Do REITs (and slicing in general) reduce risk?

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Topic Author
rca1824
Posts: 719
Joined: Tue Jun 03, 2014 9:33 am

Do REITs (and slicing in general) reduce risk?

Post by rca1824 »

One of the most common pieces of asset allocation advice I have read is to hold a slice of REIT because it's "uncorrelated with the broader market"
But REITs are already part of the broad market. They're weighted by their market cap.
So how can slicing the total market into REIT/non-REIT reduce risk?
If that were true, why not slice your portfolio into even more market segments and sectors?
My intuition tells me that a cap-weighted index must already yields the minimal risk for a given expected return. But I don't know why. It seems like a 'free lunch' to be able to increase returns and decrease risk by slicing the market and overweighted a particular slice and rebalancing regularly.
Monthly or yearly movements of stocks are often erratic and not indicative of changes in intrinsic value. Over time, however, stock prices and intrinsic value almost invariably converge. ~ WB
steve_14
Posts: 1507
Joined: Wed Jun 20, 2012 12:05 am

Re: Do REITs (and slicing in general) reduce risk?

Post by steve_14 »

Nobody knows if overweighting X and thus underweighting Y will reduce risk. That depends entirely on future, unpredictable price movements.

One thing is certain - a 5 or 10% allocation to a highly correlated asset sub-class isn't going to make much difference in your results. If you're not confident enough to "go big", don't bother. I posted the graph below of a few sector fund results in another thread, for those trying to decide which sector to overweight (which I don't recommend): VGSIX=REIT.

Image
KyleAAA
Posts: 9497
Joined: Wed Jul 01, 2009 5:35 pm
Contact:

Re: Do REITs (and slicing in general) reduce risk?

Post by KyleAAA »

rca1824 wrote:One of the most common pieces of asset allocation advice I have read is to hold a slice of REIT because it's "uncorrelated with the broader market"
But REITs are already part of the broad market. They're weighted by their market cap.
So how can slicing the total market into REIT/non-REIT reduce risk?
If that were true, why not slice your portfolio into even more market segments and sectors?
My intuition tells me that a cap-weighted index must already yields the minimal risk for a given expected return. But I don't know why. It seems like a 'free lunch' to be able to increase returns and decrease risk by slicing the market and overweighted a particular slice and rebalancing regularly.
Yes, they can. Owning any two (or more) uncorrelated assets with similar positive expected returns and rebalancing between them will tend to reduce risk, increase returns, or perhaps both. Many people DO slice their portfolios into even more market segments and sectors. Of course, whether or not REITs will end up being reasonably uncorrelated with the broader market over any given period of time is a crapshoot because correlations shift over time.

It is not the case that a cap-weighted index must already yield the minimum risk for a given expected return. In fact, diversification is usually referred to as the only free lunch in finance.
bloom2708
Posts: 9855
Joined: Wed Apr 02, 2014 2:08 pm

Re: Do REITs (and slicing in general) reduce risk?

Post by bloom2708 »

Good question. I have been considering adding REIT (VGSLX) and/or International Bonds (VTABX).

We currently have a 3 fund portfolio. Taxable (VTSAX, VTIAX, VWITX), Tax-Sheltered Rollover IRA and Roth IRA (VTSAX, VTIAX, VBTLX).

Our Roth would be where I could add one or both REIT/Inter Bonds. One day I'm convinced to add a "slice". The next I decide to stick with the core 3.

There are convincing arguments on both sides. I keep hearing Taylor L. whispering "keep it simple...keep it simple.."

Is inaction in fact action?
Call_Me_Op
Posts: 9872
Joined: Mon Sep 07, 2009 2:57 pm
Location: Milky Way

Re: Do REITs (and slicing in general) reduce risk?

Post by Call_Me_Op »

rca1824 wrote:One of the most common pieces of asset allocation advice I have read is to hold a slice of REIT because it's "uncorrelated with the broader market"
An oversimplification. Correlations change over time - including the correlation between REITs and the broader market.
rca1824 wrote: My intuition tells me that a cap-weighted index must already yields the minimal risk for a given expected return. But I don't know why.
I don't see why your intuition would tell you this. Just because each sector of the market has a specific weight, I don't see why that would lead to the conclusion that this is the weighting that yields the optimal risk-adjusted return.
rca1824 wrote: It seems like a 'free lunch' to be able to increase returns and decrease risk by slicing the market and overweighted [sic] a particular slice and rebalancing regularly.
It may or may not improve risk-adjusted return. There is no way to know ex ante.
Best regards, -Op | | "In the middle of difficulty lies opportunity." Einstein
bonaire27
Posts: 95
Joined: Sun Apr 27, 2014 12:46 pm

Re: Do REITs (and slicing in general) reduce risk?

Post by bonaire27 »

Steve,

I am a 3 funder and will be for the near future (at a minimum) so this is just a theoretical question; I have seen you and others post that if you just tilt by 5%, then it won't make enough of a difference to matter. However, I also see many posts that fret over an expense ratio of 1% vs .10%. With that in mind, can't a 5% tilt make a pretty big difference (positive or negative)?
MrMatt2532
Posts: 454
Joined: Sun Mar 15, 2009 11:58 am

Re: Do REITs (and slicing in general) reduce risk?

Post by MrMatt2532 »

Generally, I don't think its right to expect any random slice will reduce risk. For example, I wouldn't expect any benefit from taking the 500 companies in the SP500 and randomly separating them into two groups of 250, and then rebalancing between the two halves.

However, where I do expect benefit is from exposing yourself to unique risk factors, such as exposure to small companies, or value companies, which each behave differently than the market as a whole. Similarly I would expect benefit from exposure to international companies, which will behave differently than domestic companies.

Now, your question is specifically about REITS: it turns out that the 3 factor model does a poor job at explaining REIT returns, which means that REITS are like another unique risk factor: different than the market as a whole, different than small company behavior, different than value company behavior. That coupled with the fact that they have a positive real return similar to other stocks means they are probably a great diversifier to include. And to me, this would specifically mean that I would expect better risk-adjusted returns by including reits, small, value, intl exposure on the stock side of my portfolio. Then, I can use bonds to control risk.
anil686
Posts: 1316
Joined: Thu May 08, 2014 12:33 pm

Re: Do REITs (and slicing in general) reduce risk?

