Thursday, January 26, 2006

Prem Watsa & Fairfax Financial

  
The Globe and Mail, John Daly, 26 January 2006

In this war, there are no prisoners, no truces, no neutral parties. Few companies inspire deeper loyalty—and more bitter vitriol—than Fairfax Financial Holdings Inc., the largest Canadian-based property and casualty (P&C) insurance company. One camp, including a handful of shareholders who control the majority of Fairfax's shares, sees company founder, chairman and CEO Prem Watsa as a Canadian version of investing god Warren Buffett. But for his critics, including the short sellers who've been locked in trench warfare with Fairfax for years, Watsa is far short of a god. To them, he can do nothing right.

The battle has been largely waged out of the public eye because Watsa has always shunned media interviews, preferring to let Fairfax's financial results do the talking. Until now, that is. Why the change of heart? Jovial and relaxed as he sits in Fairfax's boardroom in downtown Toronto, Watsa tries to shrug off the question: "Once in 20 years. Hey, you just got lucky."

The reality, however, is that Watsa has some explaining to do. Company shares, which traded on the Toronto Stock Exchange as low as $3 in 1985, soared to peaks of $600 in 1998 and 1999. Since then, their value has deteriorated to around $175 lately as Fairfax has been battered on virtually all sides—by the shorts, by staggering claims costs from 9/11 and a wave of natural disasters, by volatile North American stock markets and by lingering hangovers from the company's ambitious acquisitions in the late 1990s.

There's more. The shorts have been almost gleeful lately as Fairfax has been swept up in the U.S. imbroglio over complex arrangements known as "finite reinsurance." Last March, the scandal felled Hank Greenberg, the chairman and CEO of American International Group Inc. (AIG), the world's largest insurance company. Now Fairfax is being probed by Washington's Securities and Exchange Commission (SEC) and investigators with the U.S. Justice Department.

Watsa insists that Fairfax's financial dealings have always been impeccably honest and fully disclosed in its annual reports and regulatory filings. Indeed, dozens of companies are also being probed in the wake of the AIG storm—Fairfax may just have been caught up in a police sweep.

Still, given Watsa's background, it's a strange turn of fate that he would end up being automatically associated in investors' and analysts' minds with the fallen titan of the North American insurance business. But then, the previous twinning—with Buffett—was pretty unlikely too.

Watsa was born in 1950 in Hyderabad, India's fifth-largest city. The place is known for its mosques and temples, but Watsa was raised an Anglican. In a cheery and nasal voice, with inflections that are part East Indian, part anglophile and part Bay Street jock talk, Watsa speaks warmly about his religious feelings. "In my story, there are tons of fortunate instances," he says. "I happen to have a spiritual faith, so I feel I'm blessed. Very hugely blessed."

Watsa's father, orphaned as a toddler, was a self-made man who became principal of a school that Watsa describes as "the equivalent of Upper Canada College." Watsa was good at chemistry, "so my dad said I should do engineering." He graduated from the Indian Institute of Technology in 1972 with a bachelor's degree in chemical engineering. He met his wife-to-be, Nalini, while he was there, but he didn't actually like his courses much. Business was far more intriguing, so he enrolled at the Indian Institute of Management.

But another paternal intervention precluded his completion of the program. Watsa's older brother, David, had settled in London, Ontario. Their father figured prospects were limited in India, so he told Prem that he, too, should go to Canada. Watsa arrived in London in 1972 and enrolled in the MBA program at the city's University of Western Ontario. Watsa recalls joking with the dean of UWO's Richard Ivey School of Business in 1999, before accepting an award for business leadership: "I didn't come here because the school was good. I came here because this was the only place I could afford to go."

