Recent Hurricanes Provide Good Lesson For Investors

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Random Walker
Posts: 1916
Joined: Fri Feb 23, 2007 8:21 pm

Recent Hurricanes Provide Good Lesson For Investors

Post by Random Walker » Sun Sep 10, 2017 12:04 pm

I was just reading some Hurricane headlines. Just read "This marks the first year on record the continental U.S. has had two Category 4 hurricane landfalls in the same year". This reminds me of Larry Swedroe's investing tenet that we should not treat the highly likely as certain or the highly unlikely as impossible.
The hurricanes also remind me of the value of diversification. I have 2% in the Stone Ridge Reinsurance fund. It's gotten hit hard by these rare events. Limiting the exposure to 2% in a diversified portfolio certainly minimizes the adverse impact on the portfolio. "Diversification is always working, whether you want it to or not." :-)

Dave

Dude2
Posts: 646
Joined: Fri Jun 08, 2007 3:40 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by Dude2 » Sun Sep 10, 2017 12:17 pm

There was a recent post looking for Livesoft which talked about three "500 year" floods that had occurred in the Houston area. Rather than become alarmist and start considering it the "end of times", I think we are better served in examining whether the data is itself valid (are themselves valid?). Many times we are reminded about the entire stock market history we have to look at amounts to only a couple of data sets statistically. I don't pretend to grasp that statement as I am not a wizz-bang statistician, but my intuition tells me it is probably highly accurate. The same can be said about global temperature and weather data. I find any data not collected within about the last 150 years to be highly dubious. Therefore, you really have a very small set of data from which to draw conclusions (like the stock market).

Valuethinker
Posts: 32637
Joined: Fri May 11, 2007 11:07 am

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by Valuethinker » Sun Sep 10, 2017 12:19 pm

[xxx edited ]
Last edited by Valuethinker on Sun Sep 10, 2017 12:35 pm, edited 1 time in total.

Valuethinker
Posts: 32637
Joined: Fri May 11, 2007 11:07 am

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by Valuethinker » Sun Sep 10, 2017 12:20 pm

Random Walker wrote:
Sun Sep 10, 2017 12:04 pm
I was just reading some Hurricane headlines. Just read "This marks the first year on record the continental U.S. has had two Category 4 hurricane landfalls in the same year". This reminds me of Larry Swedroe's investing tenet that we should not treat the highly likely as certain or the highly unlikely as impossible.
The hurricanes also remind me of the value of diversification. I have 2% in the Stone Ridge Reinsurance fund. It's gotten hit hard by these rare events. Limiting the exposure to 2% in a diversified portfolio certainly minimizes the adverse impact on the portfolio. "Diversification is always working, whether you want it to or not." :-)

Dave
And also of correlation.

After Lehman Brothers, most assets went down together. TIPS did so (for quite specific reasons regarding collateralization of commodity strategies, it appears). US Treasury Bonds did not. But just about everything else did.

Dude2
Posts: 646
Joined: Fri Jun 08, 2007 3:40 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by Dude2 » Sun Sep 10, 2017 12:24 pm

Valuethinker wrote:
Sun Sep 10, 2017 12:19 pm
Dude2 wrote:
Sun Sep 10, 2017 12:17 pm
There was a recent post looking for Livesoft which talked about three "500 year" floods that had occurred in the Houston area. Rather than become alarmist and start considering it the "end of times", I think we are better served in examining whether the data is itself valid (are themselves valid?). Many times we are reminded about the entire stock market history we have to look at amounts to only a couple of data sets statistically. I don't pretend to grasp that statement as I am not a wizz-bang statistician, but my intuition tells me it is probably highly accurate. The same can be said about global temperature and weather data. I find any data not collected within about the last 150 years to be highly dubious. Therefore, you really have a very small set of data from which to draw conclusions (like the stock market).
The proxy data is actually pretty good-- depending on the variable, takes you thousands of years.
Fair one, mate. I guess that also we have ways of taking core samples and melting glacial ice to see things. Still, if all of a sudden people started winning the lottery at a higher frequency than they should, I would question the odds-makers.

