0:04

Hi , I'm Edward Finley , a Sum-Time

0:06

Professor at the University of Virginia and a Veteran

0:08

Wall Street Investor , and you're listening

0:11

to Not Another Investment podcast

0:13

. Here we explore topics and markets

0:15

and investing that every educated

0:17

person should understand to be a good citizen

0:20

. Welcome

0:23

to the podcast . I'm Edward Finley . Well

0:26

, we take up this episode , our second

0:28

installment on real assets

0:30

, and , as you might have guessed

0:32

, today we're going to talk about real estate

0:35

. So it seems sort of fitting

0:37

that this is in the real assets

0:39

category . Real estate , I think you

0:41

can break down into really

0:44

three primary categories

0:46

. You'll recall that in my lexicon from

0:48

earlier , I describe

0:51

the real estate part

0:53

of real assets as really being those

0:56

that produce the

0:58

commodities that are thought of as the inputs

1:00

to the economy . So here we're talking primarily

1:02

about commercial real estate , we're

1:04

talking about farmland and

1:06

we're talking about timberland . I'm

1:09

also going to spend a little bit of time

1:11

talking about residential real estate , partly

1:14

because it's increasingly becoming an

1:16

interesting asset class among investors , but

1:19

also partly because it's such a

1:21

large part of the average

1:23

American's total net worth that

1:26

I felt like it's useful to consider

1:28

how to think about it as an investment

1:30

, why we have it , is it an investment

1:32

at all ? And how should we

1:34

think about the role it plays

1:37

in our portfolios ? So

1:39

let's dive right in . Let's talk

1:41

about commercial real estate . So

1:50

commercial real estate , you know here think there's

1:52

a wide , wide variety of things that

1:54

are commercial real estate . There are apartments

1:57

, there are hotels , there are office buildings

2:00

, there are places

2:02

where , increasingly , internet

2:04

companies have logistical

2:06

support as they ship goods around

2:09

the country and get them to end consumers . There

2:11

are doctors' offices , there are

2:13

communities where

2:15

people with disabilities can live

2:17

, there are old age communities . Commercial

2:20

real estate is a really

2:22

big category , very

2:24

diverse , but

2:26

despite its diversity , the

2:29

asset itself has , as

2:31

we've discussed before , extraordinary

2:34

idiosyncratic risk because

2:36

of the strong dependence on location

2:38

for returns , and

2:41

that dependence on location is non-diversifiable

2:44

. So , for example , if

2:47

you think about a logistics center that's

2:49

positioned at the intersection

2:51

of two main freeways , that's

2:54

going to have an extraordinarily

2:56

high value relative to

2:58

exactly the same structure , with

3:00

all the same technology , in the same

3:02

region , but just not easily

3:04

accessible by those two highways

3:07

, and just that alone

3:09

is going to be sufficient to make one property

3:11

worth a lot more than the other and indeed to

3:13

grow at a different rate than the other . So

3:16

highly idiosyncratic risk and

3:18

we're going to be looking at

3:20

data when we talk about this

3:23

idiosyncratic risk . We're going

3:25

to be looking at data from something called NECRIF

3:27

, which is the National

3:29

Council on Real Estate Investments

3:32

, and NECRIF is an index

3:35

of directly owned commercial

3:37

real estate , so it's not an index of funds

3:39

, it's survey-based

3:42

data , so that means there are some

3:44

glitchy things about it , it isn't perfect

3:46

and it's also not

3:48

investable , because , of course

3:51

, it's all the reported

3:53

real estate all across the country and

3:55

it would simply be impossible for any investor

3:57

at all to own a diversified

4:00

mix of those locations . And

4:03

then the third reason is because the NECRIF

4:05

index reports its values

4:07

on a non-levered

4:09

basis , but we know

4:11

, as a matter of course

4:13

, that investors in

4:15

commercial real estate always

4:17

use leverage . It's very , very

4:20

unusual to find a commercial

4:22

real estate asset that is not levered

4:24

, and so , as a consequence

4:27

, what we would expect

4:29

to see in this data is data are

4:31

returns that aren't

4:33

really going to reflect the returns of an actual

4:35

investor , because they're not going to be levered and

4:39

because it's survey data that's provided

4:41

quarterly , with lag

4:44

of about a quarter , and there's

4:46

no market comparables

4:48

. These are really just estimates of

4:51

the owners of what their asset is worth . Is

4:54

that ? It sounds a whole lot like the problem we

4:56

confronted in private equity , which is to

4:58

say that the data

5:01

itself is so lagged

5:03

and so unable

5:06

to be compared to market variables

5:08

. It's not transactions . That

5:10

means that we would expect the volatility

5:13

in this data to be really artificially

5:15

depressed by virtue of smoothing

5:18

, though , since

5:20

we know that leverage is going to be used

5:23

in these assets , I suppose we might imagine

5:26

in some respect that the volatility

5:28

will then be adjusted by the

5:30

leverage . That is , if

5:32

we computed it with leverage , we might actually get

5:34

a volatility . That's a little bit more realistic

5:37

. So what does the data

5:39

tell us about this asset ? Is

5:41

it a real asset and to remind you

5:43

