Popular in Professional

Culture and Society

The City as a Growth Machine: Toward a Political Economy of Place

New York University Harvey Moltoch

Education

The Role of Education in Reducing Racial Inequality

University of Southern California Pedro A. Noguera

Law

Diversity and Social Mobility in the UK legal profession

Law Society I. Stephanie Boyce

Internet Law

The Law of AI for Good

University of San Diego Orly Lobel

Pharmacy

Cytotoxic Activity and Anti-Migratory Effect of Berberine-Phytantriol Liquid Crystalline on Non-Small-Cell Lung Cancer

University of Technology Sydney Kamal Dua

Human Rights Law

Transforming Restorative Justice for Transitional Settings

University of Nottingham Kerry Clamp

Education

Pledgeability and Asset Prices

View more insights in Professional - Finance

View more figures from Stanford University

View the page of Zhiguo He

Read the Study

Zhiguo He discusses causal evidence for the value of asset pledgeability

Image courtesy of interviewee. January 26, 2024

Generating Workspace

Export as PDF

This video video discusses the concept of asset eligibility and its importance in financial markets, particularly in the context of borrowing against assets. It highlights how one asset may provide borrowing capabilities for investors while another similar asset may not. The author references research, particularly a study by Gallo and Peterson that explores price differences in assets based on their borrowing capabilities, termed "haircut."

The speaker presents a method to assess the impacts of changes in asset eligibility using a case study from the Chinese bond market, where they examine dual-listed corporate bonds traded in different markets with varying regulations. They analyze a specific policy change that restricted borrowing against certain bonds based on their credit ratings and assess how this change affected bond prices.

Using a difference-in-difference approach, the author explains how to quantify the effects of the policy shock on asset prices and determines that a rise in the haircut (or reduction in eligibility) from zero to one resulted in a significant drop in bond premiums. The results indicate a 30 to 40 basis point decrease in bond prices due to the loss of eligibility, providing insight into the cost of capital in China during that time.

Overall, the text emphasizes the importance of understanding asset eligibility and its implications for financial markets, as well as the need for further research in the rapidly developing Chinese bond market.

Language:

Export as PDF

X-TIMESTAMP-MAP:LOCAL=00:00:00.000,MPEGTS=900000

00:00.095 --> 00:03.485
Asset eligibility start from the long, long, long,

00:03.485 --> 00:07.485
long time ago that economists realized that, one asset,

00:07.985 --> 00:10.405
is useful, not just

00:10.405 --> 00:13.045
because it's cash flows that are brought to the investors,

00:13.505 --> 00:16.205
but also the investors could use this asset

00:16.205 --> 00:17.485
as a collateral to borrow.

00:17.905 --> 00:21.325
Simple as that. Think about the housing. Okay.

00:21.905 --> 00:26.565
Um, then there's a series of the work

00:26.675 --> 00:30.645
that has been, brought into the economic,

00:31.485 --> 00:34.965
research and, and has a huge impact on the way we

00:34.965 --> 00:38.405
understand the financial markets start from,

00:39.125 --> 00:43.765
Kentucky Moore, famous paper on the way that it

00:44.275 --> 00:45.845
explained the financial crisis

00:46.465 --> 00:48.285
and the banking crisis, all these things.

00:48.585 --> 00:52.045
And then in the financial, research is

00:52.045 --> 00:54.085
that it's always a kind of,

00:54.465 --> 00:58.525
you could talk about it like a holy grail to really try

00:58.525 --> 01:02.245
to identify how, how big is it, is the fact that,

01:02.245 --> 01:04.445
that if you have the same assets,

01:05.185 --> 01:08.805
one asset does not have the ability for the investors

01:08.865 --> 01:12.365
to borrow, but the other asset has a d ability

01:13.225 --> 01:15.885
to allow the investors to borrow how much

01:15.885 --> 01:18.725
that investors actually value This additional feature,

01:19.275 --> 01:22.285
this paper is like using some certain

01:23.085 --> 01:25.965
specific features in Chinese bond market.

01:26.145 --> 01:29.995
To answer a broad question, which was,

01:30.275 --> 01:32.635
I pointed out the asset eligibility.

01:32.695 --> 01:37.275
So in the, as a legibility side of the literature,

01:37.815 --> 01:40.955
the, the most close, framework

01:40.985 --> 01:42.835
that we are thinking about is this.

01:43.245 --> 01:48.195
Gallo Peterson, 20 12, 20 13 review

01:48.195 --> 01:49.395
of financial studies paper.

01:49.945 --> 01:53.795
What they show is that indeed they, in their model,

01:53.945 --> 01:55.515
they have, two assets.

01:56.175 --> 01:58.235
Hypothetically, it's a theoretical model.

01:59.015 --> 02:00.995
One asset does have the ability

02:01.095 --> 02:02.515
to allow the investors to borrow.

02:03.365 --> 02:05.305
The other is not. So they call this

02:05.405 --> 02:06.865
called as a margin ability.

02:07.015 --> 02:11.465
Basically, whether the investors are able to use this asset

02:11.525 --> 02:14.025
to raise margin, margin is another way

02:14.045 --> 02:15.665
to say borrowing basically.

02:16.325 --> 02:20.265
Um, so, so in that paper, that's a theoretical construction

02:20.325 --> 02:22.905
and that they show that in equilibrium.

02:23.335 --> 02:27.545
When investors, given all these asset prices, determine the,

02:27.965 --> 02:32.345
endogenously, what will be the theoretical

02:33.255 --> 02:35.265
implication of the price difference

02:35.265 --> 02:36.665
between these two A assets?

02:36.935 --> 02:39.585
This kinda depends on the importantly called haircut.

02:40.775 --> 02:41.785
What does it mean by haircut?

02:42.175 --> 02:46.305
Haircut means exactly the ability for the investors

02:46.365 --> 02:48.545
to use this assets to borrow.

02:49.085 --> 02:52.225
And that's basically another way to say legibility.

02:52.365 --> 02:54.105
We use the word p legibility.

02:54.645 --> 02:58.345
Um, some existing literature use the word of the haircut.

02:59.075 --> 03:02.005
It's basically the same thing when haircut is one.

