HomeVideos

How The Federal Reserve Could Shrink Trillions From Its Balance Sheet | Darrell Duffie

Now Playing

How The Federal Reserve Could Shrink Trillions From Its Balance Sheet | Darrell Duffie

Transcript

1677 segments

0:00

My objective is to provide options if

0:02

the Fed were to decide to reduce its

0:05

balance sheet. How could it do it?

0:06

>> The Fed's not going to send the SWAT

0:07

team, the military to the banks to

0:09

people's houses to get all the cash.

0:10

It's not not going to happen at all. So,

0:11

the only real channel to reduce it is

0:13

reducing reserves.

0:14

>> I also have the view that the Fed should

0:16

develop these options in case they're

0:18

needed, in case it becomes a break the

0:20

glass situation.

0:21

>> Today's episode is brought to you by the

0:23

Fundrise Income Fund. You'll hear more

0:25

about the income fund later in the show,

0:27

but for now, let's get into today's

0:29

interview. The new Fed Chair, Kevin

0:30

Walsh, has made it no secret that he

0:33

prefers a smaller Federal Reserve

0:35

balance sheet, perhaps a much, much

0:38

smaller Federal Reserve balance sheet.

0:40

The consequences of this range from the

0:42

mundane to the profound. But what is

0:45

without question is that in order to

0:47

reduce the Fed's balance sheet, there

0:49

needs to be additional tools. I was able

0:52

to interview someone who's the most in

0:54

demand expert on this topic. Professor

0:57

Daryl Duffy just wrote a paper proposing

1:00

four techniques the Fed can use in order

1:03

to reduce the demand of reserves,

1:06

thereby allowing it to reduce its

1:09

assets. If the Fed were to try and

1:11

reduce its balance sheet drastically

1:14

without addressing the very high demand

1:16

for reserve balances from the banking

1:19

system, then there could be a flare up

1:21

in funding markets such as the one we

1:23

experienced in September 2019 or during

1:26

the fall of 2025. I think the Federal

1:29

Reserve has a very responsible

1:31

institution that's very unlikely to

1:33

happen. So if in several years or

1:35

perhaps even a decade the Federal

1:37

Reserve's balance sheet is much much

1:39

smaller, I think it is highly likely

1:42

that the Fed will use one or some or

1:45

perhaps all of the techniques proposed

1:48

in this groundbreaking paper. I have to

1:50

say that this interview with Daryl Duffy

1:52

is very complex, much more so than our

1:55

first interview. And when preparing and

1:58

conducting this interview, there were

1:59

many times when I was very confused. So

2:02

if you feel like that, please do not be

2:04

intimidated. I bet there will be

2:06

economic PhDs watching this who are

2:09

going to watch it twice just so they

2:11

fully understand this. And we will of

2:13

course include that paper called the

2:15

payment system puts a floor on the Fed's

2:16

balance sheet as well as another paper

2:18

by Daryl Duffy. Let's get into it. We

2:21

have a very important conversation

2:23

today. I'm joined once again by Daryl

2:25

Duffy, distinguished professor of

2:27

management and professor of finance at

2:29

the Stanford Graduate School of Business

2:31

as well as the department of economics.

2:33

Professor, welcome to Monetary Matters.

2:35

Good to see you again.

2:37

>> Great to be back with you, Jack.

2:38

>> I today want to talk about a very

2:41

specific topic and I'm going to set the

2:43

stage a little bit. So, we have a new

2:46

chair of the Federal Reserve, Kevin

2:48

Worsh. He has stated a desire to reduce

2:52

the Fed's balance sheet, something that

2:53

has been going on since 2022. The issue

2:57

is that that is a very difficult thing

2:59

to do. And basically the Fed's balance

3:03

sheet bottomed uh last year at let's see

3:07

where we are about $6.5 trillion which

3:11

sounds you know is a is a huge number

3:13

and since then it's been very very

3:15

modestly growing and the question is how

3:18

is the Federal Reserve going to continue

3:20

to shrink its balance sheet and you are

3:22

on the cutting edge of making some

3:25

suggestions about how the Fed is going

3:27

to do that. You have four ideas. Before

3:29

we get into those four ideas, could you

3:32

just continue to set the stage for for

3:34

me and for our audience? Why is it that

3:38

the Fed has had such difficulty getting

3:41

below the so-called LR or the lowest

3:45

comfortable level of reserves? What are

3:47

we talking about here?

3:47

>> Terrific. You framed it ex uh with

3:50

excellence, Jack, as usual. So, in the

3:53

popular discussion of the Fed's balance

3:55

sheet size, most people are focusing on

3:58

the assets. How many how much of this

4:00

6.5 trillion dollars of assets does the

4:03

Fed really need to hold? And couldn't it

4:05

get rid of some of these assets without

4:08

blowing up the economy? And actually,

4:11

that's not the right way to look at it.

4:13

uh we should move over to the other side

4:16

of the balance sheet where the

4:18

liabilities are

4:20

because the assets by definition have to

4:22

be at least as large as the liabilities

4:25

and it's really hard to squash down the

4:28

Fed's liabilities as it turns out. So

4:31

even if the Fed didn't want to own any

4:34

of those assets, it has to own enough to

4:36

support the liabilities. And some of

4:38

them you just can't get rid of like

4:39

paper currency. What do you do? you go

4:41

out and ask people to give back their

4:43

paper currency. I don't think that's

4:45

going to work. And there's $2.5 trillion

4:47

dollar of paper currency. So that kind

4:49

of sets a baseline right there. And as

4:52

we go through this discussion, I'm just

4:54

basically going to walk through how one

4:56

would reduce the need for the Fed's

4:59

liabilities. You can't simply sell

5:01

assets today and avoid problems because

5:05

all of the liabilities on the Fed's

5:07

balance sheet are serving very important

5:09

roles. So if I if I can just let me walk

5:12

through briefly what those are, at least

5:13

the big ones, and then we'll come back

5:16

to where there is scope over time to

5:19

reduce the need for one of those

5:24

liabilities, which is reserve balances

5:26

or the deposits of commercial banks held

5:29

at the Fed. So let's set the stage. If

5:33

there's 6.5 trillion of assets, there's

5:36

got to be 6.5 trillion of liabilities.

5:39

I've already said you can't do much with

5:41

2.5 trillion of those, which is paper

5:43

money.

5:45

There's about three trillion of

5:47

reserves. These are commercial bank

5:48

deposits at the Fed. We're going to come

5:51

back to that in detail because that's

5:52

where there's some scope. There's about

5:54

a trillion or so of deposits held at the

5:58

Fed by the federal government. That's

6:01

called the Treasury General Account. So

6:03

that's when the Treasury Department

6:04

collects money or pays money, it pays it

6:07

out of that account at the Fed. And

6:10

there is a little bit of scope to reduce

6:12

that trillion and the Treasury Boring

6:14

Advisory Committee has been floating

6:16

ideas about how that could be done, but

6:18

it's not the main show because, you

6:21

know, maybe you could cut it down by

6:23

half. Well, that's 500 billion. That's,

6:26

you know, not what I think Chair Wars is

6:28

talking about. He's talking about the

6:29

big game which is the reserve balances.

6:33

Uh at least that's what I'm going to

6:35

infer because that's three trillion a

6:36

lot more scope for reduction. Now there

6:38

are there's a smattering of other things

6:40

like money held at the Fed by foreign

6:43

central banks and a few other odds and

6:45

ends. I'm not going to get into those.

6:47

>> The big three are the Treasury General

6:49

account, cash in circulation, and the

6:52

reserve balances. The Treasury General

6:53

account not big enough. The cash in

6:55

circulation, you know, the Fed is is

6:57

just not going to happen. The Fed's not

6:58

going to send the SWAT team, the

7:00

military to to to the banks to people's

7:02

houses to get all the cash. It's not not

7:03

going to happen at all. So, the only

7:04

real channel to reduce it is reducing

7:07

reserves.

7:08

>> You got it. So, starting at And by the

7:10

way, reserves are actually going up at

7:12

the moment or have been over the last 6

7:14

months because the Fed discovered late

7:17

last fall that it had gotten reserves

7:19

down as low as they could get them and

7:22

it had to start growing them again at

7:24

around 40 billion a month. By purchasing

7:26

treasury bills, the Fed can expand the

7:28

supply of reserves. Again, those are

7:31

deposits of commercial banks at the Fed

7:33

and the banks use them for a number of

7:35

purposes. Most importantly, just to make

7:37

their daily payments to each other,

7:39

which are huge. Okay, so you got around

7:42

three trillion of balances. They're

7:44

growing. How could you reduce the need

7:46

that commercial banks have to hold those

7:48

balances at the Fed?

