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

Do Credit Markets Respond to Macroeconomic Shocks? The Case for Reverse Causality

View more insights in Professional - Finance

View more figures from Nova School of Business and Economics

View the page of Giorgio Ottonello

Read the Study

The response of corporate bond credit spreads to exogenous macro shocks is large. Giorgio Ottonello suggests this countercyclicality is driven largely by credit risk premia and translates into significant return predictability.

Image courtesy of interviewee. January 10, 2024

Generating Workspace

Export as PDF

This video video discusses a research study focused on the interaction between credit markets and macroeconomic activity, specifically how macroeconomic shocks influence credit markets. The study aims to fill a gap in existing literature, which has primarily concentrated on how credit markets predict macroeconomic conditions, but not the reverse relationship.

Key points include:

1. **Research Objective**: To understand if and how macroeconomic shocks, such as changes in GDP growth or other economic disruptions, impact credit markets, particularly corporate bond markets.

2. **Importance of Study**: Understanding this relationship helps clarify how economic shocks propagate through credit markets, affecting various sectors of the economy and significantly contributing to business cycle variations.

3. **Literature Context**: Previous theoretical frameworks established that frictions in credit markets can amplify macroeconomic shocks, and empirical studies have validated the predictive power of credit markets over macroeconomic activity. This video video study complements that work by investigating the influence of macroeconomic shocks on credit spreads.

4. **Methodology**: The study utilizes standard measures of macroeconomic activity (like GDP growth) and credit spreads (such as default spreads) and analyses the effects of various macroeconomic shocks using established econometric techniques like local linear projections.

5. **Findings**:
- Macroeconomic shocks are identified as significant drivers of credit spreads, on impact and through historical variations.
- A notable finding is the delayed response of credit spreads to macroeconomic shocks, attributed to information frictions in credit markets.
- The majority of the impact on credit spreads comes from variations in credit risk premiums rather than direct losses from defaults.

6. **Implications**: The findings open avenues for further research into the two-way relationships between credit markets and the macroeconomy, emphasizing the need for models that incorporate the interplay between the two.

7. **Conclusion**: The study concludes that macroeconomic shocks significantly influence credit markets, and understanding this relationship is crucial for developing better economic models and policies. The results point to the necessity of addressing both sides of the credit market and macroeconomic activity relationship for a more comprehensive understanding of business cycles.

Language:

Export as PDF

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

00:00.185 --> 00:04.085
So the starting point, of our study is actually,

00:04.615 --> 00:06.565
let's say a longstanding goal

00:06.585 --> 00:09.085
and question in, in macro, in Macrofinance,

00:09.085 --> 00:10.485
in the macrofinance literature.

00:10.585 --> 00:14.005
And this question is, how to disentangle

00:14.505 --> 00:17.325
and how to characterize the relation between credit markets

00:17.905 --> 00:19.525
and macroeconomic activity.

00:20.185 --> 00:23.285
And, when I talk macro about microeconomic activity

00:23.345 --> 00:26.005
of, I refer, for example, to popular,

00:26.005 --> 00:27.325
measures such as GDP growth.

00:27.385 --> 00:30.445
And when I talk about credit markets, I refer to,

00:30.475 --> 00:34.285
markets where, companies, corporations get financing

00:34.305 --> 00:35.845
for, for their operations, such as,

00:36.305 --> 00:37.445
corporate bond markets.

00:38.465 --> 00:41.845
Uh, so this is a question that has received a lot

00:41.845 --> 00:43.165
of attention in the literature

00:43.265 --> 00:46.685
and the empirical work on this question is mostly focused

00:47.305 --> 00:49.605
on, one direction of the story.

00:49.905 --> 00:51.525
And I'm trying to be now,

00:51.695 --> 00:53.245
let's say go a bit more into detail into that.

00:53.265 --> 00:57.165
So, the empirical work so far has shown that,

00:57.275 --> 00:59.045
corporate bond market,

00:59.375 --> 01:03.045
based measures lead macroeconomic activity, in other words,

01:03.195 --> 01:06.565
predict, macroeconomic activity in, in the future.

01:07.185 --> 01:10.325
Uh, what is missing, from the literature so far is,

01:10.765 --> 01:13.365
analysis from an empirical standpoint on the reverse

01:14.035 --> 01:17.485
empirical relation, which is whether macroeconomic activity,

01:18.025 --> 01:21.845
proxy by, for example, macroeconomic shocks, leads,

01:22.065 --> 01:23.565
or impacts credit markets.

01:23.785 --> 01:27.165
And this is actually what our paper is going to focus about,

01:27.535 --> 01:29.965
which is understanding whether,

01:30.115 --> 01:34.005
macroeconomic shocks are a first order driver of,

01:34.025 --> 01:35.405
of credit markets.

01:36.025 --> 01:39.285
And we are going to put a lot of effort on identifying this,

01:39.345 --> 01:41.405
relationship, precisely.

01:41.945 --> 01:44.205
Uh, this topic is important,

01:44.475 --> 01:48.685
because of course, it allows us to understand how,

01:49.185 --> 01:53.685
shocks, in the aggregate economy transmits, through,

01:53.785 --> 01:57.125
through credit markets to different sectors of, of,

01:57.145 --> 01:59.165
of, of firms and in, in the economy.

01:59.785 --> 02:04.035
And of course, if we can shed light on the two-way

02:04.035 --> 02:05.595
interactions,

02:05.595 --> 02:08.195
and the highlight two ways between credit markets

02:08.335 --> 02:09.835
and macroeconomic activity, of course,

02:10.135 --> 02:13.035
we can also better understand what are the economic sources,

02:13.175 --> 02:15.115
for example, of business cycle variation.

02:15.255 --> 02:18.275
And, I think it's not controversial to say

02:18.275 --> 02:20.435
that a recent crisis episodes such

02:20.435 --> 02:24.155
as the financial crisis in 2007 and 2008,

02:24.295 --> 02:28.395
or, the more recent, covid shock in, 2020.

02:29.095 --> 02:32.515
Uh, so such episodes has shown how little yet,

02:32.515 --> 02:36.315
we understand, about how the business cycle moves

02:36.375 --> 02:39.795
and what are the, key forces, behind it.

02:40.175 --> 02:41.715
Uh, our paper aims

02:41.715 --> 02:44.835
to make a contribution into improving this understanding.

02:45.215 --> 02:48.675
And the focus is going to be exactly into shedding light.

02:49.255 --> 02:52.475
How, on how macroeconomic activity,

02:52.495 --> 02:55.395
or better macroeconomic shocks impact,

02:55.395 --> 02:56.875
credit markets in the first place.

02:57.175 --> 03:00.005
Uh, our paper mainly contributes to the literature

03:00.005 --> 03:03.565
that focuses on indeed the relation between,

03:03.565 --> 03:05.685
credit markets and macroeconomic activity.

03:06.305 --> 03:10.765
Uh, briefly, let's say this literature has,

03:10.795 --> 03:12.325
from a theoretical standpoint,

03:12.325 --> 03:16.805
really influential papers in the, let's say, early to late

03:17.485 --> 03:19.925
nineties, such as, for example, Bernanke

03:19.925 --> 03:23.605
and Gertler, KTA Moore, or Bernanke Gertler and Gilchrist.

