0:10

Hello, and welcome to another episode

0:12

of the Odd Lots podcast. I'm Joe

0:14

Wisenthal and I'm Tracy Allow. So

0:17

Tracy, we obviously, I think

0:19

have a sense of why Silicon

0:22

Valley Bank failed. We just published a really

0:24

good episode with Dan Davis, like sort of

0:26

like talking about where things went wrong

0:28

on the sort of deposit side and

0:31

failing to balance assets and liabilities

0:33

and the issues of strengths and weaknesses

0:35

of the business model. But then the

0:37

other question, there are many, many more questions

0:39

beyond just like why they failed. Yes,

0:42

I mean some of the big ones that are emerging

0:45

are where were the regulators right

0:47

it? You know, people were already

0:50

analyzing spb's balance sheet,

0:52

you know, certainly the week before

0:54

it collapsed and for some time

0:56

before that, and you could see these vulnerabilities

1:00

when it comes to duration exposure.

1:02

And again that's something we talked about with Dan Davies.

1:04

And then the other thing I guess

1:06

just in general is

1:09

it's not like bank failures

1:11

are that unusual over

1:13

the course of history. So this is the

1:16

first big one, I guess since

1:18

the two thousand and eight financial crisis.

1:20

And so it's obviously garnering a lot of

1:22

attention, But we do have

1:24

bank failures throughout history

1:27

from time to time, and it kind of begs

1:29

the question of, well, if we're going

1:31

to keep having them in different

1:33

ways, and if the government or

1:35

the Federal Reserve are going to keep coming in

1:38

and rescuing them in various

1:40

ways, should we maybe do something

1:42

to the system to make it different.

1:45

What were the failures and can we at

1:47

least you know, we're always going to be fighting the

1:49

last War cliche obviously, but

1:52

what does it tell us about weaknesses in the system?

1:55

And there may be things that we could do, And of course there's

1:57

things that people are talking about on the sort of written

2:00

law side, which is like, Okay,

2:02

maybe we need more banks to

2:05

have greater liquidity, ability

2:07

to meet withdrawals, etc. I think some

2:10

of the smaller, more regional banks don't have

2:12

a stringent requirements on that front

2:14

as the really large banks. And then people

2:16

are talking about supervision, and I

2:18

don't think supervision gets as much attention

2:21

as the sort of written laws, but it's essentially,

2:24

well, why did the supervisors the

2:26

bank regulators allow the bank

2:28

to create this confluence of risks,

2:30

this big like sort of very specific mismatch

2:33

between the nature of its assets, the

2:36

nature of its deposit base that allowed

2:38

it to unravel really quickly. Absolutely,

2:40

and then of course, with the FED announcing

2:43

this new facility, which is quite

2:45

a dramatic one, you have a question of, Okay,

2:47

if we're just going to guarantee all

2:50

the US bank deposits out there,

2:52

then should we maybe make a more

2:55

fundamental change to the banking industry

2:58

itself? Right? And actually, Matt

3:00

Klein over at the Overshoot, I

3:02

think he tweeted this, but he had that great

3:05

first line in his most recent newsletter

3:07

about how basically banks

3:09

are these private investment

3:12

funds that are grafted on top

3:14

of critical infrastructure, and

3:16

that structure is designed

3:18

to extract subsidies from the rest of

3:20

society by basically threatening

3:23

people with banking crises whenever

3:25

one of them is allowed to fail. And we saw

3:28

that last week, right, We saw especially

3:31

a bunch of vcs coming out and saying, if

3:33

you don't rescue all the SVB depositors

3:36

right now, this is going to happen to all the

3:38

banks. And so you kick off that, you know,

3:41

privatization of profits versus publicization

3:44

of loss's argument over and

3:46

over and over again. Well, that's been like really

3:49

clear in this particular episode,

3:51

like there's something about this story really

3:54

raises some like uncomfortable days because it's not

3:56

coming in like a wholesale financial

3:59

collapse that's related to the collapse

4:01

of the economy like in two thousand and eight or two

4:03

thousand nine. It's like it's this very specific

4:05

industry that sort of got it in trouble.

4:07

Anyway, we could go on and on,

4:10

but I'm excited. We do have the

4:12

perfect guest for us to talk about the

4:14

role of regulator, the role of regulatory

4:17

failure, the role of the FED in all of

4:19

this in the history of banking, and how we got here.

4:21

We're gonna be speaking to a Levmanon Here's

4:23

a professor at Columbia Law School written

4:26

a lot about the FED and regulation. So Lev,

4:28

thank you so much for joining us, Thank you so much

4:30

for having me. So just you know, very

4:33

top line view. You

4:35

know, what would you say was the main regulatory

4:38

failure with SIV with

4:41

SVB, So yeah,

4:46

yeah, we can distinguish between

4:48

maybe regulation bright line rules

4:50

yeah, put down in advance, and sort of supervision,

4:52

yes, which is discretionary

4:55

safety and soundness oversight by by

4:57

examiners and federal

5:00

federal regulators, federal officials, and

5:03

both the sort of bright line rules and the sort

5:05

of supervision failed here. Yeah, there was an overreliance

5:07

on the bright line rules and a failure to do the

5:09

discretionary oversight, the safety and soundness

5:12

oversight effectively. So on the bright line

5:14

rules side, SVB

5:16

figured out a way to take additional risk

5:18

without holding additional capital. Because

5:22

of what's called the risk capital

5:24

rules, treasury securities are

5:26

risk weighted zero. That means that a bank has

5:28

to hold zero equity against

5:31

their treasury positions. And so SEP

5:34

was able to go and buy a lot

5:36

of long dated treasuries and actually build up quite a bit

5:38

of interest rate risk without that being reflected in

5:40

the capital that was required of

5:42

them under the under the bright line rules, under

5:44

the under the regulatory framework. Now

5:47

we have a whole supervisory framework

5:49

that's designed to deal with these sorts

5:52

of holes in the rules. Everybody

5:54

knows that the rules are insufficiently precise,

5:57

and in fact, we didn't even have the rules

5:59

until the nineteen eighties in meaningful sense.