Watsa rented a room in his brother's house, and helped put himself through school by selling air conditioners and furnaces door to door. When he went looking for work after graduating in 1974, Watsa experienced one of those "fortunate instances." Three other candidates didn't show up for their interviews for an investment analyst job at Confederation Life. Watsa got the nod. At Confed, Watsa says he had his "road to Damascus moment"—his first boss, John Watson, handed him Benjamin Graham's The Intelligent Investor, a classic of value investing, the doctrine that urges investors to look for signs of underlying value in companies, rather than chasing after rapid growth.

Confed would eventually collapse because of disastrous speculation in real estate, and was liquidated in 1994. But in the 1970s, it was a solid insurer. Soon after arriving, Watsa was managing a portfolio of Canadian stocks. His value-investing discipline was put to the test during the oil and gas boom in the late 1970s and early 1980s. The value principle said: Stay out. "You had Dome Petroleum and all sorts of oil speculations, perhaps not unlike what we have today," says Watsa. "So we underperformed for about 18 months to two years." But then the energy bubble burst and Dome went bust.

At Confed, Watsa began his practice of collecting investing colleagues: Five Confed alumni are still with him today. In 1983, Watsa moved to Gardiner Watson, a Toronto boutique investment firm. There he met one of the most improbable investment managers ever, and one of the most talented: Francis Chou. At the time, Chou, another immigrant from India, was a solder-wielding Bell Canada technician who'd earned his Chartered Financial Analyst designation in his spare time. He was introduced to Watsa by one of the firm's branch managers. "I was never more impressed in a 45-minute meeting than I was with Francis," says Watsa.

Chou is now a vice-president at Fairfax. However, he's best known to investors for his own line of Chou Funds, which Watsa helped him launch in 1984. In 2004, Chou won a Canadian Investment Award as mutual fund manager of the decade. To preclude the possibility of conflicts of interest, Chou collects no salary at Fairfax, but he still has company shares he bought in the early days at around $3.25 apiece.

Soon after meeting Chou in 1984, Watsa left Gardiner Watson to found his own pension-fund management firm, Hamblin Watsa Investment Counsel, along with a former boss from Confed, Tony Hamblin, and a handful of others. They began with just $30,000 of their own money. It took Watsa almost a year to get their first client, Pratt & Whitney.

They quickly landed nine more, and Hamblin Watsa went on to become a permanent part of the Fairfax group. Watsa might have happily remained a pension manager but for Chou. He had planted a key concept in Watsa's head: that of float. "Francis tells me: Do you know how [Warren Buffett's] Berkshire Hathaway made money? And Henry Singleton's Teledyne? And Larry Tisch?" Those value-investing legends ran companies that generated lots of cash every year—pools of cash that have to be invested. If invested wisely, that investment portfolio can generate huge profits.

To that end, Buffett had started acquiring P&C insurers in the 1960s. In 1980, Watsa began making the pilgrimage to Berkshire's annual meeting in Omaha, where 20,000 acolytes now gather each year in search of homespun wisdom from Buffett.

In 1985, Watsa found a P&C of his own to buy: Markel Financial Holdings, a Canadian-based specialist in trucking insurance. Controlled by the Virginia-based Markel family, it was virtually bankrupt. But Watsa figured it just needed a capital injection. He hit it off with Steven Markel, who is still a friend. In 1987, Watsa reorganized the company and renamed it Fairfax, short for fair, friendly acquisitions.

Although Watsa says he "didn't have a clue" at the start how huge Fairfax would get, the Markel acquisition set a pattern for astonishing expansion in the late 1980s and early 1990s. Fairfax bought up beat-up P&Cs on the cheap, and turned them around by improving their core operations and by generating more profits from their investment portfolio. By 1993, Fairfax had over $1 billion in assets.

Watsa will also tell you—and tell you again and again—that as Fairfax has looked for opportunities from year to year, it has remained true to a set of 15 "guiding principles" listed in its annual report. The most important of these is to build shareholder value over the long term. There is even a specific target: a 15% annual return on shareholders' equity.

To stay focused on the long term, Fairfax says it will "always be a very small holding company and not an operating company." There are just 30 employees at head office in Toronto, although the operating divisions employ almost 8,000.