03:02.315 --> 03:04.485
That means that you cannot borrow anything

03:04.485 --> 03:08.165
because when you brought your assets to the borrower,

03:08.325 --> 03:10.765
borrower will say, I will take a haircut over one,

03:10.815 --> 03:12.365
which means 100% haircut.

03:12.675 --> 03:13.925
Take this asset to lender.

03:14.185 --> 03:16.885
Lender will say the haircut is 100%,

03:16.935 --> 03:19.645
which means the lender refuse to lend you anything.

03:19.815 --> 03:22.925
Right? When you say haircut is a zero, then lender is happy

03:23.265 --> 03:25.685
to take it as a face value and lend you a lot.

03:26.105 --> 03:30.685
So what we are doing basically is that taking this framework

03:30.785 --> 03:32.965
and ask the question that in the China context,

03:33.655 --> 03:36.765
using some institutional features that I will explain later

03:37.075 --> 03:40.485
that what if a asset, the haircut rate

03:41.015 --> 03:44.965
rises from a zero of one rises from zero to one.

03:44.985 --> 03:49.685
So legibility go down, how much is the asset price drop

03:49.795 --> 03:53.365
that that corresponded provide certain estimates?

03:53.785 --> 03:55.965
So to the, so-called legibility

03:56.825 --> 04:00.335
is really the something called endogeneity that is

04:01.515 --> 04:04.535
ev pervasive in any economic re research.

04:06.435 --> 04:07.965
What does it mean? This means

04:07.965 --> 04:10.565
that when you see a two assets in the real world,

04:11.265 --> 04:13.205
one has a higher haircut

04:13.585 --> 04:15.365
or lower haircut than the other,

04:16.275 --> 04:18.295
the haircut itself is endogenous

04:18.515 --> 04:21.455
and oftentimes that a haircut is related is

04:21.655 --> 04:22.895
determined by the fundamental.

04:23.235 --> 04:25.815
So as a result, when you compare two assets,

04:25.965 --> 04:29.735
very similar look, similar observationally, similar we call

04:29.735 --> 04:31.495
that, but the fact

04:31.495 --> 04:33.935
that they have different haircuts must

04:33.965 --> 04:35.095
mean something different.

04:35.715 --> 04:40.175
The best example in the, in the, in the existing literature,

04:40.515 --> 04:42.135
the in empirical literature that they try

04:42.135 --> 04:46.455
to answer these questions is take the European market

04:46.945 --> 04:49.735
where the policy makers basically

04:51.485 --> 04:55.005
pick certain bonds to be eligible

04:55.105 --> 04:57.045
for the borrowing from the central bank.

04:57.935 --> 05:01.035
But are you gonna worry that why these essential banks pick

05:01.045 --> 05:02.075
these bonds?

05:02.215 --> 05:03.875
Not the order. It must be

05:03.875 --> 05:05.395
that there's something different, right?

05:06.055 --> 05:09.275
So therefore that when we compare these price differential,

05:09.335 --> 05:12.515
we do not know whether this price is reacting

05:12.515 --> 05:13.715
to the haircut itself

05:14.175 --> 05:16.835
or something else that underlying to drive the haircut,

05:17.025 --> 05:21.485
haircut difference in China, fundamentally,

05:21.515 --> 05:25.965
that we are taking advantage of one institutional de detail,

05:25.965 --> 05:29.245
which is that we have so-called dual listed

05:29.705 --> 05:30.725
co corporate bonds.

05:32.205 --> 05:34.375
Once you have a dual listed, which means

05:34.375 --> 05:38.575
that exactly same bonds are traded in different markets

05:39.665 --> 05:41.525
as a result, by definition,

05:41.735 --> 05:44.205
these two bonds, they're the same.

05:44.955 --> 05:48.735
So as a result, you're gonna think about, okay, the worry

05:48.965 --> 05:50.975
that in other context is

05:50.975 --> 05:53.895
because of the fundamental difference, the qua, the quality

05:53.955 --> 05:56.455
of bonds, the recovery bonds, all these things

05:57.035 --> 05:58.045
That drives the haircut.

05:58.345 --> 06:00.525
That's not happening in China.

06:01.305 --> 06:04.485
And in China we have these two, two markets.

06:05.465 --> 06:08.525
The one two markets have the same bonds, so

06:08.525 --> 06:10.285
therefore the fundamental is the same,

06:10.745 --> 06:13.645
but it's also fortunate for us to study is

06:13.645 --> 06:17.325
because these two markets have different report rules.

06:17.745 --> 06:19.965
So as a result, we do have different

06:19.965 --> 06:21.725
haircuts on the same thing.

06:22.515 --> 06:24.055
But voila, you could do this,

06:24.235 --> 06:27.895
and then basically you need to be careful in looking for

06:28.435 --> 06:32.175
the policy shocks that are shifting one

06:32.955 --> 06:35.575
report rules, but then not the other.

06:36.335 --> 06:37.735
I already give you the background.

06:37.805 --> 06:39.815
Basically we have, two markets.

06:40.765 --> 06:44.385
The one markets, it's so-called exchange markets in China.

06:44.565 --> 06:46.425
Oh, by the way, the, the, the,

06:46.425 --> 06:50.865
corporate bonds are second, in the world only,

06:51.255 --> 06:54.045
only, only after United States.

06:54.105 --> 06:56.125
United States is a gigantic bond market.

06:56.155 --> 07:00.085
Obviously everybody going there to to, to save their,

07:00.335 --> 07:01.645
their, their money there.

07:01.945 --> 07:04.725
Uh, China's big just because it's a little bit close,

07:05.165 --> 07:07.245
although that there's a lot of investors are able

07:07.245 --> 07:08.685
to get access to through.

07:08.795 --> 07:11.725
So-called Connect. I have some research on that.

07:11.985 --> 07:16.085
So it's a big, there's a lot of, a lot of institutions,

07:16.355 --> 07:17.805
that are participating there.

07:18.385 --> 07:20.205
Um, and, and,

07:20.345 --> 07:23.285
and as a result that it's actually quite a mature.

