In Search of Distress Risk

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
User avatar
Topic Author
Robert T
Posts: 2555
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

In Search of Distress Risk

Post by Robert T » Sat Jan 30, 2010 10:15 pm

.
Here are a few recent articles by John Campbell of Harvard University.
The papers have a similar conclusion: “...[the results] are inconsistent with the conjecture that the value and size effects are compensation for the risk of financial distress.”

Quite an important statement/paper, which IMO requires some consideration. Here’s my take:

Concerns on methodology:
  • 1. The paper limits the influence of outliers on the analysis through ‘winsorizing’ the variables (“we replace any observation below the 5th percentile with the 5th percentile, and any observation above the 95th percentile with the 95th percentile”). But if we believe Taleb, in finance outliers drive the mean, and from the FF migration paper only 3% of small value stocks annually accounted for 60% of the historical SV annual returns. So ‘winsorizing’ may bias the results.

    2. The logit model (0/1 dependent variable) seems to rely on a very small share of ‘failures’ in the overall data set (0.09% of total observations in the second paper as I understand). And the subsequent explanatory power seems relatively low.
Putting the methodology concerns aside: My interpretation of the results (Table VI) in the first linked paper.
  • 1. Financial distress (probability of failure) is amplified in the smallest (micro) stocks. The portfolio sorts with highest probability of failure have size loads of 1.46, 1.55 and 1.98. These are micro-cap stocks (no surprise here).

    2. The alphas in the 4-factor model are only consistently significant (and negative) for small/micro cap stocks (particularly for size loads of 1.46, 1.55 and 1.98.).

    3. And the price-to-book on these highly distressed portfolios seems to go up. [also see figure 3: there seems to be a divergence between size and value loads towards the end of the chart]

    4. So it may be that the main focus of the result is simply on the small/micro growth (higher price-to-book) portfolios (?). i.e. small growth stocks have delivered anomalously low returns (as we known).
So why will small/micro growth stocks be also those under greatest “financial distress” (as in the paper findings)? Well I think part of the story may simply be that the result reflects the way financial distress (probability of failure) was calculated in the paper, which-in itself included measures of size and standard deviation of which small/micro growth ranks very high. So may just be highlighting the small growth anomaly rather than a ‘financial distress’ anomaly.

Again, just my take.

Here are a few other papers on the subject: Robert
.

grok87
Posts: 8585
Joined: Tue Feb 27, 2007 9:00 pm

Re: In Search of Distress Risk

Post by grok87 » Sat Jan 30, 2010 11:19 pm

Robert T wrote:.
Here are a few recent articles by John Campbell of Harvard University.
The papers have a similar conclusion: “...[the results] are inconsistent with the conjecture that the value and size effects are compensation for the risk of financial distress.”

Quite an important statement/paper, which IMO requires some consideration. Here’s my take:

Concerns on methodology:
  • 1. The paper limits the influence of outliers on the analysis through ‘winsorizing’ the variables (“we replace any observation below the 5th percentile with the 5th percentile, and any observation above the 95th percentile with the 95th percentile”). But if we believe Taleb, in finance outliers drive the mean, and from the FF migration paper only 3% of small value stocks annually accounted for 60% of the historical SV annual returns. So ‘winsorizing’ may bias the results.

    2. The logit model (0/1 dependent variable) seems to rely on a very small share of ‘failures’ in the overall data set (0.09% of total observations in the second paper as I understand). And the subsequent explanatory power seems relatively low.
Putting the methodology concerns aside: My interpretation of the results (Table VI) in the first linked paper.
  • 1. Financial distress (probability of failure) is amplified in the smallest (micro) stocks. The portfolio sorts with highest probability of failure have size loads of 1.46, 1.55 and 1.98. These are micro-cap stocks (no surprise here).

    2. The alphas in the 4-factor model are only consistently significant (and negative) for small/micro cap stocks (particularly for size loads of 1.46, 1.55 and 1.98.).

    3. And the price-to-book on these highly distressed portfolios seems to go up. [also see figure 3: there seems to be a divergence between size and value loads towards the end of the chart]

    4. So it may be that the main focus of the result is simply on the small/micro growth (higher price-to-book) portfolios (?). i.e. small growth stocks have delivered anomalously low returns (as we known).
So why will small/micro growth stocks be also those under greatest “financial distress” (as in the paper findings)? Well I think part of the story may simply be that the result reflects the way financial distress (probability of failure) was calculated in the paper, which-in itself included measures of size and standard deviation of which small/micro growth ranks very high. So may just be highlighting the small growth anomaly rather than a ‘financial distress’ anomaly.

Again, just my take.