Post by anil686 »

Not sure if this answers all of your questions - but

a) Yes - REITS are in TSM in their market weight but...

most real estate (including commercial) is private. Makes sense since some small businesses and most homeowners own their own real estate. Thus, real estate represented by publicly traded entities is only about 3% of TSM - but is 13% of the economy at large. Much of this is private capital that is not investible but if you believe in weighting of sectors based on their percentage of the US economy - you would want to overweight REITs a little more than what is in TSM - Rick Ferri has much better explanation in All About Asset Allocation - second edition.

b) Real estate typically does not follow the broader market, but that does not mean it is inversely correlated to the broader market. Therefore, just because equities zig does not mean real estate zags. So in 2008-2009, REITS and equities had difficult times. However, REITs don't always follow the market and during some recessions have done okay. They have also underperformed the broader market at times during equity bull runs.

c) You cannot simulate the presence of real estate easily through REITs due to the large preponderance of privately held real estate (again both personal and commercial). You are just extrapolating the amount held in REITs and pushing that closer to the economic totals in the domestic economy (if that makes sense). Since real estate (fundamentally) is different than equities, it will not move like equities and is not as correlated - so the theory (not that I do it) is by holding REITs to their representative weight in the US economy, will help diversify your portfolio more.

d) Not a free lunch. You are diversifying... but REITs are not sure things and you are assuming risk. I have not looked back at the performance of REITs over a long period of time (and it may not be as helpful since there have not been low cost options until the turn of the century for REITs in general) but my recollection was that they carried about the same return as equities (maybe a little less). Thus, you are not really increasing your chance of returns but rather diversifying more and thus potentially limiting your downside somewhat - I guess since they are not perfectly correlated (i.e. instead of 100% domestic equity TSM - 90% TSM and 10 % REIT).

Hope that helps, I can't explain it as well as Rick Ferri, but that was my take away from that section.
steve_14
Posts: 1507
Joined: Wed Jun 20, 2012 12:05 am

Re: Do REITs (and slicing in general) reduce risk?

Post by steve_14 »

bonaire27 wrote:Steve,

I am a 3 funder and will be for the near future (at a minimum) so this is just a theoretical question; I have seen you and others post that if you just tilt by 5%, then it won't make enough of a difference to matter. However, I also see many posts that fret over an expense ratio of 1% vs .10%. With that in mind, can't a 5% tilt make a pretty big difference (positive or negative)?
An expense ratio is a tax you pay in good times and bad - a 1% fee drag makes a dramatic difference over, say, a 30 year period. Overweighting a group of stocks, on the other hand, has a 0% expected return (we can argue about whether or not that's on a risk adjusted basis or not, but all stocks have similar risk). Could increase or decrease your results by a few basis points.
steve_14
Posts: 1507
Joined: Wed Jun 20, 2012 12:05 am

Re: Do REITs (and slicing in general) reduce risk?

Post by steve_14 »

MrMatt2532 wrote:Generally, I don't think its right to expect any random slice will reduce risk. For example, I wouldn't expect any benefit from taking the 500 companies in the SP500 and randomly separating them into two groups of 250, and then rebalancing between the two halves.
Exactly. There are cases where this wins, and where it loses, but there's no free lunch in merely having two funds instead of one. Making 50 bets on 2 roulette wheels has no greater expected return than 100 bets on 1. Overweighting one of those 250 stock halves would be pointless for the same reason.
mptfan
Posts: 7201
Joined: Mon Mar 05, 2007 8:58 am

Re: Do REITs (and slicing in general) reduce risk?

Post by mptfan »

My data tells me that REITs have a .51 correlation with the total stock market index fund over the last 30 years, and over that time, REITs have had a CAGR of 10.58% while TSM was 11.2%, while exhibiting almost identical standard deviation (18.02 vs 17.91). So my conclusion is that REITs provide reasonable diversification over time that will likely reduce portfolio volatility. So the answer to your question is yes.
anil686
Posts: 1316
Joined: Thu May 08, 2014 12:33 pm

Re: Do REITs (and slicing in general) reduce risk?

Post by anil686 »

mptfan wrote:My data tells me that REITs have a .51 correlation with the total stock market index fund over the last 30 years, and over that time, REITs have had a CAGR of 10.58% while TSM was 11.2%, while exhibiting almost identical standard deviation (18.02 vs 17.91). So my conclusion is that REITs provide reasonable diversification over time that will likely reduce portfolio volatility. So the answer to your question is yes.

Wow - that is great information. I did not realize the correlation was that low. Thanks for the info!.
mptfan
Posts: 7201
Joined: Mon Mar 05, 2007 8:58 am

Re: Do REITs (and slicing in general) reduce risk?

Post by mptfan »

rca1824 wrote:So how can slicing the total market into REIT/non-REIT reduce risk?
If that were true, why not slice your portfolio into even more market segments and sectors?
Because although REITS are stocks, and therefore part of the total stock market by definition, many people consider them to be a distinct asset class because, unlike other stocks, their performance is closely correlated with the economic cycle of the real estate market which, as the data shows, is somewhat uncorrelated with the economic cycle of the stock market. In short, REITs are different.
steve_14
Posts: 1507
Joined: Wed Jun 20, 2012 12:05 am

Re: Do REITs (and slicing in general) reduce risk?

Post by steve_14 »

mptfan wrote:My data tells me that REITs have a .51 correlation with the total stock market index fund over the last 30 years, and over that time, REITs have had a CAGR of 10.58% while TSM was 11.2%, while exhibiting almost identical standard deviation (18.02 vs 17.91). So my conclusion is that REITs provide reasonable diversification over time that will likely reduce portfolio volatility. So the answer to your question is yes.
I'd certainly be wary of comparing the 7% yielding, low risk, low volatility REITS of the 90s with a REIT fund today.
Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: Do REITs (and slicing in general) reduce risk?