Random Walker
Posts: 1916
Joined: Fri Feb 23, 2007 8:21 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by Random Walker » Sun Sep 10, 2017 12:25 pm

I think it's really important to consider the possibility of alternative histories. Our investing decisions can't necessarily be judged on the outcomes because only one history played out. Decisions need to be evaluated based on when they were made, not knowing what the future would hold.
I think it's important for US investors to appreciate that the "V shaped" rebound from the 2008 financial crisis was not at all the only possible outcome and that what has happened in Japan for the last generation could happen here as well.

Dave

User avatar
nedsaid
Posts: 8258
Joined: Fri Nov 23, 2012 12:33 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by nedsaid » Sun Sep 10, 2017 12:38 pm

Random Walker wrote:
Sun Sep 10, 2017 12:04 pm
I was just reading some Hurricane headlines. Just read "This marks the first year on record the continental U.S. has had two Category 4 hurricane landfalls in the same year". This reminds me of Larry Swedroe's investing tenet that we should not treat the highly likely as certain or the highly unlikely as impossible.
The hurricanes also remind me of the value of diversification. I have 2% in the Stone Ridge Reinsurance fund. It's gotten hit hard by these rare events. Limiting the exposure to 2% in a diversified portfolio certainly minimizes the adverse impact on the portfolio. "Diversification is always working, whether you want it to or not." :-)

Dave
Yes, this is a good lesson that things can go wrong. There was quite a lively discussion about reinsurance that was like a pie fight in a bakery. The thing is, reinsurance was the next big thing and though there was discussion about the occasional downside, no one expected the downside to happen so quickly. It shows that we get too confident in ourselves and too confident in our investments. Reinsurance will probably be an excellent investment for the long term but the hurricanes show that there are always risks. Having 2% of this in a portfolio won't hurt you much, won't help you much, but you are gaining valuable investment experience.
A fool and his money are good for business.

Valuethinker
Posts: 32637
Joined: Fri May 11, 2007 11:07 am

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by Valuethinker » Sun Sep 10, 2017 12:40 pm

Dude2 wrote:
Sun Sep 10, 2017 12:24 pm

The proxy data is actually pretty good-- depending on the variable, takes you thousands of years.
Fair one, mate. I guess that also we have ways of taking core samples and melting glacial ice to see things. Still, if all of a sudden people started winning the lottery at a higher frequency than they should, I would question the odds-makers.
[/quote]

Hi I realized after I posted what you meant so I deleted my post.

What's definitely true is that if things change, our historical data won't catch it.

For example I think major stock market crises have become more common. In that I remember 1987 (the largest single day fall every recorded). But also (vaguely) 1972-74 in the UK (deepest bear market ever recorded in a country that did not have a war or revolution at the time). And 1990 (Saddam invades Kuwait on 1st August, surprising everyone). And then 1997-98 (Asia Crash, then Long Term Capital Management). And then 2000-03 (dot com) and then 2008-09 (Mother of all bear markets). In each case there was frantic intervention by the authorities to try to stop the market fall, but fall it did. (It's forgotten that in 1987 many of the Broker-Dealers were in severe financial straits in the days after the Crash, and the Fed stepped in and provided liquidity, and may have arranged rescue Lines of Credit; Alan Greenspan did the same whilst the resolution of LTCM was being arranged, cutting interest rates).

A 1 in 100 flood will be experienced about once every 30 years, that's something else most people don't realize.

Valuethinker
Posts: 32637
Joined: Fri May 11, 2007 11:07 am

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by Valuethinker » Sun Sep 10, 2017 12:41 pm

nedsaid wrote:
Sun Sep 10, 2017 12:38 pm
. Reinsurance will probably be an excellent investment for the long term but the hurricanes show that there are always risks. Having 2% of this in a portfolio won't hurt you much, won't help you much, but you are gaining valuable investment experience.
I think the hedge funds have gone into Reinsurance in a fairly big way. Until that pool of capital is depleted, returns are not likely to be as good as they were historically.

livesoft
Posts: 55897
Joined: Thu Mar 01, 2007 8:00 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by livesoft » Sun Sep 10, 2017 12:41 pm

Just think of all that emergency fund money going directly into consumption.
This signature message sponsored by sscritic: Learn to fish.