Watsa says that maintaining focus also requires that one person has to have a controlling interest—namely Prem Watsa. His 10-for-1 multiple voting shares give him just over 50% ownership. "I put the group together," he says, "so I had control from that day onwards." And, at 55, he has no plans to give it up or retire any time soon. "I love what I do," he says. "I'm planning to be there for another 10 years at least. Another 20, to be exact."

Of course, multiple voting shares start red flags waving in the good-governance camp, as does combining the chairman and CEO's office. Ditto for boards that are small and populated partly by insiders—like Fairfax's, which consists of just Watsa and five other directors, one being his friend Robbert Hartog. But some of Fairfax's major shareholders prefer a company controlled by a strong guiding force. "I love Frank Stronach. I know he pays himself too much. And l love Gerry Schwartz. These guys are business guys," says Wade Burton, a portfolio manager at Peter Cundill & Associates. "Prem made this thing survive."

Companies run in imperial fashion can work fine, of course. Unless the leader bites off more than he can chew.

In 1998, it looked like Watsa and Fairfax could do no wrong. The company's revenues had tripled in the previous four years alone, and the share price had soared, even after Fairfax issued millions of new shares at prices as high as $500. Watsa had steadfastly maintained his media blackout, which in some eyes made him principled, and which added to his mystique. Chief financial officer Greg Taylor, who has been with Fairfax since 2002, was working for Merrill Lynch Canada during those heady days. "I remember reading the brokerage reports on [Fairfax], and the targets were $800 a share," he says.

Then came two huge U.S. acquisitions that again more than doubled the size of Fairfax. The first was Crum & Forster, a poorly performing division of Xerox that Fairfax bought for $565 million (U.S.) in August, 1998. The second was TIG Holdings of New York, which Fairfax bought in April, 1999, for $847 million (U.S.).

Both acquisitions soon went off the rails. As it turned out, 1997 proved to be the beginning of a four-year stretch of so-called soft pricing in the P&C and reinsurance business. The industry tends to go in cycles: During the soft portion, companies compete hard, often by taking on riskier business. But then they're jolted back to reality by a disaster, and premium rates harden—they shoot up.

So long as the insurer survives the claims hit. Like many other P&Cs, Fairfax staggered after the 9/11 catastrophe. The company posted its first annual loss ever: $346 million. The following year, it had to put much of TIG into "runoff." In this industry practice, a company or unit stops writing new business and, it's hoped, has enough reserves left to pay out any remaining claims.

Crum & Forster's troubles were exposed by its combined ratio, the standard measure of an insurer's operating performance. Basically, the combined ratio is the claims paid during the year, net of reinsurance, plus the company's operating expenses, including commissions and administrative costs, all divided by the premium income the company earns that year. If the ratio is greater than 100%, the company is losing money on operations. In 2001, Crum & Forster's combined ratio shot up to 131%, helping to push the average ratio for Fairfax as a whole to 121%. Matters improved in 2002, however, with Crum & Forster's combined ratio declining to 108% and Fairfax as a whole working its way down to 102%.

But Fairfax was soon hit from other directions. In December, 2002, the company listed its shares on the New York Stock Exchange. That made sense for a company whose business was increasingly U.S.-based and that needed access to capital from American investors. (In 2003, Fairfax also switched to reporting its results in U.S. dollars.) But the NYSE listing also brought increased scrutiny from merciless American analysts and short sellers. Some of them immediately jumped on Fairfax.

In January, 2003, the Memphis-based brokerage Morgan Keegan & Co. Inc. released a scathing report that circulated quickly on investor websites. It argued that Fairfax's reserves against future insurance claims were deficient by $5 billion (U.S.). The NYSE was closed for Martin Luther King Day when this broadside was fired, but in Toronto, Fairfax shares plunged to $57 within three hours, a 40% drop. Watsa quickly released a statement refuting the report; the shares rebounded and closed at $85.

