Description
Short Idea: Downey Financial (DSL)
$2.2B market cap Savings and Loan based in southern California
Consensus EPS
2005 2006
IBES EPS $7.53 $6.57
P/E 10.5X 12.0X
P/B 2.0X
Executive Summary:
DSL is in a cyclical business and its earnings have been soaring but its valuation metrics have expanded, not contracted as it hits peak EPS. Earnings power has been flattered by the housing bubble in California, but risks of stagnant or even falling home prices has not yet impacted the share price. Low quality earnings, in the form of gain on sale of Option ARM pools and non-cash negative amortization on loans, have expanded significantly as a percentage of the total, but to date, this has not negatively effected how investors value the stock. A combination of changes at the rating and regulator agencies is likely to compress earnings power in the next six months. Finally current trends in the interest rate markets could lead to a significant decline in both the assets and earnings power of DSL over the next year. Unless the company can find a buyer willing to take on these risks at a premium to the current share price, shorts will likely be richly rewarded over the next six months.
Downey’s recent success
DSL is a mid-sized thrift based in Newport Beach, CA. Earnings have been soaring and beating expectations as of late due to the strength of the residential real estate market as well as a shift in mortgage originations to DSL’s primary product, the Option ARM. There are several reasons why 2Q05 earnings will not only be the high point of the year, but likely the last positive surprise we will see through the end of 2006.
- Why has DSL seen such great EPS growth?
Quarter IBES Consensus Actual Surprise
2Q04 $.97 $.99 2%
3Q04 $.94 $.99 5%
4Q04 $1.01 $1.66 64%
1Q05 $1.19 $1.86 56%
2Q05 $1.54 $2.25 47%
Thrifts are traditionally cyclical earnings machines, and DSL has clearly been in a period where EPS is growing quickly and well ahead of expectations. Several factors have been driving this trend.
- Loss rates on mortgages have fallen to extremely low levels due to the fast pace of appreciation on single family homes in California
- Despite extremely low long term interest rates, borrowers have been shifting to short term adjustable mortgages, and DSL’s primary product, the Option ARM
- A secondary market has recently developed where DSL is able to securitze and sell for a significant gain on sale much of its production – business that otherwise would have been turned away by the company as it did not have the capital to portfolio the amount of loans being originated
- DSL has dramatically reduced the amount of documentation needed for a borrower to secure a mortgage. In a recent securitization, DSL released that 80% of all loans were made using ‘Lite’ documentation (stated vs. verified income data), 14% used ‘Alternative’ or ‘Express’ documentation, and only 5% used ‘Full’ documentation. In general, anything less than full documentation is due to a borrower attempting to make himself appear more credit-worthy so as to secure the desired mortgage
So why is EPS growth at DSL coming to an end?
1) California residential real estate prices are nearing a peak.
We have all been reading about the impending collapse of the real estate bubble. I am not willing to predict when the bubble will pop, but am convinced that prices are close to a peak in California due to how stretched borrowers are today. Let’s consider some statistics –
Alt-A Sub-Prime Jumbo
IO Lim Doc % CA IO Lim Doc % CA IO Lim Doc %CA
2000 0% 60% 35% 0% 25% 22% 0% 25% 41%
2001 3 57 41 0 27 26 5 24 46
2002 7 55 43 1 30 28 20 30 49
2003 23 58 41 2 35 31 35 30 49
2004 53 57 42 11 33 33 54 45 54
Source: Lehman Brothers, Fannie Mae (Alt-A = Non-prime, non-conforming borrowers limited credit histories and/or documentation, Sub-Prime = borrowers with poor credit histories, Jumbo = mortgages too large to conform to Fannie Mae underwriting standards, IO = Interest only loan, Lim Doc = Limited documentation on borrower, % CA = % of total loans made on properties in California)
Great places to buy a house today?
IO Share ARM Share Investor % 2004 Appreciation Appreciation
San Diego, CA 48% 65% 16% 30%
San Francisco, CA 45 83 20 16
Oakland, CA 43 79 15 18
San Jose, CA 41 83 15 15
Orange County, CA 37 79 15 18
Ventura, CA 35 82 19 27
Sacramento, CA 35 65 16 25
Las Vegas, NV 35 69 18 36
Palm Beach/Boca, FL 28 54 10 21
National Average 23 48 8 12
Source: Lehman Brothers, Fannie Mae, LoanPerformance, Inside Mortgage Finance (IO & ARM share = % of Interest only or adjustable rate mortgages as a percent of total outstanding, Investor % = % of loans made to borrowers who plan to rent out the property in hopes of capital appreciation, 2004 Appreciation = % increase in average home price in 2004)
Despite very low long term interest rates, borrowers are stretching to buy the largest home possible by avoiding traditional fixed rate amortizing mortgages and choosing the potentially temporary lower payment option of adjustable-rate amortizing, interest only, or even non-amortizing payment loans. An increasing portion of these mortgages are also being taken out by investors who will not be living in the house but are buying in the hope of future price increases.