07:23.905 --> 07:27.645
Um, so, so little bit history was

07:27.645 --> 07:30.605
that basically it started in the nineties

07:30.945 --> 07:33.645
and then, then, then in the, in the late nineties,

07:33.645 --> 07:34.765
there's a bigger reform

07:35.225 --> 07:37.885
and for historical reasons, the two markets,

07:38.425 --> 07:40.005
one market's called exchange markets,

07:40.215 --> 07:44.125
where you saw all these, exciting news about like stock

07:44.765 --> 07:46.725
movement, crash, blah, blah, blah, blah.

07:46.755 --> 07:50.465
It's there, and not only you can trade the stocks,

07:50.525 --> 07:53.065
but also bonds, corporate bonds, okay?

07:53.405 --> 07:55.665
The other is called interbank market.

07:55.965 --> 07:58.585
So these two markets is running

07:58.765 --> 08:00.825
as in a interestingly different way.

08:01.305 --> 08:03.785
Exchange markets is basically a gigantic

08:03.895 --> 08:05.385
essential limit order.

08:06.645 --> 08:08.665
It, and everybody,

08:09.215 --> 08:11.505
it's a central limit order, system.

08:11.805 --> 08:15.505
And everyone going there anonymously put the A bids

08:15.525 --> 08:19.665
and then they match the bid a, bid ask, prices.

08:20.485 --> 08:23.345
And there is basically that. It's the no longer.

08:23.495 --> 08:26.945
So-called search based OTC market, right?

08:26.945 --> 08:29.225
It's just you go there, place the bid.

08:29.325 --> 08:32.265
If you bid at other people's price, why you got it.

08:32.725 --> 08:35.065
The other market is called the OTC market

08:35.205 --> 08:38.785
or the interbank market where the people meet

08:38.965 --> 08:40.905
and just negotiate on prices.

08:41.285 --> 08:44.665
So that's the, the way you trade the bonds.

08:45.265 --> 08:46.865
I have to explain the traded the bonds

08:46.865 --> 08:49.705
before explaining the re re report rules.

08:49.935 --> 08:51.705
Okay? So now I'm mo I'm moving on

08:51.705 --> 08:54.265
to the report rule basically is the ability for you

08:54.265 --> 08:56.645
to borrow in the exchange of markets.

08:56.905 --> 08:59.885
The ability to borrow is interestingly determined

08:59.985 --> 09:01.325
by the CCP.

09:01.325 --> 09:03.365
That's the central clearing party,

09:03.575 --> 09:05.285
which is basically the exchange.

09:05.705 --> 09:09.125
So the exchange every day they will decide what, how much

09:09.125 --> 09:12.485
that you are able to bring your bonds

09:13.145 --> 09:14.285
to the exchange

09:15.285 --> 09:17.305
and how much you can borrow from the exchange.

09:18.525 --> 09:22.175
Okay? Exchange just decided certain things based on some

09:22.175 --> 09:23.735
things, some, some rules

09:23.915 --> 09:27.215
or the things while on the, in interbank,

09:27.715 --> 09:30.815
you could borrow from each other through negotiation.

09:31.115 --> 09:33.095
So that's very, very much bilateral.

09:33.835 --> 09:36.335
Um, okay, so you have this centralized system.

09:36.435 --> 09:39.855
You have very, very, very,

09:40.565 --> 09:43.575
decentralized system on the, on the interbank market.

09:43.575 --> 09:45.455
Centralized is the exchange. Okay?

09:45.875 --> 09:49.495
The policy shock we are exploiting is 24 is a

09:49.735 --> 09:54.615
December 8th, 2014, after Monday night.

09:54.885 --> 09:56.135
Usually they don't do it.

09:56.135 --> 09:58.535
Usually we do it like Friday night. Why?

09:58.645 --> 10:01.215
Because Friday night everybody go home.

10:01.485 --> 10:04.135
They have two days to absorb the shock.

10:05.075 --> 10:09.735
Um, at that time, there's a, some, some, some concern about,

10:10.155 --> 10:13.015
about a local government debt and all these things.

10:13.475 --> 10:17.775
And the exchange market is not like the biggest market

10:18.075 --> 10:20.135
for these bonds, okay?

10:20.435 --> 10:23.775
But, but they try to show, you know,

10:23.775 --> 10:24.895
allegiance to Beijing.

10:25.635 --> 10:29.935
Um, they, on the Monday night

10:31.355 --> 10:32.815
to everybody is a surprise.

10:33.605 --> 10:37.215
They basically slashed the p legibility of certain bonds

10:37.955 --> 10:41.215
and that these certain bonds has a very easy

10:41.315 --> 10:42.815
to implement rule rating.

10:43.905 --> 10:48.575
China likes just using easy to implement the rules because

10:48.575 --> 10:51.415
otherwise it's just, you always have negotiation.

10:52.195 --> 10:54.295
Is this, you know, you are supposed

10:54.295 --> 10:55.775
to be slashed or not supposed?

10:55.835 --> 10:58.975
No, no, no, no. That's just the user RA rating.

10:59.595 --> 11:03.295
And what happened was that, the, the, the, the, the,

11:03.795 --> 11:06.735
aa, you know, in China mainly there are two ratings

11:06.835 --> 11:08.935
or three four ratings that we study.

11:10.075 --> 11:12.095
aaa, AA plus aa,

11:12.195 --> 11:16.295
and AA minus what they, basically, the rule says that

11:16.715 --> 11:21.335
as long as your bonds is below aaa, starting from today,

11:22.415 --> 11:27.115
you are not allowed to use these bonds anymore

11:27.415 --> 11:28.955
to borrow from the exchange.

11:30.065 --> 11:33.675
Okay? And on the other side, no policy

11:34.335 --> 11:35.755
coming out just

11:35.755 --> 11:38.595
because that they are just managed by different,

11:39.295 --> 11:41.035
agencies.

11:41.465 --> 11:46.035
Okay? So, okay, so that's a very nice policy shock

11:46.035 --> 11:49.715
because as I explained, the fundamental are always the same,

11:50.215 --> 11:54.445
but I hit the, the p legibility rule

11:54.705 --> 11:56.565
or ripple rule on one market.