, from our last episode , we said that

5:45

a real asset we would expect

5:47

to have positive correlation

5:49

to inflation . That is , you want

5:52

to own it when inflation

5:54

expectations are dominating

5:56

markets , because during those

5:58

periods , equities and bonds

6:00

will both move in the same direction , ie down

6:02

, and so you want an asset that's

6:04

going to be somehow insulated from that

6:06

, and we would expect a positive

6:09

correlation to inflation to tell

6:11

us that that's what the asset's doing . And second

6:13

, we would expect this asset to be roughly

6:16

independent of stock and bond returns

6:18

, for a very similar reason

6:20

. That is just generally , we want to know that

6:22

this is an asset that is diversifying

6:25

the holdings , the risks that we currently

6:27

have in our portfolio and , last

6:29

but not least , is we would want to

6:31

make sure that the asset can

6:33

be expected to earn positive real

6:35

returns . So we'll want to think about each

6:38

of these in that way and we'll think about commercial

6:40

real estate in that way . So , with all of my

6:42

caveats . So

6:44

, looking at the data from 1997

6:47

to 2023 , a commercial

6:49

real estate in the US had

6:52

very modest correlation

6:54

to equity returns about 7% despite

6:58

the fact that the asset is

7:00

really exposure to the productive economy

7:02

. This is the

7:05

thing in which the efforts

7:07

of the productive economy take place . You

7:09

would expect there to be maybe more correlation

7:12

there , but despite that , it's not

7:14

the case . It's very modest correlation and

7:16

likewise very modest correlation

7:19

to interest rate risk about

7:21

7.6% . So

7:24

we've got a good independence

7:27

from equity and bond returns , and our

7:29

correlation to inflation is 33%

7:32

. Well , it's not huge , but it's really

7:34

rather material , and so

7:36

so far so good . This asset is really shaping

7:38

up to be a kind of classic real

7:42

asset Returns for

7:44

over that period were on par with

7:46

equity returns . Equity returns

7:48

were about 9.2% and

7:51

commercial real estate during that period earned

7:53

8.5% . But

7:56

due to smoothing the

7:58

volatility , here is only 7.7%

8:01

, delivering a whopping information

8:04

ratio of 1.11% . But , as we

8:06

said , we can pretty much just

8:08

discount that entirely and

8:11

not pay too much attention to it . We

8:13

might also remind ourselves that

8:16

, because we would expect this

8:18

asset to be levered , the

8:20

8.5% returns will actually

8:22

be higher . Even once we net

8:24

out the interest expense of

8:27

the debt , we

8:29

see an extraordinarily

8:31

high degree of illiquidity

8:34

, with an auto correlation of 85%

8:36

. This should surprise none

8:39

of us . Commercial real estate

8:41

, by its nature , is going to be highly

8:43

illiquid , and that is largely

8:45

due to their size of any particular

8:47

asset , but also it's going to be

8:49

a function of the fact that these assets are so

8:51

idiosyncratic that you really

8:53

might find yourself without

8:56

a buyer . And that's not because the

8:58

asset's not great , it's just there isn't

9:00

a buyer for that very unique , very specific

9:02

asset , and

9:05

so we see a high degree of illiquidity

9:07

, which causes us to then question

9:10

well , do we really think we're earning any

9:12

compensation for illiquidity ? This has

9:14

come up a few times in the podcast

9:17

, most notably when we talked about private

9:19

equity , and

9:21

at the time I told you that there's

9:23

very little evidence that we

9:25

actually earn compensation for illiquidity

9:28

. It is part of finance

9:30

theory that we should . It is a risk for

9:33

which we should be compensated , but

9:35

it's not clear that we do , and this is another

9:37

example of that . If

9:39

the unlevered returns are

9:42

in the same neighborhood as equity

9:44

returns , then it seems to

9:46

me that there's not much to

9:48

say that we earned any compensation here

9:50

for the illiquidity of the asset . Like

9:54

hedge funds , where we just spent a lot of time , we

9:56

see that there's really only

9:58

non-normal risks in this asset

10:01

. The SKU is very large

10:03

negative SKU and the kurtosis

10:06

is a very large positive

10:08

kurtosis , negative SKU of negative

10:11

2 , and kurtosis of 5.7

10:14

, both of which suggests that volatility

10:16

understates the risk . So now

10:19

volatility is sort of challenged

10:21

on two fronts . It understates the risk because

10:23

of smoothing . It also understates

10:26

the risk because owning

10:29

the asset in the real world is going to be levered

10:32

, which will make it more volatile , and

10:34

it understates volatility , understates

10:36

the risk because of negative SKU

10:38

and very large tail risk

10:40

. We can look at

10:43

state-dependent returns , though , and

10:45

see that , unlike hedge

10:47

funds , we have very

10:49

little evidence of any nonlinear

10:52

risk of , say

10:54

, concavity or convexity , which

10:56

is what we talked about a lot in

10:58

the hedge funds section . Here

11:01

, when we look at the state-dependent returns

11:03

of commercial real estate against

11:05

US equities , we see

11:08

that there's pretty much no

11:10

relationship , no concavity , no

11:12

convexity . It earned 1.5%

11:15

in the worst months of

11:18

equity returns and it earned 2.1%

11:21

in the best months of equity returns

11:23

