Dan Loeb is the founder and head of Third Point, one of the world's largest hedge funds. Loeb became famous during the dotcom bust for writing a series of inflammatory letters to corporate executives he disliked, but his epistolary career began even earlier than that. From 1997 to 2003, he was a prolific poster on the Silicon Investor message boards, where he used the screen name "Mr. Pink."
Loeb posted more than 5,000 grammatically adventurous messages on Silicon Investor. There he referred to himself in the third person, usually as Mr. Pink, but also as "the Pinkster" and "PINK DADDY." As his audience grew, he became increasingly messianic, peppering his posts with references to "He," "Him," and "His flock."
I took a day off from my usual investing research to read his Silicon Investor thread, reading hundreds of his posts and skimming a bunch more. It was time-consuming, but it was worth it because the thread offers a rare window into the past that's free of hindsight bias.
Loeb’s investing strategy
During the late '90s and early '00s, Loeb primarily invested in small-caps. His mainstay was buying the kinds of special situations that Joel Greenblatt recommended in You Can Be a Stock Market Genius: spinoffs, demutualizations, post-reorgs, etc. He also shorted small-cap frauds.
These shorts were quite successful: even during the bull market from 1997-2000, Loeb identified many companies that fell precipitously. Chromatics Color Sciences, which was trying to develop a treatment for jaundice, doubled after he first mentioned it but ultimately went to zero. Agribiotech, a seed distributor that was being hyped as a biotech, went bankrupt within two years. Sirrom, an overvalued business development company with poor loan quality, blew up.
Actrade, a commercial-lending fraud, was Loeb's most notable short. He discovered that Actrade's Israeli-born CEO had fled to the U.S. to evade his creditors from a previous failed business, which the company hadn't disclosed in its SEC filings. Actrade went bankrupt in 2002, several years after Loeb first recommended shorting it.
Loeb had only one significant miss: in 2003 he recommended shorting American Pharmaceutical Partners because he thought its CEO was shady. The stock quadrupled in the following two years.
By contrast, Loeb's special-situations investments-- his bread and butter-- performed surprisingly poorly. There were some big winners, like First Sierra, Fingerhut, and Lennar, which all doubled in less than a year, and Agribrands, which doubled in two years. But by my count, his losers outnumbered them.
As many as half of the special situations he recommend eventually went bankrupt or ran into other serious problems: Cendant (accounting scandal), Consolidated Freightways (bankrupt), Fairchild Corporation (bankrupt), Federal Mogul (bankrupt), FPIC (doubled, then fell 90%), Hayes-Lemmerz (bankrupt), Loewen (bankrupt), Reliance Acceptance (bankrupt), Solutia (bankrupt), Ventas (fell 75% within three months), and Visteon (bankrupt).
Loeb's occasional macro calls also fared poorly. He thought Brazil would collapse in 1999 the way Russia had the previous year; instead its stock market doubled. He aggressively went short near the US market's lows in 2002 and lost money that year.
Playing the internet bubble
Loeb initially misjudged the internet bubble but eventually made money off it. He shorted Compaq in 1997 before it doubled. Later that year, when Amazon.com was trading at split-adjusted $4, he told another poster, "You are very wise to be short AMZN. Mr. pink could not obtain a borrow in sufficient size so could not do it."
He expressed similar sentiments about AOL in 1998, right before its speculative blowoff: "AOL is the worst offender of accounting shenanigans. It takes balls but is a good short. Mr. pink has no position yet, but it stinks."
But by the end of the year he'd changed his opinion: "Thou shalt not short internet stocks.... It is impossible to pick the top of a bubble."
Around that time, he began buying small-caps that had low P/E ratios and hidden internet subsidiaries, assuming they would rise when day-traders discovered the internet connection. He did this successfully with four or five different stocks, making quick 30-50% profits on each one.
Playing subprime lenders
As with internet stocks, Loeb played subprime lenders as both a long and a short. His subprime investments were generally profitable.
In 1997 he lost money buying Reliance Acceptance, a subprime-lending spinoff that went bankrupt soon after he recommended it.
In 1998 he shorted FirstPlus, which had grown quickly by making 125% LTV home-equity loans. FirstPlus collapsed after the LTCM crisis and went bankrupt in early 1999. A Businessweek article gives a sense of the company's lending practices:
Gene T. O'Bryan, former president of FirstPlus's wholesale lending division, who left in March, 1997, says in some cases, mortgage brokers who sold to FirstPlus simply made a series of ever larger loans to the same customer. And by paying off the old debt, borrowers who were going deeper into hock managed to improve their credit ratings. FirstPlus barred such refinancing, but "nobody checked," he said.
Loeb recommended shorting Long Beach Financial in late '98. Washington Mutual acquired LBF at a 100% premium the next year. Regarding subprime credit-card issuer Providian, he told another poster to "Short it with impunity" around the same time. Like Long Beach, it doubled the following year.
He ultimately made money on Providian when it nearly collapsed in 2001. He also made money shorting Conseco, Metris and Americredit in 2001 and 2002.
Skill versus environment
Loeb earned 40-50% annual returns in the late 1990s when his fund had less than $100mm in assets under management. I think this success was a result of size and environment in addition to skill. For many individual companies his analysis was wrong, as demonstrated by the high failure rate of his long picks, but he was still able to earn great returns because he was exploiting a couple of niches-- special situations and small-cap frauds-- that were overlooked and mis-priced in the aggregate.