6:01

We used to just do discretionary safety

6:04

and soundness oversight. And so the real question here in

6:06

some sense is how come the supervisors

6:08

didn't pick up on the fact that SVB

6:10

had gamed the rules to take on

6:12

a lot of interest rate risk without

6:14

holding an adequate amount of capital against

6:16

it. It's a pretty obvious maneuver.

6:20

It's not nearly as complex as some

6:22

of the maneuvers the gaming

6:24

and it's not novel. It's not novel.

6:26

Yeah, this is an old move. It's

6:28

not like they camp with something new. You would

6:31

think any seasoned

6:33

supervisor looking at the balance

6:35

sheet could pick up on this pretty

6:38

quickly. So so what happened,

6:40

I think is exactly the right question to

6:42

be asking, and I think

6:44

the answer requires maybe

6:48

And this might just be my approach

6:50

to these issues twenty thirty years worth of history

6:52

to understand, because basically,

6:55

I think contemporary supervision is

6:58

broken in some sense and

7:00

this is a manifestation of that. Okay,

7:03

um, well, I'm I'm just gonna go ahead

7:06

and buy and say, please, please

7:08

give us the you know, thirty or forty years

7:10

of banking history building up to those.

7:12

Yeah, So let me. I'll say why I think it's broken, and then

7:14

I'll tell you how we got sure, how it got so broken.

7:17

So it's broken because

7:20

outside of the stress testing

7:22

framework, which I think we should definitely talk

7:24

more about. Supervisors

7:27

primarily now focus

7:30

on process and procedures.

7:33

Our insight into what actually goes on in supervision

7:35

is very limited, and that's

7:37

because supervisory

7:40

materials are all confidential

7:43

and can't be disclosed by the bank

7:45

and aren't disclosed by by the regulators,

7:48

and often never made public. So

7:50

there's a lot of opacity into into

7:53

what supervisors are actually doing. But it's it's fairly

7:55

obvious, and I'll go through a few examples

7:57

that what supervisors today tend

7:59

to focus on is the process, and so

8:02

they will look to see does the bank

8:04

have a good risk management process,

8:07

does it have the right board committees,

8:09

does it have the right management committees

8:12

looking at its risk decisions?

8:14

Are there three lines of defense? And

8:17

if the supervisors see the

8:19

requisite process, they

8:21

are very reluctant to make judgments

8:24

about the actual decisions that are coming

8:26

out of that process, and so they

8:28

don't want to impose. They are reluctant

8:30

to impose their own view. Oh,

8:33

that is excessive risk, that

8:35

is too much interest rate risk

8:38

as opposed to we like the

8:40

procedures that you're set up to manage interest rate risk.

8:42

Just to be clear, though they certainly could

8:45

under current law, they're allowed

8:47

to say they're supposed to. Arguably,

8:50

that's what current law is about, and the process

8:52

approach being grafted onto this

8:54

is an innovation. The purpose

8:57

of safety and soundness law is really very

8:59

much to address risk,

9:02

and the process angle is born

9:04

of the view that the best way to

9:06

do that, the most efficient way to do

9:08

that across a massive banking system,

9:11

is just to make sure that the procedures are good. If the procedures

9:13

are good, the problem will sort of take care

9:15

of itself. So as long as

9:17

you see the bank kind of debating

9:20

its risk exposure internally,

9:23

which you know seems to have been the case at

9:25

SVB, and I know I brought this up in the previous

9:27

episode, but you know, there were some internal

9:30

documents which I've seen where they're talking

9:32

about interest rate exposure and they're

9:34

debating it, you know, with their asset liability

9:37

committee and presumably with their risk

9:39

specialists. But if they come

9:41

to the conclusion that actually we're okay,

9:44

the regulators are just going to look at that and

9:47

take out on face value because the process

9:49

is there, and they assume that, you know, the

9:51

bank is kind of doing what it should be doing. Yeah, exactly,

9:54

if you're not one of the big gesips, the global

9:56

systemically important banks, and you're not

9:58

in the stress testing route game, I

10:01

think that is what tends

10:03

to happen. It doesn't have to happen, As Joe said,

10:05

they're certainly going to be examples where supervisors

10:08

exercise them independent judgment. But

10:10

there is a tendency still in

10:13

the supervisory process to look

10:15

at compliance with the rules, to check

10:17

for processes, and if you see

10:19

compliance with the rules and you see processes

10:21

in place, to give the bank

10:25

a clean bill of health, as it were. And

10:27

so the question is how did we get to this place, because

10:29

actually this was an innovation. At

10:32

one point, we used to do safety

10:35

and soundness substantive supervision

10:37

without much bright line rules at

10:39

all. Capital rules date

10:42

only to the mid eighties, and

10:44

this focus on process is

10:47

really an innovation from the nineteen nineties.

10:49

And so it's part of what I

10:51

think is the toxic brew of regulatory

10:54

supervisory policies that brought us the

10:56

two thousand and eight crisis. And

10:58

we still sort of have supervision

11:00

guided by this nineties

11:03

approach, and I think the SVB

11:05

failure is one of several

11:07

really significant examples of

11:10

post two thousand and eight supervisory

11:12

failure where the supervisors

11:15

are still focused on process

11:18

and too unwilling to

11:20

make substantive judgments, reflecting

11:23

the sort of approaches that were developed by

11:25

primarily the greenspan fed but also the Ludwig

11:27

occ in the nineties. Can you explain

11:30

what it was or what happened in the nineties,

11:33

was a directive that came down what

11:35

caused this philosophical shift or just

11:37

maybe mechanical shift in the approach to supervisory

11:40

Let me start with the eighties, actually just

11:42

with the birth of the capital rules. So

11:44

what's the baseline. So supervision, safety

11:47

and signed supervision dates all the way back

11:49

to the nineteenth century, and so the way that the

11:51

government's managed the incentive

11:53

misalignment between bank shareholders

11:55

and managers and bank depositors

11:57

and the public has been through

12:00

discretionary supervisory

12:03

oversight safety and soundness oversight. That's

12:05

been the bywords of federal

12:07

law since the nineteen thirties. And

12:10

the way that supervisors would do their job

12:12

is they would make judgments about the riskiness

12:14

of banks assets and the riskiness

12:16

of banks leverage and the amount

12:19

of capital and they

12:21

would write letters, and they

12:23

would job bone and they would

12:26

take enforcement actions. They would issue

12:28

cease and desist orders if they thought banks were under

12:30

capitalized, or like in the case of Silicon Valley

12:32

Bank, they would tell Silicon

12:34

Valley Bank that it had to shorten its duration risk.