I believe it is clear from these numbers that a combination of borrowers stretching to buy homes at current prices, rising short term interest rates which ratchet up payments on ARM loans, significant recent price appreciation, a large increase in the percentage of loans made with only limited documentation, rising regulatory pressure (see point 3 below), deteriorating borrower documentation and a rising percentage of total US mortgages being underwritten in California is a strong indication that home prices in the Golden State are near at least a short term peak.
DSL’s loss ratios have been extremely low – only a few bps per quarter (over the last year the non-performing loans as a % of total assets has fallen from a low 0.25% to a tiny 0.17%). As home prices stop rising (not fall – just stop rising), losses will rise and Provision Expense will increase at DSL and cut into EPS.
2) Gain On Sale on Option ARMs are under attack
For those of you who have not heard of an Option ARM – this is a mortgage that allows the borrower to make one of four different payments each month – a 30 year amortizing payment, a 15 year amortizing payment, an Interest Only payment, or the ‘Option’ payment which is currently set at an amazingly low 1% to 1.5% annual interest rate. These loans started out in California more than a decade ago as a niche product but have now become mainstream here and are spreading to other hot markets such as the northeast and Florida.
DSL originations are primarily Option ARMs. The thrift has two types – those that have an interest rate tied to an index of monthly California cost of deposits (which are held on balance sheet), and those that are tied to an index of the monthly Treasury market rates (which are securitized and sold for a gain). This later business has been of increasing importance to DSL’s P&L as you can see from this table –
Q104 Q204 Q304 Q404 Q105 Q205
GOS Amount ($M) 1.4 15.7 14.6 22.8 30.6 48.9
GOS Margin 0.7% 1.1% 0.6% 0.9% 1.4% 1.5%
Loans Sold ($M) 203 1,454 2,281 2,621 2,250 3,100
GOS % of EPS 9% 34% 36% 30% 35% 46%
First, Gain on Sale earnings are of low quality and should be given a low P/E by the Street (which has not happened to date). Second, Standard and Poor’s has become increasingly concerned about the risks of Option ARM securitizations. On 6/22/05 S&P announced that it will require additional credit enhancement on Option ARM pools. Fitch is also tightening its standards on these pools. As more capital must be retained, the Gain on Sale of these pools will inevitably decline. According to sources at bulge-bracket trading desks, prices of these issues have already fallen ½ to a 1 point which would obviously wipe out most of the GoS that DSL is recording on the loans it sells. This change came too late to effect 2Q05, but the pricing adjustment will become apparent and reduce EPS in 2H05.
3) Option ARM and I/O underwriting are under regulatory pressure
Early in June, Chairman Greenspan gave testimony on that included the following – “The apparent froth in housing markets may have spilled over into mortgage markets. The dramatic increase in the prevalence of interest-only loans, as well as the introduction of other relatively exotic forms of adjustable-rate mortgages, are developments of particular concern. To be sure, these financing vehicles have their appropriate uses. But to the extent that some households may be employing these instruments to purchase a home that would otherwise be unaffordable, their use is beginning to add to the pressures in the marketplace.”
Later in the month it was announced that the OCC is taking the lead for the bank regulatory agencies in writing new guidelines to banks and bank examiners regarding tightening lending standards in the home equity lending and Interest Only mortgage space.
Clearly the Fed wants to cool the housing market, but the ‘conundrum’ of rising short term rates leading to lower long term rates means that higher Fed Fund rates is not helping the situation. Might regulatory fiat be a better way to slow home price appreciation? A few years back, the bank regulators (OCC, OTS, FED, FDIC) became concerned with negative amortization rates on many credit card accounts. With a wave of a pen, the bank examiners forced banks to institute higher monthly payments on credit card balances that amortized the outstandings. From discussions with representatives at the OCC and the OTS, I believe something similar is soon to occur with negative amortization Option ARM mortgages (the ‘relatively exotic forms of adjustable-rate mortgages’ mentioned by Mr. Greenspan). While Greenspan recently wrote that regulatory action is not designed to pop the real estate bubble, it is quite likely that the change may help to do just that.
