a conservative estimate. So if you have a clawback of 15% from inflated prices of those contracts, Your rate of return would be 14 to 16%, even if you deflate it over a period of 3 or 4 years of implementation. So the financial rate of return is not 2%, it could be 14 to 16%. And one more variable, if you add to it, if these goods and services are being supplied by firms which have overcapacity, then just imagine the multiplier effect of the utilization of that idle capacity, uh, which, which would otherwise be a, be a very expensive financial and economic cost to the, to the, to the Chinese economy. So the haircut that you see of 2%, compare that with the cost of ideal capacity of the goods and services which are being supplied in the BRI projects. So, were you able to take any of these things into account in your calculations? Let's take 2 more questions. Hello, I'm sorry. Hello, my name is Benny Lara, um, previously a member of parliament from Nigeria. I'm here with the members of the Foundation and our heads, Solomon and Mary La Foundation, my mother, former ambassador, and Doctor Lara. So the question I have is, you did say that um there's Domestic borrowing, there's usually higher returns of, um, you know, repayment on domestic loans than external loans. What are you doing to ensure that because my country, usually Nigeria, we have, you know, huge volumes, we borrow so much money and it becomes a political issue. And it is going to be an issue in the next election. What are you doing to try to ensure that some countries such as ours, um, we do more of domestic borrowing rather than external loans, um, because that is always a, a big political issue when it comes to elections. Um, and then also, are we, is Nigeria considered a low income, medium, middle income, or What, you know, because I hear in different parlance, some say we're lower income, some say we're middle income. What exactly are we, is our income level, according to the World Bank. Thank you. Thank you. Do we have uh another question? OK. If not, let's start with this. Yes, mhm, yeah, it was, I guess a question for, uh, for me. I'm not quite sure honestly how this is classified, uh, by the World Bank. Uh, the, the, the income level for, uh, for your country, I think it's a low-income country, but honestly, I would need to, low middle income country. OK, thanks for, uh, for the, I don't work for the World Bank, so I'm not supposed to know this. So, um. In terms of domestic versus external, uh, uh, we, I mean, uh, we are not so sure that it's so straightforward to say that, uh, domestic debt is better than external one when it comes to default. I mean, we have a more nuanced, uh, actually take, uh, on, uh, on our result. Uh, for sure, uh, We see, uh, that, uh, widening, uh, the investor base, uh, is helpful for countries and is also helpful for government because government can, can actually default selectively and therefore, uh, uh, essentially, uh, you, you go from a very discrete choice of either default or I don't default to a more like, uh, Uh, continuous choice of uh a default on a fraction of the debt. So you, you essentially, you increase, uh, your, uh, your, uh, your, your, your possibilities as a government when you have, uh, uh, domestic and external debt market. I think a key issue to develop domestic, uh, debt market is, uh, the debt uh of financial, of domestic financial markets. You need, uh, to have the ability to absorb the debt domestically. Thank you. Do you want to take the question, right? Yes, sure. Uh, thank you very much for the question. These are very good points. Uh, so the way we see the paper is that it's a first step towards moving away from anecdotes and having a more systematic overview on the returns. Now, this said, this is the first step in the sense that we look at the perspective of Chinese creditor banks, right? So these are just the loans, the pure financial returns or the loans accounting for defaults and restructurings. Um. We do not compute that this would be another project, uh, very ambitious one, the returns for Chinese firms at large, for the construction firms, for the railway companies engaged there, etc. etc. So that's, you know, the equity returns or the construction returns, they are not in our paper, but it would be fascinating to move that step and, uh, the, the, the points you raised, uh, you know, raise further my interest in going in that direction, but it's hard to get systematic. Data on these projects and to compute uh return on equity or return on investments, both in the construction and the management of these projects and what it all involves. So I think in a way, this is a bigger agenda. We are doing an, I, I think an important first step to understand the lending dimension, the financial uh returns, uh, but there's much more of course to understand uh the, the, uh, additional returns involved. First, the equity, but also the non-financial returns, the geopolitical returns that I was also uh mentioning in, in my, in my talk. Thank you. I believe we have time for one or two more questions. Please I just wanted a clarification, uh, when you were presenting, I think, uh, wrong. I think it was your presentation about inequality, increasing polarization. Can you please clarify that? Because in terms of Gini coefficient, which is what came to mind as you were talking about inequality. I was wondering why you would say it increases. Doesn't polarization also worsen inequality? Because you have sort of people who control the system in terms of governance. And so when they control the system, they manipulate whatever resources comes in. To me, that seems to actually increase inequality in terms of the Gini coefficient as opposed to the reverse. Or maybe it's just a You know, maybe I got it wrong, so please clarify that. Thanks. Yeah, very good question. Uh, I was just waiting to see if there is, uh, one more question from the audience, and will that, with that will conclude the questions. OK. I think uh we are good from the audience, so over to you. Yeah, uh, this, uh, from, so, uh let me separate the, the, uh, policy implication versus in the model. In the model, uh, in the empirics, the polarization is measured as ideological difference. So in a country, the largest 3 groups was their ideology, was their seat in the government to capture this ideological or political polarization. What I mean by increasing income inequality, there's actually a recent growing literature to try to understand whether the increasing income. Equality also lead to ideological polarization. And I think I see this in working in the countries is since income inequality is so wide, only the, uh, the elites are in the government and they have a very different preference in spending decision than other groups. So, if they are more. Likely to be in making decision, they might prefer different spending or borrowing decision than if the inequality is less. So, so that's where the policy implication come from. But this, this is not in the paper per se. It's more on the reflection of how this paper, the theoretical model can be applied in the current context. Thank you. And with this, let's conclude. Thank you so much for your participation and many thanks to. Thank you so much, Manuela. Uh, we're taking a 5-minute break, right. Greater than the incoherence is the challenge of keeping us on time. So please be back in 5 minutes. There is coffee outside. That
showAllTranscripts
no
duration
PT1H17M10S
scene7File
worldbank/ABCDE2024-Session-2
scene7Domain
https://worldbank.scene7.com/
scene7FileAvs
worldbank/ABCDE2024-Session-2-AVS
title
ABCDE2024 Session 2
description
ABCDE2024 Session 2
showTimestampAndTranscript
yes
col-xs-12
col-sm-12
col-md-3
col-lg-3
col-xs-12
col-sm-12
col-md-7
col-lg-7
  • add-style
  • lp-body-content

A video recording of the second session—Day 1—of The Annual Bank Conference on Development Economics 2024 "The Great Incoherence: Growth and Human Development in An Era of Stagnation." This session discusses "Sovereign Debt and Default."

Papers discussed in this session are:

lp-heading-top-medium
lp-heading-bottom-medium
col-xs-12
col-sm-12
col-md-2
col-lg-2