Bill Miller is us­ing earth­quake fore­cast­ing to pre­dict the stock mar­ket’s next Big One

▶ He’s us­ing earth­quake mod­el­ing tools to fore­cast mar­ket risk ▶ “I’m sur­prised some­body is still mess­ing with that stuff”

Bloomberg Businessweek (North America) - - Contents -

When pitch­ing in­vestors, you nor­mally don’t want “hedge fund” and “earth­quake” in the same sen­tence. But Bill Miller, a once-dom­i­nant money man­ager known for quirky think­ing, is start­ing a fund that will make bets based in part on a com­puter model de­signed to pre­dict nat­u­ral dis­as­ters.

Called Seis­mic Value Part­ners 1, the fund marks Miller’s first foray into hedge funds af­ter decades man­ag­ing mu­tual funds at Legg Ma­son. He won ap­proval last month from the Se­cu­ri­ties and Ex­change Com­mis­sion to open Miller Value Part­ners, a money man­age­ment firm that will over­see his hedge funds. As of a De­cem­ber SEC fil­ing, no money had been raised for the ini­tial Seis­mic fund.

Miller has li­censed a com­puter model from Open­haz­ards Group, a Davis, Calif., com­pany run by en­gi­neers, math­e­ma­ti­cians, sci­en­tists, and busi­ness ex­ec­u­tives. The idea is to ap­ply the math­e­mat­ics of fore­cast­ing the prob­a­bil­ity of seis­mic ac­tiv­ity to the chances of a stock mar­ket crash.

Like se­vere drops in the ex­changes, “earth­quakes are ex­treme events,” says Lisa Ra­puano, who helped Miller run the Legg Ma­son Spe­cial In­vest­ment Trust mu­tual fund dur­ing the 1990s. “This has the po­ten­tial to help Bill avoid big mis­takes, which is one of the down­sides of his in­vest­ment style.”

Miller de­clined to com­ment. A decade ago, he was among the best­known and most-ad­mired stock fund man­agers in the world, thanks to an as­ton­ish­ing win­ning streak. His Legg Ma­son Value Trust fund beat the Stan­dard & Poor’s 500-stock in­dex for 15 con­sec­u­tive years, from 1991 through 2005. By 2008, when the cri­sis hit, the fund was heav­ily in­vested in fi­nan­cials. It lost 55 per­cent that year, vs. a 37 per­cent loss for the S&P 500.

The no­tion of sidestep­ping a big loss like that holds an ob­vi­ous ap­peal. But some schol­ars dis­miss the idea of ap­ply­ing an earth­quake model to stock­pick­ing. “I don’t think it has did­dly to do with fi­nan­cial mar­kets,” says Joseph Mccauley, a physics pro­fes­sor at the Univer­sity of Hous­ton and au­thor of Dy­nam­ics of Mar­kets: Econo­physics and Fi­nance. “I’m sur­prised some­body is still mess­ing with that stuff.”

Open­haz­ards Chair­man John Run­dle, a well-known seis­mol­o­gist who teaches physics and ge­ol­ogy at the Univer­sity of Cal­i­for­nia at Davis, has been adapt­ing his the­o­ries on nat­u­ral dis­as­ters and fi­nan­cial mar­kets with help from mem­bers of the Santa Fe In­sti­tute, a non­profit re­search or­ga­ni­za­tion where Miller is chair­man emer­i­tus and the largest donor.

Miller’s fund will use sig­nals from the model to make daily de­ter­mi­na­tions on whether to pur­chase or sell short se­cu­ri­ties that mir­ror the per­for­mance of the U.S. stock mar­ket, such as the SPDR S&P 500 ex­change-traded fund, ac­cord­ing to reg­u­la­tory doc­u­ments. The goal is to out­per­form the S&P 500 over pe­ri­ods of a year or more

with less volatil­ity. Miller will likely em­ploy the model as an “over­lay” to in­di­vid­ual stock picks, Ra­puano says.

The con­cept that seem­ingly un­re­lated catas­tro­phes, from ex­tinc­tions to avalanches to mar­ket crashes, might fol­low sim­i­lar un­der­ly­ing pat­terns dates to the 1980s, when the late Dan­ish physi­cist Per Bak be­gan writ­ing on the sub­ject. “It’s some­thing ev­ery­one would like to do, but it is in­her­ently close to im­pos­si­ble,” Gene Stan­ley, a physics pro­fes­sor at Bos­ton Univer­sity, says about trad­ing stocks based on such mod­els. Know­ing the prob­a­bil­ity of a crash is not the same as know­ing the tim­ing.

Miller has long been known for his un­usual ap­proach to in­vest­ing. At Value Trust he looked for stocks trad­ing at a dis­count to their in­trin­sic worth, a clas­sic value in­vest­ment ap­proach. But he merged that with other ideas. He was an in­vestor in Ama­zon.com at a time when most value man­agers thought the com­pany was over­priced us­ing stan­dard growth ex­pec­ta­tions. But he and Ra­puano de­vel­oped a model for how In­ter­net busi­nesses might grow based on ideas they had picked up via the Santa Fe In­sti­tute, Ra­puano says.

Af­ter 2008, the Value Trust fund trailed the in­dex in two of the three en­su­ing years. Miller stepped down as its man­ager in April 2012.

He re­bounded run­ning the $1.3 bil­lion Legg Ma­son Op­por­tu­nity Trust, gen­er­at­ing av­er­age an­nual re­turns of 26 per­cent from 2012 through 2015. In 2016, how­ever, he’s suf­fered, los­ing more than 18 per­cent so far, com­pared with a 5.3 per­cent loss for the S&P 500.

Miller runs Op­por­tu­nity Trust through LMM, a Bal­ti­more-based in­vest­ment ad­viser he owns with his son and Legg Ma­son. A Legg Ma­son spokes­woman says Miller’s ar­range­ment with LMM is un­changed.

John Seo, co-founder of Fer­mat Cap­i­tal Man­age­ment, which buys in­sur­ance com­pany bonds with pay­offs linked to catas­tro­phes, says he can see a broader use of earth­quake mod­el­ing. While traders of­ten have ex­ten­sive data on how their port­fo­lios would per­form in a mar­ket crash, they sel­dom know how likely such a down­draft would be. “That’s where the earth­quake guys come in,” Seo says. “They’re good at as­sign­ing prob­a­bil­i­ties.” �Miles Weiss The bot­tom line A fund man­ager fa­mous for big wins and a huge loss has a com­pli­cated idea for try­ing to take crashes out of the equa­tion.

Newspapers in English

Newspapers from Canada

© PressReader. All rights reserved.