03:23.785 --> 03:25.805
So this theoretical paper show,

03:25.805 --> 03:28.485
from a theoretical standpoint that,

03:28.805 --> 03:32.125
frictions in credit markets can be essential,

03:32.225 --> 03:36.485
or first order amplifier, of,

03:36.485 --> 03:37.965
macroeconomic shocks, meaning

03:37.965 --> 03:40.805
that if frictions in credit markets that prevent

03:41.605 --> 03:44.205
efficient allocation of capital across firms are large,

03:44.915 --> 03:49.045
this can, severely amplify, primitive shocks,

03:49.045 --> 03:50.645
fundamental shocks to the economy.

03:51.145 --> 03:55.445
And these papers, have been followed up by, a long

03:55.745 --> 03:57.805
and productive empirical literature

03:58.155 --> 04:01.045
that actually takes the intuition from these theoretical

04:01.105 --> 04:04.805
models into, into the empirical data

04:05.265 --> 04:07.445
and shows, convincingly

04:07.665 --> 04:10.205
and with different, let's say, methodologies that

04:10.785 --> 04:13.645
credit markets forecast microeconomic

04:13.965 --> 04:15.045
activity unambiguously.

04:15.585 --> 04:18.045
And, examples of these papers

04:18.425 --> 04:20.325
and, this is, I think, useful also

04:20.325 --> 04:22.725
to understand which type of indicators have been developed

04:22.785 --> 04:25.565
are, for example, Greenwood and Hanson in 2000

04:25.565 --> 04:27.805
and thirteens, which show that, the share

04:27.805 --> 04:29.325
of high yield bonds relative

04:29.325 --> 04:31.485
to the total corporate bonds outstanding

04:32.025 --> 04:33.845
is an extremely good predictor of,

04:33.945 --> 04:35.285
of macroeconomic activity.

04:35.665 --> 04:38.605
Uh, this measure has also been, used,

04:38.605 --> 04:41.605
to build a credit market sentiment indicator in another

04:41.605 --> 04:43.605
paper by, Lopez and Ali

04:43.605 --> 04:46.285
and co-authors, for example, as well,

04:46.385 --> 04:49.885
sch hand Taylor in 2012 showed that the balance sheet growth

04:50.045 --> 04:51.845
of financial intermediaries, which

04:51.845 --> 04:53.885
of course play an important role in credit markets,

04:54.185 --> 04:58.125
is also an important measure, when we want

04:58.125 --> 04:59.925
to forecast macroeconomic activities.

05:00.635 --> 05:04.165
More recent examples are, an average,

05:04.985 --> 05:08.005
in spread across all corporate bonds in the markets,

05:08.005 --> 05:11.605
developed by Gil Gilchrist and Za in 2012.

05:11.915 --> 05:14.605
This paper has been, extremely influential and,

05:14.705 --> 05:17.405
and has spurred even more work on that dimension.

05:17.745 --> 05:22.125
And last, but not least, in 2021, Rafa Cho

05:22.125 --> 05:26.605
and Goldstein showed that, the flows of, capital

05:27.115 --> 05:30.245
from investment grade to high yield corporate bond funds

05:30.825 --> 05:34.045
is an important predictor of,

05:34.045 --> 05:36.885
macroeconomic activities, which is, again, based on,

05:36.885 --> 05:38.365
credit markets information.

05:38.825 --> 05:43.365
So this overview shows that the direction of causality from,

05:43.865 --> 05:44.965
credit markets

05:45.105 --> 05:48.285
to macroeconomic activities has been well established

05:48.425 --> 05:50.445
and widely, widely studied.

05:50.905 --> 05:54.045
We contribute to this literature by showing the opposite,

05:54.105 --> 05:57.325
direction and focusing on identifying the strength

05:57.785 --> 06:00.845
of the impact of macroeconomic shocks on credit markets.

06:01.465 --> 06:05.685
And, this is of course, complimentary to the studies

06:05.715 --> 06:08.605
that have been, looking at the other directions.

06:08.785 --> 06:12.125
And, we will show them, with the results,

06:13.555 --> 06:16.205
more or at least as important

06:16.505 --> 06:19.165
as the direction from credit markets to,

06:19.305 --> 06:20.445
macroeconomic activity.

06:20.665 --> 06:23.205
The data that we're using in our paper are actually,

06:23.425 --> 06:24.685
quite standard.

06:24.945 --> 06:28.325
So, most of them are, actually quite,

06:28.685 --> 06:30.405
publicly available, so a anyone could,

06:30.405 --> 06:31.525
could easily retrieve them.

06:31.985 --> 06:33.965
Uh, in terms of, measures

06:33.965 --> 06:37.245
of macroeconomic activity in our paper, we focus on,

06:37.245 --> 06:40.485
real GDP growth and, the CF n ai index.

06:41.105 --> 06:43.845
Uh, credit spreads are captured by two measures.

06:44.145 --> 06:47.125
Uh, one is the default spread, which is the difference

06:47.125 --> 06:50.965
between VAA and the a, a a, rated corporate bonds.

06:51.315 --> 06:55.445
This measure spends a longer period from 1919 up

06:55.445 --> 06:56.605
to 2017.

06:56.945 --> 07:00.085
The second measure, which is a bit more precise,

07:00.105 --> 07:03.485
but spends a, a shorter sample period starting in 1973,

07:04.025 --> 07:08.765
is the, GC spread based on the georis theologic paper

07:08.765 --> 07:10.845
that I mentioned earlier, which is essentially

07:11.525 --> 07:12.605
a evaluated average

07:12.745 --> 07:15.605
of ill spreads across all corporate bonds,

07:15.665 --> 07:17.685
in the market at each point in time.

07:18.425 --> 07:20.285
And, these are the two measures.

07:20.625 --> 07:24.925
And, of course, we need, measures

07:24.925 --> 07:27.485
of macroeconomic shocks to better identify the impact

07:27.505 --> 07:30.485
of macroeconomic shocks on, on credit markets.

07:31.105 --> 07:32.125
Uh, so in this case,

07:32.145 --> 07:34.965
we leverage on the literature macroeconomics,

07:34.965 --> 07:39.365
which has done a lot of work in properly identifying,

07:39.555 --> 07:42.685
macroeconomic shocks of, diverse origin.

07:43.145 --> 07:44.205
And we use three of those.

07:44.535 --> 07:48.085
First, we use oil supply shock, which were developed by,

07:48.305 --> 07:51.205
be meister in Edmonton in, 2019.

07:51.825 --> 07:54.205
And, you can think of those shocks,

07:54.625 --> 07:57.485
as a large oil supply disruption, for example, due

07:57.485 --> 07:58.805
to conflicts in the Middle East,

07:59.065 --> 08:01.285
or, unexpected decision

08:01.425 --> 08:05.845
of the OPEC organization when it comes to setting the,

08:06.385 --> 08:08.525
supply of oil, worldwide.