User avatar
nedsaid
Posts: 8258
Joined: Fri Nov 23, 2012 12:33 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by nedsaid » Sun Sep 10, 2017 12:47 pm

Valuethinker wrote:
Sun Sep 10, 2017 12:41 pm
nedsaid wrote:
Sun Sep 10, 2017 12:38 pm
. Reinsurance will probably be an excellent investment for the long term but the hurricanes show that there are always risks. Having 2% of this in a portfolio won't hurt you much, won't help you much, but you are gaining valuable investment experience.
I think the hedge funds have gone into Reinsurance in a fairly big way. Until that pool of capital is depleted, returns are not likely to be as good as they were historically.
This was one of the points made in the pie fight that I mentioned. Pointing this out didn't go over too well. The thing is, there are always points in favor and points against any investment. You are pointing to the "late to the party" argument, the best deals have been scooped up by the time the retail investors have a crack at it. I still think reinsurance will be a good investment but it will take patience, and yes, future returns are likely to be lower than historical returns. I recall hearing the same arguments about stocks and bonds.
A fool and his money are good for business.

heyyou
Posts: 2693
Joined: Tue Feb 20, 2007 4:58 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by heyyou » Sun Sep 10, 2017 1:37 pm

by Random Walker » Sun Sep 10, 2017 12:25 pm
I think it's really important to consider the possibility of alternative histories. Our investing decisions can't necessarily be judged on the outcomes because only one history played out. Decisions need to be evaluated based on when they were made, not knowing what the future would hold.
Well said. What is the behavioral psychology term for the risk of too much linear thinking? Why are we always seeking certainty, in spite of the recurring evidence that there isn't any?

jalbert
Posts: 2075
Joined: Fri Apr 10, 2015 12:29 am

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by jalbert » Sun Sep 10, 2017 1:56 pm

What is it about investing in a reinsurance fund that is better than investing in a reinsurance company? The latter have actuaries evaluating their overall risk profile of the aggregate business they take on-- does the stone ridge fund do the same?
Risk is not a guarantor of return.

Random Walker
Posts: 1916
Joined: Fri Feb 23, 2007 8:21 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by Random Walker » Sun Sep 10, 2017 5:07 pm

Jalbert,
If one were investing in a reinsurance company's stock, he would be investing in a stock. This involves taking on lots of market beta risk. The shares would be traded on the secondary market and exposed to all the same qualities that all other stocks are subject to. This is similar to the issue of investing in the stocks of commodity producers versus commodities themselves. Moreover reinsurance companies have their own portfolios with stocks in them, making them behave more like the market. My understanding is that an investment in the Stone Ridge Reinsurance Interval fund is basically directly becoming an insurer one's self. So there is no stock market exposure. The shares are not publicly traded on any exchanges.

Dave

User avatar
nisiprius
Advisory Board
Posts: 33778
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by nisiprius » Sun Sep 10, 2017 6:21 pm

The frequency of "500-year floods" is just another example of a uniform phenomenon: we are very bad at assessing the probability of rare events, and we seem to consistently underestimate their probability.

The Rasmussen Report estimated the magnitude of event that actually occurred at Three Mile Island as being, IIRC, a one-in-27,000-year event.

The principals of LTCM asserted that the event that brought down the fund were "a ten-sigma event." Statistically, that means it was (believed to be) a 500,000,000,000,000,000,000-year event.

A bizarre, but interesting recognition of the issue occurs in quality management, in which proponents of the "Sis Sigma" methodology assert that "Six Sigma" is equivalent to 3.4 defects per million. I heard this presented in a course and it started to nag at me, but it was years before I actually looked it up in a book of table and found that (as I suspected) six sigma ought to mean one defect per billion, and that 3.4 defects per million is actually the correct value for 4.5 sigma. The "explanation" for why Six Sigma practitioners do this is... weird, but the main point is that Six Sigma methodology does acknowledge that in the real world, in process that is supposed to be under "six sigma" control, the actual defect rate is 3,400 times as high as predicted. The reason seems to be that there is a rule of thumb that processes that seem to be controlled to six sigma in the short term "typically" have an extra variation of 1.5 sigma in the longer run... or something.