12:37

That's what supervisors would do. In

12:39

the eighties. You have a

12:41

moment that's quite similar to today in

12:44

that the banking system business model comes

12:46

under a lot of pressure for macroeconomic

12:48

reasons. Inflation goes up and then

12:50

interest rates go way up because of the Vulker

12:52

shock. This causes the yield curve

12:55

to change in a way that's very ugly

12:57

for a bank. Because a bank's business is a

12:59

positive net interest margin. You earn more on your

13:01

assets than you pay on your liabilities, and

13:04

your liabilities are short duration,

13:06

and so if interest rates go way

13:09

up quickly, you can end up in a position

13:11

where you are paying more on your liabilities

13:14

than you are in your assets, which is which is going

13:16

to run right through your capital. That's not a profitable

13:18

business model. This happened in the eighties

13:20

and a lot of banks became undercapitalized

13:23

and supervisors were swamped

13:26

with cease and desist orders

13:28

and supervisory directives to banks all over

13:30

the place to raise more capital. Some banks

13:32

sued. One bank was able

13:34

to prevail in the fifth Circuit, and Congress

13:37

intervened and pass a new law

13:39

authorizing forward

13:41

looking, ex anti brightline

13:43

capital rules for the first time

13:46

and saying that capital judgments

13:48

of supervisors can't be second guests by courts.

13:51

And so, in a sort of accidental way,

13:53

you have the birth of capital regulation. The

13:56

supervisors are overwhelmed and

13:58

there are laws, suits, and you have Congress

14:01

saying, actually, just write a rule that

14:03

the banks all have to comply with so

14:05

that you don't have to get into litigation over whether

14:08

this bank or that bank is undercapitalized.

14:10

Fast forward a few years. You get

14:12

these rules, and you get you get

14:15

Basel one, you get an effort to align

14:17

these rules internationally, and

14:19

you get Alan Greenspan as fed

14:21

share. Going into the nineties, the banking

14:23

system starts to transform. You have the

14:26

emergence of large, complex banking

14:28

institutions, and

14:30

you have a lot of soul

14:33

searching in Washington

14:35

and the Federal Reserve at the occ about

14:37

whether supervisors are really up to

14:40

the task of assessing

14:42

the risk taking of

14:44

these new large complex financial

14:47

institutions, these large complex banking organizations

14:50

that we never had in this country before

14:52

through the traditional means, or whether we

14:54

should actually embrace these new rules

14:57

that have developed up and rely primarily

15:00

on the rules and shift supervisors

15:02

to a task that they are that they're more

15:04

capable of performing. This

15:06

is this is very self conscious for the policymakers

15:09

at the time. You can go back and read some of Alan

15:11

Greenspan's speeches about the changes that he's

15:13

making, and he basically thinks that, especially

15:15

for large banks, supervisors are just

15:17

not going to be able to do it the way they used to. What

15:20

we need are capital rules

15:22

that require shareholders

15:24

to have enough skin in the game, and then

15:26

the shareholders will do it. The shareholders

15:28

were supervised banks. And so Alan Greenspan

15:31

says, it's not about needing net less

15:33

regulation. It's about whether it should be public

15:35

sector regulation or private sector regulation,

15:38

and we need to reorient the banking system so

15:40

that we have more private regulation.

15:43

So that obviously goes terribly wrong in

15:46

two thousand and eight. It's not funny,

15:48

but this happens over and over and over it

15:51

does, and so by two thousand and eight you have supervisors

15:54

have more or less unilaterally disarmed.

15:56

They have shifted to enforcing

15:59

the capital rules. The banking

16:01

agencies are relying almost entirely

16:03

on the capital rules for making

16:06

judgments about whether a bank has adequate

16:08

capital for the risks that it's taking, and

16:11

instead they are looking at processes.

16:14

And there's a real theory behind this. The theory

16:16

is, in order for

16:19

market regulation to work, private regulation

16:21

to work, there has to be disclosure, and

16:23

so a bank has to be transparent about the risks

16:26

that it's taken. A bank can't be transparent

16:28

about the risks that it's taking if it doesn't have processes

16:31

to monitor and disclose those risks. And

16:33

so the job for supervisors is to

16:35

make sure banks are monitoring their

16:37

risks and disclosing them to the shareholders

16:39

so that the shareholders can discipline the banks.

16:41

And by two thousand and eight, supervisors

16:44

have stopped bringing in cease into sists.

16:46

There's no safety and soundness enforcement actions against

16:48

any of the major banks, any of the major banks that take

16:51

TART for years running up to two thousand and eight,

16:53

because if they're in compliance with the rules and they're disclosing

16:55

to the market, the judgment is that

16:58

system is going to work. What goes wrong

17:01

is that bank shareholders

17:04

have an incentive to take much more risk than is in the

17:06

interests of the government or depositors. It's sort

17:08

of basic economic stuff. The shareholders

17:10

have an incentive to extract wealth from the depositors

17:13

and from the public. And so the shareholders

17:15

is gonna be much more comfortable with a

17:17

lot more risk than the public should be. And

17:19

so if you're going to rely on them to monitor

17:21

risk, cake and you're going to get a much riskier bank.

17:24

And this is Silicon Valley Bank story.

17:26

I mean, it's very much Silicon Valley Bank story.

17:28

So you get two thousand and eight, and

17:30

you get a modification after

17:33

two thousand and eight, which is stress tests,

17:35

but a lot of ordinary day to

17:38

day supervision continues

17:40

to be I think procedurally

17:42

oriented. And you see this with the

17:44

London whale, and you see

17:46

this with the fake account scandal at Wells

17:48

Fargo, both of which I think are really

17:51

important data points for outside

17:53

observers to think about how

17:55

much did we fix supervision

17:58

after two thousand and eight, how much do we move away

18:00

from the green span nineties cocktail

18:03

of bright line capital rules and procedural

18:06

oversight oriented to shareholder discipline.