08:09.315 --> 08:11.765
Then we also use defense spending news shocks,

08:11.765 --> 08:15.845
which comes from, from Rami, a paper by

08:15.985 --> 08:17.405
by Rami in 2011.

08:17.705 --> 08:21.045
And you can think of those as, unexpected,

08:21.405 --> 08:24.045
military buildups, from, from the government.

08:24.185 --> 08:26.165
And obviously those are shocks

08:26.165 --> 08:29.365
that are particularly strong in the early part of the

08:30.045 --> 08:33.085
20th century, for example, around, the Second World War.

08:33.505 --> 08:36.645
Uh, last but not least, we use, technology shocks,

08:36.745 --> 08:41.445
from benza and k which, cover a period from,

08:42.145 --> 08:46.685
1952 in this case up to up to 2017.

08:47.225 --> 08:49.405
And as you can see from, from the description

08:49.405 --> 08:52.485
of these shocks, they are all quite diverse in,

08:52.485 --> 08:54.885
in origin, motivation, and even sample periods.

08:55.305 --> 08:58.285
And we believe that this is actually, let's say

08:58.945 --> 09:00.725
an advantage of our study, meaning

09:00.725 --> 09:03.125
that finding a consistent impact

09:03.265 --> 09:06.085
of such diverse macroeconomic shocks on credit markets

09:06.665 --> 09:10.045
is not obvious, and it sets quite a high bar in terms

09:10.045 --> 09:14.165
of identification and in terms of, the robustness

09:14.225 --> 09:17.805
and the strength of, of our finding in, in our study.

09:18.505 --> 09:22.245
Um, so in terms of other databases that we use

09:22.245 --> 09:24.325
before getting into, into the methodology,

09:24.795 --> 09:27.685
when we analyze the economic mechanisms, be, be

09:27.685 --> 09:30.965
behind our main results, we need to rely on,

09:30.995 --> 09:33.525
bond level data on, bond returns

09:33.545 --> 09:35.765
and characteristics, which we collect

09:35.825 --> 09:38.085
by merging the Lehman Brothers database

09:38.085 --> 09:40.165
and the trace database to create a time series

09:40.195 --> 09:43.765
that goes from, 1973 to 2017.

09:44.545 --> 09:48.845
And, we also need to use, measures of,

09:49.045 --> 09:52.045
expectations of credit spreads and macroeconomic activity,

09:52.465 --> 09:55.445
and we rely for that on the professional survey forecast

09:55.595 --> 09:57.125
from, the survey

09:57.185 --> 09:59.445
of professional forecasters and, and blue chip.

09:59.505 --> 10:02.405
So this is a general overview of the databases

10:02.405 --> 10:05.365
that we use in our paper in terms of empirical methodology,

10:05.665 --> 10:07.925
the first, thing we do is

10:07.985 --> 10:11.485
to understand whether macroeconomic shocks have an impact at

10:11.505 --> 10:13.245
all on, on credit spreads.

10:13.305 --> 10:17.405
And to do that, we rely on local linear projections based on

10:17.405 --> 10:20.525
the paper by, Georgia in, in 2005.

10:20.625 --> 10:22.125
And, in a nutshell,

10:22.235 --> 10:24.845
this approach is essentially regressing credit spreads on

10:24.845 --> 10:28.805
macro shocks, and of course, a wide set of, of controls

10:29.465 --> 10:31.605
and doing so at different horizons.

10:31.605 --> 10:34.005
So we are going to have one regression per, per horizon.

10:34.145 --> 10:35.885
And then analyzing the impact,

10:35.985 --> 10:38.605
the coefficient on the macros shocks at the different

10:38.605 --> 10:43.125
horizon to gorge the, the dynamic effect of,

10:43.355 --> 10:46.725
over time of, macros shocks on, on credit spreads.

10:47.225 --> 10:50.205
And this, type of exercise allows us to understand,

10:50.305 --> 10:52.685
whether when a macroeconomic shock hits,

10:52.975 --> 10:55.805
there is a large movement in credit spreads in response

10:56.065 --> 10:57.285
of, of this shock.

10:57.825 --> 11:01.685
Uh, the second, let's say methodology that we,

11:01.715 --> 11:04.805
that we use, serves to answer, a related,

11:05.025 --> 11:07.805
but different questions, meaning are the macro shocks

11:07.805 --> 11:11.405
that we use, able to explain a large fraction

11:11.405 --> 11:13.925
of the historical variation in credit spreads over time.

11:14.025 --> 11:18.645
In other words, are there enough of those shocks,

11:18.645 --> 11:21.045
to be economically relevant, o over time?

11:21.105 --> 11:22.165
It could be, for example,

11:22.165 --> 11:23.965
that there is an extremely large shock

11:24.075 --> 11:26.285
that impacts credit spreads at one point in time,

11:26.705 --> 11:28.525
but this is once in a lifetime shock.

11:28.525 --> 11:30.125
So there are no other shocks of that kind.

11:30.185 --> 11:34.325
So if you collect the total variation of credit spreads,

11:34.325 --> 11:35.565
in relation to macro shocks,

11:35.565 --> 11:37.005
it might turn out to be pretty small.

11:37.345 --> 11:39.765
So, to answer this question, we need

11:39.765 --> 11:42.525
to use a forecast error variance decomposition, tailored

11:42.785 --> 11:45.765
to our local linear projections that I described

11:45.765 --> 11:47.565
above, be before.

11:47.705 --> 11:50.605
And, to do that we apply methodology developed

11:50.605 --> 11:54.005
by Chenko and Lee in their 2020 paper,

11:54.495 --> 11:56.685
Which, exactly,

11:56.955 --> 12:00.085
Creates a forecaster variance composition tailored

12:00.225 --> 12:01.845
to local linear projections.

12:02.105 --> 12:03.105
And,

12:03.515 --> 12:06.215
Uh, a, a last, let's say, comment on,

12:06.315 --> 12:09.295
on our empirical methodologies that the type of shocks

12:09.295 --> 12:13.535
that we study, are large, but in, rather

12:13.535 --> 12:14.815
Infrequent. And

12:14.815 --> 12:16.415
Therefore, of course, the length of

12:16.415 --> 12:18.375
The sample period is essential,

12:18.375 --> 12:20.015
to get enough statistical power.

12:20.195 --> 12:22.255
So we do have quite a long sample, but

12:22.505 --> 12:24.175
Given the strength Of our results

12:24.275 --> 12:27.015
and the, the important implications of, of this type

12:27.015 --> 12:29.975
of shocks, we are also working in follow-up projects

12:29.975 --> 12:31.935
to expand even further the sample period

12:31.995 --> 12:35.735
and go back, to the 19th century, to see if we,

12:35.735 --> 12:38.335
if we can learn anything new, in terms of

12:38.555 --> 12:42.215
how macroeconomic shocks, o over over the history

12:42.595 --> 12:44.775
as, as, as impacted the, the economy

12:44.835 --> 12:47.295
and whether there is something that we can learn also for,

12:47.595 --> 12:49.815
let's say, policies. Uh, nowadays.

12:50.315 --> 12:52.975
To start with, the main finding of our paper is,

12:53.555 --> 12:55.215
Uh, that macro Shocks are

12:55.375 --> 12:57.615
A first order driver of, of credit spreads.