11:56.785 --> 11:59.405
So you'll see that if there's anything price

11:59.805 --> 12:01.055
movement, right?

12:01.435 --> 12:05.335
No, I take the difference then, then the, the change

12:05.515 --> 12:06.935
of the difference before and

12:06.935 --> 12:10.815
after should I identify the effect

12:10.995 --> 12:13.095
of the legibility change, okay?

12:13.115 --> 12:16.975
So that's so-called diff diff in the typical language

12:17.075 --> 12:18.855
of academia,

12:18.855 --> 12:20.455
because what we are doing is

12:20.455 --> 12:22.495
that across the market first, right?

12:22.495 --> 12:25.255
That's diff and then over time of the policy shock,

12:25.645 --> 12:28.695
diff diff, there's another le level

12:28.755 --> 12:31.615
of diff is quite interesting, very, very unique

12:31.795 --> 12:32.815
to this setting,

12:33.075 --> 12:37.415
and it is really, you know, make it,

12:37.835 --> 12:41.095
the results become extremely convincible is the fact

12:41.285 --> 12:44.735
that when you do this, policy, remember

12:44.735 --> 12:49.375
that it's basically true below triple A bonds was slashed

12:49.835 --> 12:51.175
of the P legibility.

12:51.365 --> 12:52.415
Okay? What does that mean?

12:53.295 --> 12:55.175
Interestingly, in before the, even

12:55.175 --> 12:56.495
before the policy shock,

12:56.885 --> 12:59.735
basically you have a four ratings explained,

13:00.275 --> 13:02.815
and the AA minus the worst bonds,

13:03.765 --> 13:06.015
investors cannot use them to pledge.

13:06.355 --> 13:08.455
And anyway, so for them,

13:09.005 --> 13:13.365
basically their legibility is zero, even before the shock.

13:14.025 --> 13:18.045
And you have AA minus, and AA has some legibility,

13:18.705 --> 13:21.445
and you have, AAA has lots of legibility.

13:21.585 --> 13:25.445
So 90%, 70%, probably 20

13:25.585 --> 13:26.885
to 30% of them.

13:26.885 --> 13:30.725
The zero, once you shock the policy shock hits,

13:30.755 --> 13:32.925
basically you only hit the middle two.

13:33.665 --> 13:37.965
Why? Because the middle two previously used to be able

13:37.965 --> 13:40.045
to borrow 30% or 70%.

13:40.385 --> 13:42.325
Now you cannot borrow anymore zero,

13:43.105 --> 13:47.405
but the top one, aaa, you're still able to borrow 90%

13:47.405 --> 13:51.405
or 90%, and the lowest part, which was

13:51.865 --> 13:53.685
before also started with zero.

13:54.425 --> 13:55.765
Now it's also zero.

13:56.225 --> 13:59.485
So as a result that this policy shock,

13:59.545 --> 14:03.385
you have a cross additional cross-sectional D

14:04.775 --> 14:07.335
impacted that we can take advantage, okay?

14:07.645 --> 14:11.935
This is the super important super, super important. Why?

14:11.935 --> 14:15.855
Because most of the time we worry that when shock hits,

14:16.325 --> 14:18.375
there's that shock is endogenous

14:18.435 --> 14:22.775
and typically will be related to certain fundamental issues.

14:23.725 --> 14:25.735
Usually, usually, you know, our,

14:26.155 --> 14:30.775
our policy is obviously also subject to, this kind

14:30.775 --> 14:34.735
of concern because that Beijing is always trying

14:34.735 --> 14:36.055
to address certain things, right?

14:36.395 --> 14:40.095
Always. There's a no no point of doing random policy

14:41.355 --> 14:42.415
for such a big scale.

14:43.625 --> 14:44.885
But why this is helps,

14:45.035 --> 14:47.925
because most of the time when these kind of shark hits,

14:48.195 --> 14:51.285
usually you have this, so-called beta hypothesis,

14:51.615 --> 14:54.765
which means that if you are safer, then you subject

14:54.825 --> 14:57.045
to this kind of concern, less aaa,

14:57.355 --> 15:01.645
then aa plus AA minus a AA minus will be subject

15:01.665 --> 15:02.685
to even more, right?

15:02.685 --> 15:05.165
Because usually you just think about the beta

15:05.165 --> 15:08.085
because they have a, increasing, beta.

15:08.545 --> 15:09.925
If you have a fundamental sharks,

15:09.925 --> 15:11.125
they should have reacted to it more.

15:11.585 --> 15:15.005
But now we have basically have the lowest beta

15:15.005 --> 15:19.885
and highest beta do not change at all from the,

15:19.955 --> 15:22.445
from the, looking at the data,

15:22.585 --> 15:25.045
but only the middle two really getting hit

15:26.115 --> 15:27.885
that just supported the hypothesis

15:27.885 --> 15:31.485
that this policy shark only hits the p legibility.

15:32.115 --> 15:34.165
Once you realize this is setting,

15:34.195 --> 15:38.245
basically we have a triple diff, then it's quite easy

15:38.465 --> 15:41.765
to do the estimation, do the measurement basically.

15:42.065 --> 15:44.165
So what you're gonna do is take the differential.

15:44.185 --> 15:45.685
So that's already three

15:45.745 --> 15:48.805
of the fundamental before the policy.

15:49.155 --> 15:53.435
What we show was that, in the, in,

15:53.975 --> 15:56.635
in the broader corporate bonds market in China,

15:57.495 --> 16:00.075
the differential we call exchange premium.

16:00.875 --> 16:05.625
Exchange premium, okay, is highly related

16:06.005 --> 16:07.905
to the haircuts itself.

16:09.225 --> 16:12.395
Okay? So that supported the idea that in this type

16:12.395 --> 16:13.835
of market, the fact

16:13.835 --> 16:16.995
that you have price differential is purely due to,

16:17.535 --> 16:19.875
or largely due to at 90%.

16:19.975 --> 16:22.915
We have the, some, some, some, some estimation on

16:22.935 --> 16:25.315
how much they can explain the,

16:25.575 --> 16:28.075
the price differential is due to this haircut.