, and it was always about 2%

11:27

to 3% per month in

11:30

each of the other sections . And

11:33

then , likewise , we see the same with interest

11:36

rate risk . When we look at the worst

11:38

months of returns of the 10-year

11:40

treasury , commercial real estate earned

11:42

1.5% a month . And when

11:44

we look at the best returns for

11:47

the 10-year treasury , it earned 2.3%

11:51

. And again , throughout the quintiles

11:53

, it's all pretty much the same . So no real evidence

11:55

for convexity . That's not a risk that

11:58

we think we're getting paid for when we own

12:00

commercial real estate . What's

12:04

the upshot ? Well , I think the upshot is

12:06

that it seems like commercial

12:09

real estate is a relatively

12:11

good real asset . It

12:13

seems like it's the kind of thing that would do

12:15

the job that we want it to do . It will have some

12:18

meaningful positive correlation to inflation . It'll

12:20

be roughly independent from stock and bond returns

12:22

. It'll earn , we think , positive

12:25

real returns over the long run . The

12:28

problem with it is that it's

12:30

just very difficult to

12:33

get a sufficiently diversified portfolio

12:35

of real estate . Now , I should mention

12:37

that there are things out there called REITs

12:39

real estate investment trusts and

12:42

these are securities that

12:45

really are issued by an entity that

12:47

owns a lot of commercial real estate , and

12:49

those securities trade usually

12:52

over the counter , but sometimes on large public

12:54

exchanges . The thing about

12:56

REITs , though , is that when you look

12:58

at the data on REITs , it stops

13:00

behaving like real

13:02

estate and it starts behaving

13:05

like equities . So the

13:07

inflation correlation goes away

13:09

, the low correlation to

13:11

US equities goes away , and

13:15

the auto correlation goes away . So what you find

13:17

is that suddenly , it trades like

13:19

US equities , but it earns a little

13:22

less than US equities , and so

13:24

, when we think about using REITs

13:26

to own commercial real estate , we don't get

13:28

the benefits of the commercial real estate . We're

13:30

really owning a kind of equity risk

13:32

, and that's why investors

13:35

who invest in this asset class typically do

13:37

it through private

13:39

funds , so funds that

13:41

would look a lot like

13:44

a private equity fund , in that it's a limited

13:46

partnership . There's a general partner

13:48

who's the expert investor , there

13:51

are limited partners who are the investors

13:53

, and not all of the investment

13:56

is called in the first moment . It's called

13:58

periodically , and over the life

14:00

of the fund , the manager , as they

14:02

sell properties , then distributes

14:05

capital back to the limited partners . So

14:07

it has a lot of the same

14:10

look , touch and feel as

14:12

private equity funds , but with a slightly different purpose

14:15

. And the reason that investors choose

14:17

that avenue is

14:19

, frankly , because if there's so

14:21

much idiosyncratic risk in commercial

14:23

real estate and we otherwise like the other

14:25

risks and we want to own it and we want to own it with leverage

14:27

then there's probably some

14:30

reward for really skilled

14:32

managers who know how to pick properties

14:35

, and that means that those returns will

14:37

be even higher than our index

14:39

returns , which don't impute any

14:41

skill . So

14:44

, altogether an interesting real asset

14:46

, somewhat tricky to access if you're just

14:48

a retail investor , but

14:50

if you're not a retail investor , then

14:53

it is not only accessible but rather

14:55

interesting . Okay

15:03

, let's turn our attention then to farmland

15:05

. So what are we talking about when we talk about

15:07

farmland ? It might seem obvious , but , just

15:09

for the avoidance of all doubt , these are

15:11

assets on which farmers

15:13

grow the soft

15:15

agricultural inputs to our economy

15:17

. They grow wheat , they grow corn , they grow

15:20

soy and the like sugar

15:22

. As

15:25

you might imagine , like commercial

15:27

real estate , farmland is going

15:29

to be subject to very high

15:32

idiosyncratic risks because

15:34

the returns on farmland

15:36

, the value of farmland , is strongly

15:38

dependent on its location . It's

15:41

strongly dependent on its location because

15:44

of weather . If you've

15:46

got farms in Indiana

15:48

, they're going to have very different weather

15:51

than farms in Iowa and

15:54

therefore that makes each of those assets

15:56

highly idiosyncratic , with risks that

15:59

are hard to diversify . We

16:01

also know that pests are going to affect

16:03

crops in different places

16:05

, so how much

16:07

crop is affected

16:09

by some particular pest

16:11

in Iowa may be very different

16:14

if that pest is not in Indiana . But

16:17

the crop yields are themselves

16:19

going to be different because a

16:22

farm in one place with

16:24

a certain kind of soil and a certain

16:26

kind of weather condition might be much

16:29

more productive in making

16:31

that commodity let's say it's corn than

16:34

a farm in another place with

16:36

a different type of soil and different

16:38

type of weather patterns , et cetera . So

16:41

, all to say , highly

16:43

idiosyncratic . Like

16:45

with commercial real estate , we're using a neck-reef

16:47

index for farmland , which is unrealistic

16:50

because with so much idiosyncratic risk