Two weeks later, Morgan Keegan lowered its estimate of Fairfax's reserve deficiency to $3 billion (U.S.). But the company still disputed that figure, and went on what, for Fairfax, was a PR offensive: Watsa actually talked to Forbes magazine, and committed himself to answering questions from analysts in quarterly conference calls.

In its rebuttal volley, Fairfax charged that short sellers were trying to manipulate its stock. Short sellers attempt to profit from share price declines by "borrowing" stock. The lender may be a brokerage or a shareholder, and typically takes a commission. The short sells the stock, hoping a wave of selling will drive down the price so he can purchase shares later at lower prices. The short returns the shares to the lender once he's taken a profit. Naturally, the shorts try hard to talk the stock price down. Morgan Keegan analyst John Gwynn, one of the authors of the damning report, responded angrily to Fairfax's accusation. "We're not in cahoots with the shorts," he said.

Analysts tend to be just as sharply divided on Fairfax as long and short investors are. No surprise, then, that the company has had more spats with analysts and bond-rating agencies over the past three years. In March, 2003, Fitch Ratings downgraded some of Fairfax's debt to junk status. Fairfax cut off Fitch from the standard private analyst briefings given by Fairfax executives. Throughout 2003 and 2004, Fitch argued that Fairfax was perilously short of cash at the holding-company level. Last September, however, Fitch took Fairfax's ratings off negative credit watch, noting that cash pressures had eased, although Fairfax's $2.2 billion (U.S.) in long-term debt was still high in relation to its earnings.

The shorts, meanwhile, haven't given an inch. Soon after Fairfax debuted on the NYSE, the total number of shares sold short swelled to two million, and it's stayed around there ever since. "We think this is a zero," veteran New York short seller Jim Chanos, President of Kynikos Associates, declared of Fairfax last summer. Other short sellers conform to the more shadowy stereotype, badmouthing the company behind the scenes. "I don't exist," says one, who nonetheless asserts that Fairfax is a huge debacle.

Fairfax has about 18 million shares outstanding, but long-term allies, including firms controlled by value-investing legends Mason Hawkins of Memphis and Peter Cundill of Vancouver, own the majority of them. Even these stalwarts are getting frustrated by the relentless attacks from the shorts. "It bothers me quite huge," says Wade Burton of Peter Cundill & Associates. A few years ago, Burton didn't mind if the shorts drove down Fairfax's price temporarily—it gave him a chance to buy more of the stock cheaply. "Now I'm prepared for that opportunity to go away," he says.

The so-called public float of shares is thin. Fairfax's allies argue that the thin float makes it easier for shorts to move the share price—more than $10 in a day on occasion. But the shorts retort that it's tough to find shares to borrow, which means it can cost them the equivalent of 40% interest.

As technical as the issues around Fairfax are, they boil down to one big question: Is the company in better or worse shape now than it was three years ago, when it was lambasted by Morgan Keegan?

Watsa likes to point to combined ratios as signs of improved performance. Take Northbridge Financial Inc., Canada's largest commercial P&C insurer. Its combined ratio was 88% for 2004. OdysseyRe, Fairfax's global reinsurance arm, came in at 98%. Fairfax as a whole averaged 98% as well. Of course, the combined ratio doesn't reflect any earnings on Fairfax's investment portfolio. The past five years have been "tough, tough, tough" for operations, Watsa concedes. "But peel the onion: Crum & Forster is looking good, OdysseyRe is looking good, Northbridge is looking good."

Morgan Keegan's Gwynn and other critics argue that many of the operational problems haven't been resolved at all. Gwynn says Watsa was hell-bent on acquiring more float in the late 1990s, and he didn't look closely enough at the problems on the books of Crum & Forster and TIG. Since then, says Gwynn, Fairfax has merely dumped much of TIG's problems into its sizable runoff operations. "The stabilization never occurred," he says. "Once you have a crappy book of business in the property-casualty business, it's a crappy book of business that you own forever." (That's a positively polite criticism compared to what gets thrown Fairfax's way in on-line investor forums such as Yahoo Finance, stocklemon and onelimestreet.)