7:51

Well, I I already mentioned you need

7:53

them the banks need them to make

7:55

payments.

7:57

Just one of the Fed's payment systems,

7:59

the largest one called Fedwire, there's

8:01

$4.5 trillion

8:04

of of payments every day going through

8:06

Fedwire. And this these are payments by

8:08

banks to each other to meet the needs of

8:10

their customers. uh you know to makes t

8:13

make tax payments for purchases of bonds

8:15

for margining of derivatives for many

8:19

many other uh purposes even small retail

8:22

payments uh ultimately are netted down

8:26

and paid with reserve balances. So this

8:29

the payment system is exceptionally

8:30

demanding of reserve balances in the

8:32

United States. How could you reduce the

8:36

need that banks have to make those

8:37

payments? And sorry D professor before

8:40

we get into that just if the Fed was to

8:44

go willy-nilly say I don't need the

8:46

professor Duffy's four recommendations I

8:48

don't need any other recommendations and

8:49

they just were to be a net seller of all

8:52

their assets. What would happen? Talk to

8:54

us what happened in 2019 and what

8:56

happened in the last year in terms of

8:58

the the spreads and stuff.

9:00

>> So what would happen if the Fed said uh

9:03

you know never mind the liabilities I'm

9:05

just going to reduce the assets. they

9:07

start selling assets. Now what happens

9:09

when they sell assets is that the

9:13

purchases by the private sector of those

9:15

assets are paid with reserve balances.

9:17

Those are and that extinguishes those

9:20

reserve balances. So the three trillion

9:22

will go down for every hundred billion

9:23

of assets that are sold 100 billion of

9:26

reserve balances are extinguished. And

9:28

what would happen if the banks had a

9:30

hundred billion or 200 billion less of

9:33

reserve balances? Then they would, you

9:35

know, they would look each day and say,

9:38

"Well, gosh, today looks like I'm going

9:40

to need a lot of balances to meet my

9:42

payment needs. I am not allowed or I

9:45

feel that I should not go below zero in

9:47

my account at the Fed. So, I'm going to

9:51

if anybody calls me and wants to do uh

9:54

wants to me to lend money in them to

9:56

them in the repo market, I'm gonna have

9:57

to charge them a very high interest rate

9:59

to make it worthwhile." And in September

10:02

2019, that's exactly what happened.

10:04

Jaime Diamond, the CEO of JP Morgan,

10:07

famously

10:08

was asked in the earnings call

10:10

immediately after that episode, you

10:14

know, I'm paraphrasing, Mr. Diamond,

10:17

uh, the repo rates were going up

10:19

hundreds of basis points on September

10:21

the 17th.

10:23

Why didn't you use some of your reserve

10:26

balances, lend them into the repo market

10:28

and get those exceptionally high

10:30

interest rates that would have quelled

10:33

the disruption in money markets if you

10:35

were to have used your balances and loan

10:37

them to the market. And Diamond said

10:40

there's a red line. We cannot go below

10:42

zero because of liquidity regulations.

10:45

We are expected to be self-sufficient in

10:48

meeting our liquidity requirements. So

10:51

if we have to go to the Fed because we

10:52

run out of reserves in the middle of the

10:54

day, that's not going to look good. Here

10:56

I'm paraphrasing still. He said, and he

10:58

he named specific regulations. One's

11:00

called Reg Y and one's called

11:02

resolution, liquidity, and planning

11:04

requirements. And those those Fed

11:06

regulations say that the globally

11:09

systemically important banks like JP

11:11

Morgan need to be self-sufficient in

11:14

meeting their own liquidity needs. Now

11:17

they don't the regs don't say and

11:18

therefore you can't go to the Fed and

11:20

get more reserves but the largest banks

11:23

have interpreted it that way and they

11:25

feel that it would not look good if they

11:28

were to go to the Fed in the middle of

11:29

the day and say we ran out of reserves

11:30

can we get some more? Uh so they so they

11:34

don't and as Jamie Diamond said they

11:36

didn't lend money into the repo market

11:38

and interest rates in the repo market

11:39

skyrocketed to about over 300 300 basis

11:43

points above the interest rate that the

11:45

Fed pays to banks and intraday in the

11:47

interdeer market around a thousand basis

11:50

points above. The Fed recognized that it

11:52

had reduced

11:54

reserve balances below where they needed

11:57

to be for banks to meet their needs and

12:00

immediately reverse course and added

12:02

reserves to the system and that quelled

12:04

the problem.

12:04

>> So if the Fed sells too many assets or

12:08

lets too many assets go down and

12:10

reserves go down, the it loses control

12:13

of interest rates which is literally its

12:15

job. So it doesn't want to do that. We

12:16

saw that in extreme amounts as you said

12:18

in September 2019. I mean, wow. You said

12:22

a thousand basis points

12:23

>> in the inter dealer market. That day,

12:25

rates went to a thousand basis points

12:28

above normal, above the deposit rate

12:31

that the Fed provides, which is crazy.

12:34

Crazy.

12:35

>> 10 10% in what's supposed to be the

12:37

safest market in the world. Totally

12:38

crazy. And then you saw uh we can put up

12:41

some charts from later in the fall of

12:44

last year, a more muted but similar

12:47

there was some stress in that market.

12:49

And as such, the Fed has done liquidity

12:52

injections essentially by buying a

12:54

certain amount of Treasury bills in the

12:55

market every month. Right.

12:56

>> Yeah. Exactly. Right. After the blow up

12:59

in September 2019, the Fed got much more

13:02

conservative about letting this happen

13:04

again. And when there were inklings last

13:07

fall that there might not be enough

13:09

reserve balances and interest rates and

13:11

the repo market were creeping above

13:13

where the Fed was targeting them,

13:15

especially at the end of October, the

13:18

Fed realized, okay, this is it. We

13:20

better add more balances to the system

13:22

and we can do that by buying Treasury

13:24

bills.

13:25

>> Yes. And the level of reserves went up a

13:28

huge amount after in in 2008. Basically

13:32

the Fed used to control interest rates

13:34

by controlling the level of reserves and

13:37

now they control interest rates by just

13:40

setting the the interest rate and paying

13:42

interest on reserves. So the lever

13:44

through which they control interest

13:45

rates is no longer the the level of

13:46

reserves. But as a consequence they

13:49

can't reduce the amount of reserves a

13:52

ton or they they literally lose control.

13:54

Okay, perfect summary. And I would just

13:56

add it's asymmetric because if the Fed

14:00

loads up too much reserves, it doesn't

14:02

really cause a problem for monetary

14:05

policy implementation because as you

14:07

said, market interest rates are guided

14:11

by the interest rate that the Fed pays

14:13

on reserve balances, not by the quantity

14:16

of money. However, if you don't have

14:19

enough reserve balances in the system,

14:22

that can get messed up and suddenly you

14:25

can get these, you know, very volatile

14:27

periods when uh core interest rates like

14:30

the repo market rates uh go go way above

14:33

where they're supposed to be.

14:34

>> Right. So, prior to the fall of 2008,

14:37

the Fed was under a so-called floor

14:39

system and now we are in a so-called

14:41

corridor system. So, people can can look

14:42

that up later. All right, professor.

14:45

What are the four recommendations that

14:48

you have? And perhaps should we start

14:50

with the simplest and least extreme sort

14:54

of least controversial ones and then

14:56

we'll move on.

14:57

>> Well, they're all kind of controversial

14:58

actually,

14:59

>> but some of them are, as you say, are

15:01

less extreme and easier to do. And I and

15:03

I wrote I wrote a paper for the

15:05

Brookings Institution

15:07

uh earlier this year in the Brookings

15:10

papers uh for econom on economic

15:13

activity and uh I gave you the floor on

15:16

the Fed's balance sheet.

15:17

>> Yeah, that's it. Okay, good.