Post by Rodc »

steve_14 wrote:
mptfan wrote:My data tells me that REITs have a .51 correlation with the total stock market index fund over the last 30 years, and over that time, REITs have had a CAGR of 10.58% while TSM was 11.2%, while exhibiting almost identical standard deviation (18.02 vs 17.91). So my conclusion is that REITs provide reasonable diversification over time that will likely reduce portfolio volatility. So the answer to your question is yes.
I'd certainly be wary of comparing the 7% yielding, low risk, low volatility REITS of the 90s with a REIT fund today.
I hold some REITS, but would add a comment that I think is really just part of what you are saying, but REITS were anything but low volatility in the Great Recession! I think that had a lot to do with their inherent leverage and the freeze up of the credit market which may not repeat, but it sure was a wild ride for a while.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
rkhusky
Posts: 17654
Joined: Thu Aug 18, 2011 8:09 pm

Re: Do REITs (and slicing in general) reduce risk?

Post by rkhusky »

rca1824 wrote: My intuition tells me that a cap-weighted index must already yields the minimal risk for a given expected return.
The benefit of cap-weighting is that it is the most efficient way to put together a basket of stocks, hence it reduces costs.
User avatar
Doc
Posts: 10606
Joined: Sat Feb 24, 2007 12:10 pm
Location: Two left turns from Larry

Re: Do REITs (and slicing in general) reduce risk?

Post by Doc »

rkhusky wrote:
rca1824 wrote: My intuition tells me that a cap-weighted index must already yields the minimal risk for a given expected return.
The benefit of cap-weighting is that it is the most efficient way to put together a basket of stocks, hence it reduces costs.
As anil686 pointed out up thread most real estate is not held in publicly traded REITs but by businesses, partnerships etc. I'm willing to bet a Starbucks that most of the equity in the buildings on Wall Street itself is not being traded on the stock exchange. Therefore if you really wanted to cap weight RE you need to add it somehow to a TSM type portfolio.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
User avatar
JoMoney
Posts: 16260
Joined: Tue Jul 23, 2013 5:31 am

Re: Do REITs (and slicing in general) reduce risk?

Post by JoMoney »

Doc wrote:
rkhusky wrote:
rca1824 wrote: My intuition tells me that a cap-weighted index must already yields the minimal risk for a given expected return.
The benefit of cap-weighting is that it is the most efficient way to put together a basket of stocks, hence it reduces costs.
As anil686 pointed out up thread most real estate is not held in publicly traded REITs but by businesses, partnerships etc. I'm willing to bet a Starbucks that most of the equity in the buildings on Wall Street itself is not being traded on the stock exchange. Therefore if you really wanted to cap weight RE you need to add it somehow to a TSM type portfolio.
I think you need to look closer at balance sheets. There's quite a bit of real estate on the books of publicly traded companies. The big difference is that REITs are companies earning their money primarily in the form of rent. REITs get you exposed to businesses focused on rental income if that's what you're looking for.
"To achieve satisfactory investment results is easier than most people realize; to achieve superior results is harder than it looks." - Benjamin Graham
placeholder
Posts: 8375
Joined: Tue Aug 06, 2013 12:43 pm

Re: Do REITs (and slicing in general) reduce risk?

Post by placeholder »

Rodc wrote:I hold some REITS, but would add a comment that I think is really just part of what you are saying, but REITS were anything but low volatility in the Great Recession!
I don't think anyone said REITs were low volatility but rather that the non-correlation with broad stocks along with similar returns lowers portfolio volatility although in 2008 all stocks were pretty correlated (in a bad way) and the only real diversifier at that time was long Treasuries.
User avatar
Index Fan
Posts: 2587
Joined: Wed Mar 07, 2007 11:13 am
Location: The great Midwest

Re: Do REITs (and slicing in general) reduce risk?

Post by Index Fan »

I've been very happy owning the REIT Index over the past ~15 years- it often performs quite differently than the Total Stock Market Index. I'm basically a 3-fund portfolio holder with the addition of TIPS and REITs and have found that the extra diversification has worked well for me over the years.
"Optimum est pati quod emendare non possis." | -Seneca
rkhusky
Posts: 17654
Joined: Thu Aug 18, 2011 8:09 pm

Re: Do REITs (and slicing in general) reduce risk?

Post by rkhusky »

Doc wrote:
rkhusky wrote:
rca1824 wrote: My intuition tells me that a cap-weighted index must already yields the minimal risk for a given expected return.
The benefit of cap-weighting is that it is the most efficient way to put together a basket of stocks, hence it reduces costs.
As anil686 pointed out up thread most real estate is not held in publicly traded REITs but by businesses, partnerships etc. I'm willing to bet a Starbucks that most of the equity in the buildings on Wall Street itself is not being traded on the stock exchange. Therefore if you really wanted to cap weight RE you need to add it somehow to a TSM type portfolio.
My guess is that most REIT index's are cap-weighted.

I do not see the benefit of overweighting REIT's in a portfolio, just because most real estate is not in a REIT. If you want to diversify into private real estate, you should buy real estate, not REIT's.
Topic Author
rca1824
Posts: 719
Joined: Tue Jun 03, 2014 9:33 am

Re: Do REITs (and slicing in general) reduce risk?

Post by rca1824 »