18:08

And I think the answer is for the small

18:11

banks that are not subject to stress tests,

18:13

and even for the big banks that are subject

18:15

to stress tests, outside

18:18

of the stress testing regime, we still

18:20

have a lot of procedural oversight.

18:23

So this is actually a very quick

18:25

mechanical question. But for

18:27

a bank like Silicon Valley Bank, is

18:30

it as supervisor? Is it a

18:33

panel like hot Like what do you know? Do you have a

18:35

sense of like how many people? I mean, because

18:37

it's someone at the SFED presumably,

18:39

Now assume there's thousands of banks in California

18:42

probably or at least hundreds, Like what kind

18:44

of just like human resources could even go

18:47

currently to paying attention to Silicon

18:49

Valley a bank like Silicon Valley Bank. So

18:51

there are thousands of supervisors

18:54

across the federal system. Okay, they're

18:56

split across three agencies, the Federal

18:58

Reserve, the Federal Deposit Insurance Corporation,

19:01

and the Controller of the Currency. They

19:03

split up responsibility

19:05

for supervising bank. Silicon

19:07

Valley Bank is a state chartered bank and

19:10

it's a member of the Federal Reserve system. So

19:13

as a result, as you say, it's the FED

19:15

that would have responsibility at the federal

19:17

level for primary responsibility

19:19

for supervision if there's always overlap,

19:21

So the FDIC has some ability to come

19:24

in because it's insuring the deposits obviously,

19:26

and then it's very involved now. But

19:28

the day to day job here is FED

19:32

personnel in San Francisco

19:34

who are really actually exercising delegated

19:37

authority of the Board of Governors, which

19:39

is the federal agency

19:42

with the power to supervise member

19:45

banks. And the Reserve Bank of San

19:47

Francisco is actually a federal bank of Federal

19:49

corporation, and so it's it's

19:52

facilitated, it's helping the board, and

19:54

the board has supervisory staff that

19:56

are supposed to sort of oversee

19:59

what is going at the reserve banks. So is

20:01

San Francisco doing a good job, And

20:03

obviously at the top of that is Michael Barr, the

20:06

Vice champer supervision. I'm

20:24

going to have some more questions on

20:26

the San Francisco FED and the fhlbs

20:29

as well. But just going back to

20:32

the evolution of bank

20:34

capital rules. So

20:36

one of the big things that

20:38

happened, and you sort of outlined it

20:40

in the lead up to the two thousand and eight crisis,

20:42

but like it definitely hardened after

20:45

two thousand and eight. Is this idea that

20:47

banks should be holding more

20:50

bonds in general, the safest bonds,

20:53

so, you know, US treasuries

20:55

in the case of US banks, maybe agency

20:58

mortgage backed securities that are

21:00

implicitly guaranteed by the US government,

21:02

things like that for their regulatory

21:05

capital and liquidity buffers.

21:07

And it seems to me like that probably

21:10

made a lot of sense in the low

21:12

inflation environment of two

21:14

thousand and eight, But now that

21:16

you have the FED raising rates, you

21:18

have a lot of volatility. It seems

21:21

like these bonds might not

21:23

be I don't think safe

21:25

is the right word, but not as

21:28

unproblematic as maybe

21:30

we imagine them to once be. Could

21:33

you talk a little bit more about

21:35

basically how we built the modern

21:37

banking system on top of a bedrock

21:40

of bonds that are presumed to be somewhat

21:42

stable in price. So

21:44

I think that you're right to sort

21:47

of highlight the

21:49

appeal of treasuries

21:51

and agencies to both bankers

21:54

and supervisors in the

21:56

wake of two and eight. Bank that loads

21:59

up on treasury, that's,

22:01

like, you know, very wholesome.

22:03

It seems very wholesome for a bank to

22:05

do right. The government likes that. The government

22:08

is like banks that buy treasuries since the Civil

22:10

War, and so it's

22:13

gonna cut against even the most ambitious

22:16

and confident, you know, safety

22:18

and soundness overseas. It's going to cut against their

22:21

their impulses to to sort of

22:23

fault a bank for loading up on treasuries.

22:25

I mean, that's that seems that seems

22:28

that's a good thing, right, And so it's

22:30

it that

22:32

helps to explain part of what's happening

22:34

here. And it's

22:36

also true that banks do have

22:39

the ability to weather usually

22:43

a fair amount of interest

22:46

rate losses on their assets. So

22:48

many banks think of themselves

22:51

as structurally hedged

22:53

against interest rate increases

22:56

because while their

22:58

assets, if they have law assets like

23:01

long dated treasuries, that's those are going to

23:03

lose value when the interest interest rates

23:06

rise, their liabilities or deposits,

23:08

and their deposits are sticky, they don't pass

23:10

through, and so actually

23:13

their deposit funding becomes much more valuable

23:15

when interest rates rise. So if in a zero

23:17

interest rate environment, interest rates

23:20

or zero deposit rates or zero

23:22

deposits aren't that useful. You're getting

23:24

a little benefit that you have deposit funding.

