Downey Financial Corp. (NYSE:DSL) recently released results for Q2, which came in very weak. Nonetheless, the stock did not experience much of a sell-off, perhaps due to the belief that because DSL is removed from the sub-prime arena, it may actually be a good hedge against the more volatile financial companies. Analyzing the Company’s latest 8-K indicates that this belief of safety may be unfounded and that while DSL is not a pure sub-prime originator, its business model has led to problems its income statement and balance sheet.
Given some recent media comments regarding the “cheap valuation” of DSL’s book value (“BV”) and its attractiveness as a potential buyout candidate, I felt compelled to demonstrate the problem with assessing DSL’s value based on its BV. Before deconstructing DSL’s BV, it’s important to review the Company’s latest summary P&L:
As illustrated in Table I, through the first half of 2007 DSL’s EBT was wholly generated by interest income from negative amortization mortgage balances. In addition, net interest income actually decelerated as did EBT, despite rising income recognition of capitalized interest from negative amortization (“CINA”). Through the halfway point of 2007, DSL’s EPS, which incorporates the highest level of CINA recognition and thus non-cash income in years, is at just $2.71 while the stock trades for approximately $60.
Current conditions in the southern California real estate market along with evidence that DSL’s NIM leverage and originations are decelerating suggest the Company will be hard pressed to earn $5.50 in full year EPS, resulting in a forward P/E valuation of 11.0x, fairly expensive when compared to larger, far more diverse banking operations whose fortunes are not wholly tied to the volatile southern California residential real estate market. However, the market seems willing to accept this risk despite consistent downward EPS revisions by various brokerage houses that still, in some cases, maintain favorable ratings on DSL.
Lehman Brothers (“LEH”) analyst Bruce Harting, for example, provided EPS estimates of $7.40 and $7.90 for 2007 and 2008, respectively, as recently as March 2007. However, a variety of Harting’s assumptions leading to these high EPS estimates were refuted in DSL’s Q1 figures, resulting in Harting revising his estimates to $6.25 for 2007 and $7.00 for 2008 this past May. Not to be outdone, just two months later, Harting once again reduced his 2007 and 2008 EPS estimates to $6.00 and $6.34, respectively, after Q2 results were reported. Interestingly enough, throughout these downward revisions he has maintained an Overweight rating on DSL. Conditions in just four months had changed significantly enough to warrant an 18% drop in EPS estimates for 2007 yet DSL still warrants an “Overweight” rating.
Financial services companies maintain financial statements that can sometimes be “illusions” as far as what’s occurring from a true economic and operational standpoint. Wall Street appears content with focusing on just cursory numbers and ratios without really digging beneath the numbers when it comes to these banks and appears willing to accept the notion that credit quality is not an issue with firms like DSL and even with subprime lenders. In fact, on July 11, 2007, the Wall Street Journal (“WSJ”) ran an article entitled “Behind a Bear Analyst’s Subprime Call” discussing how esteemed analyst Gyan Sinha erred by suggesting clients buy the ABX index before it substantially dropped in value. Sinha’s error seems to have been derived from faulty assumptions underlying his model, which indicated an ABX fair value of 90 – a value that proved to be high once the ABX shortly dropped to 63 and further slipped down to under 60 by July, according to the WSJ article.
The same optimistic assumptions still appear to be underlying DSL’s valuation resulting in some believing that DSL is an attractive value based on its BV. However, when adjusting for CINA’s impact on DSL over recent years, DSL’s true BV is far lower than what is actually stated. Table II lays out a quick analysis of what the Company’s pro forma BV would be excluding income CINA. This is a cumulative analysis going back from 2003 when CINA was just 1.6% of net interest income.
The Actual BV calculation is the reported figure on the Company’s books. To adjust for reported BV, reported EBT is adjusted for CINA and then the reported fiscal year tax rate is applied (varies from 39-42% in those year so it’s generally consistent) to generate a Pro Forma Net Income figure. From there, the prior year BV is used for 2002 in the 2003 column ($823MM) and the same calculations are used as in the Actual Book Value table with the exception that Pro Forma Net Income is substituted in place of Net Income as Stated resulting in a BV that excludes the low quality CINA earnings.
What’s interesting is that in 2003 and 2004, Pro Forma BV and Actual BV were not materially different but in the past few years it’s apparent that Actual BV has been wholly driven by CINA and not high quality earnings. In fact, through the 6 Mos ended 6/30/07, DSL cash interest income would likely have been negative considering that backing out CINA resulted in a net loss. While from a GAAP perspective, BV has increased by 60% from 2003 through 6/30/07, from an economic perspective DSL has basically treaded water and increased its overall risk profile.
The analysis in Table II only covers the impact of CINA and does not account for other potentially adverse events that could impact DSL. For example, DSL’s REO jumped from just $1.2MM from 6/30/06 to $8.5MM on 12/31/06 and now to $30.0MM on 6/30/07 based on the balance sheet data provided in the latest 8-K. With plenty of news regarding falling home prices in southern California, it is easily conceivable that DSL could write down some of this $30.0MM in REO in the coming years. NCOs and credit loss provisions should increase throughout the year further reducing eating into DSL’s EPS and thus BV growth. Examining the latest 8-K also indicates other key factors of credit quality have all worsened as NPAs as a % of Total Assets and Delinquencies as a % of Total Loans have all experienced significant increases.
Further, accretion of CINA has resulted in a negative amortization loan balance that now represents 4.2% of the $8.9B in residential loans held for investment, or $377.3MM, compared to just $229.1MM (1.7%) on 6/30/06. These balances could continue to rise throughout the year as CINA accretion outpaces the reduction of the negative amortization loan balances through any refinancings and prepayments. The reason I believe negative amortization loan balances will increase faster is primarily because those with strong enough credit quality likely refinanced early on when rates reset, meaning the current negative amortization balance may primarily be driven by weaker credit clients. As it currently stands, the current negative amortization loan balance equals about $13.48 per share and given the pace of CINA recognition should accrete to become an even larger portion as the year progresses.
What’s more, the focus on BV and valuing DSL’s BV against banks that have more diversified operations and credit instruments is highly misleading. DSL is a one trick pony in that it uses option ARMS with a major focus on residential lending in southern California. Rather than comparing it to companies like acquired GoldenWest or Washington Mutual, DSL’s valuations are better compared to small capitalization, focused option ARM originators like IndyMac (“IMB”) and BankAtlantic Bancorp (“BBX”).
IMB trades at 1.0x BV while BBX trades 0.9x BV. While IMB is in DSL’s region, BBX is similar in that it is an originator out of another highly volatile market, southern Florida. The main constant is that both rely on option ARMs in the residential market. However, while I have no position in either IMB or BBX, IMB’s average FICO score has stood above 700 in recent years while DSL’s was at 692 and 689 (2006 and 2005) based on their latest 10-Ks (IMB’s 10-Qs do not have the FICO information).
Further, IMB has set aside higher provisions for credit losses and allowances for loan losses against smaller asset bases. With a stronger FICO and more conservative accounting, IMB is valued at 1.0x BV while DSL is valued at 1.2x it’s GAAP BV of $52.58, suggesting that counter to analysts’ beliefs, DSL is actually overvalued on a relative BV basis against its true publicly traded peer group. Given all of the potential risks ahead for DSL and relative overvaluation against true comps, DSL still looks overvalued at $60.00 per share.
Disclosure: Author manages a hedge fund that is short DSL and owns puts on DSL.
DSL 1-yr chart: