return to ICG Spaces home    ICG Risk Blog    discussions    newsletters    login    

ICG Risk Blog - [ Bets are off or bets change: implications of risk curve flattening, rising probabilities of the tails ]

Bets are off or bets change: implications of risk curve flattening, rising probabilities of the tails

  #

"In a world where there's a possibility of things going extremely well or extremely badly, the tails of the bell curve are much fatter (and that can) lead to different behaviors. There's the idea that the singularity would be very successful, (causing) the biggest boom ever, or blow up the world, with nothing to invest in ever." [Peter Thiel]

Thiel's comment at the Singularity Summit on the progress of Artificial Intelligence (AI) pulled me up short. My first thought was 'all bets are off' or at least 'the bets have shifted' in ways few understand. In conditions where possible outcomes are increasingly polar, and their costs/benefits are exponentially larger, the traditional bell curve distribution flattens as its tails rise, i.e., the long tail is on its way to becoming the fat tail. If few genuinely understand statistics and probabilities today and pay the price, the future is going to create radically lopsided winners and losers.

Yes, Thiel was talking about the implications of the Singularity, the "technological creation of smarter-than-human intelligence," or "the point at which advances in artificial intelligence will bring about self-improving machines that are smarter than humans," but his comment applies equally to other fields. While AI has the nod in the race to the Singularity, the current runners-up are "direct brain-computer interfaces, biological augmentation of the brain, genetic engineering, ultra-high-resolution scans of the brain followed by computer emulation":

Some of these technologies seem likely to arrive much earlier than the others, but there are nonetheless several independent technologies all heading in the direction of the Singularity several different technologies which, if they reached a threshold level of sophistication, would enable the creation of smarter-than-human intelligence.

I submit that we don't even have to reach, or even near, the singularity to create this fat tail risk condition - and to be unaware of a fat tail climax event until it underway. Consider the complexity of embedding the products and services of many current fields into systems that are beyond the risk calculus of its users and consumers to understand. Witness the radiating meltdown in the hedge market and banking community over the securitization of subprime mortgages (here and here). All done with technologies known at least to the quants, but which escaped the understanding and control of the quants, sellers, purchasers, regulators and (collateral damage) bystanders.

And why not cascades of fat tail events where one exceptional event exceeds the program constraints or assumptions in another system and causes it to fault? Witness the recent failure of program trading "when stocks started moving not only in ways that commonly used models didn't predict, but in precisely the opposite direction from what was expected." (Continued here and here). While the description of such a Complex Adaptive System (CAS) can be somewhat dry, I particularly like this trader's description:

Complex adaptive systems are typically composed of diverse sets of many agents that are programmed to accomplish some goal. The interaction of these agents can make the population smarter than any individual agent. By emulating biological systems (e.g. - reproduction and gene mutation), such a system can "evolve" to produce emergent, very interesting behavior… each agent will act as a trader following some trading rule that can evolve with time. Agents that are good make money and survive. Agents that don't die out. It's survival of the fittest.

We, unfortunately, may not number amongst the fittest. A bioethicist, Wendell Wallach, speaking at the Singularity Summit made this prediction:

"We are just a few years away from a catastrophic disaster from an autonomous computer system making a decision." It will elicit a response similar to 9/11, [9/11 costs] he added. "We should not underestimate the political consequence of fear [of AI systems]." This kind of event wouldn’t likely stop scientific research, but it could slow it down, Wallach said.

"We need a mechanism for evaluating when thresholds that hold dangers have or are about to be crossed, and helping public policy leaders and the public at large to discrimatine real challenges from highly speculative challenges," Wallach said.

Should this error cause significant death to foreign nationals, as opposed to our own, it could be a casus belli for hostilities. If it is the malfunction of Fail-Safe or the doomsday device of Dr. Strangelove, the system itself will effect the war. Given my understanding of command and control systems, this autonomous computer system could already be in place, needing only a future 'upgrade' or an unforeseen series of events and/or interactions to cause that catastrophe.

Enter insurance, reinsurance and exotic risk management

Thiel's investment advice in the singularity - and by my extension to near-singular events - is catastrophic insurance which was said to be "the strategy, however inadvertent," of Warren Buffett. From Thiel's viewpoint, as one progresses "along the bell curve of plausible outcomes, the most likely scenarios eventually migrate toward the tails: very wonderful or very catastrophic. And of these two possible outcomes, either is acceptable to investors":

On the "wonderful" hand, the run-up to AGI creates the biggest investment boom in the history of humanity, and the positive results yield a world of never-ending promise. On the "catastrophic" hand, advanced AGI gone awry causes humanity to disappear in a cloud of [insert your favorite existential risk scenario here] dust. In that scenario, the investor has bigger problems than the lack of return anyway...