16:28.975 --> 16:32.835
So that, that's very assuring that, the, the, the,

16:32.855 --> 16:34.915
the differential is due to the legibility.

16:35.985 --> 16:37.435
Then what we could do is

16:37.485 --> 16:39.715
after the policy shock, we basically,

16:39.815 --> 16:41.995
you can even do just calculation without doing,

16:42.165 --> 16:44.395
doing the regression, right?

16:44.705 --> 16:47.275
Basically ask your question that how much

16:47.275 --> 16:48.475
that haircutter drops

16:49.375 --> 16:52.675
or rise, you know, plausibility drop, haircutter rise,

16:53.325 --> 16:56.155
let's say started from a 50% initially.

16:56.615 --> 16:59.715
Now you basically haircutter become one policy hits,

16:59.945 --> 17:02.355
then the plausibility go down by 50%.

17:03.545 --> 17:08.045
You look at the price, price movement is something, okay?

17:08.895 --> 17:13.315
And therefore that we can gauge the sensitivity, okay?

17:13.735 --> 17:18.555
And, not only one class of bonds, we also have the,

17:18.615 --> 17:20.835
the other class of A bonds, which is the AA minus.

17:21.225 --> 17:22.675
Then you do the differential

17:23.175 --> 17:25.475
and all these things when we do is,

17:25.645 --> 17:27.675
using all these dif method.

17:28.055 --> 17:31.355
So the way you do it is ba basically adding all these fixed

17:31.355 --> 17:35.865
effects, I know time fixed effects and bond fixed effect

17:35.885 --> 17:36.985
or rating fixed effects,

17:37.445 --> 17:40.385
but, the way you identify it is basically rating time

17:40.755 --> 17:41.945
times post shock.

17:42.405 --> 17:44.585
So that interaction basically picks up

17:45.285 --> 17:47.785
the same bond differential, but

17:47.835 --> 17:51.405
after the shock somehow changed it due to the policy, due

17:51.405 --> 17:52.485
to the haircutter change.

17:52.985 --> 17:56.485
The results coming out from this exercise, by the way,

17:56.485 --> 18:01.085
that we take the, six months before, six months

18:01.085 --> 18:04.085
after, because sometimes, you know, price movement take,

18:04.235 --> 18:05.325
take a little bit of time.

18:05.745 --> 18:07.645
We choose six month is

18:07.645 --> 18:11.685
because that, we, this is a, interesting period

18:11.935 --> 18:15.125
where most of the po most

18:15.125 --> 18:19.285
of the policy discussion is at, the real economy level,

18:19.385 --> 18:21.165
but it doesn't hit the,

18:21.165 --> 18:22.805
financial market yet at that time.

18:23.225 --> 18:26.445
And, as a result that that's, we think that at the,

18:26.745 --> 18:30.165
at the time, basically time fixed effect can, can,

18:30.225 --> 18:32.765
can absorb all these common shocks that are happening

18:32.905 --> 18:33.925
to the financial market.

18:34.485 --> 18:36.565
Remember, we have two markets, so this is good.

18:37.025 --> 18:42.005
Um, once you do that, you find out that, the,

18:42.305 --> 18:45.485
the, if the haircut, so here's a quantitative answer,

18:45.585 --> 18:49.605
if the haircut, rises from zero to 100%, which means

18:49.605 --> 18:53.725
that flexibility slashed from a one to zero, the

18:54.315 --> 18:58.205
bond premium goes, goes,

18:58.985 --> 19:03.285
so this is a from zero, flexibility from, one

19:03.785 --> 19:05.085
to zero.

19:05.905 --> 19:10.005
The, the premium goes down by 35%, 30

19:10.105 --> 19:11.765
to 40 basis point.

19:12.105 --> 19:16.885
And, to, to us that's, you know, we, once we get this,

19:17.025 --> 19:19.905
we can compare to other, other studies

19:20.605 --> 19:24.065
and the number is actually smaller than the usual ones.

19:25.245 --> 19:27.305
And we think that this is a very natural

19:28.135 --> 19:30.265
because that, in the previous studies,

19:30.565 --> 19:35.225
you are always being contaminated by something else, okay?

19:35.405 --> 19:37.825
So as a result that we are very happy that,

19:38.045 --> 19:42.625
the number is not like penny,

19:43.165 --> 19:45.785
so that you don't mind,

19:46.005 --> 19:49.545
but also like smaller than the existing, numbers.

19:49.805 --> 19:53.945
And this in the paper we also mentioned as academia,

19:54.045 --> 19:55.345
we need to be super careful.

19:55.375 --> 19:56.425
Like what does it mean?

19:57.095 --> 20:00.505
This is 35% in, you know, in a, in a framework.

20:01.045 --> 20:03.705
We also realized that also that we,

20:03.735 --> 20:06.505
from the very beginning when we designed the research,

20:06.825 --> 20:10.025
I knew that, hypothetic compared

20:10.025 --> 20:13.225
to a setting hypothetically where the,

20:13.225 --> 20:18.025
investors are not allowed to move bonds between two markets.

20:19.055 --> 20:22.905
Okay? You know, we have a market, so we are actually,

20:22.905 --> 20:25.465
you know, I didn't mention this detail, which means that,

20:25.465 --> 20:29.305
the investors are able to allow to move between bonds

20:30.505 --> 20:33.505
relative to a set setting that they are not allowed to.

20:34.965 --> 20:37.575
It's easy to show that,

20:37.765 --> 20:40.815
because of this, option to move around

20:41.165 --> 20:42.735
that would depress the press

20:42.735 --> 20:45.195
differential, press differential.

20:45.255 --> 20:49.645
As a result, the number we identified is indeed a little bit

20:49.645 --> 20:51.565
of a under bias, okay?

20:51.865 --> 20:55.805
So, which means that the 30 to 40 basis point differential

20:56.425 --> 21:00.485
is basically is a unbiased, so, so this is the,

21:00.485 --> 21:03.725
in the paper we made a very clear explain why or the things,

21:04.245 --> 21:07.885
although that we do not think it's such a big deal

21:08.105 --> 21:10.205
for different reasons, but unfortunately we

21:10.205 --> 21:11.485
cannot quantify that part.

