Popular in Professional

Human Rights Law

Human Rights and Corruption: Problems and Potential of Individualising a Systemic Problem

Max Planck Institute Anne Peters

Law

A statistical approach to law school citation rankings

University of Virginia Joshua Fischman

Education

High school change

Boston College Andrew Hargreaves

Constitutional Law

What is constitutional ideology?

University of Oxford Nick Barber

Civil Procedure Law

Politics, Identity, and Class Certification on the U.S. Courts of Appeals

University of Pennsylvania Sean Farhang

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

The not-so-hidden risks of ‘hidden-to-maturity’ accounting

View more insights in Professional - Accounting

View more figures from Stevens Institute of Technology

View the page of Zachary Feinstein

Read the Study

Next in Accounting

Relative Performance Evaluation and Strategic Peer‐Harming Disclosures

Matthew Bloomfield | University of Pennsylvania

On the EPA’s Radar: The Role of Financial Reports in Environmental Regulatory Oversight

Bin Li | Vanderbilt University

How can understanding ‘hidden-to-maturity’ accounting practices help prevent future bank collapses like the SVB meltdown in March 2023? This author presents a model analyzing balance sheet vulnerabilities and depositor run risks, providing insights into bank resilience and regulatory strategies.

Image courtesy of interviewee. November 20, 2024

Log-in or Sign-up to Faculti
Currently viewing this subject insight as a guest. You have insight(s) remaining for this month. Login to view 8000+ figures on the platform.

Generating Workspace...

Export as PDF

This video discusses a mathematical modeling project initiated in response to the collapse of Silicon Valley Bank (SVB) in March 2023. The authors analyze the bank's high percentage of uninsured deposits and the implications for bank runs, referring to classic models of banking behavior. They highlight that 90% of SVB's deposits exceeded the FDIC insurance limit, which contributed to a panic run on the bank, exacerbated by rapid information dissemination via social media.

Additionally, the text delves into the concept of "held to maturity" (HTM) assets, which are meant to be kept until maturity without needing to adjust their value based on market fluctuations. The authors emphasize the significance of properly managing these positions in light of the rising interest rates at the time, and the risks associated with having a large amount of these assets.

The model developed by the authors aims to evaluate the risk of bank runs by examining factors such as the percentage of uninsured deposits and the quality of the bank's assets. They found that SVB was managing its risk prudently until late 2022 but failed to adjust its HTM asset positions in time to mitigate risks associated with a changing economic environment.

The analysis extends beyond SVB to include other banks like PNC and First Republic Bank, noting that PNC effectively unwound its HTM assets and, as a result, was better positioned to withstand risks. The authors conclude that understanding a bank's accounting practices, especially regarding HTM versus available-for-sale assets, is crucial for assessing its risk level. They suggest that regulatory frameworks should consider a bank's specific context, including the proportion of insured vs. uninsured deposits, to prevent future failures like SVB’s.

Language:

Export as PDF

Export as PDF
Here are five key points from the video regarding the analysis of Silicon Valley Bank (SVB) and broader banking issues:

1. **Understanding Bank Risks**: The project began in the aftermath of Silicon Valley Bank's failure in March 2023, focusing on analyzing the high percentage of uninsured deposits and the accounting practices that may have contributed to the bank's collapse. The mathematicians aimed to structure their analysis to provide insights into what went wrong, particularly looking at the risks associated with uninsured deposits and the "held to maturity" assets.

2. **The Role of Uninsured Deposits and Bank Runs**: A significant point raised was that SVB had over 90% of its deposits uninsured, which created a vulnerability to bank runs. This situation contributed to panic and withdrawal behaviors exacerbated by social media communication, contrasting historical bank runs that spread through slower word-of-mouth.

3. **Accounting Standards Impact**: The video discusses the implications of accounting standards concerning "held to maturity" (HTM) assets versus "available for sale" assets. While HTM assets allow banks to avoid marking them to market value and thus mitigate short-term risks, this accounting can mask potential financial vulnerabilities, especially if the bank's asset composition is unbalanced.

4. **Risk Assessment Model**: The authors developed a model to assess the risk of bank runs by analyzing the composition of bank assets and the level of uninsured deposits. They noted that the optimization of HTM assets must consider both potential risks and rewards, recommending that banks keep their holdings within prudent levels to ensure stability under stress scenarios, such as rising interest rates.