16:52

, an index , while

16:54

interesting to think about the asset

16:56

class , is really hard

16:59

to think about as an investor because

17:01

you can't own all the farms

17:03

all across America , which is what the

17:05

index reports to us , like

17:09

we saw with commercial real estate

17:12

, we know that the volatility

17:14

here is going to be artificially

17:16

depressed because of the nature of the index

17:18

. The index again has quarterly

17:21

data reported

17:23

on a lag no market

17:25

comparables . These are just the best

17:27

estimates of farmers about

17:30

the value of their farm . So we would

17:32

expect volatility to be depressed

17:34

. But

17:37

, unlike commercial real estate , it's

17:39

unlikely that there's going to be much leverage

17:42

. When you own farms

17:44

in a portfolio . The index

17:46

is like commercial real estate unlevered

17:49

, but real investors tend

17:52

not to use leverage or , if they do

17:54

, not , nearly as much leverage as they

17:56

would in commercial real estate

17:58

. And

18:00

so let's take a look at what the summary

18:02

statistics tell us . First

18:06

, we see that there

18:08

is a moderate

18:11

let's call it correlation to equity risk

18:13

about 11% . Moderate

18:16

correlation to interest rate risk

18:18

about 18% , and so

18:20

that's not strong . And so we'd

18:22

say , independent of those returns

18:24

to be a good real asset . But I

18:27

don't know , it's pretty low , I'm

18:29

not going to hang it out to dry just for

18:31

that reason . But

18:34

we see that there's a strong negative

18:37

correlation to inflation . It's negative

18:39

18% , and that is definitionally

18:42

going to make farmland not a

18:44

real asset , because

18:46

if it is anything at all , it has

18:49

to be at least independent of inflation

18:51

, if not positively

18:53

correlated . In here we see negative

18:55

correlation to inflation

18:57

, which suggests why

18:59

would that be true ? It suggests

19:02

something that should be pretty obvious to us

19:04

, which is that farmers

19:06

are likely to suffer just

19:08

as other firms do when

19:11

there's inflation in the economy , because they

19:13

can't pass on all of

19:15

their higher costs . That affects

19:17

their profit , and if their profit is reduced

19:19

the value of their farm is going to be less

19:22

. And so this negative

19:24

18% correlation to inflation makes

19:26

sense when we think about this asset

19:28

and it certainly

19:31

counts against it when

19:33

thinking about it as a real asset

19:35

. The returns

19:37

for this asset over the period 1997

19:41

to 2023

19:43

were sort of more robust than equity

19:46

returns . It was a 10.9%

19:48

return per year on average

19:51

, versus equities 9.2%

19:53

. So that's much more robust

19:56

and we see

19:58

that there's little evidence

20:01

of any illiquidity . The auto

20:03

correlation here is just 4% , and

20:05

that suggests to us that there's a ready market

20:08

for buying and selling

20:10

farms , that there's not a whole lot of illiquidity

20:13

here and that prices of farms adjust

20:15

rather quickly to

20:17

changes in circumstances . So

20:19

auto correlation is super low . Not

20:21

a lot of evidence of non-normality

20:23

in that case , but

20:25

we do see non-normality

20:28

in SKU and kurtosis . The

20:30

SKU is almost 4%

20:32

, positive 4 in the kurtosis is 18%

20:36

. So this is massive tail

20:38

risks and very positively

20:40

SKU , suggesting again that

20:43

volatility is going to

20:45

be a very poor estimate of

20:47

the risk of owning this asset

20:49

. We

20:53

can also take a look at the state

20:55

dependent returns here and

20:57

what we see is that there's very

20:59

little evidence of convexity . We

21:02

can look at the state dependent returns against

21:05

US equities . In the worst

21:07

months of US equity returns

21:09

, the strategy earned 2% . In

21:12

the best months of US equity returns

21:14

, the strategy earned 3.48%

21:17

and then in between it ranged up

21:19

. There's some

21:21

linear connection here . You can see

21:23

that it increases as

21:25

equities do better , so does the

21:27

strategy , but

21:29

linearly it doesn't seem to have much in

21:32

the way of nonlinear

21:35

relationship . The same could be

21:37

said for its relationship to 10-year

21:39

treasury returns . In the worst months it

21:42

was 2.55% , in

21:44

the best months it was 2.2% and

21:46

, with an exception for the middle quintile

21:48

, the returns per

21:51

month ranged from about 1.8

21:53

to 2.8 . So it pretty

21:55

much showed very little evidence of

21:57

a relationship to the 10-year , let

21:59

alone any nonlinear

22:02

relationship . So

22:04

, on balance , what do we think about farmland

22:06

as a real asset ? And I think the answer is it's not

22:08

. It's just not a real asset

22:10

, as we saw in the last

22:12

episode . Neither are the commodities

22:15

that the farms produce . There's

22:17

very little evidence to suggest that those

22:19

commodities bear a positive relationship

22:22

with inflation , and

22:24

so I think the notion

22:26

that farmland is a real

22:28

asset is kind of a mistaken notion

22:31

. Now , it might be the case

22:33

that you want to own farmland even

22:36

though it's not a real asset

22:38

. Why ? Well , because it has very

22:41

low correlation to equity and bond returns

22:43

. So that's great . And because

22:45