23:26

But if interest rates go way up, you're

23:28

not gonna if you're the bank, you're not actually going to be

23:31

forced to raise your extracting

23:33

rents right from the deposit public. But you're

23:35

not going to be forced to raise your deposit rates. And

23:37

this is actually strengthening your business. And Silicon

23:40

Valley Bank is going to think, yeah, okay,

23:42

so we take some we take some hits

23:44

on our long assets, but actually

23:47

our net interest margin is going to position,

23:49

is going to main strong, our deposits are going to be much

23:51

more valuable, and we're just gonna we're

23:53

gonna work through, you know, year to eighteen

23:56

month period and be totally fine, right,

23:58

which it seems like they lose was sort

24:00

of the assumption that they had, like they knew they

24:02

it wasn't great and maybe even

24:04

technical and solvent, but I mean, I think

24:07

it was in this leek one of matt Leview's news

24:09

and letters, he's like, this was actually like a very profitable

24:11

time for them, like it did. It would have been fine if

24:13

everyone's did. And presumably their expectation

24:16

was, well, we're just making a lot of income

24:18

right now. So the fact that we took it a hit on the asset

24:20

side is not really well. They had

24:22

they had a specific estimate in

24:24

one of those internal documents where they said,

24:27

we could shorten duration, but

24:29

that would mean an eighteen million dollar hit

24:31

to our net interest margin in one year

24:33

alone, going up to like thirty

24:35

six million over the next three years. So they

24:38

knew that if they reduced

24:40

duration, they would be sacrificing

24:42

earnings to some extent. Yeah, I mean, I

24:44

think it's fair to say that in twenty twenty

24:46

one they were making huge profits because this strategy

24:49

was really working. Interest rates went

24:51

way up, and I think

24:53

it would have impaired their profitability,

24:56

but they were They weren't

24:58

wrong to think that they were what structurally

25:01

hedge, even though they had no interest rate hedges or anything,

25:03

by virtue of the fact that they would

25:06

slowly be able to replace their treasuries

25:08

with much higher yielding treasuries while being

25:10

able to pay depositors very little. And we also,

25:13

you know, we talked about this deposit

25:15

betas on a recent episode with Joe

25:17

Batte and why they're off at low and what if

25:19

His points was like, well, you know, you have a

25:22

bank, you have an individual has

25:24

an account at Chase or something like

25:26

that they're providing a lot of services

25:28

along with that. People are not that inclined

25:30

to move their checker account just because

25:33

the interest rate doesn't bump up a little bit.

25:35

And I imagine for Silicon Valley

25:37

depositors these companies, the whole

25:40

story about Silicon Valley Bank was

25:42

all of the products, the

25:44

startup specific products that they are that

25:46

they offered, which presumably to their

25:48

mind, insulated them to some extent

25:51

against losing deposits. Absolutely,

25:53

and I mean in the case

25:55

of SVB, there was also what an

25:57

antitrust law we call tying, where

26:00

accompany ties one product

26:02

to another product. Yeah, so the bank

26:05

would require that if you wanted

26:07

to borrow from the bank, that you would have Is

26:09

that unusual because people some people obviously some people

26:12

are like, oh, that's weird, and some people are like no, of course,

26:14

like any commercial loan. But is that unusual

26:16

in your view? I don't think it is unusual.

26:19

There are strict rules about bank tying

26:21

in other areas. But my understanding

26:24

is that banks are explicitly permitted

26:26

to tie deposit account services

26:28

to lending services. And historically

26:31

it was core to the banking business that

26:33

you were the depository institution

26:35

for the for your borrowers

26:38

that that went together, and you

26:40

know, we've moved away from that technology

26:43

lots of things have allowed people to borrow

26:46

from banks that aren't the banks where they bank

26:48

at. But a

26:50

long time, the idea was that there's a lot of synergies

26:52

between that no one's going to be in a better position

26:55

to determine how much to lend to you

26:57

than your own banker. I

27:00

imagine like there was also a bit of a

27:03

prestige element to banking

27:05

at SVOB as well, given that, you know, it

27:07

was so popular among a particular

27:09

type of tech slash VC person. But

27:12

you know, Joe touched on the episode

27:14

we did on deposit betas with

27:17

Joe Abode. But I'd love

27:19

to hear from you, like why didn't

27:21

people pull more deposits

27:24

from a bank that was essentially paying

27:27

them nothing? Because to some extent, this

27:29

is the big question, like why did

27:31

SVB have so many

27:33

deposits? Well into

27:36

twenty twenty two, at which point we started

27:39

to see some of the I guess most

27:41

interest rate sensitive parts of the

27:43

economy, ie the tech industry

27:45

lose a bunch of money and have to pull funds.

27:47

But why why were people accepting

27:49

of that for so long. So

27:52

we mentioned a couple of the sort

27:54

of rational explanations

27:56

that you know, there's

27:58

a strong brand, you want to they're they're

28:01

lending to you, they've required you to keep deposit there.

28:03

But you can't discount the fact

28:05

here that a big piece of this was

28:08

a lack of sophisticated financial

28:10

management on the part of startup companies

28:13

that maybe didn't have CFOs,

28:15

didn't have anybody on their teams with

28:18

any experience in managing cash.

28:21

They're focused on their their business

28:23

and it's very hard

28:25

to justify five hundred million dollar bank

28:27

account balance, and I think we

28:29

have one example of that that just

28:32

there's there's no reason for that's that's

28:34

very bad management. No well

28:36

run, mature company would

28:39

operate in that way. Among other

28:41

things, you you you have a

28:43

huge uninsured deposit risk that we

28:45

that we saw, but you're also just giving

28:47

up lots of return. You could have that money

28:50

invested in laddered

28:52

treasury bills or something. And

28:57

significantly more so, there's a huge amount

28:59

of money that's just been left on the floor here

29:01

there, and it's you know, it doesn't it doesn't

29:03

really make sense. So we have to understand

29:05

that these that these customers, despite

29:08

having lots of money, are not actually very sophisticated.