The Misbehavior of Markets: A Fractal View of Financial Turbulence, by Mandelbrot and Hudson, has a good deal to say about this--as do the books of Nassim Nicholas Taleb.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.

User avatar
TD2626
Posts: 335
Joined: Thu Mar 16, 2017 3:40 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by TD2626 » Sun Sep 10, 2017 8:28 pm

nedsaid wrote:
Sun Sep 10, 2017 12:38 pm
Random Walker wrote:
Sun Sep 10, 2017 12:04 pm
I was just reading some Hurricane headlines. Just read "This marks the first year on record the continental U.S. has had two Category 4 hurricane landfalls in the same year". This reminds me of Larry Swedroe's investing tenet that we should not treat the highly likely as certain or the highly unlikely as impossible.
The hurricanes also remind me of the value of diversification. I have 2% in the Stone Ridge Reinsurance fund. It's gotten hit hard by these rare events. Limiting the exposure to 2% in a diversified portfolio certainly minimizes the adverse impact on the portfolio. "Diversification is always working, whether you want it to or not." :-)

Dave
Yes, this is a good lesson that things can go wrong. There was quite a lively discussion about reinsurance that was like a pie fight in a bakery. The thing is, reinsurance was the next big thing and though there was discussion about the occasional downside, no one expected the downside to happen so quickly. It shows that we get too confident in ourselves and too confident in our investments. Reinsurance will probably be an excellent investment for the long term but the hurricanes show that there are always risks. Having 2% of this in a portfolio won't hurt you much, won't help you much, but you are gaining valuable investment experience.
Looking at this, I could see someone trying to explain their losses in one of two ways:
Option # 1. The reinsurance fund is doing exactly what it's supposed to do: having gains and losses that are independent of when the stock market has gains and losses. The low correlation is helping the portfolio. Standard deviation punishes having above average gains just as much as below average gains, so having losses in this fund during a stock bull market is good as it reduces volatility (if that's a goal).

Option # 2. Loosing money is bad. The mumbo-jumbo about correlations and such can't turn losses into gains. Larry Sweedroe may be brilliant, but those who put money in reinsurance, an unusual, unconventional, "alternative investment" lost a lot, and are learning a very expensive lesson.

Many would argue #1 is the "correct" way of describing things to a completely dispassionate, fully rational robot. Others, though, gravitate towards a more emotional explanation of #2.

Note that staying the course is something that is central to the Boglehead philosophy. One must not buy high and sell low for the kinds of strategies Sweedroe advocates to work. At some point, there'll be a large market crash in a year with few hurricanes - and the reinsurance could be a top performer that year. However, this could be decades away - (recall that it's possible that there could be a year with massive flooding and a stock crash).

We should also pause and note that discussing reinsurance fund implications of flooding is probably not the right thing to be doing at this time. The storm conditions are still ongoing and we should be focusing instead on those in need. Hopefully all forum members in affected areas are safe and sound.

GoldenFinch
Posts: 1300
Joined: Mon Nov 10, 2014 11:34 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by GoldenFinch » Sun Sep 10, 2017 8:39 pm

livesoft wrote:
Sun Sep 10, 2017 12:41 pm
Just think of all that emergency fund money going directly into consumption.
Yes, and all of the jobs created for rebuilding homes, buildings, infrastructure and landscaping.

User avatar
Dale_G
Posts: 2925
Joined: Tue Feb 20, 2007 5:43 pm
Location: Central Florida - on the grown up side of 80

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by Dale_G » Sun Sep 10, 2017 9:48 pm

Not that I want to profit from disasters, but my timber REIT is doing very well. Meanwhile I am listening to the wind howl due to Irma.

Take care folks.