15:18

>> I gave four idea, four options in kind

15:21

of increasing order of difficulty and

15:24

also impact. And the most

15:27

straightforward one and even that this

15:29

one is not a you know is not obvious and

15:31

is controversial within the Fed is

15:33

temporary open market operations. So uh

15:37

you know rather than uh setting the

15:40

course of reserve balances uh on a kind

15:44

of autopilot you know adding 40 billion

15:46

a month or keeping it flat and then

15:48

waiting to see what happens. the Fed

15:50

could on a day-to-day basis offset

15:55

uh sudden changes in reserve balances by

16:00

uh doing what are called temporary open

16:02

market operations. So when reserve

16:04

balances pop down due to something else

16:07

going on like the the Treasury

16:09

Department needs a lot of uh deposits in

16:12

the Treasury General account and they

16:14

come out of reserves suddenly the

16:15

there's not enough reserves. The Fed

16:17

could do a temporary open market

16:19

operation, which means they could uh

16:22

conduct their own repos to create more

16:25

cash to create more reserves and that

16:27

would fill in that those those missing

16:30

reserves. Think of the think of this as

16:32

you're you're you're running on a along

16:34

a smooth highway and there are potholes

16:37

and the potholes are the the level of

16:39

the highway is the quantity of reserve

16:41

balances in the system and occasionally

16:43

you're going to hit one of these

16:44

potholes and so you want to fill it in

16:46

with temporary open market operations

16:49

and that's that's the easy step and that

16:52

could reduce the average path of reserve

16:56

balances by perhaps a 100red billion or

16:58

200 billion. Not a huge amount, but um e

17:02

easy easy easy in terms of policy. The

17:06

FOMC could simply direct the market

17:09

operations group at the New York Fed to

17:11

keep an eye on reserve balances and

17:13

whenever they pop down, fill them back

17:14

up again the next day. And and if you

17:17

want to save on average, it goes the

17:20

other way. Whenever they pop up

17:21

unexpectedly, that means you don't need

17:23

as much. You can take take some out of

17:25

the system. And and the Fed is familiar

17:28

with these kinds of operations. it would

17:29

be relatively straightforward to do

17:31

them, but it's not as straightforward as

17:33

a policy matter because some policy

17:35

makers at the Fed believe that they

17:38

shouldn't be active in the markets on a

17:39

daily basis filling in these holes and

17:41

take and chipping off these bumps. They

17:44

should just set set a kind of steady

17:46

course for their operations and not

17:48

touch it until you know something big

17:50

happens. And under what circumstances

17:52

would the Fed encounter these potholes

17:56

and can you just remind us of the

17:57

figures of how much hundreds of billions

17:59

this would reduce reserve demand by?

18:01

>> Yeah. So here here are a couple of

18:03

examples. Uh one is it's tax season.

18:06

I'll give you one example. And suddenly

18:09

taxpayers are going to pay their tax

18:11

bills and that includes corporates and

18:13

individuals. And this happens a lot in

18:16

you know in the spring particularly in

18:17

April. Well, if there's an influx of

18:19

money into the Treasury general account

18:22

where people are paying taxes in, it has

18:24

to come out of reserve balances because

18:26

that's the only way ultimately that the

18:28

Treasury gets paid by banks and people

18:30

pay their taxes through the banks.

18:33

So when tax season starts gushing money

18:36

into the Treasury general account, it's

18:38

gushing money out of reserve balances

18:41

and the the Fed could say, "Oh, we

18:44

didn't really, you know, we thought we

18:46

had about the right amount, so let's

18:48

replace those missing balances with new

18:50

balances that we can create with

18:52

temporary open market operations." So

18:54

that's one example. Another example is

18:55

that at the end of every quarter,

18:58

foreign banks that have accounts at the

19:00

Fed are monitored for capital adequacy.

19:04

And uh suddenly they want to reduce

19:08

their balance sheet so they look good on

19:09

capital adequacy. And the easiest way to

19:13

make their balance sheets look smaller

19:15

for one day

19:17

is to get rid of a bunch of reserves.

19:20

And that that's happened to the extent

19:23

to the tune of 200 to500 billion dollars

19:26

over the last few years.

19:28

>> Sorry. Do you mean get rid of a bunch of

19:30

reserves or get rid of a bunch of assets

19:31

and increase reserves?

19:32

>> Well, their reserves are assets.

19:35

>> Okay. Okay. Okay.

19:36

>> Their reserves are their deposits at the

19:39

at the Fed. That's on the asset side of

19:41

their li of their balance sheet. And so

19:43

they have to have capital against all

19:45

their assets. And so an easy way to get

19:47

rid of some of their assets is to ditch

19:50

a bunch of reserves. And they do that

19:52

systematically for one day at the end of

19:54

each quarter. Now you might ask, well,

19:55

why do foreign regulators allow this?

19:57

And that's another conversation because

20:00

I mean obviously the other 90 days in

20:02

the quarter, they're not meeting their

20:04

capital requirements.

20:06

Uh but they're only monitored on the

20:08

quarter end. So that's all the only

20:09

thing they care about.

20:10

>> Yes, this I we talked about at length in

20:13

our first conversation. People can look

20:14

that up on January 2025. Professor, I'm

20:18

going to share there are two charts in

20:20

this section in your paper which we will

20:21

attach uh this chart and then the

20:24

subsequent chart. Can you explain what

20:27

we're looking at here and and the

20:29

relevance?

20:30

>> I think we should uh probably do this at

20:32

a very high level because first of all

20:35

as I said this is not that

20:37

consequential. It can save you a 100red

20:39

billion or 200 billion in the average

20:41

path of reserve balances. And so it's

20:43

not like really big game. Although it is

20:45

it is high on the list in terms of ease.

20:48

The other ones, wait till you get to the

20:49

other ones are going to be more and more

20:50

difficult. But anyway, uh briefly going

20:53

through it, the top panel, there's three

20:55

panels on this page. The top panel shows

20:58

the cumulative amount of these temporary

21:01

open market operations that would be

21:03

used to offset the bumps in the path of

21:07

reserve balances held in the system. And

21:09

I'm colorb blind, but I think one of the

21:12

colors is green. The the one that and

21:14

that's the TGA. So that that's the

21:17

offset. Those are the temporary open

21:19

market operations that would offset

21:21

fluctuations in the Treasury General

21:23

account. The red bars at the end of each

21:25

quarter, those are the ones that offset

21:27

the quarter end balances that are lost

21:29

when foreign banks reduce their

21:31

reserves. And the the blue line is the

21:34

one that offsets changes in reserve

21:36

balances caused by foreign central banks

21:39

that store money at the Fed and in in

21:43

what's called a FEMA repo uh operation,

21:48

which is technical. And I I think that I

21:50

think that's probably enough from this

21:52

page.

21:53

And then uh these are bar charts showing

21:56

on the left the largest reductions in

22:00

reserve balances with these temporary

22:02

open market operations and then without

22:05

them. And if you have really good eyes,

22:08

you can probably see that the ones with

22:10

the reserve balances, you know, they're

22:12

not so big on the ta in the big tail on

22:15

the right. the very large uh reductions

22:17

in reserve balances are much smaller

22:19

when you when you uh conduct temporary

22:22

open market operations. And the right

22:25

the right hand bar chart is a histogram

22:27

of the sizes of these temporary open

22:29

market operations. So you can see they

22:31

range from increasing reserve balances

22:33

by about $300 billion on a given day to

22:36

reducing it by about 300 billion on a

22:38

given day

22:40

reserve. So these are sometimes pretty

22:42

big operations. I hope you're enjoying

22:44

today's episode. One of the most

22:46

powerful vehicles in finance for

22:47

generating reliable, consistent cash

22:49

flow is real estate income. It's a yield

22:52

generating strategy that's backed by

22:53

more than just cash flows. It's backed

22:55

by hard assets like property and land.

22:57

Investors in real estate credit tend to

22:59

hold priority positioning, meaning they

23:00

typically get paid first. Real estate

23:03

income is a hands-off, highly structured

23:05

investment designed to deliver

23:06

consistent and insulated returns. No

23:08

operational friction, no management

23:10

headaches, just precision, priority, and

23:12

predictable yield. Historically, many of

23:14

these real estate credit strategies were

23:16

reserved exclusively for institutional

23:17

investors, but now they're available to

23:20

all investors thanks to the Fundrise

23:21

Income Fund, which has more than $600

23:23

million invested at a 7.9% distribution

23:26

rate. The income fund aims to deliver

23:28

high current yields from a diversified

23:30

portfolio of Fundrise's most favored

23:32

real estatebacked fixed income

23:33

strategies. Visit fundrise.com/m

23:36

to learn more and invest in the Fundrise

23:38

Income Fund. That's fundrise.commm.

23:41

Past performance does not guarantee

23:42

future results. Current distribution

23:43

rate as of June 30th, 2026. Carefully

23:45

consider the investment material before

23:47

investing, including objectives, risks,

23:48

charges, and expenses. This and other

23:50

information can be found in the income

23:51

funds perspectus at fundra.com/income.

23:54

This is a paid advertisement. Thanks for

23:56

listening. Let's get back to today's

23:57

interview. Yes. and you describe them

23:59

as, oh, we're going to start with this

24:01

because these are the least

24:02

consequential and it's not big. It's

24:04

quote unquote only a few hundred

24:05

billion, but even for the Fed, a few

24:07

hundred billion is a lot. All right,

24:08

that's number one. What's the second

24:10

one?