21:11.945 --> 21:13.565
So as a result, we,

21:14.105 --> 21:17.005
we also provide a overestimation.

21:18.295 --> 21:20.195
So for the over estimation,

21:20.425 --> 21:22.595
basically we take the same setting,

21:23.055 --> 21:24.995
but they use a similar methodology

21:25.815 --> 21:27.115
in the existing literature,

21:29.135 --> 21:31.115
do not use the match upon.

21:31.465 --> 21:36.075
Okay? As a result, we can show that.

21:36.255 --> 21:39.755
So basically what we did was basically saying that what if

21:39.755 --> 21:44.355
that they, we take the bonds there, look, look identical,

21:45.785 --> 21:49.715
look identical, and when you say look identical, okay,

21:50.135 --> 21:51.515
you know, look at their price,

21:51.625 --> 21:53.675
look at their yield, blah, blah, blah, blah, blah.

21:53.785 --> 21:56.075
Then take the differential, okay?

21:56.585 --> 21:59.635
Then we'll see that it's basically the number coming out is

21:59.635 --> 22:00.995
about an 85 basis point.

22:01.655 --> 22:05.795
So as a result that we can, you know, happily conclude that,

22:06.015 --> 22:08.155
you know, we have underestimates, which,

22:08.245 --> 22:10.475
which is pretty tight, like 30, 30

22:10.535 --> 22:12.595
to 40 basis point, exactly number.

22:12.675 --> 22:15.515
I forgot. Then the other side is about 80 basis point.

22:15.655 --> 22:20.155
So we have a pretty good sandwich to give a, a gauge of

22:20.255 --> 22:22.195
how big is this flexibility is

22:23.135 --> 22:27.475
one implication is on the financial market side, is that

22:27.695 --> 22:31.595
to basically give a sense that how big is

22:32.135 --> 22:36.225
the cost of a capital, in China during

22:36.225 --> 22:37.785
that period, okay?

22:38.165 --> 22:42.385
And we found that, that this number is very, very,

22:43.445 --> 22:47.185
in accordance, with, with the other method

22:47.375 --> 22:51.945
that we can use, for instance, the liquidity effect, to,

22:52.285 --> 22:55.105
during that time in 2019.

22:55.405 --> 22:59.505
So that's basically, it's kind of, just give a sense

22:59.925 --> 23:03.105
of the entire financial, financial,

23:03.105 --> 23:05.705
financial research community on the first time

23:05.705 --> 23:07.705
that we are able to tell them

23:07.735 --> 23:09.425
that this is gonna be that big.

23:09.895 --> 23:13.705
Okay? And the second one is, is more

23:13.705 --> 23:16.305
or less like, from my own,

23:17.465 --> 23:19.505
I wouldn't say my own agenda,

23:19.685 --> 23:23.145
but I think that I have the motivation to

23:24.275 --> 23:27.895
introduce the smart scholars

23:28.795 --> 23:31.855
to look into the Chinese bond markets.

23:32.725 --> 23:35.935
It's just so much of interesting

23:36.525 --> 23:37.935
institutional details.

23:38.795 --> 23:40.735
So much effort has been done

23:41.235 --> 23:45.845
by the policy makers in China in develop the,

23:46.105 --> 23:49.005
corporate bonds market in the past, decade.

23:49.505 --> 23:53.165
And they are, they, they were doing very, very well in the,

23:53.465 --> 23:55.525
you know, from, I would say from five

23:56.065 --> 23:58.645
to probably nine 19, right before the Covid.

23:59.345 --> 24:02.285
At this point. They're facing lots of, challenges,

24:02.465 --> 24:04.885
but it's really just, you know, for,

24:04.945 --> 24:07.165
for obvious the other parts of the reasons.

24:07.625 --> 24:08.925
But I wish that, more

24:08.925 --> 24:11.245
and more researchers who can look into these markets.

00:00 - Asset eligibility start from the long, long, long, long time ago that economists realized that, one asset, is useful, not just because it's cash flows that are brought to the investors, but also the investors could use this asset as a collateral to borrow.

00:17 - Simple as that. Think about the housing. Okay. then there's a series of the work that has been, brought into the economic, research and, and has a huge impact on the way we understand the financial markets start from, Kentucky Moore, famous paper on the way that it explained the financial crisis and the banking crisis, all these things. And then in the financial, research is that it's always a kind of,

00:54 - you could talk about it like a holy grail to really try to identify how, how big is it, is the fact that, that if you have the same assets, one asset does not have the ability for the investors to borrow, but the other asset has a d ability to allow the investors to borrow how much that investors actually value This additional feature, this paper is like using some certain specific features in Chinese bond market. To answer a broad question, which was,

01:30 - I pointed out the asset eligibility. So in the, as a legibility side of the literature, the, the most close, framework that we are thinking about is this. Gallo Peterson, 20 12, 20 13 review of financial studies paper. What they show is that indeed they, in their model, they have, two assets. Hypothetically, it's a theoretical model. One asset does have the ability

02:01 - to allow the investors to borrow. The other is not. So they call this called as a margin ability. Basically, whether the investors are able to use this asset to raise margin, margin is another way to say borrowing basically. so, so in that paper, that's a theoretical construction and that they show that in equilibrium. When investors, given all these asset prices, determine the, endogenously, what will be the theoretical

02:33 - implication of the price difference between these two A assets? This kinda depends on the importantly called haircut. What does it mean by haircut? Haircut means exactly the ability for the investors to use this assets to borrow. And that's basically another way to say legibility. We use the word p legibility. some existing literature use the word of the haircut. It's basically the same thing when haircut is one.