29:11

So I want to go back to the supervisory

29:13

question and ask about it, kind of come at it

29:15

from a different angle. You know. Obviously

29:18

the two thousand and eight two thousand and nine

29:21

crisis was very focused

29:23

on the asset side of the business

29:26

and were these really high quality assets? And

29:28

then part of the reasonable bunch of banks failed

29:30

is because the assets like weren't very good that

29:32

they held. And as people have discussed

29:35

with Silicon Value Bank,

29:37

a lot of the issues. Yes, maybe they made

29:39

a wrong bet on treasuries or they put

29:41

too much, but is the flightiness

29:43

of the deposits? Can you talk a little

29:46

bit more. I know that regulators

29:48

do bucket deposits

29:51

from the most sticky to the least sticky, but

29:53

could you talk a little bit about like how

29:55

good supervisory in

29:57

a sort of like active pre nineties

30:00

way might have approached the

30:02

uniformity of SVB deposits

30:05

and the risk of them all leaving at once. Yeah,

30:07

I mean I think that if you showed svb's balance

30:09

sheet to supervisor brought up during the new

30:11

deal system. So let's say it's nineteen seventy

30:14

five, they would be horrified

30:17

at the enormous concentration

30:19

of uninsured deposits

30:21

controlled by a group of businesses

30:25

with very similar

30:27

risks to their business and so all

30:30

of your depositors, and this is something that and

30:32

this is something that a new deal era supervisor

30:35

would be familiar with from press experience. Well, I

30:37

think I think it would have been. They

30:39

would have been unfamiliar with it in the sense

30:42

that it would have been so unusual back then everyone

30:44

would have looked at it and said, WHOA, this

30:47

bank has a very unstable deposit

30:49

base. It would not have been novel to

30:52

view this with concern. It would be it

30:54

would be even more concerning because

30:56

of how risk it would have been to operate

30:59

a bank in this way at that point in time,

31:01

when when, of course, people still remembered the

31:03

bank runs of the thirties much more

31:05

than they do than they do today. And

31:08

you know, part of what went wrong at SVP is it's

31:10

not just that they had ninety seven percent

31:12

or something in uninsured deposits,

31:15

but that all of their depositors

31:17

were going to withdraw at the same time. And so

31:19

there's sort of a classic issue in the banking

31:22

business always is

31:25

in what circumstances am I going to be subject to a

31:27

deposit train? You know, you get

31:29

to model your deposits as

31:32

sticky if you're a

31:34

bank, because over time, for the

31:36

banking system, the deposit base is always sort

31:38

of growing. I mean, with the exception of over the last

31:40

year where monetary policy is trying to shrink the

31:42

money supply, but you know, over time it's

31:44

a constant growing base, and so deposits

31:46

are really, in some simes a very long duration asset,

31:49

except if you're the one bank that experiences

31:51

a drain to the rest of the system where everyone withdraws

31:54

from you, and so if your customers are

31:56

all going to face hardship your depositories

31:59

or depot it was hard. The same

32:01

time, you really can't treat yourself as

32:03

structurally hedged. You're the opposite

32:05

of structurally head And that's what that's what Silicon

32:08

Valley Bank found out, is it thought that,

32:10

oh, you know, when when my assets

32:12

lose value, my deposits will become more valuable.

32:15

But actually all their depositors started to draw

32:17

down their accounts and the opposite happened, and

32:19

so they were just very very

32:22

long low interest rates, silicon valuing.

32:24

The whole business model was tied to low interest

32:26

rates, I think to an extent that they did not appreciate,

32:29

and to an extent supervisors clearly didn't

32:31

depreciate, but maybe weren't even

32:34

thinking as hard about as they might have in

32:36

an earlier period where they

32:38

were more empowered to make those sorts of judgments.

32:41

Yeah, this is exactly what I said on our

32:43

episode with Dan Davies. It was interest rate

32:46

exposure kind of squared, but just

32:48

on the deposit side. Because to me, this

32:51

is kind of the most novel or

32:53

interesting thing about all of this,

32:55

because we know that a lot of banks have unrealized

32:58

losses on bonds, and you

33:00

know too. Seems

33:02

like broadly they've been managing their

33:04

interest rate risk so far. But

33:07

with SVB, the big difference was

33:10

that group of highly concentrated,

33:12

extremely unreliable depositors

33:15

who themselves had significant interest rate

33:18

exposure. And we're pulling money over the

33:20

past years. So what could

33:22

regulation do on that

33:25

front? So I guess instead of the asset

33:27

side, looking more at the liability

33:29

side. Yeah, So what

33:33

you want to see is a coherent

33:35

asset liability management

33:38

strategy for a bank, and so a bank

33:40

that anticipates deposit trains

33:44

for a bank that has flighty

33:46

deposits, and there are many banks that

33:48

can fall into this category. This is something that regulator

33:50

supervisors do think a lot about. If you're

33:52

in that category, then you need to hold

33:55

liquid assets that you can

33:57

sell and that at their fair market

34:00

value to cover the withdrawals. And

34:03

so part of the problem here is that Silicon

34:06

Valley Bank did not actually have available

34:10

for sale securities at

34:12

fair market values sufficient to cover the

34:14

withdrawals, and so the fix

34:17

for this would have been to

34:20

have much less duration in the in

34:23

the in the asset portfolio, or many

34:25

many more reserves. This is the same problem,

34:28

by the way, that took down silver

34:30

Gate Bank and to some extent, Signature

34:32

Bank. They had deposit

34:35

bases that were flighty,

34:38

that their depositors suffered

34:40

and their deposit the banks

34:42

experienced deposit drains because

34:45

they were concentrated in a group of people that were

34:47

exposed to interest rate hiking. I

35:06

just realized I promised to ask about discount

35:09

lending and the fhlb's

35:11

the Federal Home Loan Bank. So you

35:13

know, in theory, when you have this type

35:16

of banking crisis or you know, some sort

35:18

of liquidity issue with the financial

35:20

institution. You would expect them to either

35:22

go to the FED, to the discount window

35:25

and for all blots listeners, we recorded

35:27

an episode on this a month or two ago, or

35:30

they can borrow from the FHLBS.

35:33

And some of the talk out there is

35:35

that SVB got

35:38

cut off by FHLB.

35:42

Why would that have happened and why

35:44

wouldn't those two lenders of last

35:46

resort do everything they can in

35:49

order to step in and support the bank.

35:51

Or is it the case that at some point,

35:53

you know, maybe they're talking to the FDIC and they

35:55

just say this is untenable and no matter

35:58

how much money we provide, like the bank is

36:00

not going to be able to get up and running again. So

36:03

I'm speculating a bid here, and you

36:05

might want to talk to the FHLB

36:08

expert in the legal academy, K Judge,

36:10

who's my colleague. But the

36:13

FHLBS are not a

36:15

lender of last resort in the way that the FRBs

36:18

are. Right, the FHLBS

36:21

they do provide sort of lender

36:23

of second to last resort services

36:26

to their members, but they

36:28

are much more operated by

36:31

their members, and they pay dividends

36:33

to their members than the FRBs.