Here are a few other papers on the subject: Robert
.
THanks Robert,
Re your question
Robert T wrote: So why will small/micro growth stocks be also those under greatest “financial distress” (as in the paper findings)?
You may be interested in the "Accruals Anomaly". This was first documented by Sloan in 1996. Here is a more recent paper
http://www.columbia.edu/~dn75/The%20Per ... cation.pdf
I think the connection is that often firms with lower quality earnings (earnings where accruals make up a large portion of earnings rather than cash flow) tend to concentrated in small growth
cheers,
RIP Mr. Bogle.

User avatar
Topic Author
Robert T
Posts: 2555
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Post by Robert T » Mon Feb 01, 2010 11:10 pm

.
Thanks grok,

Perhaps accruals have something to do with it, don’t really know. FWIW here are return comparisons over the study period: 1981-2003 (for the first paper). So significant underperformance of small growth.

Code: Select all

1981-2003
                  Annualized    Standard
                    Return      Deviation 

FF Small Growth       6.5         23.5      
FF Small Value       19.4         20.7 
FF Large Growth      12.3         18.9
FF Large Value       14.9         16.1

Source: Ken French website
The second Campbell et al. paper which is easier to read, defines firm failure as “chapter 7 or 11 bankruptcy filing, de-listing due to performance related reasons, and a default or selective default rating by a rating agency.” If these are mainly ‘small growth’ firms then perhaps Bridgeway’s approach of avoiding stocks “passing through to bankruptcy” has some merit (although it seems to have hurt last year).

Robert
.

rogue409
Posts: 78
Joined: Mon Jan 28, 2008 11:19 pm

Re: In Search of Distress Risk

Post by rogue409 » Tue Feb 02, 2010 8:46 am

Robert T wrote: The logit model (0/1 dependent variable) seems to rely on a very small share of ‘failures’ in the overall data set (0.09% of total observations in the second paper as I understand). And the subsequent explanatory power seems relatively low.
1. The model doesn't really "rely" on just the failures; the "successes" contribute to the model. The number of failures is small, but nothing can be done about that. There isn't a better model - the alternative is ignorance.
2. The explanatory power ranges from reasonably good to pretty bad.

Predicting success or failure of a single firm is not a good use of this model.

Understanding why a firm might fail IS a good use of the model, as well as predicting how many firms might fail from a fairly large subsection of all firms. A manager just needs to be slightly better than everyone else to get above normal returns. Although, now that this information is very public, it might be hard to do that.

User avatar
wbond
Posts: 1073
Joined: Wed Dec 10, 2008 6:55 pm

Post by wbond » Tue Feb 02, 2010 5:37 pm

Robert,

Thanks for the interesting links (I haven't read them yet, but will).

A couple of thoughts that came to mind:

Small (and possibly value) risk may be associated with a liquidity risk premium (such as discussed here).

I am also reminded of a five-year old W. Bernstein piece that you've probably seen: The Nature of Risk. In it he also sites John Campbell's work that makes the same point you do in your post about distress risk, and, in addition, he describes value as being a decent surrogate for the opposite of speculative risk, rather than as a "risk story" itself. Call this behavioral if you will, but I like the idea, i.e. if correct, it implies that the free lunch aspect of the value premium will be "arbitraged away" only when speculation is excised from investing - and from human nature one might add.

grok87
Posts: 8585
Joined: Tue Feb 27, 2007 9:00 pm

Post by grok87 » Tue Feb 02, 2010 7:43 pm

wbond wrote:Robert,

Thanks for the interesting links (I haven't read them yet, but will).

A couple of thoughts that came to mind:

Small (and possibly value) risk may be associated with a liquidity risk premium (such as discussed here).

I am also reminded of a five-year old W. Bernstein piece that you've probably seen: The Nature of Risk. In it he also sites John Campbell's work that makes the same point you do in your post about distress risk, and, in addition, he describes value as being a decent surrogate for the opposite of speculative risk, rather than as a "risk story" itself. Call this behavioral if you will, but I like the idea, i.e. if correct, it implies that the free lunch aspect of the value premium will be "arbitraged away" only when speculation is excised from investing - and from human nature one might add.
nice post wbond!
RIP Mr. Bogle.

Gekko
Posts: 3779
Joined: Fri May 11, 2007 5:00 pm
Location: USA

Post by Gekko » Wed Feb 03, 2010 4:35 pm

hmmmn. maybe the Harvard Endowment people should read it.

richard
Posts: 7961
Joined: Tue Feb 20, 2007 3:38 pm
Contact:

Post by richard » Wed Feb 03, 2010 5:00 pm

For those of us who like to find quotes which agree with things we often say,
this is rather satisfying:
This result is a significant challenge to the conjecture that the value and size effects are proxies for a financial distress premium
Would someone please remind me of the economic risk for which small and value proxy?