12:57.795 --> 12:58.795
And

12:59.035 --> 13:02.215
We, we can show this both when the shock hits.

13:02.555 --> 13:04.815
So there is a large impact on credit spreads when the

13:04.815 --> 13:07.695
shock hits, but also, as I mentioned earlier,

13:07.845 --> 13:11.855
Equally important macros shocks can explain a large

13:11.855 --> 13:13.815
Variation of credit spreads historical.

13:14.115 --> 13:17.055
So even in the forecast error variance decomposition,

13:17.055 --> 13:19.415
we can show that credit spreads indeed are,

13:19.515 --> 13:21.655
are a major driver of historical variation.

13:21.675 --> 13:23.455
Uh, macros shocks are a major driver

13:23.475 --> 13:25.855
of historical variation in, in credit spreads.

13:26.555 --> 13:28.215
Uh, to give an idea of,

13:29.035 --> 13:31.575
Why am I saying large, what are the magnitudes?

13:31.835 --> 13:33.855
Uh, for example, a one standard

13:33.855 --> 13:37.815
Deviation negative macro Shocks on impact,

13:38.135 --> 13:42.015
Increases credit spreads by about 40 basis points, which

13:42.555 --> 13:43.615
Is relatively large.

13:43.995 --> 13:46.335
Uh, if you think about the fact that,

13:46.335 --> 13:48.615
40 basis points is around

13:48.615 --> 13:49.615
42%

13:49.615 --> 13:52.055
of the unconditional standard deviation in credit spreads

13:52.115 --> 13:53.215
in, in our sample,

13:53.765 --> 13:55.855
When we move to, the forecast

13:55.855 --> 13:57.055
Error variance decomposition,

13:57.595 --> 14:00.655
We see, Always from a magnitude standpoint that,

14:01.155 --> 14:04.415
Uh, macroeconomic shocks can explain between 40 to

14:04.415 --> 14:08.415
50% of, the historical variation in credit spreads.

14:08.515 --> 14:09.855
And I'm referring to

14:10.115 --> 14:13.295
The aggregate Impact of all of our shocks together to,

14:13.295 --> 14:14.575
to reach that, that number.

14:15.075 --> 14:17.615
And this is obviously a large number,

14:17.615 --> 14:19.495
it's definitely indicating that

14:19.585 --> 14:21.495
Macro shocks are a first order driver

14:21.595 --> 14:22.695
of credit spreads. And

14:22.695 --> 14:27.575
Importantly, if we look at how much, variation in

14:27.575 --> 14:30.375
Macroeconomic activity, macro shocks explain,

14:30.875 --> 14:33.655
for example, how much variation in GDP growth is explained

14:33.655 --> 14:34.775
by macro shocks, we

14:34.775 --> 14:36.655
Get about The same, the same quantity.

14:36.995 --> 14:38.375
So macroeconomic

14:38.395 --> 14:41.695
Shocks are as important in explaining

14:41.695 --> 14:43.055
The variation in credit spreads

14:43.075 --> 14:45.455
as they are in explaining the variation in macroeconomic

14:45.775 --> 14:48.335
activity, which are the thing they first impact

14:48.355 --> 14:49.455
in, in, in the first place.

14:49.755 --> 14:51.645
So this is actually, we believe

14:51.645 --> 14:53.605
that this result really speaks to the fact that

14:53.885 --> 14:55.445
whenever, you want

14:55.445 --> 14:59.095
to study the variation in credit spreads, you need to turn

14:59.095 --> 15:01.935
to Mhh shocks as a first order order effect.

15:02.395 --> 15:03.455
Uh, for sure.

15:03.725 --> 15:05.095
This is, one of the main,

15:05.155 --> 15:06.935
the main conclusions of our study.

15:07.675 --> 15:11.375
Uh, in the second part of our paper, we try

15:11.375 --> 15:12.615
to disentangle,

15:12.755 --> 15:14.895
and better understand the economic mechanism

15:14.895 --> 15:16.015
behind our findings.

15:16.075 --> 15:18.135
And we do so by answering two questions.

15:18.705 --> 15:23.535
First, which component of credit spreads, is,

15:23.765 --> 15:26.455
driving the reaction to micro shocks?

15:26.875 --> 15:31.635
And second, why do we observe a delayed response

15:31.635 --> 15:33.715
of credit spreads to micros Shocks, I,

15:33.835 --> 15:35.315
I haven't specified it earlier,

15:35.415 --> 15:38.035
but the impacts that,

15:38.065 --> 15:40.955
I've shown the 40 basis point impact is large,

15:40.975 --> 15:44.035
but comes with a slight delay relative to the impact

15:44.035 --> 15:46.035
of macros, shocks on macroeconomic activity.

15:46.175 --> 15:49.715
So we want to understand why, why this is the case,

15:49.735 --> 15:52.595
and this probably can, can also help

15:52.735 --> 15:55.875
to shed light on the economic mechanism behind,

15:56.135 --> 15:57.235
behind such effects.

15:57.695 --> 16:00.435
So, coming to the first question, which component

16:00.435 --> 16:02.675
of credit spreads impact the,

16:03.055 --> 16:04.835
is most impacted by micro shocks?

16:05.495 --> 16:09.275
Uh, we decompose, in this case, credit spreads into

16:10.415 --> 16:13.825
cash flows, meaning how much losses accrued

16:13.825 --> 16:16.505
to investors when, a company defaults,

16:17.005 --> 16:21.205
and the discount rates, meaning, what is the variation

16:21.205 --> 16:24.845
of the risk premium that the investors require

16:24.945 --> 16:28.085
to hold the same amount of risk at, at, at each,

16:28.085 --> 16:29.245
at a certain point in time.

16:29.425 --> 16:33.585
And, to decompose this empirically,

16:33.605 --> 16:35.905
we apply a decomposition of credit spreads,

16:35.905 --> 16:38.585
which is based on a methodology developed by,

16:38.585 --> 16:41.065
no zawa in in his 2017 paper.

16:41.685 --> 16:44.705
And the, once we do that, we essentially

16:45.275 --> 16:47.905
study the impact of micros shocks on each

16:47.905 --> 16:49.025
of the components separately.

16:49.085 --> 16:51.705
So this is, again, a new variable, for,

16:51.705 --> 16:54.345
for our local linear projection that is based only on one

16:54.345 --> 16:55.745
of the component of credit spread.

16:56.205 --> 16:59.905
And the answer is, again, pretty clear.

17:00.445 --> 17:03.425
The large part of, the impact

17:03.425 --> 17:05.745
of macroeconomic shocks in credit spread is driven.

17:06.365 --> 17:08.585
And we're talking about 80%

17:08.605 --> 17:12.145
or more of the vari of, of the impact is driven by,

17:12.145 --> 17:13.505
variation in credit risk premium,

17:13.505 --> 17:16.265
meaning credit spreads move, not because,

17:17.005 --> 17:20.345
or, not mainly because when, a bond

17:20.605 --> 17:21.905
or a company defaults.