5. **Regulatory Implications and Solutions**: The analysis suggests that regulatory changes, like ensuring all deposits or regulating the size of uninsured accounts, can mitigate the risk of bank runs. Notably, Massachusetts has laws mandating insurance on all bank deposits, contrasting with California, where SVB operated. This highlights the potential effectiveness of regulatory mechanisms in preventing bank failures.

Export as PDF
Here are some key quotes extracted from the video, focusing on the main ideas presented:

1. **On the impetus to study SVB:**
- "We want to understand how can we contribute to understanding what happened to SVB, what happened to a couple of other kind of mid-sized US banks at that time."

2. **Regarding uninsured deposits:**
- "SVB had very high percentage of uninsured deposits...90 plus percent of their deposits back in March, 2023 were uninsured."

3. **On the concept of bank runs:**
- "If there is no deposit insurance...people will run on the bank just based on fear."

4. **Describing the held-to-maturity (HTM) accounting:**
- "The idea is these are assets that you have...it's supposed to be intent and ability to hold these assets until they mature."

5. **Explaining the risk associated with HTM:**
- "There's also a risk to doing it...so we try to optimize between the risk and reward."

6. **Insights on Silicon Valley Bank's actions:**
- "Up until through 2021 Silicon Valley Bank was acting prudently...but it was not hitting the critical thresholds ultimately at the time of the collapse."

7. **Comparing other banks to SVB:**
- "What we found is PNC prior to the fourth quarter of 2022, unwound its entire held to maturity position...they were, according to our model, in incredibly safe compared to first Republic for Silicon Valley Bank."

8. **On the importance of accounting practices:**
- "How do the banks actually do their accounting? Because ultimately held to maturity versus available for sale is just accounting standards."

9. **Aftermath of the SVB failure:**
- "One key...is to ensure all deposits. If you don't limit the amount of the size of these accounts...you're immediately removing that desire to run on a bank."

10. **On regulatory considerations:**
- "There are jurisdictions within the United States that do actually prevent such a problem...but Silicon Valley Bank...is not subject to Massachusetts law."

These quotes capture the essence of the analysis regarding Silicon Valley Bank’s failure, the dynamics of bank runs, the implications of accounting practices, and regulatory considerations that could help prevent such failures in the future.

Export as PDF
Here are some articles that discuss themes similar to those in your video, focusing on the collapse of Silicon Valley Bank (SVB), issues related to uninsured deposits, bank runs, accounting standards (specifically held-to-maturity assets), and risk management in banking:

1. **The Collapse of Silicon Valley Bank: What We Know** - A comprehensive analysis of the events leading up to SVB's collapse, examining the high percentage of uninsured deposits and the role of social media in accelerating bank runs.

2. **Bank Runs: Lessons from Silicon Valley Bank** - An exploration of the historical convideo of bank runs, comparing SVB's situation to previous banking crises, and discussing theoretical models of bank runs based on the work of economists like Diamond and Dybvig.

3. **Risk Management in Banking: The Case of SVB and Other Mid-Sized Banks** - This article delves into the risk management practices of mid-sized banks, contrasting SVB's strategies with those of more stable institutions, and emphasizing the importance of understanding asset composition.

4. **Accounting Standards and Their Impact on Banking Stability** - An analysis of how different accounting treatments (held-to-maturity vs. available-for-sale) can influence the stability of banks, particularly in times of economic stress.

5. **Uninsured Deposits and Financial Instability: A Study of Recent Bank Failures** - This article investigates the implications of high levels of uninsured deposits for banks, focusing on SVB and similar institutions in the convideo of regulatory frameworks.

6. **The Role of Social Media in Modern Banking Crises** - A discussion on how social media has transformed the dynamics of bank runs, providing real-time information that can exacerbate financial panic among depositors.

7. **Regulatory Responses to Bank Failures: The Aftermath of SVB** - An overview of the regulatory changes implemented following SVB's collapse, including measures to insure all deposits and prevent future bank runs.

8. **Understanding Held-to-Maturity Assets: Risks and Rewards** - A detailed look at the implications of holding assets to maturity, including challenges in accurately assessing risk and the potential pitfalls during periods of rising interest rates.

