We have a great deal of respect for short sellers – it can be a brutal and dangerous game to try to fight the long-term upward trend of the market and the Wall-Street-and-management-fueled promotion machine – so when we find a stock with a high short interest, we typically give it a wide berth. So it might come as somewhat of a surprise that the funds we independently manage are both long Canada's Fairfax Financial Holdings (FFH).
In the case of T2 Partners, this is the first time we've ever been long a stock that we were once short. Sometimes the shorts are wrong, especially when the story around a company changes and becomes more attractive over time, which we believe to be the case here. First some history. Fairfax was started in 1985 by its current CEO, Prem Watsa, and expanded rapidly by buying beatendown property and casualty (P&C) insurers on the cheap and then improving their core operations and investment results. This formula proved extraordinarily successful: from 1985 to 1998, Fairfax compounded book value per share at more than 40% per year, while the stock rose nearly 50% per year. Beginning in 1998, however, Fairfax suffered a series of calamities that nearly sunk the company. It doubled its size with two ill-fated acquisitions of U.S. insurance companies, Crum & Forster and TIG Holdings, just as the P&C insurance industry entered one of its worst cyclical downturns ever, from 1998 to 2001.
It also soon became clear that both TIG and Crum & Forster were significantly under-reserved, as growing claims for asbestos and workers compensation turned out to be far worse than originally expected. Then came 9/11 and, a few years later, the terrible hurricane seasons of 2004 and 2005, which hammered Fairfax along with its peers. The resulting losses, combined with a complex corporate structure and a CEO who rarely spoke to the media, analysts or shareholders, led short sellers to Fairfax's door. Soon after the company debuted on the NYSE in late 2002, the total number of shares sold short swelled to two million, and it's stayed around that level ever since.
According to shortsqueeze.com, as of August 29, 2007, 10.7% of Fairfax's float was sold short, equal to 28 days of average trading volume. "We think this is a zero," veteran short seller Jim Chanos, President of Kynikos Associates, declared of Fairfax in the July 29, 2005 issue of Value Investor Insight. That was a great call for about a year, as the shares sank below $100, but they have since rebounded, currently trading near $205. The absence of major hurricane activity since 2005 has been a well-timed bless-ing to Fairfax, given that another record hurricane year might have put the company under. But hard pricing following the 2005 hurricanes led to record premiums for Fairfax of $6.8 billion last year and the company earned a net profit of $227.5 million.
In fact, excluding a non-cash (and non-economic) accounting charge, Fairfax had a monster year – net income would have been over $630 million, and return on equity would have been 22.8%. As its performance has improved, the company has also made significant strides in deleveraging its balance sheet, extending maturities on existing debt and simplifying what had become an overly complex corporate structure that served as fuel for the short sellers' suspicions. Today, the short thesis that seemed plausible two years ago no longer appears as convincing. In particular, it does not appear that one key pillar of the short thesis – that throughout the 1990s Fairfax overpaid for a range of bad insurers that are destined to be chronic moneylosers – is an accurate explanation for Fairfax's poor performance from 1998 to 2005. Rather, it appears that two bad acquisitions (Crum & Forster and TIG) and some bad luck were the culprits, and that those problems have now for the most part been fixed.
For comparison, consider another large insurance-company acquisition made with richly priced stock by a well-known value investor at roughly the same time: the acquisition of General Re by Warren Buffett's Berkshire Hathaway. While Gen Re was not as broken as C&F and TIG, it nevertheless lost billions of dollars – yet now has finally returned to health. Similarly, Fairfax's underwriting businesses now appear to be in good shape. Its overall cost of float has been negative in three of the past five years (2003, 2004 and 2006) and even C&F is now consistently posting underwriting profits. Approximately 74% of Fairfax's equity value is represented by its ownership in two publicly traded subsidiaries, Northbridge and Odyssey Re, both high quality operations that trade at meaningful premiums to book value. Fairfax's runoff operations, including TIG, reinsurer Sphere Drake and other historical underwriting liabilities, have been worked down steadily over the past several years. Reinsurance receivables, the collectability of which was a big part of the short thesis, have been significantly reduced, and much of the remainder is with highly credit worthy counterparties.
The holding company is no longer highly leveraged – Fairfax had just over $750 million in cash and total debt of $1.1 billion at the parent company level as of June 30. Virtually none of the debt matures before 2012, and none of it is callable bank debt. In short, then, an investor in Fairfax at the current price gets a diverse collection of high-quality insurance businesses that appear positioned to once again compound per-share book value at a meaningful clip – its internal target is at least 15% per year over time – all at a slight premium to stated book value. In Prem Watsa, shareholders are represented by an ethical and shareholder-oriented CEO with proven capital-allocation skills.
In addition, there are at least two potentially valuable hidden assets for which today's investor pays essentially nothing. The first is a 26% stake in ICICI Lombard, the largest private general insurance company in India, with a 12.5% market share. For the fiscal year ended in March, the company earned over $700 million in gross written premiums, grew book value more than 150% and turned a profit despite investing heavily in rapid growth. Fairfax is carrying this minority interest on its balance sheet at $60 million, which appears to be a significant discount to intrinsic value.
This became obvious recently when ICICI Bank, which owns the 74% of ICICI Lombard not owned by Fairfax, requested government approval to sell a 5.9% stake in itself for $600 million to a group of private equity investors led by Goldman Sachs. This price implies a value of more than $10 billion for all of ICICI. While it’s not clear how much of this value should be ascribed to ICICI’s stake in ICICI Lombard, Watsa suggested $2-3 billion in a recent conference call. To be conservative, we’re using $1 billion – just over 4x book value – which values Fairfax's stake at $260 million and doesn’t seem unreasonable for a market leader in an attractive emerging market with high barriers to entry.
The second hidden asset, an enormous credit-default swap [CDS] book, is starting to pay off handsomely. At June 30, Fairfax and its subsidiaries held CDSs with $18 billion of notional value and an average term to expiration of 4.4 years on 25 to 30 companies, the majority of which (bond insurers, mortgage lenders, etc.) are significantly exposed to the unfolding subprime debacle. Prior to the credit crunch, the market value of these CDSs – which originally cost $341 million – had fallen to $198 million as of June 30. Just in the month of July, however, their market value soared to a total of $537 million.
If you share our belief that the subprime train wreck is still in its early stages, it would be hard to find a better way to play this than Fairfax. When short sellers figure out that in shorting Fairfax they are short $18 billion of credit default swaps on mostly subprime-related companies, there could be a squeeze of rather impressive proportions.
Given all of this, what do we think Fairfax is worth? We believe a fair valuation for a solid insurance company growing per-share book value at 15% annually is closer to 1.5x book, 25% greater than the 1.2x multiple today. Just adding July's $339 million increase in the CDS portfolio would increase book value by 11%, while valuing Fairfax's stake in ICICI Lombard at $260 million rather than the $60 million on its balance sheet would increase book value another 6%. Adding both of these to book value and putting a 1.5x multiple on the total yields a stock price of $310, more than 50% above today's level.
Disclosure: Author has a long position in FFH