17:21.905 --> 17:24.465
There are larger losses to investors, but mostly

17:24.465 --> 17:28.265
because investors when a negative shock,

17:28.265 --> 17:31.465
macroeconomic shock hits require a higher compensation

17:31.485 --> 17:32.625
for the same amount of risk.

17:32.845 --> 17:34.825
So their, compensation

17:34.935 --> 17:36.465
that they require for risk increases.

17:36.465 --> 17:41.385
So time variation in credit risk premium is the main,

17:41.885 --> 17:44.185
driver of the reaction of credit spreads

17:44.185 --> 17:45.465
to macroeconomic shocks.

17:45.485 --> 17:47.985
So this is an important, result also

17:47.985 --> 17:50.565
because it can speak to,

17:51.095 --> 17:54.885
let's say theoretical models that try to understand how,

17:54.885 --> 17:57.325
the variation in credit spreads,

17:57.745 --> 17:59.285
can be explained over time.

17:59.425 --> 18:03.325
So turning to time varying credit risk premia is,

18:03.605 --> 18:05.725
a really good direction given the,

18:05.785 --> 18:07.725
and promising direction given the empirical

18:07.725 --> 18:09.005
findings that, that we have.

18:09.745 --> 18:12.805
Um, the second question is that, is targeting

18:12.805 --> 18:15.565
to understand why the response of credit spreads is delayed.

18:16.185 --> 18:19.925
And, to answer this question, we,

18:20.865 --> 18:24.005
we try to provide evidence of the presence

18:24.005 --> 18:27.045
of information frictions in, in credit markets

18:27.075 --> 18:28.925
that could explain this delayed response.

18:29.065 --> 18:32.385
And, in simple words,

18:32.495 --> 18:35.105
information frictions just reflect the, the idea

18:35.215 --> 18:37.665
that there is not a full response

18:37.845 --> 18:39.545
of credit spreads when a shock hits,

18:39.695 --> 18:41.385
because the exact information,

18:41.385 --> 18:44.465
content is not immediately available, to all investors,

18:44.885 --> 18:46.025
at that point in time.

18:46.085 --> 18:48.185
So there is, there are frictions

18:48.185 --> 18:51.545
that prevent investors from accessing, the,

18:51.645 --> 18:54.065
the complete set of information at that point in time.

18:54.205 --> 18:56.785
It might require, a bit longer,

18:56.845 --> 18:58.825
and that's why, the response is delayed.

18:59.285 --> 19:00.785
Uh, to test for this empirically,

19:00.805 --> 19:03.145
we follow a methodology developed by Kion

19:03.225 --> 19:05.345
and Chenko in their 2012 paper.

19:05.525 --> 19:07.665
And essentially, the idea is pretty simple.

19:08.285 --> 19:12.185
Uh, you need to turn to measures of expectations of,

19:12.365 --> 19:16.385
of credit spreads, meaning, what investors expect about,

19:16.645 --> 19:18.385
credit spreads in the future,

19:19.005 --> 19:22.785
and check whether those measures respond indeed,

19:22.785 --> 19:24.665
with a delay to micro shocks relative

19:24.665 --> 19:26.145
to the realized credit spreads.

19:26.525 --> 19:30.065
If this is the case, the, the implication

19:30.085 --> 19:33.305
that's forecast error, meaning the error in the forecast

19:33.305 --> 19:36.505
between the realization relative to the realization,

19:36.565 --> 19:38.985
if this is the case forecast errors will be predictable, and

19:38.985 --> 19:41.185
therefore this will be a sign of the presence

19:41.245 --> 19:43.745
of information frictions in, in

19:43.745 --> 19:46.025
that specific, dimension.

19:46.485 --> 19:50.375
And, once we turn to the data, we do indeed find

19:50.405 --> 19:53.415
that, the impact of,

19:53.705 --> 19:56.335
macro shocks on expectations of forecast

19:56.335 --> 19:59.375
of credit spreads is indeed delayed relative

19:59.455 --> 20:02.055
to the realization, which is consistent with the idea

20:02.055 --> 20:04.175
that there, the, there is a scope,

20:04.265 --> 20:05.615
there is an important role

20:05.615 --> 20:07.335
of information frictions in credit markets,

20:07.385 --> 20:10.855
which can explain the delayed response of credit spreads.

20:11.315 --> 20:15.015
Uh, so this is essentially, wrapping up the,

20:15.015 --> 20:19.415
the part on, on why credit spreads are delayed.

20:20.015 --> 20:23.455
I want to mention, really briefly, other two tests

20:23.605 --> 20:27.095
that we do to better, provide, an understanding

20:27.155 --> 20:29.175
of the economic mechanisms and,

20:29.195 --> 20:31.135
and the robustness of, of our findings.

20:31.305 --> 20:33.975
First, we are able to show that, the response

20:33.975 --> 20:37.335
of credit spreads to macro shocks is much stronger when we

20:37.335 --> 20:40.535
focus on credit spreads of firms with high leverage, meaning

20:40.535 --> 20:44.255
that if you, if you, restrict the sample to firms,

20:44.285 --> 20:46.935
with high leverage, you see that the response is,

20:47.365 --> 20:50.605
a double, of credit spread to micros shocks relative to,

20:50.605 --> 20:52.165
for example, firms with low leverage.

20:52.355 --> 20:54.045
This is important because it indicates

20:54.045 --> 20:56.085
that firm leverage is an important

20:56.835 --> 21:00.165
amplifi amplification mechanism when it comes

21:00.165 --> 21:02.565
to understanding the impact of micros shocks on,

21:02.745 --> 21:04.765
on the economy through credit markets.

21:05.425 --> 21:09.005
And, the last exercise I wanna mention is,

21:09.285 --> 21:11.205
a sort of external validity exercise.

21:11.305 --> 21:12.925
We, we take our macro shocks

21:12.925 --> 21:16.965
and we, underst we try to figure out what is the impact

21:16.965 --> 21:20.085
of these shocks on popular proxies of equity risk premium.

21:20.185 --> 21:21.565
So we turn to the equity market

21:22.025 --> 21:25.515
and check whether the impact

21:25.515 --> 21:28.955
that we find on credit markets is actually reflected also in

21:28.955 --> 21:32.515
other, financial assets, which if these are truly,

21:32.565 --> 21:33.715
let's say exogenous

21:33.855 --> 21:36.395
and primitive macros shocks, it should be the case.

21:36.775 --> 21:39.755
And we did find that the impact of, of these macros,

21:39.755 --> 21:43.955
shocks on popular, equity risk pri proxies is,

21:43.955 --> 21:46.795
really similar to what we observe both in direction

21:46.795 --> 21:49.115
and magnitude in, in, in credit markets.

21:49.375 --> 21:50.915
So this is important, first of all,

21:50.915 --> 21:53.395
because it, let's say, reinforces the idea

21:53.425 --> 21:55.955
that we are looking at, a proper set

21:55.955 --> 21:59.915
of microeconomic shocks, but also, importantly

21:59.915 --> 22:04.035
because it shows a common source of variation of,

22:04.275 --> 22:07.475
discount rates in, credit markets and equity markets.