24:11

>> Second one is uh the liquidity

24:13

regulations that Jamie Diamond spoke

24:15

about in that earnings call. So he

24:17

mentioned a couple uh one is called Reg

24:20

Y the other is called RLAP which state

24:24

as principle as a principle that the the

24:27

globally systemically important banks

24:29

should have enough liquidity to meet

24:31

their own needs without going to the

24:33

Fed. What does that mean? It means

24:35

they're reluctant to go to the discount

24:38

window to go to the Fed for repos to get

24:42

more reserves and they're also reluctant

24:45

to overdraft their reserve balance

24:48

account at the Fed. They have this

24:50

deposit account. Think of it as a

24:51

checking account. They're in they're

24:53

allowed to overdraft, but they don't

24:55

want to because it would look like they

24:57

didn't have enough liquidity of their

24:58

own. Uh so these liquidity regulations

25:04

could be

25:06

the Fed could emphasize in

25:08

communications to the largest banks, hey

25:11

look, yeah, we want you to be

25:12

self-sufficient, but on the other hand,

25:14

we don't if you need more reserves,

25:16

please come and see us. We have all

25:19

these facilities that are available. We

25:21

don't want you to go short and we don't

25:23

want to mess up our monetary policy

25:25

implementation by having rates skyrocket

25:28

just because

25:30

you're reluctant to come to the discount

25:32

window or use standing repo operations

25:35

or overdraft your deposit account at the

25:38

Fed. Just go ahead and use those when

25:40

you need them. Now it sounds a little

25:43

bit contradictory because on the one

25:44

hand we the Fed's saying you have to be

25:46

self-sufficient which implies don't come

25:48

to the Fed and the other hand you're

25:50

getting they're getting the message do

25:52

come to the Fed.

25:54

Uh and you know if you ask uh in my view

25:56

if you ask the leadership of the Fed

25:58

would you want these banks to come to

25:59

the Fed when they need more balances

26:01

they would all say absolutely that's

26:03

what those facilities are for. But if

26:05

you ask us, let's say a supervisor down

26:08

in the middle ranks of the supervisory

26:09

staff at the Fed,

26:11

um if they are ever concerned when they

26:15

see a bank go to the discount window to

26:17

get more balances, they might say,

26:19

"Yeah, well, I mean, of course we're

26:21

concerned."

26:22

Uh we're monitoring that situation very

26:25

carefully, and these banks are supposed

26:26

to be self-sufficient. There's a

26:27

regulation about that.

26:30

So, you know, it if you're the manager

26:33

at the bank that's responsible for uh

26:36

ensuring there's enough liquidity in the

26:38

middle of each day, you're probably

26:41

don't want to get uh a nasty memo from

26:44

your regulatory supervisor saying, "What

26:47

happened today? Why why did you have to

26:49

go to the discount window? That doesn't

26:50

look very good to us." So, it's kind of

26:53

like if you know there's a consumer who

26:56

had and they have a say a $12,000 credit

26:59

limit officially on paper from a bank,

27:02

but then anytime they spend more than

27:04

$300, they get these annoying calls from

27:07

the bank and they say, "What are you

27:08

doing? What are you What's going on?"

27:10

>> You know, I've never had one of those.

27:12

[laughter] So, so on paper they they say

27:16

we want you to use these liquidity tools

27:18

where you can borrow in times of stress

27:20

but they are kind of haunted by the

27:22

ghost of this regulation from after 2008

27:24

which I'm not saying is a bad thing of

27:26

they they need to hold a absolutely

27:29

massive amount of reserves and the

27:31

people who are supervising the banks at

27:33

the Fed and perhaps the FDI too whenever

27:35

they see the the banks use these tools

27:39

that they're supposed to be using in

27:41

their times of stress they get nervous.

27:44

>> That's a good summary, Jack.

27:45

>> Okay.

27:46

>> And you know, even if uh the supervisors

27:49

would just hold fire and say, "Okay,

27:51

well, I understand these things happen.

27:53

I'm not going to say anything." What

27:55

really matters is the perception at the

27:57

banks. If they perceive that, they might

27:59

get a black mark or a memo might be sent

28:02

to their boss saying, "Uh, what went

28:04

wrong?" If they have that perception,

28:06

that's enough for them to demand a lot

28:09

of reserves at the start of each day so

28:11

they don't go near zero.

28:13

>> So, how do we solve this stigma issue

28:17

and where where the Fed's standing repo

28:19

facility or in the discount window,

28:21

these tools to provide liquidity under

28:22

times of stress has barely been touched

28:25

even when the market repo rates blow

28:27

past the administered rate. And you

28:30

you've pointed out that there are other

28:31

central banks around the world such as

28:32

the Bank of England that have

28:34

transitioned to a demand driven system

28:37

without the stigma. How do we remove

28:40

this stigma? It sounds like a very hard

28:42

thing to do because the issue is

28:44

something that's intangible. It's kind

28:45

of like how how do you if you're a

28:47

marketing person, how do you change this

28:49

perception of a brand in a consumer's

28:51

head? Like it's it's not something

28:53

physical you can change. It's it's in

28:55

people's heads. There's two approaches

28:58

and one is to force the issue by

29:00

reducing the quantity of reserve

29:02

balances where the banks don't have a

29:04

choice. They have to go to the Fed

29:06

because there's no other way to meet

29:08

their required you know payment needs

29:11

and maintain their customer

29:12

relationships and you know do run their

29:15

businesses. So the the problem with that

29:18

forcing it by that approach is that

29:22

interest rates will get very volatile on

29:25

days when banks are running short

29:27

because they're not sure it's okay and

29:30

they're not, you know, they're going to

29:31

hold back. They're going to and they're

29:33

going to charge higher interest rates to

29:34

their counterparties and rates will get

29:36

quite volatile and banks will eventually

29:38

say, "Wow, these rates are exceptionally

29:40

high. I'm going to take advantage of

29:41

these. I need to borrow the money

29:44

anyway." rates will get volatile. people

29:46

will get used to that and then over time

29:49

the the those managing liquidity at the

29:51

largest banks will say hey you know what

29:55

it's been okay I mean there's not that

29:57

much stigma other banks are doing it we

29:59

can do it too and nobody seems to be

30:01

getting in trouble

30:03

uh so it's new the new business as usual

30:06

and that's what the bank of England was

30:07

able to do they have actually quite

30:08

volatile repo rates and they're okay

30:12

with that and the banks in the UK system

30:14

are okay with that there's sort of an

30:16

understanding that if you need the money

30:19

you you go to the bank of England but in

30:20

the meantime rates might be volatile so

30:22

it's a different regime if the US

30:24

financial system could adapt to that

30:26

it's called demand-driven regime then

30:29

rates would be more volatile but the

30:31

demand for reserve balances would also

30:33

go down the Fed could have a smaller

30:35

balance sheet so that's approach one

30:38

approach two is communication the Fed

30:40

could communicate very clearly and often

30:44

to the banks and to their own

30:45

supervisory staff. These facilities are

30:48

there to be used. We want

30:52

all the banks to feel free that they can

30:54

use these for their legitimate liquidity

30:57

needs. So, you know, we're open for

31:00

business. That that communication

31:02

strategy could also be successful. Of

31:05

course, the Fed has to get comfortable

31:07

with the idea that if these facilities

31:09

do get used a lot, then technically

31:11

these banks are not self-sufficient for

31:13

liquidity. they're violating some of the

31:15

principles of these regulations. I guess

31:17

there's a third approach which is change

31:19

the regulations

31:21

uh so that the you know the banks are

31:24

not required to be self-sufficient at

31:25

all times. There are exceptions.

31:27

>> Okay. And I believe this is the only

31:30

section in your paper that does not have

31:33

a chart. And that is indicative to me

31:35

that this is the most as I said before

31:38

intangible thing. It's hard to measure.