These articles will provide a broader understanding of the issues discussed in your video, including theoretical models of bank behavior, real-world case studies of bank runs, and the implications of accounting practices on financial stability.

X-TIMESTAMP-MAP=LOCAL:00:00:00.000,MPEGTS:90000

1
00:00:00.075 --> 00:00:04.005
Then we really started this project, this was shortly

2
00:00:04.095 --> 00:00:07.765
after March, 2023, when SVB

3
00:00:08.295 --> 00:00:11.245
quite spectacularly, blew up.

4
00:00:11.665 --> 00:00:15.405
Uh, so we were reading all these reports about,

5
00:00:15.405 --> 00:00:18.885
the uninsured deposits that were held at SVB, at

6
00:00:19.425 --> 00:00:22.125
the actual accounting standards they were using about

7
00:00:22.265 --> 00:00:24.285
how they were hedging their risks or

8
00:00:24.395 --> 00:00:26.365
otherwise not hedging their risks.

9
00:00:27.185 --> 00:00:29.765
And we tried to say, we are mathematicians.

10
00:00:29.855 --> 00:00:31.725
We're coming at this from a

11
00:00:31.725 --> 00:00:33.645
mathematical modeling standpoint.

12
00:00:33.645 --> 00:00:37.325
We want to understand how can we contribute

13
00:00:37.905 --> 00:00:41.725
to understanding what happened to SVB, what happened

14
00:00:41.725 --> 00:00:43.445
to a couple of other kind

15
00:00:43.445 --> 00:00:45.725
of mid-sized US banks at that time.

16
00:00:46.625 --> 00:00:51.005
And also to understand what could be done on other,

17
00:00:51.265 --> 00:00:52.405
institutions.

18
00:00:52.465 --> 00:00:56.005
How could we look at other institutions more generally?

19
00:00:56.745 --> 00:00:59.685
So as we're looking at SVB,

20
00:00:59.995 --> 00:01:03.245
they had very high percentage of uninsured deposits,

21
00:01:03.245 --> 00:01:06.245
which kind of the classical bank run models.

22
00:01:06.425 --> 00:01:08.165
If you go back to Diamond Dip Vig,

23
00:01:08.455 --> 00:01:13.365
where they won the Nobel Prize in 2022 based on their,

24
00:01:13.915 --> 00:01:16.845
bank run, study back in the eighties.

25
00:01:17.385 --> 00:01:20.205
Uh, if we're really looking at that,

26
00:01:20.275 --> 00:01:23.885
that was saying there's no in, deposit insurance.

27
00:01:24.305 --> 00:01:27.365
So people will run on the bank just based on fear.

28
00:01:28.185 --> 00:01:31.485
Uh, then in, subsequent works,

29
00:01:31.745 --> 00:01:34.325
we really found if there are,

30
00:01:34.425 --> 00:01:37.325
if there is deposit insurance, then

31
00:01:37.905 --> 00:01:41.165
we completely eliminate this risk of a bank run

32
00:01:41.165 --> 00:01:44.725
because ultimately, I will keep my money if it,

33
00:01:44.725 --> 00:01:45.965
if it's in the bank and the bank

34
00:01:45.975 --> 00:01:47.405
fails, I still have my money.

35
00:01:47.845 --> 00:01:50.005
I have no need to run on the bank

36
00:01:50.265 --> 00:01:52.085
and get my money out first.

37
00:01:53.255 --> 00:01:56.315
But when we look at SVB, all of a sudden,

38
00:01:56.775 --> 00:02:01.555
90 plus percent of their deposits back in March, 2023

39
00:02:02.145 --> 00:02:03.235
were uninsured.

40
00:02:03.305 --> 00:02:06.155
Were above the $250,000 limit

41
00:02:06.305 --> 00:02:09.755
that is in the FDIC insurance within the United States.

42
00:02:10.775 --> 00:02:13.635
So this contributed back to those

43
00:02:14.235 --> 00:02:15.995
original models on bank runs,

44
00:02:16.005 --> 00:02:19.395
where now you did have the lines of people trying

45
00:02:19.395 --> 00:02:20.515
to get their money out

46
00:02:20.735 --> 00:02:25.515
before SVB collapsed, really in that last week.

47
00:02:26.095 --> 00:02:28.995
Uh, and this was spread through social media as opposed

48
00:02:29.195 --> 00:02:33.715
to previously just being word of mouth slower almost.