Watsa rejects the charge that he bought indiscriminately to accumulate investable assets. He says he drove hard bargains by negotiating directly with the vendors, rather than participating in inflation-prone auctions. And he points out that he paid substantially less than book value. The goal in all of Fairfax's acquisitions, he adds, is to generate long-term returns from investments and operations. "The combination of those two has made money for our shareholders," he says.

Indeed: Despite the recent reversals, Fairfax's 20-year record is formidable: from $17 million (Canadian) in annual revenue in 1985 to $5.8 billion (U.S.) in 2004 (making Fairfax the 31st-largest publicly traded North American insurer). Average annual return on shareholders' equity has been about 15%.

Still, runoff operations remain a nagging cash drain on Fairfax. The company has been anticipating that it would lose about $100 million (U.S.) a year for several years, excluding unusual items. But there have been a lot of those items lately. In 2004, run-off losses totalled $194 million (U.S.). And in the third quarter of last year alone, they hit $181 million (U.S.). "A runoff book of property-casualty insurance is almost, by definition, an extremely difficult proposition. You have no cash inflow and all cash outflows—claims," says Gwynn. "That's usually what catches with an under-reserved company. That's what I think is gonna happen here."

Watsa admits that putting much of TIG's operations into runoff "was a tough decision for us." But he says part of the reason for the big cash drain since then is that the runoff group is settling remaining claims quickly and efficiently. "When we put it in runoff, [there were] 55,000 claims," says Watsa. "Today, it's 15,000. It's a dull roar. It's almost over."

In the third quarter of last year, however, Fairfax was slammed by Hurricanes Katrina and Rita, and it posted a $220-million (U.S.) quarterly loss. But the impact of the hurricanes could be short-lived. And with that factor excluded, Fairfax says it would have made $91 million (U.S.). The hurricanes have also prompted big regional price increases, with some premium rates tripling.

To get through the tough period, Fairfax has also raised cash by selling new shares, $300 million (U.S.) worth in December, 2004, and the same amount last October. That Fairfax needed the money showed it's been under strain. That it was able to raise it, particularly right after last year's hurricanes, showed it still has support in the market. "The third quarter served as a real-world 'field test' of the financial strength of Fairfax, a test the company comfortably passed," says Mark Dwelle, an equity analyst with Ferris, Baker Watts, Inc., a Washington, D.C.-based investment bank that has long rated Fairfax a "buy."

Well, yes, it passed, but not without a lot of bobbing and weaving. The major buyers of those Fairfax issues were firms controlled by Hawkins and Cundill, and Steve Markel's Markel Corp., which bought $100 million (U.S.) worth in December 2004. As with several other issues in recent years, Fairfax also had to sell shares at less than book value. Yet P&Cs trade on average at a premium to book value. What's more, selling for less violates one of Warren Buffett's guiding principles: "We will issue common stock only when we receive as much in business value as we give."

The runoffs, the hurricanes, the money-raising moves—presumably, all of these have been digested by the market. So, one answer to the question of whether Fairfax is better off than in 2003 is simply the price of its shares: At around $175 on the TSX lately, the price is a far cry from the glory days, though it's more than double the $85 close on Martin Luther King Day in January, 2003. But there is one factor on which the jury is definitely still out: the complicated, topsy-turvy world of finite reinsurance.

Ever since Morgan Keegan's report in January, 2003, John Gwynn and other critical analysts have honed in on the second-largest single asset on Fairfax's balance sheet, an item called amounts "recoverable from reinsurer." That is money Fairfax expects to collect from reinsurers. As of last September 30, those recoverables were $7.6 billion (U.S.). By industry standards, that's very high—28% of Fairfax's assets, and more than double Fairfax's shareholders' equity of $3 billion (U.S.).