31:40

It's either they press a button saying

31:42

your reserves have to go down or it's

31:44

via communication, which is a very soft

31:47

tool that at least with regards to the

31:49

press, incoming Fed Chair Kevin Wars

31:51

does not seem to love that that tool. So

31:54

I I'm not going to I'm not calling this

31:56

a little wishy-washy, but would you

31:57

would you agree with me that it's a a

31:59

little harder to implement? Maybe

32:01

>> it's hard to model for sure to quantif

32:03

quantify. Uh the reason it's hard to

32:05

quantify is

32:06

>> as I said, it's about perceptions. It's

32:08

about perceptions at the Fed. Do we

32:09

really want the banks to use these

32:11

facilities in a way that demonstrates

32:13

that they're not self-sufficient? And if

32:15

so, can we communicate that? And then

32:17

there's perceptions at the globally

32:20

systemically important banks, you know,

32:22

are is it really okay uh uh kind of

32:26

reminds me of the movie Jaws. Is it okay

32:28

to go into the water or am I going to

32:31

get bitten by a shark? Uh if everybody's

32:33

else is in the water, I'm probably going

32:35

to go in, too. And as long as there's no

32:37

shark bites, I'll probably stay in the

32:39

water. So that's basically a perceptions

32:41

thing, right? And it's an equilibrium.

32:43

Uh there's an equilibrium to it because

32:44

if everybody else is doing it, I'm going

32:46

to be willing to do it as a bank. But if

32:48

nobody else is doing it, I don't want to

32:49

stick out like a sore thumb and be

32:52

drawing on the Fed's liquidity and

32:53

standing out like um you know, maybe I'm

32:56

a weaker bank.

32:57

>> Okay. So door number one is temporary uh

33:00

temporary market operations from the

33:02

Fed. Door number two, suggestion number

33:04

two is liquidity regulations.

33:07

What is tool number three that you

33:09

suggest for reducing reserve demand as a

33:11

way for the Fed to reduce its balance

33:13

sheet?

33:14

>> So the third uh approach in in my paper

33:18

is what's called the liquidity savings

33:20

mechanism which sounds like plumbing and

33:22

it is plumbing. So most, in fact, all

33:25

large developed market central banks

33:27

other than the Fed have a way to

33:30

conserve

33:32

on the amount of reserves that a bank

33:34

needs to make its payments every day. So

33:37

let's do a little role play. You're you

33:41

owe me uh10 billion today. I mean,

33:44

you're going to pay me$10 billion in

33:46

cash. It's going to it's going to come

33:48

to me two different ways. One is you're

33:51

just going to take 10 billion out of

33:53

your current stash of reserves and send

33:55

it to me. And by the way, it gets sent

33:57

through a payment system called Fedwire,

33:59

which is the world's largest payment

34:01

system. And the way you send it to me is

34:04

just send a message saying, "Take 10

34:05

billion out of Jack's account and put it

34:08

into Daryl's account at the Fed."

34:10

>> I don't like that. I like the other way.

34:11

>> Yeah. Okay. Well, sorry. You're going to

34:13

get it's that's going to come up

34:14

shortly.

34:15

>> Okay.

34:16

>> Okay. Now, it turns out uh that

34:20

unknown to you, there's a $12 billion

34:24

payment coming from JP Morgan to you.

34:27

>> Sounds good.

34:28

>> But you're not sure when it's going to

34:30

get there, but you're kind of hoping

34:32

maybe it'll get there in time. Uh that

34:34

you could use that money to pay the 10

34:36

billion to me. So rather than take the

34:38

10 billion out of your current stash,

34:41

you could send a message to a piece of

34:44

software at the Fed if it builds it, a

34:48

software called a liquidity savings

34:50

mechanism. And this the software says

34:54

uh Jack's message indicates he will be

34:58

paying Daryl 10 billion

35:01

in the next round, which could be 10:00

35:03

this morning.

35:06

And now the LSM, the liquidity savings

35:08

mechanism has that message in place.

35:10

Then it get gets a message from the

35:14

payments manager at JP Morgan saying we

35:17

are going to be paying Jack 12 billion

35:20

at 10:00 by 10:00 this morning. And then

35:24

the LSM sees a message from me saying

35:26

I'm going to be paying JP Morgan $8

35:30

billion

35:32

by 10:00 this morning. But you don't

35:33

know about these other messages. Now the

35:36

LSM has three messages. One saying Jack

35:38

is paying Daryl 10. The other saying

35:41

Daryl is paying JP Morgan 8 and the

35:44

another message saying JP Morgan is

35:46

paying Jack 12. And you see that little

35:48

circle of payments. 10 81 12. So the LSM

35:52

can see that loop of payments and say,

35:54

you know what, Jack doesn't really need

35:56

to take 10 billion out of his his

35:58

account to make that payment to Daryl.

36:00

What we could do is we could credit his

36:03

account with the 12 billion from JP

36:05

Morgan at the same time that we debit

36:08

his account for the 10 billion paid to

36:10

Daryl. So net he gets 2 billion coming

36:12

in at 10:00 this morning. Did you notice

36:16

that? Uh did you notice that uh

36:19

cancellation your outgoing payment never

36:21

came out of your reserve balances. It

36:23

came it was cancelled against an

36:26

incoming payment to you of 12 billion

36:27

from JP Morgan. If you're showing

36:29

charge, yeah, the right hand side one is

36:31

the situation I just described with

36:33

different numbers. So there are three

36:34

banks, A, B, and C, and they're paying

36:36

each other these amounts. The LSM

36:38

notices there's this circle of payments,

36:41

and it simply cancels the largest amount

36:43

that it can cancel in this loop, which

36:46

is 20 billion. And that that means that

36:48

the banks didn't have to pay 20 billion

36:50

out of their out of their account. They

36:53

got 20 billion, and they paid 20 billion

36:55

by just cancelelling incoming against

36:57

outgoing. And you might say, well, kind

36:58

of that's kind of obvious. I mean, why

37:00

wouldn't why wouldn't you do that? Uh,

37:02

it's kind of natural, right? Well, you

37:04

need the software that detects these

37:05

loops. And the Fed doesn't have that

37:07

today. And so, all the banks, they can't

37:10

plan on these cancellations. They have

37:11

to start each day knowing that they have

37:13

enough to do it on their own with some

37:15

guesswork about incoming payments. And

37:18

what happens on a stress day like

37:19

September the 17th, 2019, there was a

37:22

record late payments. payments came into

37:25

the largest banks over 150 minutes later

37:29

than normal because everyone was holding

37:31

back, not knowing if they were going to

37:32

have enough to make their payments that

37:34

day. that was the that was the all-time

37:36

record in the in the surrounding years

37:38

uh for the data set that from some work

37:41

that I did with Adam Adam Copeland at

37:42

the New York Fed and Elen Yang who's at

37:45

the University of Minnesota we showed

37:47

that these stresses in money markets

37:50

arise when the banks are delaying their

37:52

payments because they don't have enough

37:53

balances to meet all their payment

37:55

needs. So if we were to if the Fed were

37:57

to introduce a liquidity savings

37:59

mechanism like the one illustrated uh

38:02

which is from the Bank of Japan then

38:04

banks wouldn't be as stressed. They

38:07

would be able to meet their payment

38:08

needs with much less reserve balances.

38:10

The Bank of England has a system that

38:12

does this also. It's called CHAPS. And

38:15

they've estimated a 20 to 30% saving in

38:20

the amount of reserve balances needed to

38:22

make payments. And whether that that

38:25

would be a bigger saving or a smaller

38:27

saving for the Fed, we don't know. But

38:29

I'm hoping to do some research on this.

38:31

I already have a theoretical uh paper

38:33

that shows why banks would want to do

38:36

this.

38:36

>> That's the piece that came your most

38:38

recent piece that came out earlier in

38:39

June.

38:39

>> Uh just came out. Yes. It's called

38:41

something like optimal liquidity savings

38:44

mechanisms and that's with uh Chao Wong

38:46

at the University of Pennsylvania and

38:48

Shrist Singh at the University of

38:50

Toronto.

38:50

>> We will link that paper as well. Darl,

38:53

as you referenced, you said people might

38:56

think this is obvious. I had that exact

38:58

same thought because when reading this

39:00

paper in preparation, I'll be honest, I

39:02

got a little nervous. These are like

39:04

very very complex ideas that are on the

39:07

cutting edge. Like many of them haven't

39:08

been implemented at least in the US.

39:10

Perhaps they've been implemented around

39:11

the world with other central banks of

39:13

foreign that that you work with and

39:15

advise as well. But this is deceptively

39:19

simple simple. Like isn't this the same

39:20

thing as like let's say we were coming

39:22

off a poker game and I owed you $100 and

39:25

you owed another professor at Stanford.

39:27

You owed Hannah Lustig $100. We would

39:30

just net it out and I would pay Hano

39:31

$100 instead of you me paying you a

39:33

hundred and you paying him a hundred.

