The comprehensive low-risk long/short financial institution arbitrage strategy which we identified as offering the strongest risk-adjusted returns through 2008 is a sector neutral strategy that involves going long a basket of “thrift and home finance companies” [Freddie Mac (FRE): $67.9, Fannie Mae (FNM): $59.4, and Countrywide Financial Corporation (CFC): $42.5] and short a basket of three property/casualty insurance stocks [Philadelphia Consolidated Holding Corp (PHLY):$44.6, W.R. Berkley Corporation (BER); $34.5, and SAFECO Corporation (SAF); $62.6]. Assumptions for our “base” and “best” case scenarios are summarized.
Significant Research Conclusions
The recent relative strength in the property/casualty stocks due to overall sector re-weighting (and benign hurricane season) is masking the developing poor fundamental outlook for the intermediate term and long term. Our basic thesis assumes that the near-term consensus earnings estimates (next two quarters or so) are reasonably accurate for PHLY, BER, and SAF. But the ROEs should degrade to the low double digit level or below by late 2007/early 20008, under our “base case” scenario. The complacency surrounding expectations at this stage of the property/casualty pricing cycle beyond 2006 is unusual. It is reflected in the consensus earnings estimates that assume a high teens ROE through 2008; This is a level of profitability consistent with (even below) results reported in 2005 and through most of 2006, but is unusually optimistic, given the inherent cyclicality of the p/c business and absence of barriers to entry. We don’t believe the conviction surrounding analysts’ earnings expectations beyond the next several quarters is high. In our view, the potential error in the forecasts looking out two years rises substantially during profit “peaks” or “troughs”. The expected ROE implied for CFC is below that reflected in consensus estimates, suggesting near term expectations are still being low balled and continued volatility in the whole mortgage sector. In fact, from the current stock quotes, one can impute expectations of an ROE for CFC sufficiently below consensus estimates for the 2007 to make this trade worthwhile. For FRE and FNM, the market is no longer discounting material adverse developments, though the relative valuations suggest the market views the recurring earnings stream as riskier than that of even the big name banks. There is a high probability that ‘intermediate-term” consensus estimates (beyond mid 2007) and prospective ROEs for FRE, and FNM are being “low balled” here as well.
Base Case: Short, medium, and long-term rates stay between 4.25% and 5.25% through 2007, and early 2008. That is what the market is discounting, as viewed through the lens of financial stocks. In a +5% to +10% market (S&P through 2007), the upside to the “mortgage” basket is an estimated +10% to +20%. The downside to the P/C insurance basket is -5% to -15% (2.0 average price/book for the 3 stocks) Best Case: Long-term interest rates back up closer to the 5.25% level (high end of our range) concurrent with deterioration in the insurance stock’s ROE to low double digits by early 2008. In fact, the possibility of breakeven or (even) money losing operations for property/casualty insurers at some point in the cyclical downturn through 2010 is not out of the question. In a flat market (with the 10-year Treasury backing up to closer to 5.25%), the upside for the basket of mortgage related stocks is an estimated +5% to +15%, and downside for the p/c sector is an estimated -15% to -25% to an average price/book of about 1.7 (average for the three stocks)
Property/Casualty Insurance (PHLY, BER, SAF)
While supply and demand for insurance “product” are a function of numerous variables (including company specific and aggregate industry capital levels, cash flow, and ratings), the innate cyclicality of this commodity-type business suggests picking “spots” in the cycle is crucial in evaluating opportunities for investing. We believe the current environment offers one of the “spots” due to the excess capital and liquidity in the insurance and capital markets, the flush profit and cash flow levels (albeit deteriorating recently), and unusually long period of “excess returns” in this commodity business. Exposure to excess and surplus lines is particular troublesome by some players, given the experience in the late Nineties. Our view is that the probability of a more serious cyclical downturn is a “when” not “if” question and believe the prospective degradation in earnings and reserve quality has not been fully discounted, recent reserve releases by PHLY notwithstanding. If we’re correct, a contraction of the price/book valuation to a a still historically high premium to book value seems like a reasonable outcome.
Freddie and Fannie
In contrast to the prevailing conventional wisdom, the security of the revenue and earnings stream for FRE and FNM has improved (on a relative basis) with short, medium and long term rates in the current 4.25% to 5.25% range) versus a few years back with much lower 10 year Treasury and Fed Funds rates. The higher rates reduce the volatility in the refinancing market (near multi-year lows), and gives spread income more of an annuity stream quality. Modestly higher rates would even be more beneficial to the companies’ revenue and earnings streams. There is also minimal credit risk in the mortgage portfolios, while the 25%-plus ROE for the insurance guaranty business continues to be low-risk. Our base case thesis on FRE/FNM assumes consensus numbers and the status quo through 2007 (i.e. mid-teens ROE and no capital releases for FRE). While one shouldn’t underestimate the importance of Fannie and Freddie returning to regular and accurate audited reporting, our own methodology for evaluating fair value for financial stocks has always centered on a smoothed ROE (relative to the price/book). That strategy has always allowed for moderating expectations of “potential stock returns” during periods of “earnings” growth. Such a methodology also minimizes risk due to earnings “misses”, since earnings have a fair amount of variability in the estimation process (be it insurance reserves, credit losses, and mark to market derivatives). The likely “realization” of additional unrecognized losses on investments or derivatives for both Freddie and Fannie would like have a minimal impact on their book value/share growth, investors should be mindful of the development of the FRE/FNM story over the last few years. If the past record of the Street in assessing risk and reward in these poorly misunderstood institutions is anything to go by, investors would do well to take a contrarian approach. The changes in the business model and GSE environment precludes the possibility of returns (ROE’s) getting back to the average 25% average level of the late Nineties and the early part of this decade. But the business model should support a mid teens ROE s for the foreseeable future, with various “options” that add some upside potential.
Countrywide’s discounted P/E reflects heightened near-term uncertainty surrounding the ability of the company’s diversified insurance, capital markets and banking businesses to offset the drop in revenues, margins, and profits in origination. There is added uncertainty (more recently) in the ability of management to hedge the Mortgage Servicing Rights [MSR] asset. But to a better extent than most, CFC has navigated this terrain rather well (in fact, the company’s track record goes back more than thirty years).
Our view is that any fourth quarter earnings miss or lumpiness is likely to have less of an impact on the two-year average ROE. It’s hard to imagine a scenario in which Countrywide’s ROE would stay much below 15% level (from an estimated 20% based on consensus numbers for the four quarters ending in December) for more than a few quarters or so. Barring a collapse in the mortgage market (unlikely with rates well below 5%), a return to mid-to high teens ROE seems like a safe bet. In fact, for our price targets to be missed, the below trend ROE would have to persist well beyond 2007.