over this long period of time it

22:47

earned returns far in excess of

22:50

the returns that you could have earned for equity

22:52

risk in that period , and

22:54

so that might suggest it's a might be

22:56

a very interesting thing to own , but

22:59

it's not going to be because it's a real asset

23:01

. It's going to be simply on the basis

23:03

of the kind

23:05

of profile that farmland represents

23:08

. Okay

23:14

, next let's turn our attention to timberland

23:17

. So timberland

23:19

, just to set the record straight , here

23:21

are our commercial

23:23

farms in which

23:25

the farmer grows trees , that's

23:28

all it's . So it's like a farm of any other kind , except

23:31

that they grow trees . Typically

23:33

, the division in timberland is between

23:35

hardwoods and softwoods . Hardwoods

23:38

are the kinds of things that are used to make

23:40

furniture and they those

23:42

trees tend to grow in the

23:44

northeast of the United States there are some

23:46

exceptions , for

23:49

example in the Pacific Northwest . And

23:51

then softwoods are the sorts of things that

23:53

are used in building materials

23:56

are used to produce paper

23:58

, cardboard , which in a modern

24:01

e-commerce economy becomes

24:03

increasingly more important , and

24:06

those tend to grow in the southeast

24:08

of the United States and also in the Pacific

24:11

Northwest . So farmland

24:13

is again another

24:15

way in which we see a place

24:18

where an input to the productive

24:20

economy gets produced . So the input

24:22

is going to be lumber , and

24:24

lumber gets produced in

24:26

a timberland farm . These

24:30

farms are going to be highly

24:32

idiosyncratic in terms of

24:34

their risk . Not surprisingly

24:37

, this is going to be a theme for the real

24:39

estate assets Highly idiosyncratic

24:41

. Why ? Because there's a the returns

24:44

on the asset . There's going to be a very strong dependence

24:46

on the location of the asset

24:48

for returns . So weather

24:51

is not going to be the same in

24:53

the northeast of the United States , as the southeast

24:56

, as the Pacific Northwest , and so if

24:58

you own a timberland only

25:00

in the northeast , weather in the northeast

25:02

is going to have a significant impact

25:04

on your returns , but may not have an impact

25:06

on the returns in other regions , pests

25:09

likewise , and in the category

25:11

of pests I would also put things

25:13

like wildfire . These

25:16

are sort of risks that are not going to be

25:18

systematic , that you're not going to find anywhere . They're

25:20

going to be highly idiosyncratic and

25:23

it's it's yield is

25:26

not really diversifiable . It's

25:28

really very difficult to own

25:31

all of the different kinds

25:33

of timberland that are out there . Again

25:36

, we're going to be using a neck grief

25:38

index for timberland

25:40

and we're going to think in

25:43

terms of it as an entire asset

25:45

class , but I want to just remind us that the

25:47

timberland index is

25:49

pretending like we can own

25:51

timberland everywhere and

25:54

take away these idiosyncratic risks . We

25:56

can't . They're going to be very much present

25:58

, but we have to . We have to

26:00

at least think about whether they're

26:03

something that we see in the numbers or not

26:05

in there , and they're not

26:07

. Unlike farmland , though , timberland

26:10

doesn't need to harvest its crop every

26:12

year and , if you think about it , that creates

26:15

a rather interesting dynamic . So

26:17

in the case of farmland , we said

26:19

that the prices of farmland

26:21

are going to be highly dependent on

26:24

the economic cycle , because if

26:26

you produce corn , if you do produce

26:28

sugar , et cetera , these are inputs to

26:30

the productive economy , and the better

26:33

the economy is doing , the higher the price

26:35

you can get for your crop

26:37

, and the higher the price you can get for your crop

26:39

, the more valuable your farm is . All

26:41

of which makes sense . But in reverse

26:44

it's also pretty terrible , because

26:46

if the economy is in the

26:48

daldrums and you

26:51

have wheat and corn and so

26:53

on , then the

26:55

price that you're going to get for your commodity is going to

26:57

be rather low , and that means the value

26:59

of your farm is going to be rather low . But

27:02

when we talk about timberland , in

27:04

the case of timberland , we see

27:07

that the tree doesn't have to get

27:09

harvested every year , like corn

27:11

or wheat or soy or any

27:13

of the other soft agricultural commodities

27:15

. You can just let the tree grow

27:17

another year . Moreover , it

27:19

will grow every year . It's

27:22

not like other sorts of growth

27:24

in the economy that we think of . It's growth

27:27

in nature , and so not only

27:29

do you not have to harvest

27:31

each year , but you

27:33

know with a certainty that

27:35

the value of your asset will

27:37

grow over the course

27:40

of one year , because the trees will all

27:42

grow over the course of

27:44

a year . And so

27:46

there is a lot about Timberland

27:49

that makes it rather interesting , and

27:51

so we want to think very carefully about

27:53

what the profile then of that looks

27:55

like in our usual frame

27:57

. Well , first

28:00

, volatility , as I said

28:02

, is going to be something that's going to be highly

28:04

suspect , primarily because

28:06

of the index that we're using to

28:09

understand the data . The index