36:36

So the FRBs were set up in a

36:38

similar model, but today basically

36:41

function as public banks, so they have no

36:43

interest in profits or anything like that, and

36:46

they're willing to sort of take one

36:48

for the team in a way that the fhlbs

36:51

are not. So I think it's a mistake to look

36:53

at the FHLBS and say, oh, well, you

36:55

really ought to have lent into an insolvent

36:58

institution and took on that potential. Rik k.

37:00

The FRBs are are are are wary

37:02

of that for various reasons we could

37:04

get into. But the flhol bes

37:06

have even more reason to sort of to pull

37:09

back. We definitely have to do

37:11

an FHLB episode at some point with

37:13

k Judge, because I don't know much about them at all,

37:15

and it definitely it's been too long or

37:17

it's far too long without having yeout

37:21

having her on. I want to ask another dimension.

37:24

You know, people pointing to the

37:26

twenty eighteen law

37:29

change to Dodd Frank that

37:31

seemed to exempt banks like Silicon

37:34

Valley Bank from some of these liquidity

37:36

requirements that you were talking about, like do you have enough liquid

37:38

assets? Hey? Could you sort of characterize

37:40

the change that was made there in b had

37:43

that not been in place, like is

37:45

that is that change that was made in twenty

37:48

eighteen? Does does that tell the story of the demise?

37:50

Had the old Dodd Frank laws

37:52

remained in place for a bank the size of Silicon

37:55

Valley Bank, would they have been able to weather

37:57

the storm? You can never know for sure, but I would

38:00

suggests yes, that is

38:02

decisive. And so this

38:04

brings us back to how the government responded

38:07

to the two thousand and eight failure of supervision

38:09

regulation. And they responded with a

38:11

set of tiered new requirements. And

38:14

for the very biggest banks you had the

38:16

CCR stress testing regime, and

38:19

for all of the banks with more than

38:22

fifty billion dollars of assets, you had

38:24

stress testing as well as collection

38:27

of other enhanced prudential standards.

38:30

And this new cocktail

38:33

of regulation and supervision was geared

38:35

towards preventing a repeat of something

38:37

on the scale of two thousand and eight. And so

38:39

we weren't going to impose this on

38:42

the whole banking system, on all of the sort of smaller

38:44

banks. But the thinking was, if

38:46

we could just really get back to

38:48

serious government oversight of banks

38:51

over fifty billion dollars that would

38:53

really go a long way towards preventing another

38:56

another calamity like two

38:58

thousand and eight. And what happened

39:00

was immediately there was

39:02

litigation over the threshold for

39:05

this new regulatory supervisory

39:08

cocktail. And so this fifty billion dollars

39:10

threshold came under a lot of political

39:12

pressure from the banking agencies and from

39:14

various people in Washington, and the

39:17

bank lobby fought a battle over

39:19

many years to raise the threshold.

39:22

One of the important

39:24

figures lobbying for raising the threshold

39:27

was the CEO of Silicon Valley Bank.

39:30

He was growing his bank and he did not want

39:32

to grow his bank into additional

39:35

regulatory and supervisory requirements.

39:37

And in twenty eighteen,

39:40

after winning over people in both

39:42

parties to this cause of raising

39:44

the threshold, Congress changed

39:47

the law and the threshold moved up. A

39:49

bunch of thresholds moved around, but the relevant threshold,

39:52

I think, moved up to two hundred and fifty billion

39:54

dollars. And the result was Silicon Valley

39:56

Bank and its peers had successfully

39:59

exempted themselves from the

40:02

enhanced prudential standards that Congress had created.

40:04

After two thousand and eight, and the

40:06

result was you didn't have the stress

40:09

testing, which is the primary means

40:11

by which supervisors now exercise

40:13

substantive judgment about risks, and

40:16

so you had a much more

40:18

I think light touch process focused

40:21

oversight that allowed rule

40:23

compliant balance sheet configurations

40:26

like svbs to go relatively

40:28

unchallenged. But if you had been in the

40:31

in the Enhanced Prudential Standards

40:33

bucket, I think that they would have been challenged.

40:35

They would have been challenged through all of these additional

40:38

rules and also supervisory programs.

40:41

And it's unlikely. You

40:43

can never know, but it's unlikely that

40:45

they could have taken so much duration

40:48

risk and not raised

40:50

capital earlier, been permitted

40:52

to go for so long in a position

40:55

where their liquidation value was possibly

40:57

negative. Yeah, since we're on the topic

40:59

of the blame game, I mean, one of

41:01

the things that you see people saying now

41:04

is that, well, it's the

41:06

Fed's fault. The FED kept interest rate

41:08

slow for too long, and it basically

41:11

forced people to assume additional

41:14

duration risk in order to seek

41:16

out yield. And I think, you

41:18

know, financial repression is a

41:20

real thing. But on the other hand, you cannot

41:22

ignore the individual actions of one

41:25

or a few specific banks, their managers,

41:27

their shareholders, and their depositors. But

41:30

how would you describe the

41:32

overall monetary policies role

41:35

in the current predicament.

41:38

So overall monetary policy

41:40

is of course central to the current predicament.

41:43

But there's a sort of false dichotomy

41:45

underline, the view that somehow

41:48

it's like monetary policy happening

41:50

up here at the FED that's then causing

41:53

problems down here at the banking system. The

41:55

whole thing is monetary policy. The whole reason we have banks

41:57

is monetary policy. Banks are creating the money supply,

42:00

and the question is how much money do we want to

42:02

be created. So it would be the tail wagging

42:05

the dog if we had to change

42:07

our judgment about how much money should be created because

42:10

like somehow the system couldn't create that amount

42:12

of money safely and stably. We have

42:14

a broken system if it can't

42:16

create the amount of money that the FMC says

42:18

is appropriate for macroeconomic conditions.