49
00:02:34.455 --> 00:02:39.235
Um, but the other big aspect was the HTM,

50
00:02:39.255 --> 00:02:42.755
the held to maturity or hidden to maturity assets.

51
00:02:43.295 --> 00:02:46.515
Uh, so officially it is held to maturity, though,

52
00:02:46.655 --> 00:02:48.995
in the academic literature, people are starting

53
00:02:49.055 --> 00:02:51.955
to use this hidden to maturity, pejorative.

54
00:02:52.215 --> 00:02:56.875
The idea being when you have assets on your balance sheet,

55
00:02:57.185 --> 00:03:01.125
typically in finance, you what's called mark to market.

56
00:03:02.005 --> 00:03:05.325
Whatever the current outstanding price of that asset is, is

57
00:03:05.325 --> 00:03:09.485
what you need to say it is worth in your balance sheet,

58
00:03:09.635 --> 00:03:11.845
held to maturity is different.

59
00:03:12.065 --> 00:03:15.005
The idea is these are assets that you have.

60
00:03:15.185 --> 00:03:19.245
The intent and ability according to the rules, it's supposed

61
00:03:19.245 --> 00:03:23.005
to be intent and ability to hold these assets

62
00:03:23.295 --> 00:03:24.365
until they mature.

63
00:03:24.865 --> 00:03:26.725
So if this is government debt,

64
00:03:26.825 --> 00:03:28.565
it could be 30 year government debt.

65
00:03:28.625 --> 00:03:29.805
So the idea is you are going

66
00:03:29.805 --> 00:03:31.365
to hold this asset for 30 years.

67
00:03:31.865 --> 00:03:35.605
You have no intent to sell it

68
00:03:36.025 --> 00:03:37.645
before that time is up.

69
00:03:39.025 --> 00:03:42.125
The reason for this, the reason that we allow these banks

70
00:03:42.145 --> 00:03:46.245
to do this is it allows 'em to smooth out,

71
00:03:46.295 --> 00:03:47.525
their cash flows.

72
00:03:47.945 --> 00:03:51.365
So now as prices fluctuate a little, they're not going

73
00:03:51.365 --> 00:03:53.365
to have to keep remarking these assets.

74
00:03:53.365 --> 00:03:56.165
They're not going to have to hedge these risks

75
00:03:56.165 --> 00:03:59.045
because they say, I don't care about this short term risk.

76
00:03:59.405 --> 00:04:03.365
I am just worried about this long term, problem.

77
00:04:03.635 --> 00:04:07.285
There's really two aspects to the model that we developed.

78
00:04:07.785 --> 00:04:10.645
The first aspect is really looking at

79
00:04:10.835 --> 00:04:12.405
what is the size of the bank run.

80
00:04:12.405 --> 00:04:15.245
So you have the balance sheet of these,

81
00:04:15.625 --> 00:04:19.085
of whatever bank you're looking at Silicon Valley Bank,

82
00:04:19.225 --> 00:04:22.325
or you can look at a more stable bank like PNC.

83
00:04:22.375 --> 00:04:25.725
We've done some, subsequent work where we've started

84
00:04:25.725 --> 00:04:27.925
to look at other institutions,

85
00:04:28.225 --> 00:04:31.645
and you can basically say, what is the run risk

86
00:04:31.875 --> 00:04:34.365
that they have based on how much

87
00:04:34.365 --> 00:04:37.765
of their deposits are uninsured, how much of it, how much

88
00:04:37.765 --> 00:04:40.205
of their assets are held

89
00:04:40.205 --> 00:04:42.045
to maturity versus available for sale?

90
00:04:42.045 --> 00:04:43.645
What are the quality of these assets?

91
00:04:44.225 --> 00:04:46.045
And we can really use that to say,

92
00:04:46.275 --> 00:04:48.285
what are the risks of the run?

93
00:04:48.395 --> 00:04:51.485
What would happen during this event?

94
00:04:52.225 --> 00:04:55.125
Uh, the second aspect of the model is saying,

95
00:04:55.195 --> 00:04:58.125
what if we just know the total amount of assets

96
00:04:58.315 --> 00:05:00.005
that this bank holds?