39:35

>> Yeah. you probably wouldn't need to

39:36

bring as much cash to the game because

39:38

on average you're if you netted all this

39:40

out, you wouldn't need as much cash. And

39:43

and that's right. Uh if you do it in a

39:46

sequence of steps without doing these

39:47

netting cycles, you need more cash. And

39:49

that's been demonstrated both

39:51

theoretically and in empirically. The

39:53

Bank of Canada has this, the Bank of

39:54

England has this, the European Central

39:56

Bank has this, the Bank of Japan has

39:58

this, other central banks have it, but

40:00

the Fed, which runs the world's largest

40:02

payment system, does not have it. And so

40:05

it would be good at a minimum for the

40:07

Fed to investigate how much could they

40:09

save on reserve balances by introducing

40:13

a liquidity savings mechanism. I'm

40:15

presenting that paper at the New York

40:16

Fed next week and that's the main theme

40:18

of my presentation that the Fed should

40:21

investigate using their own data for

40:24

payments

40:25

and using you know the kinds of

40:28

approaches maybe maybe the one I

40:30

suggested in our paper or maybe one of

40:31

the conventional approaches used at the

40:33

other central banks and just do a test

40:35

run. Yes, it you know not not to add too

40:39

much commentary but I think people who

40:40

are not familiar with the banking system

40:42

at all when they first find out that the

40:44

way a bank makes a loan is just by

40:47

crediting the borrower with cash and

40:50

then suddenly a loan appears on the

40:52

asset side of the bank and a liability

40:54

in terms of a deposit. The cash that

40:56

they gave them that they lent them

40:57

appears as a deposit. Like I think it it

40:59

it blows people's minds. It blew my

41:01

mind, you know, when I first found that

41:02

out. But it seems like the deeper you

41:04

dive into the the sort of the guts of

41:06

the banking system, literally going to

41:07

the most, you know, priority level of

41:10

the of the Fed, it does seem like we

41:12

still are on a very hard money system,

41:15

particularly when it comes to reserves

41:16

of like, no, I need that. I need the

41:18

gold in my vault right now.

41:20

>> Yeah. Well, I mean, and these liquidity

41:22

savings mechanisms ultimately you still

41:24

need the gold in the vault. You uh and

41:26

the gold in this case is reserves held

41:28

at the Fed, but they're electronic

41:30

balances. And so with the with the the

41:32

miracles of software, you can economize

41:35

a lot on how much of that kind of gold

41:37

you need.

41:38

>> Okay. And you can you talk about other

41:39

you said Canada, I think talk about the

41:41

other central banks who have this and

41:43

just how far behind is the Fed in not

41:46

having this and then of course we have

41:47

some a few other charts that that may be

41:49

relevant here.

41:49

>> Well, it's one of those things where you

41:51

know the leading countries go first and

41:53

the other countries follow and leapfrog

41:55

and that's what's happened here. The Fed

41:56

was the first to introduce

41:59

real-time gross settlement, which means

42:01

this the kind of payment system that

42:03

we've been talking about, but they did

42:05

it before the introduction of liquidity

42:08

savings mechanisms. Then the other

42:10

central banks introduced this real-time

42:12

growth settlements uh coming after the

42:14

Fed and they realized this is kind of

42:16

expensive for in terms of the quantity

42:20

of uh reserves that are necessary to run

42:22

the system. They introduced liquidity

42:24

savings mechanisms. And the Fed, which

42:27

has had since the crisis at least, tons

42:29

and tons of reserves, never had to think

42:31

about it. But now that it's trying to

42:33

reduce its balance sheet, it should

42:35

start to think about maybe we should do

42:38

what the other central banks have done

42:39

and add these this new software.

42:42

>> Okay. So, professor, if people want to

42:44

and they should check out that that

42:47

again that is in the first paper which

42:49

is called the payment system puts a

42:52

floor on the Fed's balance sheet that is

42:54

in the section about the liquidity

42:56

mechanism. Let's talk about the final

42:59

one which is called the taring the

43:02

remuneration of reserves section six.

43:05

What's going on here?

43:06

>> Okay. Uh well, as you noted before the

43:09

financial crisis, the Fed did not pay

43:12

interest to banks on their on their

43:14

reserve balances. And and here's think

43:17

of yourself again as liquidity manager

43:20

at a very large bank. Maybe you're at a

43:23

city bank and you're looking at your

43:26

holdings of balances at the Fed. This is

43:28

2007 before the crisis. The Fed's not

43:31

paying you interest. And you've got, you

43:33

know, uh $5 billion sitting there in

43:36

reserve B. Well, actually in those days

43:37

you had much less. Say you have 500 $500

43:40

million sitting there. You're not

43:42

getting any interest on that money. What

43:45

are you going to do? Well, you're going

43:46

to try to cash it in for something that

43:49

pays interest. So, buy treasury bills.

43:51

Uh, you know, invest in something that

43:54

pays interest because why why would you

43:56

sit on on reserves that are not paying

43:58

interest? Okay. So with that motive,

44:00

banks hated to hold reserve balances

44:03

more than they need the more than the

44:04

minimum needed to run business in the

44:06

precrisis era. They hated to have

44:08

reserve balances. Now let's move to the

44:11

post crisis world where as you noted the

44:14

Fed is now doing monetary policy by

44:17

paying by setting the interest rate they

44:19

pay to banks and they set it at a rate

44:23

where they want the market to be

44:24

roughly. So if they say we want uh

44:27

market rates to be at 3 and a half%

44:30

today then they set the interest rate

44:32

they pay to banks at about three and a

44:34

half%.

44:37

Now you're you know it's it's uh 20

44:40

years later you're you're still the guy

44:43

running liquidity at City Bank and

44:45

you're saying well you know I've got5

44:47

billion sitting at the Fed but you know

44:50

that's fine because I'm earning three

44:51

and a half% interest. If I were to cash

44:53

it in for something in money markets,

44:55

I'd be getting about 3 and a half%. So,

44:58

no urgency. I'm happy to own those

45:00

reserve balances. That's great. And I

45:02

can use them to make payments or I can

45:03

just use them as an investment.

45:06

And in my paper, I note that JP Morgan

45:08

was holding hundreds of billions as an

45:10

investment that they did not need to

45:12

make payments a few years ago. They've

45:15

subsequently got rid of those because

45:16

it's no longer a great investment given

45:19

the term structure of interest rates.

45:21

But the data show that the banks really

45:25

don't worry much about having a lot of

45:27

reserve balances held to the Fed because

45:30

they're getting lots of interest.

45:32

So that that means there's a healthy

45:34

demand for reserve balances. They're

45:36

like the Swiss Army knife of finance.

45:37

They're good for everything. Meet

45:39

liquidity rags, make payments, great

45:41

investment, you're getting market rates.

45:44

What you know why not have them and load

45:47

up on them? Well, here's a way to

45:50

discourage that the demand for reserve

45:53

balances without messing up the system.

45:55

It was invented by the Reserve Bank of

45:57

New Zealand in 2007. That's the same

45:59

central bank that brought us the 2%

46:01

inflation target. So, they're an

46:03

innovator. And they said in 2007, you

46:05

know, if we want to implement monetary

46:07

policy by setting the deposit rate on

46:10

reserves, fine. But we don't have to pay

46:12

the same high rate no matter how much

46:14

balances a bank has. We could pay Jack's

46:17

bank the full market rate for the first

46:20

let's say required amount that he needs

46:22

to run his business and then once he has

46:24

more than he needs to run his business

46:26

we could drop the interest rate on the

46:27

rest to a lower level and in the case of

46:30

RBNZ was 100 basis points lower recently

46:34

as you can see in this chart on the left

46:35

hand side when the Reserve Bank of New

46:37

Zealand did that the banks in New

46:39

Zealand said well yeah I mean we need a

46:42

certain amount to run our business and

46:43

we're getting a market rate on that but

46:45

let's not hold any more than that

46:47

because why would we we can get you know

46:49

we could we could get rid of those extra

46:51

reserves and earn a higher interest rate

46:53

on on some uh securities. So by doing

46:56

this, the Reserve Bank of New Zealand

46:57

lowered the demand for reserve balances

47:00

without impairing its ability to

47:02

implement monetary policy and without

47:04

messing up the ability of banks to meet

47:06

their payment needs because they you

47:10

know they got full freight interest on

47:11

the amount they needed to run their

47:13

business and the rest they got a lower

47:15

rate but that they didn't need that

47:17

extra money anyway. I think you got the

47:19

idea. Several other central banks have

47:21

done this. uh Norisbank which is the

47:23

central bank of Norway

47:25

the South Africa central bank other

47:28

central banks have done this

47:30

and it's it's worked and it's enlivened

47:33

also the market for overnight borrowing

47:36

and lending between banks because if you

47:39

want to get rid of your extra balances

47:41

the easiest way is to lend them to

47:42

another bank that needs them and so you

47:45

get a better distribution of reserve

47:47

balances in a system you don't have to

47:49

load up every bank you just put as much

47:51

as you need in the system if you're the

47:53

central bank and then the banks will

47:55

sort it out by borrowing and lending in

47:57

the secondary market. That's not been

47:59

working well in the United States. The

48:02

secondary market for reserves which is

48:03

called the federal funds market is just

48:05

not working uh hardly at all. The

48:07

volumes are extremely low. So in order

48:10

to enliven the secondary market, get

48:12

better sharing of reserves and reduce

48:14

the demand for unneeded reserves, the

48:16

Fed could do what the Reserve Bank of

48:18

New Zealand did and just reduce the

48:20

interest rate on the unneeded part of

48:23

the reserve balances. That's easier said

48:25

than done because now you have to find

48:28

out for each type of bank, how much do

48:30

they actually need?