As almost all of DSL’s mortgages are Option ARMs, greater regulator restrictions on this product would have an immediate and potentially dramatic negative impact on EPS at the thrift.
So what could the regulators do? They could mandate no more negative amortization is allowed in future loans. They could demand much lower LTV ratios on new Option ARMs, or they could mandate much better documentation. Any or all of these changes would be very painful to DSL’s EPS stream.
4) Non Cash Earnings are growing quickly as a % of DSL’s Earnings
Now let’s delve a little deeper into DSL’s business. The thrift underwrites Option ARMs up to 80% LTV (loan to value – the portfolio average is currently in the 70s) and will allow the borrower to pay the ‘option’ rate of 1% and accept negative amortization to take place until the loan grows to 110% of its original amount. However, due to GAAP accounting, even though the borrower is only paying 1%, DSL records on its books the negative amortization as current interest income. This non-cash earnings stream has been growing quickly –
Q104 Q204 Q304 Q404 Q105
Interest Income ($M) 120.8 131.2 149.2 166.5 180.4 Neg Amortization ($M) 1.6 3.4 6.6 11.4 18.7 Net Interest Income($M) 69.4 76.6 81.9 90 94.1 Neg AM % of Net Int Inc. 2% 4% 8% 13% 20%
Q205 Neg Am amounts will be released in DSL’s 10-Q, From talking to management of the company and other Option ARM lenders we can expect another major increase when the data is released. Clearly negative amortization of the loan book is likely to lead to higher losses later, and is a low quality source of earnings. Also, if the loans default it could lead to a restatement on interest already booked as revenue over the last several quarters.
So over the last year we have seen an explosion in combined non-cash and gain on sales earnings at DSL from 20% to almost 70% of earnings. This growth accounts for more than 100% of the recent earnings surprises reported vs. Street expectations. 2Q05 saw a one quarter jump in non-cash and gain on sale earnings from 55% to around 70% of EPS and this will should be the high water mark as the above mentioned changes in the securitization market and regulatory pressure hits in the second half of the year.
Despite this severe decline in earnings quality, we have not seen the valuation multiples on DSL compress– in fact the opposite has occurred. Since 12/31/03, the Price/Book on DSL has increased from 1.4X to its current level of 2.0X (a 40% increase), and its dividend yield has declined from 0.75% to 0.5% today. DSL’s forward P/E remains at 12.1X despite this decline in earnings quality. Mortgage companies with large amounts of gain on sale and/or non-cash earnings generally trade on 2006 P/Es of 5-9X and significantly lower P/B values than the 2X currently at DSL. Examples include CFC, RGF, NEW, DRL, LEND, & NDE.
As regulators clamp down on Option ARMs and the new S&P collateralization requirements hit the securitization market we will likely see EPS and ROE fall at DSL by at least 20% YoY. As earnings pressure occurs, I believe consensus EPS for 2006 could drop from today’s $6.57 by 25% to $5.00 by December. Assuming that this drop in profitability leads to a decline in forward P/E to 11X yields a share price of $55, or a 30% drop from current share price by the end of the year. Of course, if home prices in California start to flatten out, and loss rates begin to tick up, then even this $5.00 EPS for 2006 could be too high.
5) Asset quality can only get worse
DSL has outstanding asset quality at the moment:
6/30/05 12/31/05 6/30/04
Prime Mortgages ($M) $12,481 $11,509 10,721
Nonperforming ($M) $12 $21 $25 Nonperforming % 0.1% 0.2% 0.23% Subprime mortgages ($M) $1,262 $1,255 $1,076 Nonperforming ($M) $11 $11 $13 Nonperforming % 0.9% 0.9% 1.2%
Much has already been written about the potential for housing prices to fall in the US. Suffice to say, that asset quality can’t really get much better than 10bps of problem assets in DSL’s core portfolio. If home prices simply flatten out and short term rates rise over the next year, there will be a large jump in problem assets requiring higher provision expense at DSL. If home prices in California actually start to fall then the carnage could be significant.