28:11

is also providing

28:13

its data quarterly , and

28:15

that quarterly data is on a lag

28:18

and it's not market comparables

28:20

. It's merely what the owner expects

28:24

the price or the value of the asset

28:26

to be , and so that's

28:28

going to smooth returns , and because

28:30

we're smoothing returns , it makes volatility lower

28:33

and that makes the information ratio

28:35

unreliable . But

28:38

let's talk about what it is . So

28:40

the average annual return here on

28:43

Timberland was about

28:45

7% 6.8%

28:48

over the period that we looked at , with volatility

28:51

of 5.8 . Again

28:53

, that volatility is highly depressed

28:55

. It gives us therefore an unrealistic

28:58

information ratio of 1.2 . We

29:02

see that the returns are

29:05

roughly independent

29:07

from equity returns . There's a 3%

29:10

correlation there , only modestly

29:12

correlated to bond returns

29:14

, and 8% correlation there . But

29:17

what we see that's the most damning

29:20

is that the correlation to inflation

29:23

is basically zero . It

29:25

comes in as negative 0.96%

29:27

, but it's basically independent of

29:30

inflation . And so , while

29:33

the returns were lower than equity

29:35

returns 6.8% as

29:37

opposed to equity returns of about 9.2%

29:40

a year during the period it's

29:43

still positive real returns but

29:45

on balance it doesn't sound terribly

29:47

much like a real asset . It doesn't

29:49

sound like a real asset because it has

29:51

no positive correlation with inflation

29:54

but it otherwise fits the bill . That

29:56

may make it a good asset to own , but

29:58

it might not be a good asset to own because

30:01

of its quote real and

30:03

quote nature . We

30:05

see that Timberland has the same

30:08

non-normality that we saw in

30:10

commercial real estate . So there is

30:12

a positive skew of about 1.2

30:14

. So it's significantly positively skewed

30:16

. And the kurtosis is also quite large

30:18

at 5 , which is of

30:21

course around the same as

30:23

commercial real estate , not

30:26

terribly massive , but

30:28

it's not nearly as high as

30:30

in farmland . We

30:33

also see that illiquidity is

30:35

present here with an auto correlation

30:38

of about 24% . That's

30:41

a lot less auto correlated

30:43

than commercial real estate but

30:45

a lot more correlated than , say , equities

30:47

or farmland . So there's some evidence

30:49

of illiquidity and it begs the

30:51

question whether investors are

30:53

really getting paid any compensation

30:56

for that illiquidity . Turning

30:59

our attention to the state dependent returns

31:01

, we see again that there's very

31:04

little evidence of non-linearity

31:06

. In the worst monthly

31:08

returns for US equities , the

31:10

strategy earned 1.5% . In

31:13

the best monthly returns it earned 1.5%

31:15

and in between it ranged from

31:17

1.2 to 2.2 . So

31:20

pretty much not

31:23

really showing any convexity in

31:25

a relationship to equity returns . Nor

31:28

is it showing any convexity in

31:30

its relationship to the tenure treasury . If anything

31:32

, it's showing a sort of a linear

31:35

relationship with rates . So

31:37

in the worst month for returns

31:39

on interest rates , the strategy earned

31:42

1.5% a month and

31:44

in the best months of interest rates the

31:46

strategy earned 2.2% a month

31:49

. And it's not perfectly linear

31:51

, but you can see the sort of upward

31:53

trend . And so , if anything , that

31:55

correlation

31:58

with tenure of 8% doesn't tell us the

32:00

whole story . Yes , it's not terribly

32:02

correlated , but it seems to reflect

32:04

a sensitivity to interest

32:06

rate risk . So

32:08

, again , it doesn't seem

32:10

like it's a particularly good real

32:12

asset . It's not positively correlated

32:15

to inflation , but one might

32:17

want to own it given its low correlation

32:19

to equities and interest rates . One

32:23

might want to own it because of

32:25

this feature of growth that

32:27

will occur regardless of

32:29

growth in the economy . And

32:32

one might want to own it because

32:34

it is some kind of asset

32:36

that seems probably able

32:38

to support most goals , its nominal

32:41

returns being around 7%

32:43

. It's not as high as equities but it's

32:45

not really low like some of the

32:47

hedge fund strategies we saw that

32:49

were good diversifiers but didn't earn

32:51

robust real returns . So

32:54

it could be an interesting asset

32:56

to own , but not a real

32:58

asset . All

33:07

right , let's wrap up the episode , then , by talking

33:09

about residential real estate . As I mentioned

33:11

, it is the

33:13

case that , increasingly , there

33:16

are investors interested in

33:18

investing in residential

33:20

real estate , but , by and large , residential

33:23

real estate is not an asset

33:25

class for investors

33:28

. However , it is probably

33:30

the biggest asset on the

33:32

average American's balance

33:34

sheet , and

33:39

so , as a result , I think

33:41

it's worth thinking about residential real estate because

33:43

, for the average investor and

33:45

, I think , probably the average listener to this

33:47

podcast , if you're

33:50

talking about your portfolio and what should

33:52

you buy and what should you own and that sort of

33:54

thing , don't ignore the

33:56

asset that you live in . And

33:59

so , if you own your home , that's going to be a

34:01

pretty big deal and you're going to want to make