48:33

And my paper discusses that. And it is

48:36

it's not an easy problem, but it is

48:37

surmountable. And there are approaches

48:40

uh in the literature that describe how

48:41

to do it. And in fact, the Fed did it uh

48:44

for a brief period of time after the

48:46

failure of Lehman. It's not well known

48:48

that for a number of weeks after the

48:51

failure of Leman, the Fed ran exactly

48:54

this two-tiered system of remuneration.

48:57

You pay a higher interest rate on

48:59

required reserves and then a low

49:00

interest rate on the rest. And uh that

49:04

experiment was squashed by the financial

49:08

crisis because the Fed had to drop all

49:10

interest rates to zero, the zero lower

49:12

bound. And so the high tier rate and the

49:14

low tier rate both went to zero. And

49:17

then once rates reemerged into positive

49:20

territory, this

49:22

this two-tiered approach disappeared.

49:24

It's never come back, but it could come

49:26

back. And if it did, it would cause a

49:28

dramatic reduction in the demand for

49:31

reserve balances in the US system.

49:33

>> How much?

49:33

>> I don't know because I haven't I' I've

49:36

collaborating uh with two of our PhD

49:38

students.

49:40

Uh Franchesco Spitz Spitzu and Thanowat

49:43

Sornwani. We built a theoretical model

49:46

showing that it really works

49:47

theoretically, [laughter]

49:49

but we haven't calibrated it to the US

49:51

economy. So I can't tell you how much

49:54

the reduction would be. theoretically

49:56

you can get a very substantial reduction

49:59

in the demand for reserve balances which

50:02

is totally makes sense right because

50:04

when I ran through this little uh

50:06

vignette where you were the guy running

50:08

liquidity at your at City Bank it was

50:10

kind of obvious you didn't want to hold

50:12

more than required reserves because the

50:15

rest were getting compensated at a low

50:17

interest rate that and you could do

50:19

better by just uh converting the rest to

50:22

interestbearing securities. So, it

50:25

sounds like you're a little reticent to

50:27

give numbers, but theoretically, like

50:30

honestly, I would do it if I if I could,

50:32

but I can't. Uh, you know, do you leave

50:35

any breadcrumbs in the paper as to just

50:36

how large the potential reductions are

50:38

theoretically?

50:39

>> Not that I would be willing to put

50:40

numbers on uh and not other than to say

50:43

theoretically it could be quite

50:45

significant.

50:46

>> I see. Okay. Well, we'll say you started

50:48

with the quoteunquote least powerful

50:50

tool and you said that was maybe, you

50:52

know, a hundred billion or a few hundred

50:54

billion and so we're now at the last

50:56

tool. So, just to give people a sense of

50:58

that.

50:59

>> Yeah. Well, I mean, if I ordered them

51:00

correctly, this one is a lot more than

51:02

one or two hundred billion dollars.

51:04

>> Yes.

51:05

>> Yeah. But I'm not going to uh definitely

51:07

not going public with an estimate of the

51:09

quantity. There are there are other uh

51:13

subsequent research uh papers like for

51:16

example former governor Stephen Myron of

51:19

the Federal Reserve put out a paper and

51:22

I forget how much the number was but he

51:24

put a big number on this.

51:25

>> Yeah.

51:26

>> I mean a really big number.

51:28

>> Okay. So, professor because you are uh

51:32

very proficient in this topic as well as

51:35

these are you know your ideas that you

51:36

wrote about when you make these ideas

51:39

and you propose these to me and probably

51:41

to our audience they sound so simple and

51:44

uh like a very good idea and why would

51:45

you not do this you know Mr. Mr. Kevin

51:47

Worsh, why would why would you you not

51:49

do this? What could be the potential

51:51

drawbacks or push back that the Federal

51:54

Reserve, whether it's the new chair

51:55

Kevin Worsh or the old guard who kind of

51:58

put in these rules, particularly like

52:00

the Fed staff who have, you know, unlike

52:02

governors have been there for, you know,

52:03

perhaps over, you know, two decades or

52:06

something like that in terms of why

52:08

these four tools are so controversial to

52:13

uh to propose and difficult to to

52:15

implement.

52:16

>> Okay. So I'm very careful in my paper

52:18

and and anytime I speak about this

52:20

publicly to say that I am not saying the

52:24

Fed should reduce the size of its

52:27

balance sheet. My objective is to

52:30

provide options. If the Fed were to

52:34

decide to reduce its balance sheet, how

52:36

could it do it? And there are really

52:37

good arguments by others saying, well,

52:40

the Fed doesn't really need to reduce

52:42

its balance sheet. Nothing goes wrong

52:45

when there's extra reserves in the

52:48

system. Banks have lots of liquidity.

52:50

Markets are not going to go haywire.

52:52

Uh there's, you know, payment uh timing

52:55

is not an issue. We wouldn't need a

52:57

liquidity savings mechanism. Everything

52:59

would go great in the minds of of many

53:04

commenters with a large balance sheet

53:06

for the Fed. So many would say, well,

53:09

why go to all the trouble building a

53:11

liquidity savings mechanism? That's a

53:12

lot of retooling. Uh tearing down uh

53:15

reserve balances. That's, you know, you

53:17

have to figure out where you're going to

53:18

tear them down. The banks are not going

53:20

to be that happy because they're going

53:21

to make less profits on their on their

53:24

interest at uh on reserves held at the

53:26

Fed. They might lobby against this. Why,

53:29

you know, why go through all this

53:30

stress? Why not just keep the balance

53:32

sheet large? Many would say. And then

53:34

those that speak in the other direction

53:36

point to the fact that a very large Fed

53:39

balance sheet is kind of a political

53:40

lightning rod. those that are, you know,

53:43

not deep into the weeds are going to

53:45

say, "Well, why does the Fed need to own

53:46

so much assets?" This is where we

53:48

started the conversation. It doesn't

53:50

really need to own $6.5 trillion of

53:53

treasuries and mortgage back securities.

53:56

It's just it's going too far. It's it's

53:59

it's expanded beyond its needs. It's and

54:04

maybe they would even uh suspect that

54:07

the Fed is meddling in fiscal policy,

54:10

>> right? I certainly he heard those

54:12

arguments and yes, so thank you for

54:14

saying that and to be clear, you are not

54:16

making a prescriptive claim that the Fed

54:18

should do this, the Fed should do that.

54:20

It's like if you know there's a new

54:22

president or or uh you know person in in

54:25

the country who says we need to build

54:27

every single semiconductor that America

54:28

consumes, we need to build here. You're

54:30

not saying that's a good idea or that's

54:32

a bad idea. You're saying if the

54:33

president has this agenda, here is how

54:36

much electricity we need. Here's how

54:37

many fabs we need. That that's what

54:38

you're doing. Okay. Now, with that being

54:40

said, the new Fed chair, Kevin Worsh,

54:44

has been extremely open about his desire

54:47

to reduce the Fed's balance sheet and

54:50

the fact that he thinks that the Fed

54:52

should not be in the business of having

54:54

a large balance sheet or I believe to to

54:56

paraphrase that the you know, Fed is

54:58

involved in credit provision and that

55:00

it's it's its um what's it called? Its

55:01

profile in the market should be a a lot

55:04

more reduced. So, you're not on the

55:06

record saying the Fed should do this,

55:07

the Fed should do that. the guy running

55:08

the Fed is on the record, at least

55:09

before he ran the Fed, that this is what

55:11

the Fed should do. So, these are your

55:14

four proposals. How receptive do you

55:18

think the leadership of the Fed as well

55:21

as the other uh board of governors, the

55:25

presidents, the research staff are to

55:27

these ideas generally?