22:08.055 --> 22:09.115
And this is important

22:09.115 --> 22:12.875
because empirically, the correlation between equity risk,

22:12.875 --> 22:16.035
premier proxies and credit risk premier proxies,

22:16.035 --> 22:18.995
have been quite low, empirically,

22:19.345 --> 22:22.115
despite theoretically that should be, pretty high.

22:22.575 --> 22:25.195
Our paper shows that this is an identification problem.

22:25.305 --> 22:29.875
Once you can identify, a well, let's say shocks

22:29.875 --> 22:33.155
that affect, both markets at the same point in time

22:33.215 --> 22:36.235
and hold constant other confounding factors,

22:36.235 --> 22:38.155
this can actually resolve this puzzle

22:38.265 --> 22:40.795
that has been shown empirically when you look at the

22:40.795 --> 22:43.475
unconditional correlation between credit markets

22:43.575 --> 22:44.595
and, and equity markets.

22:45.055 --> 22:47.195
And we believe actually that this is,

22:47.615 --> 22:50.195
an interesting avenue also for, for future work.

22:50.415 --> 22:53.835
The findings that, that we have in our paper, can be,

22:54.255 --> 22:56.835
used,

22:57.095 --> 22:59.555
and extended in, in future research.

22:59.575 --> 23:02.835
And I'm going to give you just here, a few examples.

23:02.895 --> 23:07.355
So, the first part that, we believe is, is,

23:07.415 --> 23:09.675
is actually, useful for, for future,

23:09.815 --> 23:12.315
future research is the finding that the response

23:12.315 --> 23:14.195
of credit spreads to,

23:14.295 --> 23:16.355
to macroeconomic shocks is a combination

23:16.575 --> 23:20.435
of information rigidities, which explains, the, the low,

23:20.455 --> 23:23.915
the slow response and leverage related amplifications firms

23:23.915 --> 23:26.115
with higher leverage will, respond more to,

23:26.115 --> 23:27.355
to macroeconomic shocks.

23:27.975 --> 23:32.345
Uh, there are, to our knowledge,

23:32.605 --> 23:35.145
no macroeconomic models that include both

23:35.685 --> 23:39.705
of these components, when it comes to explaining the,

23:40.125 --> 23:43.025
the, the pricing of, of credit, of,

23:43.125 --> 23:44.985
in credit markets, and also the evolution

23:45.085 --> 23:46.445
of credit risk premium over time.

23:46.545 --> 23:49.365
So we believe that, developing models

23:49.365 --> 23:52.445
that incorporate these two features can go a long way into

23:52.455 --> 23:55.645
explaining the empirical facts, that we document,

23:56.265 --> 23:57.325
o overall.

23:58.065 --> 24:02.685
And of course, more broadly the fact that,

24:02.685 --> 24:04.885
we find an important,

24:04.885 --> 24:07.085
let's say a large impact on macroeconomic shocks on credit

24:07.085 --> 24:08.845
markets speaks to the fact

24:08.845 --> 24:11.125
that there are probably large feedbacks

24:11.125 --> 24:13.645
between credit markets and macroeconomic activity.

24:13.705 --> 24:15.885
As I said, the direction between credit markets

24:16.345 --> 24:18.165
and the macro economy has been well documented.

24:18.185 --> 24:22.725
We now show that also the reverse causality, direction,

24:22.865 --> 24:24.205
works pretty strongly.

24:24.425 --> 24:27.565
So models that incorporate the two-way interaction

24:27.565 --> 24:30.125
between the macro economy and the credit markets,

24:30.825 --> 24:34.805
and use this to understand, the contribution

24:34.805 --> 24:36.365
to business cycles, we believe

24:36.365 --> 24:38.165
that this remains a really important

24:38.195 --> 24:39.445
area for future research.

24:39.905 --> 24:43.565
And last, but not least, the fact that these shocks

24:43.565 --> 24:44.965
that we document,

24:45.155 --> 24:47.645
have a large impact on other markets too in equity

24:47.645 --> 24:48.805
markets that we have shown.

24:49.285 --> 24:52.365
I think it, it can open interesting avenues, to,

24:52.785 --> 24:55.085
to empirically measure common source of,

24:55.345 --> 24:57.325
of risk across different markets.

24:57.665 --> 25:00.885
And I think there is a, there is a lot of work

25:00.885 --> 25:03.645
that can be done, empirically in, into that direction.

00:00 - So the starting point, of our study is actually, let's say a longstanding goal and question in, in macro, in Macrofinance, in the macrofinance literature. And this question is, how to disentangle and how to characterize the relation between credit markets

00:17 - and macroeconomic activity. And, when I talk macro about microeconomic activity of, I refer, for example, to popular, measures such as GDP growth. And when I talk about credit markets, I refer to, markets where, companies, corporations get financing for, for their operations, such as, corporate bond markets. so this is a question that has received a lot of attention in the literature

00:43 - and the empirical work on this question is mostly focused on, one direction of the story. And I'm trying to be now, let's say go a bit more into detail into that. So, the empirical work so far has shown that, corporate bond market, based measures lead macroeconomic activity, in other words, predict, macroeconomic activity in, in the future. what is missing, from the literature so far is, analysis from an empirical standpoint on the reverse

01:14 - empirical relation, which is whether macroeconomic activity, proxy by, for example, macroeconomic shocks, leads, or impacts credit markets. And this is actually what our paper is going to focus about, which is understanding whether, macroeconomic shocks are a first order driver of, of credit markets. And we are going to put a lot of effort on identifying this, relationship, precisely. this topic is important,

01:44 - because of course, it allows us to understand how, shocks, in the aggregate economy transmits, through, through credit markets to different sectors of, of, of, of firms and in, in the economy. And of course, if we can shed light on the two-way interactions, and the highlight two ways between credit markets and macroeconomic activity, of course, we can also better understand what are the economic sources, for example, of business cycle variation.

02:15 - And, I think it's not controversial to say that a recent crisis episodes such as the financial crisis in 2007 and 2008, or, the more recent, covid shock in, 2020. so such episodes has shown how little yet, we understand, about how the business cycle moves and what are the, key forces, behind it. our paper aims to make a contribution into improving this understanding. And the focus is going to be exactly into shedding light.

02:49 - How, on how macroeconomic activity, or better macroeconomic shocks impact, credit markets in the first place. our paper mainly contributes to the literature that focuses on indeed the relation between, credit markets and macroeconomic activity. briefly, let's say this literature has, from a theoretical standpoint, really influential papers in the, let's say, early to late nineties, such as, for example, Bernanke

03:19 - and Gertler, KTA Moore, or Bernanke Gertler and Gilchrist. So this theoretical paper show, from a theoretical standpoint that, frictions in credit markets can be essential, or first order amplifier, of, macroeconomic shocks, meaning that if frictions in credit markets that prevent efficient allocation of capital across firms are large, this can, severely amplify, primitive shocks, fundamental shocks to the economy.