03:02 - That means that you cannot borrow anything because when you brought your assets to the borrower, borrower will say, I will take a haircut over one, which means 100% haircut. Take this asset to lender. Lender will say the haircut is 100%, which means the lender refuse to lend you anything. Right? When you say haircut is a zero, then lender is happy to take it as a face value and lend you a lot. So what we are doing basically is that taking this framework

03:30 - and ask the question that in the China context, using some institutional features that I will explain later that what if a asset, the haircut rate rises from a zero of one rises from zero to one. So legibility go down, how much is the asset price drop that that corresponded provide certain estimates? So to the, so-called legibility is really the something called endogeneity that is ev pervasive in any economic re research. What does it mean? This means

04:07 - that when you see a two assets in the real world, one has a higher haircut or lower haircut than the other, the haircut itself is endogenous and oftentimes that a haircut is related is determined by the fundamental. So as a result, when you compare two assets, very similar look, similar observationally, similar we call that, but the fact that they have different haircuts must

04:33 - mean something different. The best example in the, in the, in the existing literature, the in empirical literature that they try to answer these questions is take the European market where the policy makers basically pick certain bonds to be eligible for the borrowing from the central bank. But are you gonna worry that why these essential banks pick these bonds? Not the order. It must be

05:03 - that there's something different, right? So therefore that when we compare these price differential, we do not know whether this price is reacting to the haircut itself or something else that underlying to drive the haircut, haircut difference in China, fundamentally, that we are taking advantage of one institutional de detail, which is that we have so-called dual listed co corporate bonds. Once you have a dual listed, which means

05:34 - that exactly same bonds are traded in different markets as a result, by definition, these two bonds, they're the same. So as a result, you're gonna think about, okay, the worry that in other context is because of the fundamental difference, the qua, the quality of bonds, the recovery bonds, all these things That drives the haircut. That's not happening in China. And in China we have these two, two markets.

06:05 - The one two markets have the same bonds, so therefore the fundamental is the same, but it's also fortunate for us to study is because these two markets have different report rules. So as a result, we do have different haircuts on the same thing. But voila, you could do this, and then basically you need to be careful in looking for the policy shocks that are shifting one report rules, but then not the other.

06:36 - I already give you the background. Basically we have, two markets. The one markets, it's so-called exchange markets in China. by the way, the, the, the, corporate bonds are second, in the world only, only, only after United States. United States is a gigantic bond market. Obviously everybody going there to to, to save their, their, their money there. China's big just because it's a little bit close,

07:05 - although that there's a lot of investors are able to get access to through. So-called Connect. I have some research on that. So it's a big, there's a lot of, a lot of institutions, that are participating there. and, and, and as a result that it's actually quite a mature. so, so little bit history was that basically it started in the nineties and then, then, then in the, in the late nineties,

07:33 - there's a bigger reform and for historical reasons, the two markets, one market's called exchange markets, where you saw all these, exciting news about like stock movement, crash, blah, blah, blah, blah. It's there, and not only you can trade the stocks, but also bonds, corporate bonds, okay? The other is called interbank market. So these two markets is running as in a interestingly different way.

08:01 - Exchange markets is basically a gigantic essential limit order. It, and everybody, it's a central limit order, system. And everyone going there anonymously put the A bids and then they match the bid a, bid ask, prices. And there is basically that. It's the no longer. So-called search based OTC market, right? It's just you go there, place the bid. If you bid at other people's price, why you got it.

08:32 - The other market is called the OTC market or the interbank market where the people meet and just negotiate on prices. So that's the, the way you trade the bonds. I have to explain the traded the bonds before explaining the re re report rules. Okay? So now I'm mo I'm moving on to the report rule basically is the ability for you to borrow in the exchange of markets. The ability to borrow is interestingly determined

08:59 - by the CCP. That's the central clearing party, which is basically the exchange. So the exchange every day they will decide what, how much that you are able to bring your bonds to the exchange and how much you can borrow from the exchange. Okay? Exchange just decided certain things based on some things, some, some rules or the things while on the, in interbank,

09:27 - you could borrow from each other through negotiation. So that's very, very much bilateral. okay, so you have this centralized system. You have very, very, very, decentralized system on the, on the interbank market. Centralized is the exchange. Okay? The policy shock we are exploiting is 24 is a December 8th, 2014, after Monday night. Usually they don't do it. Usually we do it like Friday night. Why?

09:58 - Because Friday night everybody go home. They have two days to absorb the shock. at that time, there's a, some, some, some concern about, about a local government debt and all these things. And the exchange market is not like the biggest market for these bonds, okay? But, but they try to show, you know, allegiance to Beijing. they, on the Monday night to everybody is a surprise.

10:33 - They basically slashed the p legibility of certain bonds and that these certain bonds has a very easy to implement rule rating. China likes just using easy to implement the rules because otherwise it's just, you always have negotiation. Is this, you know, you are supposed to be slashed or not supposed? No, no, no, no. That's just the user RA rating. And what happened was that, the, the, the, the, the, aa, you know, in China mainly there are two ratings

11:06 - or three four ratings that we study. aaa, AA plus aa, and AA minus what they, basically, the rule says that as long as your bonds is below aaa, starting from today, you are not allowed to use these bonds anymore to borrow from the exchange. Okay? And on the other side, no policy coming out just because that they are just managed by different, agencies.

11:41 - Okay? So, okay, so that's a very nice policy shock because as I explained, the fundamental are always the same, but I hit the, the p legibility rule or ripple rule on one market. So you'll see that if there's anything price movement, right? No, I take the difference then, then the, the change of the difference before and after should I identify the effect of the legibility change, okay?

12:13 - So that's so-called diff diff in the typical language of academia, because what we are doing is that across the market first, right? That's diff and then over time of the policy shock, diff diff, there's another le level of diff is quite interesting, very, very unique to this setting, and it is really, you know, make it, the results become extremely convincible is the fact

12:41 - that when you do this, policy, remember that it's basically true below triple A bonds was slashed of the P legibility. Okay? What does that mean? Interestingly, in before the, even before the policy shock, basically you have a four ratings explained, and the AA minus the worst bonds, investors cannot use them to pledge. And anyway, so for them,

13:09 - basically their legibility is zero, even before the shock. And you have AA minus, and AA has some legibility, and you have, AAA has lots of legibility. So 90%, 70%, probably 20 to 30% of them. The zero, once you shock the policy shock hits, basically you only hit the middle two. Why? Because the middle two previously used to be able to borrow 30% or 70%. Now you cannot borrow anymore zero,