55:29

>> Well, I've had conversations, but I'm

55:31

not going to report on those. But there

55:33

are public remarks by several uh members

55:37

of the FOMC and including governors and

55:40

presidents on this topic and they are

55:42

it's a mixed bag of views about how

55:45

important it is to reduce the balance

55:46

sheet. Some uh you know much more

55:49

interested than others. Uh Governor

55:51

Chris Waller for example has said uh

55:53

having abundant reserves is not a

55:55

problem. If we want to reduce the

55:56

balance sheet, I'm I'm paraphrasing,

55:58

then we shouldn't do it without reducing

56:00

the demand for reserves first because

56:03

why would you cause the costs of

56:05

insufficient reserve balances? He

56:07

likened it to banks. He used the

56:10

interesting metaphor. He said, why would

56:12

you have banks scrging and under the

56:15

couch cushions for money? He said,

56:17

quote, that would be massively stupid

56:19

and inefficient.

56:22

And that and what he meant by that is if

56:24

we're going to reduce the balance sheet,

56:26

we shouldn't make it hard for banks to

56:28

get the reserve balances they need. We

56:30

would have to reduce the amount of

56:31

reserve balances that they need. Uh

56:35

other uh you know uh uh President Lori

56:38

Logan, the president of the Dallas Fed,

56:39

wrote a paper with Sam Schuler Woler

56:42

describing technically what would need

56:44

to be done. It's sort of analogous to

56:46

the paper that I wrote. Uh former

56:49

Governor Steven Myron also wrote a paper

56:52

uh going through the options. Uh

56:54

Governor Michael Bar has ventured

56:57

opinions on this. He's less concerned

57:00

about the size of the Fed's balance

57:02

sheet, but he's also indicated, you

57:04

know, that if we're if if this is going

57:06

to be done, there are better and worse

57:07

ways to do it.

57:09

Um and so on. So there are mixed there's

57:11

a mis mixed range of views at the Fed.

57:15

My my view is it's not simply providing

57:18

options uh in my paper. It's not purely

57:20

a clinical kind of plumbing paper. I

57:23

also have the view that the Fed should

57:26

develop these options in case they're

57:28

needed. So even if the Fed doesn't

57:30

currently have any plans to reduce its

57:32

balance sheet, it should have these

57:34

options available. What if, for example,

57:35

what if Congress said to the Fed, "We

57:38

think you're too big for your britches.

57:40

We want you to cut down your balance

57:42

sheet by 20%."

57:43

The Fed currently has no options that

57:45

would allow that. And if Congress said,

57:48

"Oh, you can't do that, why not? Uh, and

57:53

uh, if you're going to hold all these

57:54

assets, maybe you should hold some

57:56

assets that we want you to hold instead

57:58

of the assets that you want to hold."

58:00

And that would be eroding the Fed's

58:03

independence, right? if if Congress were

58:05

to use the size of the Fed's balance

58:08

sheet uh as kind of leverage or as a way

58:13

to conduct fiscal policy. So the Fed

58:16

needs to have these options available in

58:18

case, you know, it becomes a break the

58:21

glass situation and they have to reduce

58:23

their their balance sheet. And now it

58:26

may, you know, with chair chair wars has

58:28

impanled a task force on the Fed's

58:30

balance sheet, it may become a matter of

58:31

policy eventually that it will reduce

58:33

its balance sheet. Uh so having these

58:36

options, developing these options, doing

58:38

research on them, measuring how

58:40

effective they are, answering some of

58:41

the questions that you asked me uh

58:44

today, how many hundred billion for

58:46

this, how many hundred billion for that?

58:48

Th you know that research should be done

58:51

at the Fed in my view.

58:52

>> Are you going to be involved in the task

58:54

force on the Fed's balance sheet? Would

58:55

you like to be if you're not?

58:57

>> Don't know.

58:57

>> Maybe they should. I mean, you're you're

58:59

kind of the guy. I mean, you've got some

59:01

uh people should people should read this

59:03

paper. Uh are there any ideas that

59:07

people have, other academics,

59:10

economists, Fed people have that you

59:12

didn't include because you think there

59:15

are drawbacks that maybe they weren't

59:16

considering? There's a there's a longer

59:18

list in uh Governor Myron's paper with

59:21

collaborators at the Federal Reserve

59:22

Board when he was when he uh when he was

59:25

there and uh I don't remember what they

59:28

all are but you know I think I think the

59:30

ones that I've listed with you today are

59:32

the are probably the most important ones

59:34

to consider.

59:35

>> Okay. So you think if you know hopefully

59:38

we do another interview and um the

59:40

reserves of the Fed in let's say 2028

59:43

instead of 3.07 07 trillion are 1.07

59:47

trillion. You think it is likely that

59:49

they will have used

59:52

several or all of the tools that we

59:54

talked about today?

59:55

>> I'm I'm going to venture that by 2028

59:57

you that that's not going to happen.

60:00

It's that's way more reduction than

60:02

could happen by that time. As I

60:04

mentioned, the easier ones are not that

60:07

potent. If you really want to get a lot

60:10

of reserve balances, it goes down to

60:12

these things like liquidity savings

60:13

mechanisms and tiering the remuneration

60:16

of reserves and those take years to

60:18

develop. So you know maybe add a few

60:20

years to what what you said and and then

60:23

yes maybe

60:24

>> and the liquidity saving mechanisms and

60:26

the tiering reserves to summarize other

60:28

central banks have done this

60:31

successfully.

60:32

>> Correct.

60:33

>> Okay.

60:34

>> But other central banks are not the Fed.

60:36

Fed is special. It's big and it's

60:38

different. Uh so I'm not suggesting that

60:41

the Fed is going to go this way, but if

60:43

it really wants to get its balance sheet

60:45

down a lot, it may well need to consider

60:47

uh going this way.

60:49

>> And what if it's not 2028, but it's 2033

60:53

and there's a substantial reduction in

60:56

reserves. What does that world look

60:59

like? The Fed has achieved its goal.

61:00

It's reduced reserves and therefore

61:01

reduced its assets, reducing

61:04

liabilities, reducing assets. But what

61:06

might the consequences be? I think you

61:07

said one of heightened interest rate

61:09

volatility. Are there other

61:10

consequences?

61:11

>> Yeah, I mean the Fed will have a smaller

61:13

balance sheet. It'll be less at risk of

61:16

coming into

61:19

uh under a bright light uh from

61:22

commentators who are concerned about the

61:24

size of the Fed's balance sheet because

61:25

they don't understand what's causing it

61:28

to be so big. If it's smaller, then that

61:30

just basically alleviates those

61:31

concerns. the Fed will be in a better

61:33

position, at least politically, if

61:35

another crisis comes along and it needs

61:37

to expand its balance sheet

61:39

dramatically, let's say, to to manage a

61:42

dysfunction in the Treasury market, if

61:43

it needs to buy a few trillion of of

61:46

treasuries, that would be much more

61:48

politically palatable when the balance

61:50

sheet starts at a small level than when

61:52

the balance sheet starts uh you know at

61:55

6.5 trillion and growing with the

61:57

current uh framework. So yeah, I mean

62:00

the Fed will be in a much better

62:01

position at that time with respect to

62:04

those issues.

62:06

>> Daryl, thank you so much for joining. We

62:08

will link to your papers which are

62:10

available at daryl duffy.com. The one we

62:13

discussed mainly came out in the spring

62:15

of 2026, the payment system puts a floor

62:17

in the Fed's balance sheet and your

62:18

piece that came out just came out in

62:20

June of this year, an efficient

62:22

liquidity savings mechanism. Thank you.

62:25

And uh sounds like the Feds wants to be

62:27

involved. Maybe maybe they should uh

62:29

they should contact you. Thank you.

62:30

Thank you again, Professor

62:32

>> Jack. It's always a pleasure. Thank you

62:34

for having me.

62:34

>> I hope you enjoyed today's interview.

62:36

Remember to check out the Fundrise

62:38

Income Fund. Click the link in the

62:39

description to learn more about its

62:41

assets and strategy. Until next time.

62:46

Thank you. Just close the door.

Interactive Summary

This video features an in-depth conversation with Professor Daryl Duffy regarding the challenges the Federal Reserve faces in reducing its balance sheet. Duffy discusses the structural importance of reserve balances and proposes four key mechanisms the Fed could adopt to reduce demand for these reserves—and consequently its overall balance sheet—without triggering market disruptions like those seen in 2019. The discussion covers temporary open market operations, addressing regulatory stigma, implementing liquidity savings mechanisms, and tiering the remuneration of reserves.

Suggested questions

3 ready-made prompts