Thank you all of you for the great feedback which gave me a lot to think about.
mptfan wrote:Because although REITS are stocks, and therefore part of the total stock market by definition, many people consider them to be a distinct asset class because, unlike other stocks, their performance is closely correlated with the economic cycle of the real estate market which, as the data shows, is somewhat uncorrelated with the economic cycle of the stock market. In short, REITs are different.
anil686 wrote:Thus, real estate represented by publicly traded entities is only about 3% of TSM - but is 13% of the economy at large. Much of this is private capital that is not investible but if you believe in weighting of sectors based on their percentage of the US economy - you would want to overweight REITs a little more than what is in TSM - Rick Ferri has much better explanation in All About Asset Allocation - second edition.
Ah these are both great explanations have convinced me that REITs are a good diversifying investment, but for different reasons. One reason is to use REITs merely restore real estate to its rightful place at 13% of your portfolio to represent the economy. Another reason is that they are another asset class with unique risk, so diversifying across different risk elements reduces risk.
The first argument, which is to hold them according to their true cap weight, doesn't seem to sit right with me. What if bonds were 20% of the US economy. Would that make it optimal to always hold 20% bonds? Of course not. We don't hold things to their arbitrary representation of the US economy, we hold things according to their unique risk factors. What if gold was 10% of the us economy? Wouldn't necessarily make gold a good investment. Or commodities, which are also a large part of the economy. Also not necessarily good.
Of course if you chose to diversify to every possible risk factor, things could get out of control. For example, imagine holding equal weights in all the 500 companies in the S&P 500. This would be the most "diversified" in a sense because you have equally weighted the unique risks associated with each company. But this would also be a disaster because if just one company went bankrupt, it would become a "black hole" and you would go bankrupt with it.
So we should be careful not to choose sectors or segments that are so small they can become black holes. But I think REITs and small cap value are not so narrow that they could potentially become black holes...although you never know. What if some huge sweeping change eliminated our need for space and real estate prices crashed 90%? That would quickly become a black hole and as you keep rebalancing into REIT, you would just keep sinking with it. Similarly if you held a regional index fund in say Japan, then Japan was annihlated by a nuclear meltdown, but you kept rebalancing into Japan as its index crashed, you would go bankrupt.
I hope these are not stupid questions but they are burning me right now.
steve_14 wrote:Nobody knows if overweighting X and thus underweighting Y will reduce risk. That depends entirely on future, unpredictable price movements.
Call_Me_Op wrote:It may or may not improve risk-adjusted return. There is no way to know ex ante
I really want to address this category of comments because there seems to some confusion over the meaning of "risk" and "ex ante".
Risk is the ex-ante variance in expected returns. We do know the ex-ante expected returns and ex-ante variance from looking at historical data and long term trends, or can at least estimate them to some confidence interval. From these estimates, we then can calculate the risk and ex-ante returns.
The rational way to make decisions is to integrate a utility function over all possible outcomes and maximize the expected utility. If the utility function is strictly concave to represent positive risk aversion, then variance in returns will decrease utility. There will be an optimal trade-off between maximizing returns and minimizing variance, which will be dependent upon the individual's investment horizon and that individual's personal preferences over consumption smoothing.
If you haven't taken an upper level microeconomics course and have questions on the above explanation please ask and I will elaborate or find a good reference.
anil686 wrote:Not a free lunch. You are diversifying... but REITs are not sure things and you are assuming risk.
Nothing is a sure thing but risk is the variance in expected returns. If the expected returns of REITs/stocks have less variance than the expected returns of pure stocks, then it can be said that diversifying into REITs reduces risk. This is in a sense a free lunch.
KyleAAA wrote:Of course, whether or not REITs will end up being reasonably uncorrelated with the broader market over any given period of time is a crapshoot because correlations shift over time.
Call_Me_Op wrote:An oversimplification. Correlations change over time - including the correlation between REITs and the broader market.
While we can't predict the ex-post correlation, we do know the ex-ante correlation from long term trends.
KyleAAA wrote:It is not the case that a cap-weighted index must already yield the minimum risk for a given expected return. In fact, diversification is usually referred to as the only free lunch in finance.
But is slicing a market really diversification? Why is it better, for example, to hold US and ex-US stock separately, as opposed to a single global fund? Many people believe the global fund is already optimal, expense ratios aside.
MrMatt2532 wrote:However, where I do expect benefit is from exposing yourself to unique risk factors, such as exposure to small companies, or value companies, which each behave differently than the market as a whole. Similarly I would expect benefit from exposure to international companies, which will behave differently than domestic companies.
So exposing to international companies is good just from the principal of diversification. That's adding a slice from another pie. The question is whether you can take one slice from a cap-weighted pie, blow it up, stuff it back into the original pie, and get a less risky portfolio?
For example, in reference to this thread by Rodc in 2007: http://www.bogleheads.org/forum/viewtopic.php?t=9445
The 4-way LB/LV/SB/SV, what looks like a so-called "diversified" slice and dice portfolio, had higher risk than TSM.
Monthly or yearly movements of stocks are often erratic and not indicative of changes in intrinsic value. Over time, however, stock prices and intrinsic value almost invariably converge. ~ WB
User avatar
Doc
Posts: 10606
Joined: Sat Feb 24, 2007 12:10 pm
Location: Two left turns from Larry

Re: Do REITs (and slicing in general) reduce risk?

Post by Doc »

rca1824 wrote:But is slicing a market really diversification? Why is it better, for example, to hold US and ex-US stock separately, as opposed to a single global fund? Many people believe the global fund is already optimal, expense ratios aside.
I think the concept of diversification is often abused. Is a 1000 element random sample really any more diversified than a 2000 element random sample taking from the same universe? Slicing and dicing is usually done for two reasons. 1) Because one segment has a higher expected return and you want more return despite risk concerns so you overweight it. 2) Because you believe that different segments have different risk/return/correlation attributes and you want to keep them constant irrespective of what the market itself does. So if the guv'ment doubles its public debt you don't buy twice as many Treasuries and if some index publisher changes its definition of emerging markets you don't necessarily go out and buy more. Same applies to US / ex-us stock. I don't consider either of these reason as "diversification".
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: Do REITs (and slicing in general) reduce risk?

Post by Rodc »

placeholder wrote:
Rodc wrote:I hold some REITS, but would add a comment that I think is really just part of what you are saying, but REITS were anything but low volatility in the Great Recession!
I don't think anyone said REITs were low volatility but rather that the non-correlation with broad stocks along with similar returns lowers portfolio volatility although in 2008 all stocks were pretty correlated (in a bad way) and the only real diversifier at that time was long Treasuries.
They certainly did before 2008. :)

Which is why steve_14 mentions past belief in low volatility.
Last edited by Rodc on Thu Jun 12, 2014 11:14 am, edited 1 time in total.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.
steve_14
Posts: 1507
Joined: Wed Jun 20, 2012 12:05 am

Re: Do REITs (and slicing in general) reduce risk?

Post by steve_14 »

rca1824 wrote:So exposing to international companies is good just from the principal of diversification. That's adding a slice from another pie.
Yes. Owning Ford and Mercedes is better than owning just one or the other.
rca1824 wrote:The question is whether you can take one slice from a cap-weighted pie, blow it up, stuff it back into the original pie, and get a less risky portfolio? For example, in reference to this thread by Rodc in 2007: http://www.bogleheads.org/forum/viewtopic.php?t=9445
The 4-way LB/LV/SB/SV, what looks like a so-called "diversified" slice and dice portfolio, had higher risk than TSM.
Or, continuing my example, can we squeeze a free lunch out of overweighting Volkswagen and underweighting Ford, betting we can switch back and forth between them at just the right times? In other words - have we got the market all figured out? I wouldn't bet my retirement on it.
Valuethinker
Posts: 48954
Joined: Fri May 11, 2007 11:07 am

Re: Do REITs (and slicing in general) reduce risk?