97
00:05:00.545 --> 00:05:03.405
So not saying how much of it is available for sale

98
00:05:03.425 --> 00:05:06.525
or how much of it is held to maturity as the bank are,

99
00:05:06.525 --> 00:05:07.605
is reporting, but just saying,

100
00:05:07.605 --> 00:05:09.805
what's the sum total of the assets?

101
00:05:10.585 --> 00:05:13.045
And then we wanted to figure out, we wanted to study

102
00:05:13.435 --> 00:05:16.405
what is the, in some sense, optimal amount

103
00:05:16.505 --> 00:05:20.605
of health maturity that really gets the aspect of the intent

104
00:05:20.665 --> 00:05:23.645
and ability to hold those assets to maturity.

105
00:05:24.385 --> 00:05:28.045
So we wanted to look at what kind of stress,

106
00:05:28.985 --> 00:05:33.465
so the bank has some reasonable stress so that we kind

107
00:05:33.465 --> 00:05:35.625
of get this idea of the intent and ability.

108
00:05:35.725 --> 00:05:37.545
So it's a reasonable stress scenario

109
00:05:37.615 --> 00:05:39.705
that hits the financial market.

110
00:05:39.805 --> 00:05:41.385
In the case of Silicon Valley Bank,

111
00:05:41.575 --> 00:05:43.785
that would be the rising interest rate environment

112
00:05:43.785 --> 00:05:47.305
that was prevailing at the time, to combat the inflation

113
00:05:47.375 --> 00:05:49.425
that we had in the post covid economy.

114
00:05:49.755 --> 00:05:53.625
Based on that stress, we want to see that they would not,

115
00:05:53.695 --> 00:05:56.825
that no bank would have that, this bank would not have to

116
00:05:57.805 --> 00:06:01.045
Remark its assets, it's held to maturity assets

117
00:06:01.425 --> 00:06:04.205
and make them available for sale like Silicon Valley

118
00:06:04.235 --> 00:06:05.445
Bank was forced to do.

119
00:06:05.905 --> 00:06:09.085
Uh, and other banks, potentially would be forced to do.

120
00:06:09.145 --> 00:06:12.285
So. Again, there's the advantage to held to maturity,

121
00:06:12.285 --> 00:06:15.965
but we wanted to say that there's also a risk to doing it.

122
00:06:16.745 --> 00:06:20.925
Uh, so we try to optimize between the risk and reward.

123
00:06:21.155 --> 00:06:25.125
When we ran this, this data on Silicon Valley Bank,

124
00:06:25.595 --> 00:06:28.365
what we found is that up until

125
00:06:28.995 --> 00:06:33.965
through 2021 Silicon Valley Bank was acting prudently.

126
00:06:34.225 --> 00:06:36.165
The amount of held to maturity assets

127
00:06:36.195 --> 00:06:40.285
that they had on their balance sheet was actually

128
00:06:40.895 --> 00:06:43.325
below the amount that our optimization was saying

129
00:06:43.355 --> 00:06:45.125
that they could hold.

130
00:06:45.665 --> 00:06:48.885
Um, in contrast, once we got into 2022

131
00:06:48.945 --> 00:06:50.685
and early 2023

132
00:06:50.685 --> 00:06:51.965
before they collapsed,

133
00:06:52.945 --> 00:06:55.125
the optimization was saying they needed

134
00:06:55.125 --> 00:06:56.645
to unwind that position.

135
00:06:57.385 --> 00:07:01.045
It was much too large an amount of health maturity.

136
00:07:01.625 --> 00:07:04.005
And in contrast, they were not doing that in practice.

137
00:07:04.785 --> 00:07:08.925
So while they were prudent, there was really,

138
00:07:09.385 --> 00:07:12.245
we found that there was not this big,

139
00:07:12.245 --> 00:07:15.405
like you could see the runup of this risk,

140
00:07:15.465 --> 00:07:17.845
but it was not hitting the critical thresholds

141
00:07:18.605 --> 00:07:20.525
ultimately at the time of the collapse

142
00:07:20.755 --> 00:07:22.645
that threshold was being exceeded.

143
00:07:22.985 --> 00:07:25.885
Uh, so we did find in the case of Silicon Valley Bank

144
00:07:26.115 --> 00:07:30.005
that there were these big risks, that

145
00:07:30.875 --> 00:07:34.325
were happening just before their collapse did occur.