13:43 - but the top one, aaa, you're still able to borrow 90% or 90%, and the lowest part, which was before also started with zero. Now it's also zero. So as a result that this policy shock, you have a cross additional cross-sectional D impacted that we can take advantage, okay? This is the super important super, super important. Why? Because most of the time we worry that when shock hits, there's that shock is endogenous

14:18 - and typically will be related to certain fundamental issues. Usually, usually, you know, our, our policy is obviously also subject to, this kind of concern because that Beijing is always trying to address certain things, right? Always. There's a no no point of doing random policy for such a big scale. But why this is helps, because most of the time when these kind of shark hits, usually you have this, so-called beta hypothesis,

14:51 - which means that if you are safer, then you subject to this kind of concern, less aaa, then aa plus AA minus a AA minus will be subject to even more, right? Because usually you just think about the beta because they have a, increasing, beta. If you have a fundamental sharks, they should have reacted to it more. But now we have basically have the lowest beta and highest beta do not change at all from the,

15:19 - from the, looking at the data, but only the middle two really getting hit that just supported the hypothesis that this policy shark only hits the p legibility. Once you realize this is setting, basically we have a triple diff, then it's quite easy to do the estimation, do the measurement basically. So what you're gonna do is take the differential. So that's already three of the fundamental before the policy.

15:49 - What we show was that, in the, in, in the broader corporate bonds market in China, the differential we call exchange premium. Exchange premium, okay, is highly related to the haircuts itself. Okay? So that supported the idea that in this type of market, the fact that you have price differential is purely due to, or largely due to at 90%. We have the, some, some, some, some estimation on

16:22 - how much they can explain the, the price differential is due to this haircut. So that, that's very assuring that, the, the, the, the differential is due to the legibility. Then what we could do is after the policy shock, we basically, you can even do just calculation without doing, doing the regression, right? Basically ask your question that how much that haircutter drops

16:49 - or rise, you know, plausibility drop, haircutter rise, let's say started from a 50% initially. Now you basically haircutter become one policy hits, then the plausibility go down by 50%. You look at the price, price movement is something, okay? And therefore that we can gauge the sensitivity, okay? And, not only one class of bonds, we also have the, the other class of A bonds, which is the AA minus. Then you do the differential and all these things when we do is,

17:25 - using all these dif method. So the way you do it is ba basically adding all these fixed effects, I know time fixed effects and bond fixed effect or rating fixed effects, but, the way you identify it is basically rating time times post shock. So that interaction basically picks up the same bond differential, but after the shock somehow changed it due to the policy, due to the haircutter change.

17:52 - The results coming out from this exercise, by the way, that we take the, six months before, six months after, because sometimes, you know, price movement take, take a little bit of time. We choose six month is because that, we, this is a, interesting period where most of the po most of the policy discussion is at, the real economy level, but it doesn't hit the, financial market yet at that time.

18:23 - And, as a result that that's, we think that at the, at the time, basically time fixed effect can, can, can absorb all these common shocks that are happening to the financial market. Remember, we have two markets, so this is good. once you do that, you find out that, the, the, if the haircut, so here's a quantitative answer, if the haircut, rises from zero to 100%, which means that flexibility slashed from a one to zero, the bond premium goes, goes,

18:58 - so this is a from zero, flexibility from, one to zero. The, the premium goes down by 35%, 30 to 40 basis point. And, to, to us that's, you know, we, once we get this, we can compare to other, other studies and the number is actually smaller than the usual ones. And we think that this is a very natural because that, in the previous studies, you are always being contaminated by something else, okay?

19:35 - So as a result that we are very happy that, the number is not like penny, so that you don't mind, but also like smaller than the existing, numbers. And this in the paper we also mentioned as academia, we need to be super careful. Like what does it mean? This is 35% in, you know, in a, in a framework. We also realized that also that we, from the very beginning when we designed the research,

20:06 - I knew that, hypothetic compared to a setting hypothetically where the, investors are not allowed to move bonds between two markets. Okay? You know, we have a market, so we are actually, you know, I didn't mention this detail, which means that, the investors are able to allow to move between bonds relative to a set setting that they are not allowed to. It's easy to show that, because of this, option to move around that would depress the press

20:42 - differential, press differential. As a result, the number we identified is indeed a little bit of a under bias, okay? So, which means that the 30 to 40 basis point differential is basically is a unbiased, so, so this is the, in the paper we made a very clear explain why or the things, although that we do not think it's such a big deal for different reasons, but unfortunately we cannot quantify that part. So as a result, we,

21:14 - we also provide a overestimation. So for the over estimation, basically we take the same setting, but they use a similar methodology in the existing literature, do not use the match upon. Okay? As a result, we can show that. So basically what we did was basically saying that what if that they, we take the bonds there, look, look identical, look identical, and when you say look identical, okay,

21:50 - you know, look at their price, look at their yield, blah, blah, blah, blah, blah. Then take the differential, okay? Then we'll see that it's basically the number coming out is about an 85 basis point. So as a result that we can, you know, happily conclude that, you know, we have underestimates, which, which is pretty tight, like 30, 30 to 40 basis point, exactly number. I forgot. Then the other side is about 80 basis point.

22:15 - So we have a pretty good sandwich to give a, a gauge of how big is this flexibility is one implication is on the financial market side, is that to basically give a sense that how big is the cost of a capital, in China during that period, okay? And we found that, that this number is very, very, in accordance, with, with the other method that we can use, for instance, the liquidity effect, to, during that time in 2019.

22:55 - So that's basically, it's kind of, just give a sense of the entire financial, financial, financial research community on the first time that we are able to tell them that this is gonna be that big. Okay? And the second one is, is more or less like, from my own, I wouldn't say my own agenda, but I think that I have the motivation to introduce the smart scholars

23:28 - to look into the Chinese bond markets. It's just so much of interesting institutional details. So much effort has been done by the policy makers in China in develop the, corporate bonds market in the past, decade. And they are, they, they were doing very, very well in the, you know, from, I would say from five to probably nine 19, right before the Covid. At this point. They're facing lots of, challenges,

24:02 - but it's really just, you know, for, for obvious the other parts of the reasons. But I wish that, more and more researchers who can look into these markets.

Faculti AI can make mistakes