Post by Valuethinker »

rca1824 wrote: Thank you all of you for the great feedback which gave me a lot to think about.
mptfan wrote:Because although REITS are stocks, and therefore part of the total stock market by definition, many people consider them to be a distinct asset class because, unlike other stocks, their performance is closely correlated with the economic cycle of the real estate market which, as the data shows, is somewhat uncorrelated with the economic cycle of the stock market. In short, REITs are different.
anil686 wrote:Thus, real estate represented by publicly traded entities is only about 3% of TSM - but is 13% of the economy at large. Much of this is private capital that is not investible but if you believe in weighting of sectors based on their percentage of the US economy - you would want to overweight REITs a little more than what is in TSM - Rick Ferri has much better explanation in All About Asset Allocation - second edition.
Ah these are both great explanations have convinced me that REITs are a good diversifying investment, but for different reasons. One reason is to use REITs merely restore real estate to its rightful place at 13% of your portfolio to represent the economy. Another reason is that they are another asset class with unique risk, so diversifying across different risk elements reduces risk.
The first argument, which is to hold them according to their true cap weight, doesn't seem to sit right with me. What if bonds were 20% of the US economy. Would that make it optimal to always hold 20% bonds? Of course not. We don't hold things to their arbitrary representation of the US economy, we hold things according to their unique risk factors. What if gold was 10% of the us economy? Wouldn't necessarily make gold a good investment. Or commodities, which are also a large part of the economy. Also not necessarily good.
I have been making a related argument (that in an efficient market, the market knows the holdings of RE of the listed companies in the US or anywhere else and there is no value to be created just by overweighting on the assumption that somehow you are seeing something the market doesn't know).

Somewhat contrary to my argument there *do* appear to be some gains to market value from spinning out corporate property assets as a separate vehicle. But then spinoffs generally do create value for shareholders-- so it's a broader information problem.

I don't think it warrants a c. 4.0x overweighting in REITs. This ignores the fact that REITs are also stocks.

BTW my argument never shifts anyone ;-). See if your formulation does ;-).

various comments about risk.
if you believe REITs have less than perfect correlation with bonds and equities then they can add to portfolio efficiency. Empirically they seem to.

HOWEVER

- I think there are things in the historical data that won't repeat in the future (it was a small sector, no one understood it that well, changes in tax law etc.)-- if you look at REIT performance it seems to have a very distinct trend (serial correlation of annual returns)-- that makes naive calculation of numbers a bit hazardous (first thing they always say in econometrics 'graph the data'!).

- it's clear from 2008-09 that you are taking on significant financial risk, arising from the financial leverage of REITs. In a financial crash, they get hammered, and worse than stocks. The flipside is what Japan-REITs have been doing (after a decades long slump) which is turbocharged--if monetary policy ever works, then REITs benefit

I see adding REITs then as risking 'deworsification': higher volatility to the portfolio and not addint to average returns. For most investors, I suspect the volatility shown in 08-09 is too much to stomache.

I also think that valuation matters. If you are going to buy into REITs, you want to make sure you think the sector is cheap (on a dividend yield and discount to NAV basis). The economic cycle matters more in real estate than it does in the market as a whole (and Commercial RE seems to have this 10 to 14 year cycle).

It is clear (at least to me) that if we have unexpected inflation, then REITs are likely to outperform, by the nature of the underlying asset and its greater correlation with inflation (double whammy: rents rise with inflation, and they pay back their debts with inflated dollars). They are a greater inflation hedge than the stock market as a whole.
placeholder
Posts: 8375
Joined: Tue Aug 06, 2013 12:43 pm

Re: Do REITs (and slicing in general) reduce risk?

Post by placeholder »

Rodc wrote:They certainly did before 2008.
I can't comment as to what people believed prior to 2008 but the growth charts don't support the notion that REITs (at least VGSIX) were less volatile than large caps.
User avatar
Ketawa
Posts: 2521
Joined: Mon Aug 22, 2011 1:11 am
Location: DC

Re: Do REITs (and slicing in general) reduce risk?

Post by Ketawa »

Robert T has a great post in this thread that discusses whether REITs should be considered a separate asset class from a Fama/French factor loading perspective. FWIW I don't think there is a good reason to hold a separate allocation to REITs, especially for small value tilters.
steve_14
Posts: 1507
Joined: Wed Jun 20, 2012 12:05 am

Re: Do REITs (and slicing in general) reduce risk?

Post by steve_14 »

Rodc wrote:
placeholder wrote:
Rodc wrote:I hold some REITS, but would add a comment that I think is really just part of what you are saying, but REITS were anything but low volatility in the Great Recession!
I don't think anyone said REITs were low volatility but rather that the non-correlation with broad stocks along with similar returns lowers portfolio volatility although in 2008 all stocks were pretty correlated (in a bad way) and the only real diversifier at that time was long Treasuries.
They certainly did before 2008. :)

Which is why steve_14 mentions past belief in low volatility.
Indeed, investment books from the 1990s and 2000s (Including I believe Bill Bernstein's books) describe REITS as halfway between stocks and bonds on the risk/return scale. Basing your REIT allocation off data from those days isn't just skating where the puck has been, it's confusing the puck with a soccer ball.
Topic Author
rca1824
Posts: 719
Joined: Tue Jun 03, 2014 9:33 am

Re: Do REITs (and slicing in general) reduce risk?