Dale
Volatility is my friend

jalbert
Posts: 2075
Joined: Fri Apr 10, 2015 12:29 am

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by jalbert » Sun Sep 10, 2017 11:46 pm

The likely problem with the Stoneridge reinsurance fund is that the counterparties on the other side of its transactions are likely large insurance companies with large teams of very experienced actuaries. The Stoneridge prospectus does not inspire confidence in their ability to price these securities:
Risk-Modeling Risk. The Adviser, in selecting investments for the Fund, will generally consider risk models created by independent third parties, the sponsor of a reinsurance-related security or a broker. The Adviser may also consider its own risk models based on comparable prior transactions, quantitative analysis, and industry knowledge. Risk models are designed to assist investors, governments, and businesses understand the potential impact of a wide variety of catastrophic events and allow such parties to analyze the probability of loss in regions with the highest exposure. The Adviser will use the output of the risk models before and after investment to assist the Adviser in assessing the risk of a particular reinsurance-related security or a group of such securities. Risk models are created using historical, scientific and other related data, and they may use quantitative methods. Because such risk models are based in part upon historical data and averages, there is no guarantee that such information will accurately predict the future occurrence, location or severity of any particular catastrophic event and thus may fail to accurately calculate the probability of a trigger event and may underestimate the likelihood of a trigger event. Securities or other investments selected using quantitative methods may perform differently from the market as a whole or from their expected performance for many reasons, including factors used in building the quantitative analytical framework, the weights placed on each factor, and changing sources of market returns, among others. In addition, any errors or imperfections in a risk model (quantitative or otherwise), analyses, the data on which they are based or any technical issues with the construction of the models (including, for example, data problems and/or software or other implementation issues) could adversely affect the ability of the Adviser to use such analyses or models effectively, which in turn could adversely affect the Fund’s performance. Risk models are used by the Adviser as one input in its risk analysis process for Fund investments. There can be no assurance that these methodologies will help the Fund to achieve its investment objective.
Relying on the risk model of the issuer of the security or broker conducting the transaction or serving as market maker is a significant red flag for me-- would need to understand more than they disclose in the prospectus.

Also, the ER of the reinsurance fund is 2.13%.
Risk is not a guarantor of return.

james22
Posts: 1151
Joined: Tue Aug 21, 2007 2:22 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by james22 » Mon Sep 11, 2017 8:35 am

The real trouble with this world of ours is not that it is an unreasonable world, nor even that it is a reasonable one.

The commonest kind of trouble is that it is nearly reasonable, but not quite.

Life is not an illogicality, yet it is a trap for logicians.

It looks just a little more mathematical and regular than it is; its' exactitude is obvious; but its' inexactitude is hidden; its' wildness lies in wait.

G. K. Chesterton
This whole episode is likely to end so badly that future children will learn about it in school and shake their heads in wonder at the rank stupidity of it all... Hussman

Random Walker
Posts: 1916
Joined: Fri Feb 23, 2007 8:21 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by Random Walker » Mon Sep 11, 2017 8:40 am

The real trouble with this world of ours is not that it is an unreasonable world, nor even that it is a reasonable one.

The commonest kind of trouble is that it is nearly reasonable, but not quite.

Life is not an illogicality, yet it is a trap for logicians.

It looks just a little more mathematical and regular than it is; its' exactitude is obvious; but its' inexactitude is hidden; its' wildness lies in wait.

G. K. Chesterton
Good reason to diversify across as many sources of return as possible!

Dave

jalbert
Posts: 2075
Joined: Fri Apr 10, 2015 12:29 am

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by jalbert » Mon Sep 11, 2017 12:20 pm

Moreover reinsurance companies have their own portfolios with stocks in them, making them behave more like the market. My understanding is that an investment in the Stone Ridge Reinsurance Interval fund is basically directly becoming an insurer one's self. So there is no stock market exposure.
You would be a reinsurer with unfunded actuarial liabilities (or funded by the rest of the portfolio). If you tilt a stock market portfolio away from the market index portfolio and towards reinsurance stocks you are replacing beta with reinsurance exposure diluted with beta. I think you can accomplish the same thing but you will just need a little more of it to get the same reinsurance exposure, and there won't be a 2.13% ER.
Risk is not a guarantor of return.

garlandwhizzer
Posts: 1559
Joined: Fri Aug 06, 2010 3:42 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by garlandwhizzer » Mon Sep 11, 2017 2:49 pm

It has been suggested in the past that reinsurance sits at some unique place on the risk/reward spectrum with bond-like security and equity-like return. It may well have a unique risk/reward profile, not really correlating with the stock market while offering higher expected returns than the bond market. Bonds, however, are primarily for safety. Although reinsurance does not correlate closely with equity it clearly has risk of precipitous declines in value over a short period of time. The lack of correlation with stocks was counterproductive recently, as the stock market went up in the face of massive weather disasters while reinsurance took a big hit. Recent losses in reinsurance suggest that if one chooses to invest there, it should be considered equity risk rather than bond risk in the overall portfolio.