It's routine for P&Cs to buy reinsurance from other companies. But Gwynn and other critical analysts have long argued that Fairfax's recoverables are "soft" assets. Some of the reinsurers may be troubled, they argue, and may not be willing or able to pay the full amounts.

Regulators can also be suspicious of "finite re." A highly simplified example shows why. Suppose that an insurer is worried that it may have to pay out more in claims over the next several years than it had anticipated because courts are awarding higher settlements in, say, environmental cases. The insurer estimates the amount will be $1 billion. If it chooses to keep the risk itself, it must set aside $1 billion in reserves now. Instead, however, it could purchase a limited, or finite, amount of reinsurance (rather than unlimited protection). The reinsurer is then responsible for up to $1 billion worth of claims. The insurer agrees to pay a premium of, say, $520 million for the coverage.

If the reinsurer doesn't have to pay out the $1 billion within 10 years, for instance, and earns more than 7% by investing the money, it will earn a profit. Meanwhile, the insurer gets benefits on its financial statements. Yes, the insurer's revenue drops by $520 million, which, technically, is premium revenue "ceded" to the reinsurer. But the insurer is off the hook for $1 billion in claims, so its net income for the year is $480 million more than it would have been. The capital on its balance sheet is also $480 million more.

Two other wrinkles: Suppose the events covered by the insurer's original policies have already happened but it appears the claims costs may yet escalate. (This is the case with decades-old asbestos claims, for example.) And suppose the insurer hangs on to the $520 million, and instead guarantees the reinsurer a 7% return.

Such arrangements are controversial. Unless there's a transfer of a meaningful amount of risk to the reinsurer, the arrangement looks an awful lot like a loan rather than an asset, albeit one that improves the insurer's balance sheet immediately. Gwynn says finite re is like "cocaine"—it can provide a quick fix. In Canada and the United States, the regulators' test for legitimate risk transfer is a greater-than-10% chance of a greater-than-10% loss for the insurer.

Although the above example is simplified, it has similarities to a controversial arrangement Fairfax made in 1999, after it bought TIG. Fairfax turned around and bought $1 billion (U.S.) worth of coverage from a subsidiary of Swiss Re of Zurich. The coverage was against "adverse development" on all of Fairfax's subsidiary business written before 1999, including TIG.

Watsa and Fairfax CFO Taylor say the Swiss Re "cover" and their other finite reinsurance arrangements are entirely legitimate risk transfers, and in keeping with one of Fairfax's fundamental guiding principles: "We always look at opportunities but emphasize downside protection and look for ways to minimize loss of capital."

As for the SEC and U.S. Justice Department investigations into finite reinsurance, Watsa and Taylor can't comment on any specifics. But they point out that Fairfax's numbers are reviewed by its own actuaries and accountants, as well as by external auditors, ratings agencies and government regulators.

Ground zero in the industry investigation is AIG. In November, 2004, AIG agreed to pay a $126-million (U.S.) fine to settle an allegation that, as a reinsurer, it helped Pittsburgh-based PNC Financial Services polish its books. The firing of company founder Hank Greenberg came four months later, after the company's board reviewed a suspicious 2000 transaction with Warren Buffett's General Re Corp.

New York State Attorney General Eliot Spitzer then filed a civil suit accusing Greenberg of using fraudulent finite reinsurance contracts to understate AIG's liabilities and overstate its profits. By industry standards, Fairfax is a relatively heavy user of finite re, so it was a logical target in the U.S. regulators' sweeping investigations of so-called non-traditional insurance and reinsurance products. Thus Fairfax's announcement last fall that it was also being probed by the SEC and officials with the U.S. Justice Department.

The adverse publicity from the AIG case is a danger for Fairfax. At the end of March, 2005, Merrill Lynch Canada analyst Brad Smith downgraded Fairfax's shares to a "sell" rating—and caused the share price to dip by more than 10% within a week—by saying the negative stories and the investigations may prompt some reinsurers to get out of the business. If that happens, Fairfax will be squeezed.