Post by rca1824 »

There was a really good point in this topic () that shatters the idea of investing in REITs:
Not so unique risk? REITs seem more like a sector rather than an asset class. Even though market, value, size (and fixed interest) exposure only explains about 25 percent of the movement of the REIT index, with about 75 percent of the returns being unique to REITs, the same argument can be made for most sectors or industries, with the three factor model having as low or even lower explanatory power. The three factor model results do not seem to make a special case for REITs as a separate asset class beyond other sectors.
So REITs are uncorrelated with the broader market. But couldn't the same be said for any other sector such as Tech or Pharma?
Or is the idea that REITs are just a proxy for real estate, i.e. land, which is an asset class?
What makes investing in land any different from investing in gold? Because the demand for land is grounded in a real need that increases with population growth?
Could land price stagnate when population reaches a steady state?
Could land prices crash if populations decline or people find substitute goods, e.g. space colonies, underground colonies, sky colonies, or just really tall skyscrapers?
For the above reasons, it seems like the REIT sector might be at risk of becoming a "black hole", though probably not in my lifetime.
Monthly or yearly movements of stocks are often erratic and not indicative of changes in intrinsic value. Over time, however, stock prices and intrinsic value almost invariably converge. ~ WB
pascalwager
Posts: 2312
Joined: Mon Oct 31, 2011 8:36 pm

Re: Do REITs (and slicing in general) reduce risk?

Post by pascalwager »

In his new book, W. Bernstein considers REITs an inflation hedge, but doesn't recommend their purchase at high current price.
VT 60% / VFSUX 20% / TIPS 20%
Topic Author
rca1824
Posts: 719
Joined: Tue Jun 03, 2014 9:33 am

Re: Do REITs (and slicing in general) reduce risk?

Post by rca1824 »

pascalwager wrote:In his new book, W. Bernstein considers REITs an inflation hedge, but doesn't recommend their purchase at high current price.
Isn't that market timing? How can you know whether an asset is over- or under-valued? People have been making similar claims about there being a long-run average P/E ratio that the market will revert to, but real strategies designed around buying and selling based on the P/E ratio tend not to be very successful as passive investing plus there is always uncertainty as to what exactly the true long run average P/E ratio is that the market is allegedly reverting to.

Any expectations of future price movements should already be built into the current price under efficient market hypothesis, right??
Monthly or yearly movements of stocks are often erratic and not indicative of changes in intrinsic value. Over time, however, stock prices and intrinsic value almost invariably converge. ~ WB
Valuethinker
Posts: 48954
Joined: Fri May 11, 2007 11:07 am

Re: Do REITs (and slicing in general) reduce risk?

Post by Valuethinker »

rca1824 wrote:
pascalwager wrote:In his new book, W. Bernstein considers REITs an inflation hedge, but doesn't recommend their purchase at high current price.
Isn't that market timing? How can you know whether an asset is over- or under-valued? People have been making similar claims about there being a long-run average P/E ratio that the market will revert to, but real strategies designed around buying and selling based on the P/E ratio tend not to be very successful as passive investing plus there is always uncertainty as to what exactly the true long run average P/E ratio is that the market is allegedly reverting to.

Any expectations of future price movements should already be built into the current price under efficient market hypothesis, right??
There's a thread (his ident is wbern) here and I think it is very illuminating. Suffice it to say he's not a pure EMH guy. If you use the Gordon Growth model on REITs at the yields then prevalent, you did not get a good picture.
Valuethinker
Posts: 48954
Joined: Fri May 11, 2007 11:07 am

Re: Do REITs (and slicing in general) reduce risk?

Post by Valuethinker »

rca1824 wrote:There was a really good point in this topic () that shatters the idea of investing in REITs:
Not so unique risk? REITs seem more like a sector rather than an asset class. Even though market, value, size (and fixed interest) exposure only explains about 25 percent of the movement of the REIT index, with about 75 percent of the returns being unique to REITs, the same argument can be made for most sectors or industries, with the three factor model having as low or even lower explanatory power. The three factor model results do not seem to make a special case for REITs as a separate asset class beyond other sectors.
So REITs are uncorrelated with the broader market. But couldn't the same be said for any other sector such as Tech or Pharma?
Or is the idea that REITs are just a proxy for real estate, i.e. land, which is an asset class?
What makes investing in land any different from investing in gold? Because the demand for land is grounded in a real need that increases with population growth?
Could land price stagnate when population reaches a steady state?
Could land prices crash if populations decline or people find substitute goods, e.g. space colonies, underground colonies, sky colonies, or just really tall skyscrapers?
For the above reasons, it seems like the REIT sector might be at risk of becoming a "black hole", though probably not in my lifetime.
Way too much crystalballing in that to be able to make definitive statements.

The country with the most land (developed country) per person has the tallest skyscrapers and some of the highest land values (the USA)-- probably the highest in spots. Canada and Australia aren't much different (Vancouver has the distinction of having the most overvalued homes in the world by several measures).

Office accomodation in Moscow, in a country which has declining population and not a great economy or political situation, was the most expensive in the world I believe.

Couple of secular trends going on:

- there is a war on office space per employee- it's shrunk by a lot (down around avg 95 square feet ie at least a 40% fall) in the last 30 years: 'paperless' offices and filing, removing walls, teleworking etc. Yet companies are also putting in canteens, playrooms etc- a countervailing force

- whilst population might be a factor (see Japan, or former German Democratic Republic where outmigration has been high) J REITs are blazing right now-- in other words, there are other factors going on

- valuation is of course on top of this
placeholder
Posts: 8375
Joined: Tue Aug 06, 2013 12:43 pm

Re: Do REITs (and slicing in general) reduce risk?

Post by placeholder »

Valuethinker wrote:There's a thread (his ident is wbern) here
Quick note he changed his ID to Bill Bernstein a little while back:
http://www.bogleheads.org/forum/memberl ... ile&u=2464
Topic Author
rca1824
Posts: 719
Joined: Tue Jun 03, 2014 9:33 am

Re: Do REITs (and slicing in general) reduce risk?

Post by rca1824 »

What I'm trying to get at is why do we overweight REIT but other sectors?
What is so special about investing land, per se? As opposed to gold or other commodities?
Monthly or yearly movements of stocks are often erratic and not indicative of changes in intrinsic value. Over time, however, stock prices and intrinsic value almost invariably converge. ~ WB
steve_14
Posts: 1507
Joined: Wed Jun 20, 2012 12:05 am

Re: Do REITs (and slicing in general) reduce risk?

Post by steve_14 »

rca1824 wrote:What I'm trying to get at is why do we overweight REIT but other sectors?
What if they had a 9% yield with low volatility, while all the other sectors had a 1% yield, with high volatility? That's the way things were at the beginning of this century. REITs were clearly something special. Unfortunately it seems a lot of folks are investing based on books they read around that time, instead of looking at the way things are in 2014.
User avatar
Taylor Larimore
Posts: 32839
Joined: Tue Feb 27, 2007 7:09 pm
Location: Miami FL

The "Efficiency" of Total Market Index Funds.