03:51 - And these papers, have been followed up by, a long and productive empirical literature that actually takes the intuition from these theoretical models into, into the empirical data and shows, convincingly and with different, let's say, methodologies that credit markets forecast microeconomic activity unambiguously. And, examples of these papers and, this is, I think, useful also

04:20 - to understand which type of indicators have been developed are, for example, Greenwood and Hanson in 2000 and thirteens, which show that, the share of high yield bonds relative to the total corporate bonds outstanding is an extremely good predictor of, of macroeconomic activity. this measure has also been, used, to build a credit market sentiment indicator in another paper by, Lopez and Ali

04:43 - and co-authors, for example, as well, sch hand Taylor in 2012 showed that the balance sheet growth of financial intermediaries, which of course play an important role in credit markets, is also an important measure, when we want to forecast macroeconomic activities. More recent examples are, an average, in spread across all corporate bonds in the markets, developed by Gil Gilchrist and Za in 2012. This paper has been, extremely influential and,

05:14 - and has spurred even more work on that dimension. And last, but not least, in 2021, Rafa Cho and Goldstein showed that, the flows of, capital from investment grade to high yield corporate bond funds is an important predictor of, macroeconomic activities, which is, again, based on, credit markets information. So this overview shows that the direction of causality from, credit markets to macroeconomic activities has been well established

05:48 - and widely, widely studied. We contribute to this literature by showing the opposite, direction and focusing on identifying the strength of the impact of macroeconomic shocks on credit markets. And, this is of course, complimentary to the studies that have been, looking at the other directions. And, we will show them, with the results, more or at least as important as the direction from credit markets to, macroeconomic activity.

06:20 - The data that we're using in our paper are actually, quite standard. So, most of them are, actually quite, publicly available, so a anyone could, could easily retrieve them. in terms of, measures of macroeconomic activity in our paper, we focus on, real GDP growth and, the CF n ai index. credit spreads are captured by two measures. one is the default spread, which is the difference

06:47 - between VAA and the a, a a, rated corporate bonds. This measure spends a longer period from 1919 up to 2017. The second measure, which is a bit more precise, but spends a, a shorter sample period starting in 1973, is the, GC spread based on the georis theologic paper that I mentioned earlier, which is essentially a evaluated average of ill spreads across all corporate bonds, in the market at each point in time.

07:18 - And, these are the two measures. And, of course, we need, measures of macroeconomic shocks to better identify the impact of macroeconomic shocks on, on credit markets. so in this case, we leverage on the literature macroeconomics, which has done a lot of work in properly identifying, macroeconomic shocks of, diverse origin. And we use three of those. First, we use oil supply shock, which were developed by,

07:48 - be meister in Edmonton in, 2019. And, you can think of those shocks, as a large oil supply disruption, for example, due to conflicts in the Middle East, or, unexpected decision of the OPEC organization when it comes to setting the, supply of oil, worldwide. Then we also use defense spending news shocks, which comes from, from Rami, a paper by by Rami in 2011.

08:17 - And you can think of those as, unexpected, military buildups, from, from the government. And obviously those are shocks that are particularly strong in the early part of the 20th century, for example, around, the Second World War. last but not least, we use, technology shocks, from benza and k which, cover a period from, 1952 in this case up to up to 2017. And as you can see from, from the description of these shocks, they are all quite diverse in,

08:52 - in origin, motivation, and even sample periods. And we believe that this is actually, let's say an advantage of our study, meaning that finding a consistent impact of such diverse macroeconomic shocks on credit markets is not obvious, and it sets quite a high bar in terms of identification and in terms of, the robustness and the strength of, of our finding in, in our study. so in terms of other databases that we use before getting into, into the methodology,

09:24 - when we analyze the economic mechanisms, be, be behind our main results, we need to rely on, bond level data on, bond returns and characteristics, which we collect by merging the Lehman Brothers database and the trace database to create a time series that goes from, 1973 to 2017. And, we also need to use, measures of, expectations of credit spreads and macroeconomic activity, and we rely for that on the professional survey forecast

09:55 - from, the survey of professional forecasters and, and blue chip. So this is a general overview of the databases that we use in our paper in terms of empirical methodology, the first, thing we do is to understand whether macroeconomic shocks have an impact at all on, on credit spreads. And to do that, we rely on local linear projections based on the paper by, Georgia in, in 2005. And, in a nutshell,

10:22 - this approach is essentially regressing credit spreads on macro shocks, and of course, a wide set of, of controls and doing so at different horizons. So we are going to have one regression per, per horizon. And then analyzing the impact, the coefficient on the macros shocks at the different horizon to gorge the, the dynamic effect of, over time of, macros shocks on, on credit spreads. And this, type of exercise allows us to understand, whether when a macroeconomic shock hits,

10:52 - there is a large movement in credit spreads in response of, of this shock. the second, let's say methodology that we, that we use, serves to answer, a related, but different questions, meaning are the macro shocks that we use, able to explain a large fraction of the historical variation in credit spreads over time. In other words, are there enough of those shocks, to be economically relevant, o over time? It could be, for example,

11:22 - that there is an extremely large shock that impacts credit spreads at one point in time, but this is once in a lifetime shock. So there are no other shocks of that kind. So if you collect the total variation of credit spreads, in relation to macro shocks, it might turn out to be pretty small. So, to answer this question, we need to use a forecast error variance decomposition, tailored to our local linear projections that I described

11:45 - above, be before. And, to do that we apply methodology developed by Chenko and Lee in their 2020 paper, Which, exactly, Creates a forecaster variance composition tailored to local linear projections. And, a, a last, let's say, comment on, on our empirical methodologies that the type of shocks that we study, are large, but in, rather

12:13 - Infrequent. And Therefore, of course, the length of The sample period is essential, to get enough statistical power. So we do have quite a long sample, but Given the strength Of our results and the, the important implications of, of this type of shocks, we are also working in follow-up projects to expand even further the sample period and go back, to the 19th century, to see if we,

12:35 - if we can learn anything new, in terms of how macroeconomic shocks, o over over the history as, as, as impacted the, the economy and whether there is something that we can learn also for, let's say, policies. nowadays. To start with, the main finding of our paper is, that macro Shocks are A first order driver of, of credit spreads. And We, we can show this both when the shock hits.

13:02 - So there is a large impact on credit spreads when the shock hits, but also, as I mentioned earlier, Equally important macros shocks can explain a large Variation of credit spreads historical. So even in the forecast error variance decomposition, we can show that credit spreads indeed are, are a major driver of historical variation. macros shocks are a major driver of historical variation in, in credit spreads. to give an idea of,

13:29 - Why am I saying large, what are the magnitudes? for example, a one standard Deviation negative macro Shocks on impact, Increases credit spreads by about 40 basis points, which Is relatively large. if you think about the fact that, 40 basis points is around 42% of the unconditional standard deviation in credit spreads in, in our sample,

13:53 - When we move to, the forecast Error variance decomposition, We see, Always from a magnitude standpoint that, macroeconomic shocks can explain between 40 to 50% of, the historical variation in credit spreads. And I'm referring to The aggregate Impact of all of our shocks together to, to reach that, that number. And this is obviously a large number, it's definitely indicating that