Thiel believes [Buffett's shift to insurance] is a classic adoption of the wonderful vs. catastrophic philosophy of Singularity investors. For Buffet, Thiel sees 4 possible outcomes for his investments:

  1. Nothing bad happens: Buffet happily collects the premiums from his policy holders and maybe even lives to 1000.
  2. A Mild disaster like 9/11 occurs: This helps the insurance industry in that it allows them to raise rates and drives more policy purchases.
  3. A big catastrophe happens [Thiel showed a rendering of nukes exploding over Manhattan]: In this scenario, the rules get changed. The Government steps in and makes things whole for the insurance industry.
  4. [Again, insert your favorite Existential Risk scenario] – No one is around to collect.

I thought Buffett's approach to be more financially driven than propelled by singularity. Buffett appears to employ an expanded value investment strategy pioneered by Benjamin Graham:

  • Fundamentals: undervalued securities possessing margin of safety to meet expected return-to-risk criteria
  • Arbitrage events: M&A, liquidations, et al, unrelated to market shifts
  • Controlling positions: unilateral, allied or proxy positions to force change
  • Wide moats: quality firms with significant barriers to competitive entry

I was under the impression that Buffett entered the insurance market, and subsequently reinsurance (or catastrophic insurance), for its cash flow and the float. (For me, reinsurance (re, cat) was merely a bookmaker's laying off of the bet.):

[Float] is money we hold but don't own. In an insurance operation, float arises because premiums are received before losses are paid, an interval that sometimes extends over many years. During that time, the insurer invests the money. Typically, this pleasant activity carries with it a downside: The premiums that an insurer takes in usually do not cover the losses and expenses it eventually must pay. That leaves it running an "underwriting loss," which is the cost of float. An insurance business has value if its cost of float over time is less than the cost the company would otherwise incur to obtain funds. But the business is a lemon if its cost of float is higher than market rates for money.

A caution is appropriate here: Because loss costs must be estimated, insurers have enormous latitude in figuring their underwriting results, and that makes it very difficult for investors to calculate a company's true cost of float. Estimating errors, usually innocent but sometimes not, can be huge. The consequences of these miscalculations flow directly into earnings. An experienced observer can usually detect large-scale errors in reserving, but the general public can typically do no more than accept what's presented, and at times I have been amazed by the numbers that big-name auditors have implicitly blessed.

Buffett, float and super-cat insurance

Presumably the greater the risks, the greater the premiums and the float, but I wondered how Buffett saw matters. Indicative of Buffett, Berkshire Hathaway issues what are the benchmark for clear, plain-spoken summaries of operation. They are worth the read; other financial statement are wanting in comparison. The 2006 Chairman's Letter notes the "40th anniversary of our entrance into the insurance business" where it has become dominate among the operating sectors of Berkshire. My amateur review indicates great attention to float and super-cat performance:

At the end of 2006, our float had grown to $50.9 billion, and we have since written a huge retroactive reinsurance contract with Equitas [that] boosts float by another $7 billion. Much of the gain we’ve made has come through our acquisition of other insurers, but we’ve also had outstanding internal growth, particularly at Ajit Jain’s amazing reinsurance operation…

Whatever [the size of the float], the all-important cost of Berkshire’s float over time is likely to be significantly below that of the industry, perhaps even falling to less than zero. Note the words "over time." There will be bad years periodically. You can be sure of that.

In 2006, though, everything went right in insurance really right. Our managers… simply shot the lights out… Of course, the overall insurance industry also had a terrific year in 2006. But our managers delivered results generally superior to those of their competitors…

In 2007, our results from the bread-and-butter lines of insurance will deteriorate, though I think they will remain satisfactory. The big unknown is super-cat insurance. Were the terrible hurricane seasons of 2004-05 aberrations? Or were they our planet’s first warning that the climate of the 21st Century will differ materially from what we’ve seen in the past? If the answer to the second question is yes, 2006 will soon be perceived as a misleading period of calm preceding a series of devastating storms. These could rock the insurance industry. It’s naïve to think of Katrina as anything close to a worst-case event.

Neither Ajit Jain, who manages our super-cat operation, nor I know what lies ahead. We do know that it would be a huge mistake to bet that evolving atmospheric changes are benign in their implications for insurers.