Post by Taylor Larimore »

One day I'm convinced to add a "slice". The next I decide to stick with the core 3.

There are convincing arguments on both sides. I keep hearing Taylor L. whispering "keep it simple...keep it simple.."
Bloom:

You hear me whispering "keep it simple" because a simple portfolio is lower cost (including taxes), and easier to understand and manage by its owner, spouse, future caregivers and heirs.

The Three Fund Portfolio of total market index funds has another important advantage: Its Efficiency (highest return with the least amount of risk).

Three Proof's That TSM is Efficient

Best wishes.
Taylor
"Simplicity is the master key to financial success." -- Jack Bogle
User avatar
Doc
Posts: 10606
Joined: Sat Feb 24, 2007 12:10 pm
Location: Two left turns from Larry

Re: Do REITs (and slicing in general) reduce risk?

Post by Doc »

rca1824 wrote:What I'm trying to get at is why do we overweight REIT but other sectors?
It been addressed:

1) Adding REITs to Total Stock Market does not overweight real estate investments with respect to the total US equity market because a large portion of the real estate investment is not publically traded on the stock exchanges.

2) REITs unlike many other equity sectors have different basic capitalization/profit factors than most other business represented on the stock exchanges and therefore are expected to react differently at time to the general stock market.

Other sectors might have some differences on their own but not to the extent of REITs.

Now you may not buy into these factors but the people who add REITs to a TSM portfolio do.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
mptfan
Posts: 7201
Joined: Mon Mar 05, 2007 8:58 am

Re: Do REITs (and slicing in general) reduce risk?

Post by mptfan »

rca1824 wrote: What makes investing in land any different from investing in gold? Because the demand for land is grounded in a real need that increases with population growth?
Yes, and because land produces income in the form of rent payments that gold does not. Land is an income producing asset and gold is not.
mptfan
Posts: 7201
Joined: Mon Mar 05, 2007 8:58 am

Re: Do REITs (and slicing in general) reduce risk?

Post by mptfan »

rca1824 wrote:The first argument, which is to hold them according to their true cap weight, doesn't seem to sit right with me. What if bonds were 20% of the US economy. Would that make it optimal to always hold 20% bonds? Of course not. We don't hold things to their arbitrary representation of the US economy, we hold things according to their unique risk factors. What if gold was 10% of the us economy? Wouldn't necessarily make gold a good investment. Or commodities, which are also a large part of the economy. Also not necessarily good.
:thumbsup
Some people think that just because "the market" as defined by the publicly traded stock market allocates a certain amount to a certain sector, then that is gospel on what is the correct asset allocation for everyone, and I think that is wrong. Interestingly, some of these same people will refuse to allocate their bond investments based on the market allocations to bonds. For example, the bond market may allocate 20% to high yield corporate bonds (a made up number), but the same people who worship at the altar of market allocations as being optimal for stocks will refuse to follow the market allocation for bonds.
steve_14
Posts: 1507
Joined: Wed Jun 20, 2012 12:05 am

Re: Do REITs (and slicing in general) reduce risk?

Post by steve_14 »

mptfan wrote:Some people think that just because "the market" as defined by the publicly traded stock market allocates a certain amount to a certain sector, then that is gospel on what is the correct asset allocation for everyone, and I think that is wrong. Interestingly, some of these same people will refuse to allocate their bond investments based on the market allocations to bonds. For example, the bond market may allocate 20% to high yield corporate bonds (a made up number), but the same people who worship at the altar of market allocations as being optimal for stocks will refuse to follow the market allocation for bonds.
One, diversifying within an asset class, and choosing which classes to own, are entirely separate concepts. I take my risk on the stock side, so I wouldn't hold risky bonds to begin with. Nor am I interested in the market weightings of, say, Chinese real estate, since I've decided a level up from there not to own this class.

Two, nobody's worshiping at any alters. There may be valid reasons to hold other than TSM weightings in general. This thread isn't about why TSM portfolios are good ideas, but why cluttering your portfolio with REITs is misguided. IMO TSM should be the default position if you believe markets are efficient, since it agrees with the market values. IOW if Company A has 1 oil well and Company B has 9, and the market is split 10%/90% between the two, that's probably your best allocation. That said, any broadly diversified portfolio should be expected to perform the same as any other going forward, so overweighting X or Y isn't a big deal.
mptfan
Posts: 7201
Joined: Mon Mar 05, 2007 8:58 am

Re: Do REITs (and slicing in general) reduce risk?

Post by mptfan »

steve, thank you for proving my point for me. I hear what you are saying, but I will never understand why people argue that TSM is the default or optimal allocation for stocks, while at the same time, they reject the market allocation to bonds. Sorry, that is inconsistent to me no matter how many times you repeat the "I take my risks on the stock side" mantra.
User avatar
Doc
Posts: 10606
Joined: Sat Feb 24, 2007 12:10 pm
Location: Two left turns from Larry

Re: Do REITs (and slicing in general) reduce risk?

Post by Doc »

steve_14 wrote: IMO TSM should be the default position if you believe markets are efficient, since it agrees with the market values.
Efficient markets have little to do with Total Stock Market. All that an efficient market hypothesis says is that within a given market where everyone participating in that market has the same access and information then all the assets in that market are priced appropriately according to their risk profile. There is a significant amount of investable real estate in the US that cannot be accessed on stock exchanges. If you want to weigh your portfolio to the market weight of investible real estate you can't just buy TSM. One way is to invest in real estate through partnerships or individually. This is not practical for retail investors. As a (perhaps poor) substitute we can add more REITs than is represented in the TSM. All EMH tells us is that those REITs are priced correctly given their risk. If you don't want a market weight in investable real estate don't add REIT's to TSM. But don't make up artificial reasons for not doing so.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
steve_14
Posts: 1507
Joined: Wed Jun 20, 2012 12:05 am

Re: Do REITs (and slicing in general) reduce risk?