Any take on Fairfax has to consider Watsa's personality as well as the numbers that surrounds him. Even some of his critics concede it would be out of character for him to misuse finite reinsurance. He's a straight arrow who is dedicated to Fairfax. "More than 99%" of his net worth is tied up in the company, he says, and he collects a meagre salary by CEO standards: $600,000. He and Nalini do live in Toronto's tony Rosedale neighbourhood, but theirs is a relatively modest house.

The couple has three children, all in their 20s: Ben, an investment banker in New York; Christine, an investment analyst; and Stephanie, who recently earned a BA in business and marketing. "First of all, I told them, 'You can't get into Fairfax,'" Watsa says with a smile. "'You'd never get a sense of accomplishment. They'd always say it was your dad.'"

Apart from family, the biggest influence in Watsa's life, as both a friend and business adviser, is mutual fund legend Sir John Templeton, now 93 years old. Watsa keeps a bust of Templeton in Fairfax's boardroom, and visits him at least once a year at his home in the Bahamas. "What I always liked about him was that he had a spiritual bent to his personality, and to his work," says Watsa. "He's a wonderful guy. Just a really wise man."

In his spare time, Watsa has helped Toronto's St. Paul's Anglican Church and the Hospital for Sick Children greatly in recent years by running their investment committees. He took up golf about 10 years ago, but has not joined any of the city's fusty established clubs, opting instead for the Devil's Pulpit in Caledon, where he also has a country home.

Of course, whether or not the head guy's a straight arrow, it would be irresponsible to speculate on the guilt or innocence of a company that is merely under investigation by regulators. But analysts do have to construct risk scenarios. One of them is Mark Dwelle of Ferris, Baker Watts. U.S. regulators often file civil suits in complex financial cases, rather than serving harder-to-prove criminal charges. "This seems to be the way these things have been resolved," says Dwelle. "The litigator says, 'We're gonna look at it like this, and if you want to go to trial to try to defend this arcane, esoteric product, you can do that. Or you can settle for X.' "

"Everybody chooses to settle for X," says Dwelle.

Just about the only thing that Fairfax's supporters and critics agree on are the investing abilities of Watsa and his team. The return on Fairfax's portfolio often bounces up and down from year to year. But the yearly average from 1985 to 2004 was a healthy 9.5%.

Yet Watsa's ultra-cautious value investing discipline has held back Fairfax's returns lately. Even when stocks are at their most alluring, Fairfax never puts more than 25% of its money in equities, and lately it's been a lot less. The company had actually started reducing the percentage of stocks in its portfolio by 1995, even though the Dow Jones Industrial Average would go on to more than double its value, before peaking in early 2000. By then, Watsa was sure the markets had gone mad—how else to explain Yahoo trading at 1,425 times earnings in 1999?

So Fairfax has kept most of its portfolio in bonds. But analysts and economists say the upside on bonds now looks limited. Meanwhile, the NYSE and the TSX have climbed back to near-record or record heights. No matter to Watsa: Last September 30, the end of Fairfax's third quarter, its portfolio consisted of $4.4 billion (U.S.) in cash, $8.3 billion (U.S.) in bonds and preferred shares, and just $2.6 billion worth of equities. As a result, Fairfax earned 9.1% over the preceding 12 months, including paper gains and losses.

As with insurance operations, Watsa says one of Fairfax's main objectives with investments is protecting itself from catastrophes. "If there's a big drop in the stock market, if the economy weakens, sort of a 1-in-50- or 1-in-100-year event, we protect ourselves on the investment side from that," he says.

Spoken as only a diehard conservative investor can. For a sunny forecast from Watsa, one has to turn the subject to the only sort of event that will bring the long war with the shorts to an end.

"Once we perform, once we make a lot of dough, believe me, the stock's going up."
;