42:21

And so I would not blame the

42:24

FOMC for thinking that we need to adjust

42:26

the amount of money that the banking system

42:29

and the financial system are creating. I would

42:31

blame the banking and financial system and the banking and

42:33

financial laws. If they're incapable

42:36

producing the amount of money and changing the amount

42:39

of money they're producing overtime consistent

42:41

with the fomc's directives and the needs of the economy,

42:44

that's a really big problem. And it does look

42:46

like we're facing that problem now, where

42:49

in the coming months we could be in quite a predicament

42:51

where the FOMC may make the judgment

42:54

and we can brack it whether it would be the correct judgment

42:56

that the economy needs less money and

42:59

the banking system may be incapable

43:02

of functioning properly under

43:05

that directive. And that's the flip side

43:07

of, you know, the suggestion that the banking

43:09

system should also be able to function under the judgment

43:11

that interest rate should be zero and

43:13

function in a way that is sustainable over time.

43:16

And so to the extent that is what's going on, I

43:18

think it's a real indictment not of monetary

43:21

policy, like the high level decision about how much

43:23

money we need, but the structure being

43:26

unable to follow through to execute on those

43:28

decisions. Leveman on that was an amazing

43:30

answer, That was an amazing conversation that

43:33

was so helpful that was so helpful

43:35

and so clear. I've said it every time

43:37

we talked to I'm like, oh, I've finally

43:39

had There's been a few more, but I so that

43:42

was a that was very good. I'll

43:44

just leave it there. I really I

43:46

really appreciate that. So thank you so much for coming

43:49

back on up. Thank you so much for having me that,

44:04

Tracy. I love that whole conversation, starting

44:07

from the very end, which I think is a really

44:10

excellent way to sort of reconceptualize

44:13

the monetary policy problem, which

44:15

is that if the FED is going to be tasked

44:18

with like the sort of like big sweep

44:20

macro management, right, getting

44:22

employment inflation at its targets

44:25

and so forth, in theory, we

44:27

want to have a financial system that

44:30

can operate under any you know,

44:32

operate relatively safely under whatever

44:35

rates the FED deems to be appropriate. Absolutely,

44:37

I mean, I do think there is a fundamental

44:40

tension between monetary policy,

44:42

which you know, like the big

44:44

thing monetary policy does is basically

44:46

impact the price of bonds and then having the

44:49

financial system and banks specifically

44:51

have to hold a bunch of bonds as

44:53

part of their capital and liquidity mandates.

44:56

Like that tension is there,

44:58

but there are way is to manage

45:01

it such that we can avoid failures

45:03

and also provide for the effective

45:06

implementation of monetary policy

45:08

as a whole. The other thing that really

45:10

struck me is this is kind of an

45:12

incentives episode, right,

45:15

And I thought Lev's point about basically

45:17

outsourcing a lot of bank supervision

45:20

to private shareholders

45:23

and expecting them to you

45:25

maybe press the brakes on risk

45:28

when things start to get out of hand, that

45:31

was a really interesting one. And SVB,

45:33

I think, is going to end up as a classic case

45:36

where you know, there was an acknowledgement

45:38

that there was an issue here. There was

45:40

the asset liability duration mismatch,

45:43

and there was too much exposure to long bonds

45:46

and too much of an assumption

45:48

that deposits would be around

45:50

forever or that they might even return

45:53

or start growing again, and it

45:56

was a conscious decision to pick

45:59

up net interest margin, or at least it

46:01

looks like that. I'm sure more will come

46:03

out over the course of all of this, but for now,

46:05

it certainly seems like it was a decision to

46:08

do that. That's really interesting, like thinking about

46:10

like, it's pretty fascinating that a lot

46:12

of these like capital requirements and ratios

46:15

and regulations that we think of as

46:17

courtA how we manage the banking system are

46:19

all pretty young, and that for the most part,

46:21

for a long time in history it was like active

46:24

super advisory people

46:26

making judgments based on the

46:28

operations of the bank, whether their decisions

46:31

on loans and deposits were healthy.

46:33

But it does make total sense that it

46:35

is sort of like you know, I hate usually

46:38

where it's like neoliberal, but that like

46:40

that the role of supervisors

46:43

would essentially transform to making

46:46

sure that shareholders were

46:48

getting adequate information. That you start with

46:50

the assumption that the market is the best regulator,

46:53

and then what is the role of the government well, and

46:55

the role of the government, and that point is to make

46:57

sure that market regulators get good information.

47:00

Like that seems like a very like big theme

47:03

that like you could like characterize across a lot

47:05

of different industries, a lot of different government and

47:07

then the failure of just like well, the problem

47:10

is like shareholders can lose everything and not

47:12

be accountable for the spill over when

47:14

a bank fails. And so yeah,

47:16

the need perhaps to get back to a type of supervisory

47:19

that actually takes decisions into own

47:21

hands and rather than just outsourcing it. You

47:24

know, I realized we didn't even get into

47:26

FED checking accounts, which is one of them.

47:28

Now that's the next step. So then I think the series

47:33

is if all deposits post SVB

47:36

are presumed to be ensured, which almost

47:38

seems like implicitly the case, now why

47:41

do we have private deposit taking institution?

47:44

So I kind of think that's the next one in this series

47:47

to look at, Well, what does this tell us

47:49

now about even the point of private deposit

47:51

taking institutions? And should there be a point? I mean,

47:54

that is what I was kind of hinting at in the intro.

47:56

But we'll just have to leave that. You know,

47:59

it was a trailer not for this episode but for

48:01

the next one, so plenty more to come,

48:03

but shall we leave it there? Let's leave it there. This

48:06

has been another episode of the Odd

48:08

Thoughts podcast. I'm Tracy Alloway. You

48:10

can follow me on Twitter at Tracy Alloway

48:13

and I'm Joe wisn't All. You can follow me

48:15

on Twitter at the Stalwart, follow

48:17

our guest Levmnond on Twitter at levmanand

48:20

follow our producers Carmen Rodriguez

48:22

at Carmen Arman and Dash Bennett

48:24

at dashbot Check out Bloomberg's

48:26

podcast son to the handle at podcasts,

48:29

and for more Odd Lots content, go to bloomberg

48:31

dot com slash odd Lots, where we

48:33

post all the transcripts. Tracy and I have a blog

48:35

and weekly newsletter that comes out every Friday.

48:38

Go there and sign up. Thanks for listening.