34:03

sure you think about it in the context

34:05

of your portfolio

34:07

. So

34:09

let's take a look . We have some

34:11

data here that's courtesy

34:14

of the S&P . That's

34:17

the same company standard and pours that provides

34:20

the index for the 500

34:22

largest stocks and is one

34:25

of the providers we mentioned when we discussed

34:27

bonds , of credit ratings . On bonds

34:29

, s&p also runs

34:31

something called the Case Shiller

34:33

National Home Price Index

34:35

Case Shiller named

34:37

for the two academics who developed

34:40

the idea of maintaining

34:42

a database on the

34:44

prices of homes in

34:46

the US . The database is , like

34:48

all of the others that we looked at , unlevered

34:50

, and so we're going to want to account

34:52

for that in a minute . So , at least initially

34:55

, I'm going to talk about it unlevered

34:57

, but then we'll pivot and we'll think a little

34:59

bit more carefully about it as

35:02

a levered asset because , like commercial

35:04

real estate , the average American

35:06

owns their home with

35:08

a mortgage , so we'll want to take that into consideration

35:11

. So

35:13

let's do the same analysis as always how

35:15

is this characterized as a real asset

35:17

? Is it a real asset and do we want to own

35:19

it regardless ? Inflation

35:22

, with inflation is 22% . That

35:25

seems like a real asset to me . That

35:28

seems like something you would own , so

35:31

that in the times when

35:33

inflation expectations dominate asset prices

35:35

and your stocks and your bonds

35:38

are going down , this asset is

35:40

not necessarily going to be going down . So

35:43

positive correlation to inflation

35:45

. But in

35:47

the short term , we know that residential

35:49

real estate is much more

35:51

affected by economic

35:54

growth , by interest rates and

35:56

by demographics . I

35:59

mean , the value of homes

36:01

is going to be so much more

36:03

dependent on those things than on inflation

36:05

. I don't want to overstate the idea

36:08

that a positive correlation to inflation

36:10

is going to be the be

36:12

all and end all . But I think

36:14

it's worth noting here that in the aggregate

36:17

it exists , and in a minute

36:19

we're going to look at the data , sort of tweaked for

36:21

a specific reason , and we'll see that correlation

36:23

to inflation go away . So , yes

36:26

, you might think of it as a real asset . The

36:28

data suggests that it might be bought in

36:30

general over shorter time horizons

36:32

. Economic growth , interest rates

36:34

and demography are far more

36:36

important to the returns to this asset than

36:39

inflation is . There's

36:42

very modest correlation to US

36:44

equities about 7% , and

36:46

very modest inverse

36:49

correlation to the tenure negative

36:52

8% . So

36:54

roughly independent to stock and bond

36:56

returns . As we've discussed

36:59

, it's a much more normal distribution . So

37:02

there's no skew , it's

37:05

symmetric like a bell curve and

37:07

the kurtosis is 0.76 , which

37:09

, while in theory

37:12

, is not a normal distribution , only

37:14

a zero would be a normal distribution

37:16

. 0.76 , compared

37:18

to equity markets where

37:20

the kurtosis is normally around 2

37:22

, suggests that this is a lot more

37:25

normally distributed than equity

37:27

returns . So there's really very

37:29

little evidence that there's much non-normality

37:33

risk that you get paid for here

37:35

. But the auto correlation

37:37

is 90% and

37:40

that suggests that what you really own is

37:42

a highly , highly illiquid asset

37:44

and , as anybody can tell

37:46

you who has owned a home and

37:49

tries to sell it when they need

37:51

to sell it , that is usually

37:53

a really , really bad situation

37:56

Because the kind

37:58

of asset is so highly idiosyncratic

38:00

. There's no way to diversify that

38:02

risk and if you happen

38:05

to be needing to sell

38:07

in a market where there are limited numbers

38:09

of buyers , that is going to significantly

38:12

affect the value of your home

38:14

. So , yes

38:17

, a normal distribution , but beware

38:19

, highly , highly illiquid asset

38:22

. It earned

38:24

lower returns than equities

38:26

. Over the period . It earned a little shy

38:28

of 5% a year compared

38:32

to equities 9.2% , so about half

38:34

. Nevertheless , there

38:36

are positive real returns here

38:38

and so I think

38:40

, on balance , positive

38:42

correlation to inflation , independence

38:44

from stock and bond returns , positive

38:47

real returns . Okay , it's

38:49

probably a real asset with

38:51

some of the caveats that I mentioned

38:54

, but let's take

38:56

into consideration now , as we didn't

38:58

do before , because I want to make this example kind

39:01

of real for the average

39:03

homeowner is let's take a

39:05

look at what returns were like during

39:07

this period if the

39:09

asset were levered . So

39:11

if you assume that you

39:14

put 25% down and you

39:16

borrow 75% of the

39:18

asset , that's equivalent

39:20

to leverage of about 4x

39:22

. If I recompute

39:24

returns and volatility , taking into account

39:27

leverage , then what I get is a

39:29

return of a little shy of 21%

39:31

per year with volatility

39:33

of about 18%

39:36

per year , so that the information

39:38

ratio doesn't change terribly much . But , like

39:41

in the unlevered case , there's very little

39:43

to suggest here that volatility