Garland Whizzer

Random Walker
Posts: 1916
Joined: Fri Feb 23, 2007 8:21 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by Random Walker » Mon Sep 11, 2017 3:39 pm

My impression is that reinsurance has expected equity like returns with expected SD of about 10. Bonds have SD about 5 and equities SD about 20. Insurance does have big left tail risk, but diversification across different types of insurance risks, each with its own left tail, should minimize the left tail of the overall reinsurance portfolio.
I agree that the safest asset is bonds. But in the effort to create a more efficient portfolio, one can take from either the bond side or the equity side to make the reinsurance position. If one takes from the equity side, he is keeping expected return the same, but doing so by adding a noncorrelated asset with lower SD. If one takes from the bond side, he is increasing expected return, adding an asset that is not correlated to either stocks or bonds, increasing expected portfolio SD. In both cases, portfolio efficiency, Sharpe ratio, is expected to increase.

Dave

jalbert
Posts: 2075
Joined: Fri Apr 10, 2015 12:29 am

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by jalbert » Mon Sep 11, 2017 10:56 pm

My impression is that reinsurance has expected equity like returns with expected SD of about 10.
The securitized insurance market has only been around since 2001 the 1990s and more or less completely collapsed in 2008. I don't think a large enough sample is possible to get a reasonable estimate for standard deviation of return.

Expected SD is not really a probabilistic or statistical concept. There is the standard deviation of the actual probability distribution of return (which is usually not known for asset returns) and there is sample variance or sample standard deviation from a sample of actual outcomes. The latter may or may not be a good estimator for the former.

You could design a random variable defined to be the sample standard deviation of a time series of returns and talk about the expected value of it, but this is a different random variable with a different distribution than the random variable of return.
Risk is not a guarantor of return.

TravelforFun
Posts: 720
Joined: Tue Dec 04, 2012 11:05 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by TravelforFun » Tue Sep 12, 2017 12:33 am

Recent hurricanes also provide great investment opportunities. Those who bought USG, the company who makes drywall, a few days after Harvey hit is making a nice profit. Home Depot, Lowes, car companies will do well as well.

Valuethinker
Posts: 32637
Joined: Fri May 11, 2007 11:07 am

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by Valuethinker » Tue Sep 12, 2017 3:38 am

nisiprius wrote:
Sun Sep 10, 2017 6:21 pm
The frequency of "500-year floods" is just another example of a uniform phenomenon: we are very bad at assessing the probability of rare events, and we seem to consistently underestimate their probability.

The Rasmussen Report estimated the magnitude of event that actually occurred at Three Mile Island as being, IIRC, a one-in-27,000-year event.
So then Chernobyl and Fukushima didn't happen? ;-).

Browns Ferry, and Davis Besse (no actual accident, but the shielding had been eaten away) were of course not serious by those standards, but they could have been serious.
The principals of LTCM asserted that the event that brought down the fund were "a ten-sigma event." Statistically, that means it was (believed to be) a 500,000,000,000,000,000,000-year event.
Reflexivity to use the Soros term. In other words, what LTCM ignored was its own impact on the market.

Portfolio Insurance, which played a major role in the 1987 Crash, had a similar problem-- the assumption of continuous trading was violated.

With the 2008 Crash, there was the "Gaussian Copula" which was used to estimate the joint probability of default of the mortgages in bonds. Of course it turned out to be incorrect in practice, due to the systemic effects of the cheap mortgage finance the bonds enabled.
A bizarre, but interesting recognition of the issue occurs in quality management, in which proponents of the "Sis Sigma" methodology assert that "Six Sigma" is equivalent to 3.4 defects per million. I heard this presented in a course and it started to nag at me, but it was years before I actually looked it up in a book of table and found that (as I suspected) six sigma ought to mean one defect per billion, and that 3.4 defects per million is actually the correct value for 4.5 sigma. The "explanation" for why Six Sigma practitioners do this is... weird, but the main point is that Six Sigma methodology does acknowledge that in the real world, in process that is supposed to be under "six sigma" control, the actual defect rate is 3,400 times as high as predicted. The reason seems to be that there is a rule of thumb that processes that seem to be controlled to six sigma in the short term "typically" have an extra variation of 1.5 sigma in the longer run... or something.