146
00:07:34.625 --> 00:07:36.725
So we've subsequently,

147
00:07:36.995 --> 00:07:40.325
used the exact same model taking data from other banks.

148
00:07:40.745 --> 00:07:45.205
Uh, we've been looking at First Republic Bank, PNC,

149
00:07:45.205 --> 00:07:47.765
Pacific West, and a couple of other institutions,

150
00:07:48.545 --> 00:07:51.605
and just to try to see do similar,

151
00:07:51.715 --> 00:07:54.285
effects occur in banks, other banks

152
00:07:54.285 --> 00:07:55.805
that failed like First Republic

153
00:07:56.545 --> 00:08:00.405
or banks that did not seem to be at risk like PNC.

154
00:08:00.825 --> 00:08:04.885
And what we found is PNC prior to the fourth quarter

155
00:08:04.905 --> 00:08:09.365
of 2022, unwound, its entire held to maturity position.

156
00:08:10.025 --> 00:08:12.845
So they were entirely unavailable for sale.

157
00:08:12.955 --> 00:08:17.685
They were, according to our model, in incredibly,

158
00:08:17.915 --> 00:08:21.325
safe compared to, first Republic

159
00:08:21.745 --> 00:08:26.165
for Silicon Valley Bank, which had these growing,

160
00:08:26.825 --> 00:08:29.805
hidden to maturity, positions.

161
00:08:30.275 --> 00:08:33.445
What this is really telling us when we look at these models,

162
00:08:33.515 --> 00:08:37.165
when we really try to build these constructions about risk

163
00:08:37.745 --> 00:08:40.645
is to look at not just, well,

164
00:08:40.645 --> 00:08:42.325
what's the total asset composition?

165
00:08:42.905 --> 00:08:45.885
But how do the banks actually do their accounting?

166
00:08:46.105 --> 00:08:49.005
Uh, because ultimately held to maturity versus available

167
00:08:49.145 --> 00:08:51.565
for sale is just accounting standards.

168
00:08:52.025 --> 00:08:54.485
Uh, so the way that this accounting is done

169
00:08:55.065 --> 00:08:57.325
really should be bank dependent,

170
00:08:57.345 --> 00:09:01.325
should depend on the different institution, based on

171
00:09:01.425 --> 00:09:04.205
how much uninsured versus insured deposits they have.

172
00:09:04.555 --> 00:09:07.765
Because if most of the deposits are insured,

173
00:09:08.275 --> 00:09:10.725
then there is not the same risk of a run.

174
00:09:11.145 --> 00:09:14.405
So there is easier to have this ability

175
00:09:14.945 --> 00:09:17.365
to keep the assets until maturity.

176
00:09:17.905 --> 00:09:20.325
Uh, in practice, what we found

177
00:09:20.385 --> 00:09:23.205
and others have reported, the intent

178
00:09:23.265 --> 00:09:27.485
and ability to hold held to maturity, to maturity, seems

179
00:09:27.485 --> 00:09:29.205
to be on paper,

180
00:09:29.205 --> 00:09:33.165
but not in practice in terms of regulatory changes

181
00:09:33.275 --> 00:09:35.365
that we were kind of investigating.

182
00:09:36.105 --> 00:09:38.765
One key, one that like jumps out

183
00:09:38.765 --> 00:09:40.405
that actually was done immediately

184
00:09:40.405 --> 00:09:42.805
after the failure of Silicon Valley Bank is

185
00:09:42.825 --> 00:09:44.605
to ensure all deposits.

186
00:09:44.705 --> 00:09:48.685
If you don't limit the amount of, the size

187
00:09:48.685 --> 00:09:51.645
of these accounts in, the amount that they'd be able

188
00:09:51.645 --> 00:09:55.325
to recover, you're immediately removing that desire

189
00:09:55.425 --> 00:09:56.485
to run on a bank.

190
00:09:56.865 --> 00:09:59.605
Uh, and actually within the United States, Massachusetts,

191
00:09:59.845 --> 00:10:02.245
I believe is the only state in the United States

192
00:10:02.515 --> 00:10:06.245
that actually requires, insurance on all bank deposits,

193
00:10:06.625 --> 00:10:08.605
not just the FDIC insurance.