39:46

is going to

39:49

be distorted by non-normal

39:51

or nonlinear risks , and

39:54

so in that case , with a 21%

39:56

return , that's clearly

39:58

a lot higher than equity risk . But

40:01

you really won't earn the 21%

40:04

because you're going to have to pay some interest

40:06

on your mortgage . So if we adjust that

40:08

, on average , during that period mortgage

40:10

rates were about 5.25%

40:12

, and so let's just

40:14

adjust then for

40:17

the leverage . When we adjust for

40:19

the leverage , then the average annual return

40:21

is a little more like 17% . In

40:24

addition , we want to take into consideration

40:27

the fact that the average American lives

40:29

inside their home , and so that

40:31

means that they get something

40:33

. There's some return for owning the home

40:36

that's not captured by changes

40:38

in price . Some people call

40:40

that rental yield . It would be

40:42

the amount that you could earn if you

40:44

rented the asset , but instead you

40:46

live in it , and that has some value to you , and

40:49

so it's an important part of the returns . We should take

40:51

it into consideration , but

40:54

in general , they're excluded from the data . One

40:57

economist estimates that the rental

41:00

yields from 1960 to

41:02

2008 in the United States were about

41:04

5% a year , and

41:06

so what we might imagine is that that pretty

41:08

much offsets whatever

41:11

the interest expense is on the mortgage

41:13

. If the rental yields about 5%

41:15

, the average interest rate was about

41:17

5% . Maybe they cancel each

41:20

other out and you're left with

41:22

at the same sort of maybe the same

41:24

place . In addition

41:26

, we want to make sure that we take into

41:28

consideration that , unlike a lot of other assets

41:30

, you have to maintain your house . You

41:32

got to make capital repairs , and

41:35

that is an investment in

41:37

the asset . Most

41:40

estimates of maintenance

41:42

and capital expenses are

41:44

about 5% a year as well . So

41:47

if I adjust then for

41:50

the rental yield

41:52

and I adjust for maintenance

41:54

and capital expenses , then I come up

41:56

with a net return

41:59

of around 11% , and

42:02

11% , as we mentioned a little

42:04

while ago , is about 2% a year

42:07

higher than equity returns . And

42:09

you get to live in your house , and

42:11

so I would say not

42:13

just a real asset , but

42:16

a really good asset , and

42:18

you can see why so many

42:20

Americans who can afford to do it

42:22

want to own their home . Because

42:26

if you have limited resources and

42:28

you have to live somewhere , owning your

42:30

home is necessarily going

42:33

to be a very , very prudent thing

42:35

to do in

42:37

your wealth accumulation and wealth accretion

42:39

. As a little

42:41

bit of a thought experiment , though , I wanted

42:44

to sort of tease out of the

42:46

data . The housing bubble

42:48

, which most people agree

42:50

, runs from January 2004

42:53

until August 2006

42:55

, when US home prices kind

42:57

of went nutty and they

43:00

came crashing down , and

43:02

that , you'll remember from earlier episodes

43:04

, is one of the major impetus

43:06

for the global financial crisis

43:09

. It's the uniform

43:11

reduction in the value of

43:13

houses everywhere in the US all

43:16

at once . So what

43:18

if we took out of the case-shiller

43:20

data the returns

43:22

during those months ? We take out the

43:25

bubble , the run-up , and

43:27

we take out the crash in the

43:29

returns , what

43:31

do we end up with ? It's not

43:33

terribly different . So the

43:36

annual returns then are

43:38

4.3% as opposed to 4.99

43:41

, unleveored . The volatility

43:43

remains around the same , sq

43:46

and kurtosis remain the same , auto-correlation

43:49

remains the same . Correlation

43:51

to equities , correlation to bonds remains

43:54

the same . What changes is

43:56

correlation to inflation ? And it drops

43:58

from 22% to 7%

44:01

. And there's the point that I made just

44:03

a little while ago when I said

44:05

that things like economic

44:08

growth , interest rates and

44:10

demographics are far more important to

44:12

this asset than inflation . When

44:15

you take out that period of the housing

44:17

bubble , you find then

44:19

suddenly the correlation to inflation

44:21

really drops radically

44:24

, and then we can do the

44:26

same exercise by levering

44:28

it 25% down . I

44:30

won't recite all of the information , but just

44:32

to say it's roughly the same , and

44:36

so the reason I did that is because

44:38

I think , when thinking about residential real

44:40

estate and I'm using these aggregate numbers and

44:43

it's just an index , and maybe

44:45

you know somebody who didn't have that experience

44:47

it's easy to sort of say well

44:50

, you had the housing bubble and that changed everything . It

44:52

skews the numbers and whatnot , but when

44:54

you remove the housing bubble , it doesn't

44:56

change the numbers by terribly much . Well

44:59

, that's it for real assets

45:01

. As we wrap up our discussion

45:03

of real estate , next time

45:05

we're going to come back and not talk about asset

45:07

classes but instead talk about what we do with them

45:09

. So

45:13

we're going to start talking about portfolio construction

45:15

and understand the notion of strategic asset allocation

45:18

. As

45:24

always , thanks for listening . Look

45:26

forward to talking to you next time .