The Misbehavior of Markets: A Fractal View of Financial Turbulence, by Mandelbrot and Hudson, has a good deal to say about this--as do the books of Nassim Nicholas Taleb.
Interesting and thank you.

mmcmonster
Posts: 153
Joined: Sun Jan 12, 2014 1:18 pm

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by mmcmonster » Tue Sep 12, 2017 5:05 am

Getting back to Houston and the '500 year flood' statistic. From what I understand, that number just means that it's felt that in any particular year that there is a 1/500 chance of getting a flood.

In other words, there is a 499/500 chance of not getting a flood in any year. In any hundred year period there is a (499/500)^100 chance of not getting a flood. That's 82%.

So a 500 year flood has a 18% chance of happening in any 100 year period. And Houston has been a city for ~180 years.

Then you have to get back to the data and determine both the assumptions made in it and how old the data is.

Makes me nervous building anything not on high ground. :confused

Valuethinker
Posts: 32637
Joined: Fri May 11, 2007 11:07 am

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by Valuethinker » Tue Sep 12, 2017 5:35 am

mmcmonster wrote:
Tue Sep 12, 2017 5:05 am
Getting back to Houston and the '500 year flood' statistic. From what I understand, that number just means that it's felt that in any particular year that there is a 1/500 chance of getting a flood.

In other words, there is a 499/500 chance of not getting a flood in any year. In any hundred year period there is a (499/500)^100 chance of not getting a flood. That's 82%.

So a 500 year flood has a 18% chance of happening in any 100 year period. And Houston has been a city for ~180 years.

Then you have to get back to the data and determine both the assumptions made in it and how old the data is.

Makes me nervous building anything not on high ground. :confused
I believe that you are correct in your probability calculations, and your correct surmise that most people hear "1 in 100" and think it will happen once every 100 years. One should understand that with Houston, there are few (any?) cities in the developed world that could take that kind of rainfall in that time frame, without significant flooding -- and that would hit people on high ground (but perhaps local minima) as well. It's not just about your position relative to sea level, unfortunately.

For those living near the coast, though, this problem of storm surges is very real.

A related problem is that the underlying frequency and/or magnitude of these events appears to be increasing. At least, there is broad agreement on the magnitude/ intensity. On frequency, for Atlantic hurricanes Kerry Emanuel thinks they are, and he's a world-recognized expert. But not everyone agrees with him, by any means.

https://emanuel.mit.edu/publications-0

The general problem is the assumption that distributions of outcomes are stable, when in fact the parameters are changing all the time.


jalbert
Posts: 2075
Joined: Fri Apr 10, 2015 12:29 am

Re: Recent Hurricanes Provide Good Lesson For Investors

Post by jalbert » Wed Sep 13, 2017 2:42 am

I think the main problem with getting reinsurance exposure through reinsurance companies is that many of them are large conglomerates with a variety of business lines beyond reinsurance.

Here is some data about the CAT bond market in 2008/2009. Credit spreads on CAT bonds did not widen as much as credit spreads for junk bonds, but they widened dramatically, despite no triggering catastrophic event. Of course this means the price of CAT bonds fell significantly. It was shifted a little beyond the Lehman bankruptcy in calendar time, but it Is not clear that the diversification worked in 2008/2009. See Exhibit 1 on page 3:

https://partnerre.com/wp-content/upload ... ricing.pdf

This may have been the effect of liquidity in the secondary CAT bond market drying up from insurers deleveraging-- still a major concern though.

Inflation is another concern I have, perhaps unjustified. Because much casualty insurance pays for replacement cost, premiums are nominal cash flows while claims are real cash flows. Will CAT bonds hold up if there is a spike in inflation?
Risk is not a guarantor of return.

Post Reply