194
00:10:09.385 --> 00:10:12.765
Uh, so there are, jurisdictions within the United States

195
00:10:12.795 --> 00:10:16.205
that do actually prevent such a, problem.

196
00:10:16.625 --> 00:10:20.045
But Silicon Valley Bank, famously, primarily in California,

197
00:10:20.635 --> 00:10:24.045
does not, is not subject to Massachusetts, law.

198
00:10:24.465 --> 00:10:27.365
Uh, but otherwise it really comes down to,

199
00:10:27.505 --> 00:10:30.125
and it's hard to judge what intent

200
00:10:30.145 --> 00:10:33.925
and ability, how you can enforce, those rules.

201
00:10:34.745 --> 00:10:36.965
But since they are the official rules

202
00:10:37.025 --> 00:10:39.565
for when you should be doing held to maturity,

203
00:10:39.565 --> 00:10:43.245
accounting, there should be some way to actually,

204
00:10:43.615 --> 00:10:47.045
check whether someone can have that intent

205
00:10:47.185 --> 00:10:51.565
and ability, to follow the rules and regulations.

Analyzing the SVB Collapse: Insights from Mathematical Models

00:00:00 - Then we really started this project, this was shortly after March, 2023, when SVB quite spectacularly, blew up. so we were reading all these reports about, the uninsured deposits that were held at SVB, at the actual accounting standards they were using about how they were hedging their risks or otherwise not hedging their risks. And we tried to say, we are mathematicians. We're coming at this from a

00:00:33 - We want to understand how can we contribute to understanding what happened to SVB, what happened to a couple of other kind of mid-sized US banks at that time. And also to understand what could be done on other, institutions. How could we look at other institutions more generally? So as we're looking at SVB, they had very high percentage of uninsured deposits,

00:01:06 - If you go back to Diamond Dip Vig, where they won the Nobel Prize in 2022 based on their, bank run, study back in the eighties. if we're really looking at that, that was saying there's no in, deposit insurance. So people will run on the bank just based on fear. then in, subsequent works, we really found if there are,

Understanding Bank Runs: The Case of SVB and Uninsured Deposits

00:01:37 - we completely eliminate this risk of a bank run because ultimately, I will keep my money if it, if it's in the bank and the bank fails, I still have my money. I have no need to run on the bank and get my money out first. But when we look at SVB, all of a sudden, 90 plus percent of their deposits back in March, 2023 were uninsured. Were above the $250,000 limit

00:02:10 - So this contributed back to those original models on bank runs, where now you did have the lines of people trying to get their money out before SVB collapsed, really in that last week. and this was spread through social media as opposed to previously just being word of mouth slower almost. but the other big aspect was the HTM,

00:02:43 - so officially it is held to maturity, though, in the academic literature, people are starting to use this hidden to maturity, pejorative. The idea being when you have assets on your balance sheet, typically in finance, you what's called mark to market. Whatever the current outstanding price of that asset is, is what you need to say it is worth in your balance sheet,

Assessing Bank Run Risks and Asset Management

00:03:12 - The idea is these are assets that you have. The intent and ability according to the rules, it's supposed to be intent and ability to hold these assets until they mature. So if this is government debt, it could be 30 year government debt. So the idea is you are going to hold this asset for 30 years. You have no intent to sell it before that time is up. The reason for this, the reason that we allow these banks

00:03:46 - their cash flows. So now as prices fluctuate a little, they're not going to have to keep remarking these assets. They're not going to have to hedge these risks because they say, I don't care about this short term risk. I am just worried about this long term, problem. There's really two aspects to the model that we developed. The first aspect is really looking at what is the size of the bank run.

00:04:15 - of whatever bank you're looking at Silicon Valley Bank, or you can look at a more stable bank like PNC. We've done some, subsequent work where we've started to look at other institutions, and you can basically say, what is the run risk that they have based on how much of their deposits are uninsured, how much of it, how much of their assets are held to maturity versus available for sale? What are the quality of these assets?