Don’t think, however, that we have lost our taste for risk. We remain prepared to lose $6 billion in a single event, if we have been paid appropriately for assuming that risk. We are not willing, though, to take on even very small exposures at prices that don’t reflect our evaluation of loss probabilities. Appropriate prices don’t guarantee profits in any given year, but inappropriate prices most certainly guarantee eventual losses. Rates have recently fallen because a flood of capital has entered the super-cat field. We have therefore sharply reduced our wind exposures. Our behavior here parallels that which we employ in financial markets: Be fearful when others are greedy, and be greedy when others are fearful.

My takeaway

Buffett is paying the closest attention to super-cat performance while maintaining float and minimizing float cost. Implicit in that position is aggressive risk analysis. I presume Buffett's attention to fat tail risk will manifest itself in the pricing of risk assumption in each of its major insurance sectors (General Re, B-H Reinsurance, GEICO et al).

Firms that do not define and factor their risks appropriately for an increasingly complex and less predictable environment will find themselves at an operational disadvantage, will find risk coverage unaffordable, or both.

What is the Singularity?
Why Work Toward the Singularity?
Singularity Institute for Artificial Intelligence (SIAI)

Abstracts
Singularity Summit 2007

Is Warren Buffet Betting on The Singularity?
posted by Jonas Lamis
Singularity News
September 10, 2007

Facebook backer Thiel's investment strategy for singularity
Posted by Stefanie Olsen
CNET
September 9, 2007 12:42 PM PDT

Google's director of research talks AI
Posted by Stefanie Olsen

CNET
September 9, 2007 11:43 AM PDT

A call for machine morality
Posted by Stefanie Olsen
CNET
September 8, 2007 2:13 PM PDT

Liveblogging Singularity Summit 2007
Michael Anissimov 10:12 am
Accelerating Future
Saturday, Sep 8 2007

Singularity Summit Talk: Openness and the Metaverse Singularity
Posted by Jamais Cascio
Open the Future
September 8, 2007 2:00 PM

Counting the cost of the home loan crisis
By Greg Wood
BBC Radio 4 Today programme
Last Updated: Thursday, 30 August 2007, 11:10 GMT 12:10 UK

Ten Reasons
Michael Anissimov 8:35 am
Accelerating Future
Saturday, Aug 18 2007

Q&A: World stock market falls
BBC News
Last Updated: Thursday, 16 August 2007, 08:15 GMT 09:15 UK

Quant "Bloodbath" - The continuation
CASTrader Blog
August 16, 2007

Quant "Bloodbath" - The lessons learned
CASTrader Blog
August 12, 2007

One 'Quant' Sees Shakeout For the Ages -- '10,000 Years'
By KAJA WHITEHOUSE
Wall Street Journal
August 11, 2007

Second Earth
The World Wide Web will soon be absorbed into the World Wide Sim: an environment combining elements of Second Life and Google Earth.
By Wade Roush
Technology Review
July/August 2007
Contains additional content not published in the print edition. Print version
here.

Metaverse Roadmap Report
Posted by Jamais Cascio
Open the Future
June 25, 2007 3:26 PM

Metaverse Roadmap Overview (PDF)
Pathways to the 3D Web
A Cross-Industry Public Foresight Project
John Smart, Jamais Cascio, Jerry Paffendorf
Acceleration Studies Foundation
2007

Peter Thiel on Bloomberg.
Marco Hunter
Land of Black Gold
Sunday, April 15, 2007

2006 Chairman's Letter
Warren Buffett
Berkshire Hathaway
2007

Previous shareholder letters

Peter Thiel says be careful out there.
Marco Hunter
Land of Black Gold
Posted on 11/08/2006 08:42 AM

Twelve Things Journalists Need To Know to be Good Futurist/Foresight Reporters
Jamais Cascio
Open the Future
2006

Originally published (in somewhat modified form) June 14, 2006

Peter Thiel - Tar Baby.
Marco Hunter
Land of Black Gold
March 21, 2005

Economic Costs to the United States Stemming From the 9/11 Attacks
by Robert Looney
Strategic Insights
Volume I, Issue 6 (August 2002)

Question 5: What is float and how does it impact Berkshire?
Berkshire Hathaway Frequently Asked Questions (FAQ)
Source: 1997 Berkshire Hathaway Annual Report

Gordon Housworth



InfoT Public  Risk Containment and Pricing Public  Strategic Risk Public  

discussion

  discuss this article


<<  |  August 2019  |  >>
SunMonTueWedThuFriSat
28293031123
45678910
11121314151617
18192021222324
25262728293031
1234567
view our rss feed