14:19 - Macro shocks are a first order driver of credit spreads. And Importantly, if we look at how much, variation in Macroeconomic activity, macro shocks explain, for example, how much variation in GDP growth is explained by macro shocks, we Get about The same, the same quantity. So macroeconomic Shocks are as important in explaining The variation in credit spreads

14:43 - as they are in explaining the variation in macroeconomic activity, which are the thing they first impact in, in, in the first place. So this is actually, we believe that this result really speaks to the fact that whenever, you want to study the variation in credit spreads, you need to turn to Mhh shocks as a first order order effect. for sure. This is, one of the main,

15:05 - the main conclusions of our study. in the second part of our paper, we try to disentangle, and better understand the economic mechanism behind our findings. And we do so by answering two questions. First, which component of credit spreads, is, driving the reaction to micro shocks? And second, why do we observe a delayed response of credit spreads to micros Shocks, I,

15:33 - I haven't specified it earlier, but the impacts that, I've shown the 40 basis point impact is large, but comes with a slight delay relative to the impact of macros, shocks on macroeconomic activity. So we want to understand why, why this is the case, and this probably can, can also help to shed light on the economic mechanism behind, behind such effects. So, coming to the first question, which component

16:00 - of credit spreads impact the, is most impacted by micro shocks? we decompose, in this case, credit spreads into cash flows, meaning how much losses accrued to investors when, a company defaults, and the discount rates, meaning, what is the variation of the risk premium that the investors require to hold the same amount of risk at, at, at each, at a certain point in time. And, to decompose this empirically,

16:33 - we apply a decomposition of credit spreads, which is based on a methodology developed by, no zawa in in his 2017 paper. And the, once we do that, we essentially study the impact of micros shocks on each of the components separately. So this is, again, a new variable, for, for our local linear projection that is based only on one of the component of credit spread. And the answer is, again, pretty clear.

17:00 - The large part of, the impact of macroeconomic shocks in credit spread is driven. And we're talking about 80% or more of the vari of, of the impact is driven by, variation in credit risk premium, meaning credit spreads move, not because, or, not mainly because when, a bond or a company defaults. There are larger losses to investors, but mostly because investors when a negative shock,

17:28 - macroeconomic shock hits require a higher compensation for the same amount of risk. So their, compensation that they require for risk increases. So time variation in credit risk premium is the main, driver of the reaction of credit spreads to macroeconomic shocks. So this is an important, result also because it can speak to, let's say theoretical models that try to understand how,

17:54 - the variation in credit spreads, can be explained over time. So turning to time varying credit risk premia is, a really good direction given the, and promising direction given the empirical findings that, that we have. the second question is that, is targeting to understand why the response of credit spreads is delayed. And, to answer this question, we, we try to provide evidence of the presence

18:24 - of information frictions in, in credit markets that could explain this delayed response. And, in simple words, information frictions just reflect the, the idea that there is not a full response of credit spreads when a shock hits, because the exact information, content is not immediately available, to all investors, at that point in time. So there is, there are frictions

18:48 - that prevent investors from accessing, the, the complete set of information at that point in time. It might require, a bit longer, and that's why, the response is delayed. to test for this empirically, we follow a methodology developed by Kion and Chenko in their 2012 paper. And essentially, the idea is pretty simple. you need to turn to measures of expectations of, of credit spreads, meaning, what investors expect about,

19:16 - credit spreads in the future, and check whether those measures respond indeed, with a delay to micro shocks relative to the realized credit spreads. If this is the case, the, the implication that's forecast error, meaning the error in the forecast between the realization relative to the realization, if this is the case forecast errors will be predictable, and therefore this will be a sign of the presence of information frictions in, in

19:43 - that specific, dimension. And, once we turn to the data, we do indeed find that, the impact of, macro shocks on expectations of forecast of credit spreads is indeed delayed relative to the realization, which is consistent with the idea that there, the, there is a scope, there is an important role of information frictions in credit markets, which can explain the delayed response of credit spreads.

20:11 - so this is essentially, wrapping up the, the part on, on why credit spreads are delayed. I want to mention, really briefly, other two tests that we do to better, provide, an understanding of the economic mechanisms and, and the robustness of, of our findings. First, we are able to show that, the response of credit spreads to macro shocks is much stronger when we focus on credit spreads of firms with high leverage, meaning that if you, if you, restrict the sample to firms,

20:44 - with high leverage, you see that the response is, a double, of credit spread to micros shocks relative to, for example, firms with low leverage. This is important because it indicates that firm leverage is an important amplifi amplification mechanism when it comes to understanding the impact of micros shocks on, on the economy through credit markets. And, the last exercise I wanna mention is, a sort of external validity exercise.

21:11 - We, we take our macro shocks and we, underst we try to figure out what is the impact of these shocks on popular proxies of equity risk premium. So we turn to the equity market and check whether the impact that we find on credit markets is actually reflected also in other, financial assets, which if these are truly, let's say exogenous and primitive macros shocks, it should be the case. And we did find that the impact of, of these macros,

21:39 - shocks on popular, equity risk pri proxies is, really similar to what we observe both in direction and magnitude in, in, in credit markets. So this is important, first of all, because it, let's say, reinforces the idea that we are looking at, a proper set of microeconomic shocks, but also, importantly because it shows a common source of variation of, discount rates in, credit markets and equity markets. And this is important

22:09 - because empirically, the correlation between equity risk, premier proxies and credit risk premier proxies, have been quite low, empirically, despite theoretically that should be, pretty high. Our paper shows that this is an identification problem. Once you can identify, a well, let's say shocks that affect, both markets at the same point in time and hold constant other confounding factors, this can actually resolve this puzzle that has been shown empirically when you look at the

22:40 - unconditional correlation between credit markets and, and equity markets. And we believe actually that this is, an interesting avenue also for, for future work. The findings that, that we have in our paper, can be, used, and extended in, in future research. And I'm going to give you just here, a few examples. So, the first part that, we believe is, is, is actually, useful for, for future,

23:09 - future research is the finding that the response of credit spreads to, to macroeconomic shocks is a combination of information rigidities, which explains, the, the low, the slow response and leverage related amplifications firms with higher leverage will, respond more to, to macroeconomic shocks. there are, to our knowledge, no macroeconomic models that include both of these components, when it comes to explaining the,

23:40 - the, the pricing of, of credit, of, in credit markets, and also the evolution of credit risk premium over time. So we believe that, developing models that incorporate these two features can go a long way into explaining the empirical facts, that we document, o overall. And of course, more broadly the fact that, we find an important, let's say a large impact on macroeconomic shocks on credit

24:07 - markets speaks to the fact that there are probably large feedbacks between credit markets and macroeconomic activity. As I said, the direction between credit markets and the macro economy has been well documented. We now show that also the reverse causality, direction, works pretty strongly. So models that incorporate the two-way interaction between the macro economy and the credit markets, and use this to understand, the contribution

24:34 - to business cycles, we believe that this remains a really important area for future research. And last, but not least, the fact that these shocks that we document, have a large impact on other markets too in equity markets that we have shown. I think it, it can open interesting avenues, to, to empirically measure common source of, of risk across different markets.

24:57 - And I think there is a, there is a lot of work that can be done, empirically in, into that direction.

Faculti AI can make mistakes