Analyzing Bank Asset Management Risks and Strategies

00:04:46 - what are the risks of the run? What would happen during this event? the second aspect of the model is saying, what if we just know the total amount of assets that this bank holds? So not saying how much of it is available for sale or how much of it is held to maturity as the bank are, is reporting, but just saying, what's the sum total of the assets? And then we wanted to figure out, we wanted to study what is the, in some sense, optimal amount

00:05:20 - and ability to hold those assets to maturity. So we wanted to look at what kind of stress, so the bank has some reasonable stress so that we kind of get this idea of the intent and ability. So it's a reasonable stress scenario that hits the financial market. In the case of Silicon Valley Bank, that would be the rising interest rate environment that was prevailing at the time, to combat the inflation

00:05:49 - Based on that stress, we want to see that they would not, that no bank would have that, this bank would not have to Remark its assets, it's held to maturity assets and make them available for sale like Silicon Valley Bank was forced to do. and other banks, potentially would be forced to do. So. Again, there's the advantage to held to maturity, but we wanted to say that there's also a risk to doing it. so we try to optimize between the risk and reward.

SVB's Prudent Management and the Shift Before Collapse

00:06:25 - what we found is that up until through 2021 Silicon Valley Bank was acting prudently. The amount of held to maturity assets that they had on their balance sheet was actually below the amount that our optimization was saying that they could hold. in contrast, once we got into 2022 and early 2023 before they collapsed,

00:06:55 - to unwind that position. It was much too large an amount of health maturity. And in contrast, they were not doing that in practice. So while they were prudent, there was really, we found that there was not this big, like you could see the runup of this risk, but it was not hitting the critical thresholds ultimately at the time of the collapse that threshold was being exceeded.

00:07:26 - that there were these big risks, that were happening just before their collapse did occur. So we've subsequently, used the exact same model taking data from other banks. we've been looking at First Republic Bank, PNC, Pacific West, and a couple of other institutions, and just to try to see do similar, effects occur in banks, other banks

Assessing Bank Risk Through Asset Management and Accounting Practices

00:07:56 - or banks that did not seem to be at risk like PNC. And what we found is PNC prior to the fourth quarter of 2022, unwound, its entire held to maturity position. So they were entirely unavailable for sale. They were, according to our model, in incredibly, safe compared to, first Republic for Silicon Valley Bank, which had these growing,

00:08:30 - What this is really telling us when we look at these models, when we really try to build these constructions about risk is to look at not just, well, what's the total asset composition? But how do the banks actually do their accounting? because ultimately held to maturity versus available for sale is just accounting standards. so the way that this accounting is done really should be bank dependent,

00:09:01 - how much uninsured versus insured deposits they have. Because if most of the deposits are insured, then there is not the same risk of a run. So there is easier to have this ability to keep the assets until maturity. in practice, what we found and others have reported, the intent and ability to hold held to maturity, to maturity, seems

Regulatory Changes Post-SVB Collapse

00:09:29 - but not in practice in terms of regulatory changes that we were kind of investigating. One key, one that like jumps out that actually was done immediately after the failure of Silicon Valley Bank is to ensure all deposits. If you don't limit the amount of, the size of these accounts in, the amount that they'd be able to recover, you're immediately removing that desire to run on a bank. and actually within the United States, Massachusetts,

00:10:02 - that actually requires, insurance on all bank deposits, not just the FDIC insurance. so there are, jurisdictions within the United States that do actually prevent such a, problem. But Silicon Valley Bank, famously, primarily in California, does not, is not subject to Massachusetts, law. but otherwise it really comes down to, and it's hard to judge what intent

00:10:47 - But since they are the official rules for when you should be doing held to maturity, accounting, there should be some way to actually, check whether someone can have that intent and ability, to follow the rules and regulations.

Faculti AI can make mistakes

Faculti

Popular in Professional

Human Rights and Corruption: Problems and Potential of Individualising a Systemic Problem

A statistical approach to law school citation rankings

High school change

What is constitutional ideology?

Politics, Identity, and Class Certification on the U.S. Courts of Appeals

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

The Role of Education in Reducing Racial Inequality

The not-so-hidden risks of ‘hidden-to-maturity’ accounting

Next in Accounting

Guide

Platform and Category Pages

Contact Us

FAQs

What is Faculti?

How do you select interviews?

Interview Process

Page Structure

Popular in Professional

Guide

Platform and Category Pages

Contact Us

FAQs

What is Faculti?

How do you select interviews?

Interview Process

Accessibility Options

Contrast Adjustment

Text Size Adjustment

Text Spacing Adjustment

Line Height Adjustment

Cursor Size Adjustment

Page Structure Options

Page Structure