One more dark cloud on the horizon for DSL
To make money on this short, I could be wrong on everything written above, and still win if Mr. Greenspan’s ‘conundrum’ continues. The 10-year Treasury is currently hovering in the 4.0- 4.3% range. Were it to fall to 3.50-3.60% in yield, most industry watchers expect another massive refinance boom to occur. The wholesale and correspondent mortgage channels (where DSL sources most of its loan production) would become clogged with fixed rate borrowers attempting to refinance. This would temporarily crowd out much of the Option ARM originations occurring and reduce DSL’s gain on sale earnings. Also, this refinance wave would lead to much of DSL’s $13B portfolio loans prepaying as borrowers switch from ARMs into fixed rate mortgages resulting in falling net interest income.
Even if the 10-year bond yield remains flat and doesn’t decline, as the Fed raises overnight rates, we could still see a major loss of income at DSL. This is because as the yield curve flattens, there may be significant refinance activity as many large banks with significant California operations (including Bank America, US Bank, and Wells Fargo) are rolling out long-term fixed rate interest only mortgages. Rational Option ARM borrowers making interest only payments will refinance from a 1-month adjustable interest only rate loan into a fixed rate interest only mortgage if given the chance to do so at very close to the same current monthly payment.
Thus, a continuing flattening of the yield curve could hurt DSL’s production and portfolio, and falling 10-year yields could trigger another refi boom that would be even more destructive to the EPS stream.
The Big Risk to a Short – DSL sells
DSL stock has moved up more than 25% over the last three months on takeover speculation. There has been significant insider buying in this period of time, and a sale certainly could be imminent. The aging Chairman of the thrift, Maurice Mcalister owns more than 20% of DSL and Cheryl Olson, the Vice Chairman, owns an additional 2%. Others have noticed that DSL’s earnings power may be under pressure and the short interest has moved up from 500K shares to almost 1.5M over the last few months (out of 28M outstanding). Management might see the writing on the wall and try to sell out for the best possible price before EPS collapses.
So how much would the company go for in a sale? It is difficult to say that DSL has a fantastic franchise – it makes very few mortgage loans through its branch network, and these offices are primarily used to source checking accounts and CDs (May 05 average cost of deposits was 2.4%). It is also difficult to say that DSL’s mortgage origination unit, which mostly sources assets from the correspondent and wholesale channel, is of the highest quality. So how about the implied premium to deposits ratio? In May of ’05, DSL had $10.6B in deposits and a market capitalization of $2.2B yielding a deposit premium of around 11%. If you strip out the value of the headquarters building in Newport Beach, the premium drops to around 10%.
This is not a big premium for sure, but it is still pretty pricey for DSL when you consider that stronger franchises such as SUSQ, CMA, FRC, and CBH trade on similar deposit premiums with higher-value commercial banking activities, and very low levels of gain on sale or non-cash earnings. The large French bank, BNP recently announced its intention to acquire a mortgage-oriented thrift, Commercial Federal Corp (CFB). Pricing for this deal was around an 11% deposit premium, 1.7X P/B and 14.5X forward P/E. Certainly it is possible that DSL could go for more than its current high $70s share price, but much of the premium of an acquisition of the thrift is already in the stock after its strong recent run.
Conclusion:
DSL is in a cyclical business and its earnings have been soaring but its valuation metrics have expanded, not contracted as its moves towards peak EPS. Earnings power has also been flattered by the housing bubble in California, but risks of stagnant or even falling home prices has not yet impacted the share price. Low quality earnings, in the form of gain on sale of Option ARM pools and non-cash negative amortization on loans, have expanded significantly as a percentage of the total, but to date, this has not negatively effected how investors value the stock. A combination of changes at the rating and regulator agencies is likely to compress earnings power in the next six months. Finally current trends in the interest rate markets could lead to a significant decline in both the assets and earnings power of DSL over the next year. Unless the company can find a buyer willing to take on these risks at a premium to the current share price, shorts will likely be richly rewarded.
Regards, Grumpy
"
Catalyst
DSL is in a cyclical business and its earnings have been soaring but its valuation metrics have expanded, not contracted as it hits peak EPS. Earnings power has been flattered by the housing bubble in California, but risks of stagnant or even falling home prices has not yet impacted the share price. Low quality earnings, in the form of gain on sale of Option ARM pools and non-cash negative amortization on loans, have expanded significantly as a percentage of the total, but to date, this has not negatively effected how investors value the stock. A combination of changes at the rating and regulator agencies is likely to compress earnings power in the next six months. Finally current trends in the interest rate markets could lead to a significant decline in both the assets and earnings power of DSL over the next year. Unless the company can find a buyer willing to take on these risks at a premium to the current share price, shorts will likely be richly rewarded over the next six months.