Post by steve_14 »

mptfan wrote:steve, thank you for proving my point for me. I hear what you are saying, but I will never understand why people argue that TSM is the default or optimal allocation for stocks, while at the same time, they reject the market allocation to bonds. Sorry, that is inconsistent to me no matter how many times you repeat the "I take my risks on the stock side" mantra.
Well, what about Chinese real estate? Do I have to own the market weighting there, too? Why or why not?
Topic Author
rca1824
Posts: 719
Joined: Tue Jun 03, 2014 9:33 am

Re: Do REITs (and slicing in general) reduce risk?

Post by rca1824 »

mptfan wrote:
rca1824 wrote: What makes investing in land any different from investing in gold? Because the demand for land is grounded in a real need that increases with population growth?
Yes, and because land produces income in the form of rent payments that gold does not. Land is an income producing asset and gold is not.
This is starting to make more sense. Land is something that everything need, so land can always be rented out for profit.
I just worry that just because land has always been rented out for profit since the dawn of civilization, means that it will continue to be rented out.
If some unforeseeable event reduces the demand for land, then land prices will plummet and REITs will plummet with them. I don't know what kind of event would hit land but not corporate equity, but any sort of positive supply shock or negative demand shock would drastically alter the price of land.
One time a ship carrying gold sank in the ocean, and the price of gold skyrocketed and there was huge deflation.
So the question is, "do you expect any large supply or demand shocks for land that would drastically alter its price?"
That sounds far-fetched, but if some day we start building sky colonies or space colonies, then the demand for land will plummet, land prices will plummet, and REITs will become a "black hole" in everyone's portfolio.

Perhaps it won't happen in our lifetime, but I do challenge the idea that real estate is a core asset class the same as stocks or bonds. It seems more like a sector within stocks. It may be a sector with a long history and and the most unique risk factors, but it is still just a sector nonetheless. It may be an underweighted sector, due to the prevalence of private land holdings, in which case it would be rational to overweight REITs in a well-diversified portfolio. But you would have to be careful to adjust your allocation to match its true weight in the economy so that it doesn't become a "black hole". If land ever shrinks from 20% to 5%, you should NOT keep rebalancing to 20% but rebalance according to the new market cap. The same would hold true for anyone holding cap-weights of US/intl equity. If you hold 50/50 US/intl and keep rebalancing to 50/50, but the US becomes a black hole and crashes while the world economy flourishes, rebalancing to the same old 50/50 is a losing strategy. You want to rebalance according to the true market cap.
Monthly or yearly movements of stocks are often erratic and not indicative of changes in intrinsic value. Over time, however, stock prices and intrinsic value almost invariably converge. ~ WB
steve_14
Posts: 1507
Joined: Wed Jun 20, 2012 12:05 am

Re: Do REITs (and slicing in general) reduce risk?

Post by steve_14 »

Doc wrote:
steve_14 wrote: IMO TSM should be the default position if you believe markets are efficient, since it agrees with the market values.
Efficient markets have little to do with Total Stock Market. All that an efficient market hypothesis says is that within a given market where everyone participating in that market has the same access and information then all the assets in that market are priced appropriately according to their risk profile.
Sure they do. See my oil field example above. If the market's efficient, you can't do better than a 90%/10% weighting. You've got a 10% bet on each oil field. That's maximum diversification across the elements of value. Now, maybe you can do "different", with for example, an 80/20 bet to account for management risk. Or maybe you could bet 100% on the larger stock, to account for its economies of scale. But you can't do "better".

If the market's not efficient, you simply find the better stock and make a big bet on it. You don't care how the market values each stock.
Doc wrote:There is a significant amount of investable real estate in the US that cannot be accessed on stock exchanges. If you want to weigh your portfolio to the market weight of investible real estate you can't just buy TSM.
But that goes back to my example of overweighting PF Chang stock because there are a million Chinese restaurants in the world. The existence of those million restaurants is irrelevant to the risk, return, correlations, profits, growth, quality of the egg rolls, etc. Of PF Chang. Many are probably "investible" if you put in enough effort, but they're a different type of investment.
dbr
Posts: 46137
Joined: Sun Mar 04, 2007 8:50 am

Re: Do REITs (and slicing in general) reduce risk?

Post by dbr »

mptfan wrote:My data tells me that REITs have a .51 correlation with the total stock market index fund over the last 30 years, and over that time, REITs have had a CAGR of 10.58% while TSM was 11.2%, while exhibiting almost identical standard deviation (18.02 vs 17.91). So my conclusion is that REITs provide reasonable diversification over time that will likely reduce portfolio volatility. So the answer to your question is yes.
Yes, this simple answer is about as much as you are going to get to answer your question. All of this depends on properly understanding what it means to estimate a future outcome.
User avatar
Doc
Posts: 10606
Joined: Sat Feb 24, 2007 12:10 pm
Location: Two left turns from Larry

Re: Do REITs (and slicing in general) reduce risk?

Post by Doc »

steve-14 wrote:IOW if Company A has 1 oil well and Company B has 9, and the market is split 10%/90% between the two, that's probably your best allocation.
And if both A and B are independent drillers it has nothing to do with TSM or EMH either. On the other hand if A and B are Exxon and Chevron they do have something to do with TSM and EMH. But that doesn't mean that is best for you or me.

And how do you determine what is the "best" allocation for yourself. The only way that TSM is best is that if you have exactly the same risk/return goal as the the average investor in that specific market. I may define minimum risk as having enough dividends coming in each month to pay my retirement expenses so that I don't risk my health by losing sleep worrying about my portfolio value going down. That may not be the same group of stocks that maximizes my total return which may be the goal of a twenty something.
A scientist looks for THE answer to a problem, an engineer looks for AN answer and lawyers ONLY have opinions. Investing is not a science.
mptfan
Posts: 7201
Joined: Mon Mar 05, 2007 8:58 am

Re: Do REITs (and slicing in general) reduce risk?

Post by mptfan »

steve_14 wrote:
mptfan wrote:steve, thank you for proving my point for me. I hear what you are saying, but I will never understand why people argue that TSM is the default or optimal allocation for stocks, while at the same time, they reject the market allocation to bonds. Sorry, that is inconsistent to me no matter how many times you repeat the "I take my risks on the stock side" mantra.
Well, what about Chinese real estate? Do I have to own the market weighting there, too? Why or why not?
You don't have to own anything you don't want to own.
Post Reply