User avatar
BlueEars
Posts: 3657
Joined: Sat Mar 10, 2007 12:15 am
Location: West Coast

Post by BlueEars » Wed Feb 03, 2010 6:12 pm

richard wrote:...(snip)...
Would someone please remind me of the economic risk for which small and value proxy?
If returns count as economic risk, then maybe one recent example is for Mar-07 through Feb-08:

Code: Select all

VFINX (sp500)  -45.4%
VISVX (SV)     -53.4%
VISGX (SG)     -47.0%

richard
Posts: 7961
Joined: Tue Feb 20, 2007 3:38 pm
Contact:

Post by richard » Wed Feb 03, 2010 6:21 pm

Les wrote:
richard wrote:...(snip)...
Would someone please remind me of the economic risk for which small and value proxy?
If returns count as economic risk, then maybe one recent example is for
Return does not really count as an underlying economic risk. Return is what we are trying to explain, so it can't also be the explanation.

The argument becomes circular:
1) Why do some stocks perform worse than others?
2) Risk
3) What is the risk?
4) The fact that some stocks perform worse than others

User avatar
BlueEars
Posts: 3657
Joined: Sat Mar 10, 2007 12:15 am
Location: West Coast

Post by BlueEars » Wed Feb 03, 2010 7:04 pm

Perhaps the negative performance numbers during periods of distress illustrate that the positive premium for small and value comes with a risk. The bet is that the risk will not show up for very long and will not bring down the system. The Great Depression and recent Recession didn't do us in but illustrated the tip of the iceberg's risk. The ship is still afloat for now.

dumbmoney
Posts: 2304
Joined: Sun Mar 16, 2008 8:58 pm

Post by dumbmoney » Wed Feb 03, 2010 8:05 pm

It seems most equity risk factors are unrewarded (or even penalized), which is a challenge to efficient market theory. For example, you'd think that stocks with steady earnings would return less than stocks with volatile earnings. Nope.

One risk factor that is rewarded is liquidity risk.
I am pleased to report that the invisible forces of destruction have been unmasked, marking a turning point chapter when the fraudulent and speculative winds are cast into the inferno of extinction.

User avatar
Topic Author
Robert T
Posts: 2555
Joined: Tue Feb 27, 2007 9:40 pm
Location: 1, 0.2, 0.4, 0.5
Contact:

Post by Robert T » Wed Feb 24, 2010 7:45 pm


grok87
Posts: 8585
Joined: Tue Feb 27, 2007 9:00 pm

Post by grok87 » Wed Feb 24, 2010 11:00 pm

Robert T wrote:.
May be of interest:

Fama-French Q&A: Bankrupt Firms: Who's buying?
Thanks that was interesting- I didn't know about the minimum $1 per share rule for collateral.
cheers,
RIP Mr. Bogle.

peter71
Posts: 3768
Joined: Tue Jul 24, 2007 8:28 pm

Re: In Search of Distress Risk

Post by peter71 » Thu Feb 25, 2010 12:43 am

rogue409 wrote:
Robert T wrote: The logit model (0/1 dependent variable) seems to rely on a very small share of ‘failures’ in the overall data set (0.09% of total observations in the second paper as I understand). And the subsequent explanatory power seems relatively low.
1. The model doesn't really "rely" on just the failures; the "successes" contribute to the model. The number of failures is small, but nothing can be done about that. There isn't a better model - the alternative is ignorance.
2. The explanatory power ranges from reasonably good to pretty bad.

Predicting success or failure of a single firm is not a good use of this model.

Understanding why a firm might fail IS a good use of the model, as well as predicting how many firms might fail from a fairly large subsection of all firms. A manager just needs to be slightly better than everyone else to get above normal returns. Although, now that this information is very public, it might be hard to do that.
Haven't read the paper but if anyone's interested people do sometimes modify logit models for situations in which one of the two outcomes is "rare" . . .

http://ideas.repec.org/a/jss/jstsof/08i02.html

All best,
Pete

dumbmoney
Posts: 2304
Joined: Sun Mar 16, 2008 8:58 pm

Post by dumbmoney » Thu Feb 25, 2010 4:24 am

A bankrupt, delisted company (the GGP REIT) was recently bought out with value for the stock. But there were some sophisticated investors betting on the stock, so it was not a complete surprise.
I am pleased to report that the invisible forces of destruction have been unmasked, marking a turning point chapter when the fraudulent and speculative winds are cast into the inferno of extinction.

Post Reply