Description
I am recommending a short of FirstFed Financial Corp (FED), a $1B market cap thrift based in southern California. The thesis is based on the fact that FED, Downey Savings (DSL), and GDW (Golden West) have very similar business models operating in the same markets (California)—purchasing adjustable rate mortgages (ARMs) from wholesale mortgage brokers and funding them with a combination of deposits raised through their branch networks (low interest rate paid) and the brokered CD/wholesale borrowing/FHLB market (higher interest rate paid). VIC readers interested in this space should be familiar with Grumpy922’s excellent write-up of DSL a few weeks ago. It provides a lot of good background on the option ARM market and why it is potentially so dangerous. Since then, DSL has plummeted from $80 to $64. (A front page article in the WSJ on Downey and a downgrade from Moors & Cabot shortly after the write-up were unanticipated, but very timely and welcome catalysts.) I would recommend reading it if you haven’t done so already. And because of the avalanche of recent reports on negative amortization and option ARMs, this note will not cover the basics of those topics.
While I also think DSL is still a short, I have decided to highlight FED against DSL to show why FED should trade at a discount to DSL, not the premium it currently does. (FED trades at 10.9 times 2006 EPS vs. only 9.9 for DSL and at 1.9 times book value vs. only 1.6 for DSL.) There are four reasons:
1) FED’s particular type of option ARMs are more risky because of extended low fixed payment periods.
2) It’s branch network is far smaller. This fact has been ignored because the recent growth in earnings from negative amortization has been so fantastic. However, over the longer term, branch operations have proven to be highly valued, particularly from an M&A perspective.
3) Negative amortization now makes up 30% of earnings. If we exclude negative amortization from the last three quarters, earnings growth has stalled despite the large increase in assets and leverage.
4) Underwriting standards have deteriorated rapidly.
Ultimately, I think FED should trade at least at DSL’s P/B multiple of 1.6, which translates into a $49.50 price target, or 16% below its current $59 price.
1) FED’s Peculiar Option ARMs
The 2Q/0510Q (page 10) indicated that $2.6B of its $6.3B in single family option ARMS are “3 or 5 year fixed payment”. That is 41% of the current single family mortgage portfolio and 30% of the total loan book.
Long Duration Fixed Payment Loans Are Increasing Rapidly
Q2/05 Q4/04 Q1/04
Total single family loans ($M) 6,302 4,586 2,961
3 or 5 year fixed payment loans ($M) 2,600 1,600 767
Total loans ($M) 8,719 6,883 4,998
Total Assets ($M 9,279 6,992 5,525
3&5 yr fixed pmt as % of 1-4 family 41% 35% 26%
3&5 yr fixed pmt as % of growth in '05 58%
3&5 yr fixed pmt as % of growth YOY 55%
Note: Q1/05 #s not available
When I first read this, I thought it was talking about the plain vanilla hybrid rate mortgages we hear about all the time (i.e. fixed rate mortgage for five years and then floating rate afterwards). But this is not so. The three and five year fixed term refers to the payment, not the rate. I must admit I was rather shocked when I understood. Most option ARMs have a minimum monthly payment that a borrower can make for one year, which then typically increases by a maximum of 7.5% each year afterwards (the cap rate). While this product is dangerous enough, FED is retaining an even more risky product that extends the term of the initial low payment for as long as five years. As far as I know, FED is the only lender who retains this particular type of option ARM. I asked Downey’s CFO about the product and he said they could not get comfortable with the potential amount of negative amortization that could occur over the five year period. Even Wall Street, which appears to have an endless appetite for ARMs, has not developed a secondary market for this particular type of option ARM. This means that FED is taking all the risk of this particular product, which could certainly create a problem of adverse selection because of the extremely attractive payment terms. To repeat, these loans now constitute 41% of FED’s total single family mortgage portfolio, up from 26% a year ago; they also constituted 55% of the loan growth in the first six months of 2005—clearly it is FED’s hottest product and one of the reasons why its portfolio is still growing and DSL’s has stalled. (FED’s assets grew 11% from Q1 to Q2 while DSL was down 2%). Without a doubt, FED is betting on continued real estate appreciation in one of the more “frothy” markets in the country.
2) FED’s Smaller Branch Network
FED’s branch network is far less developed with only 36 branches vs. DSL’s 172 (92 of which are in-store). DSL is not even twice as big on an asset basis ($16.6B vs. $9.3B), and while in-store branches probably aren’t quite as valuable as stand alone branches, DSL’s network is still over 4 times as large. DSL’s employee base is also about 4 times as large (2,500 vs. 630). If you believe that branches and people provide both tangible and intangible “franchise value”, then DSL is about 4 times more valuable (yet its market cap is only 80% larger at $1.8B vs. $1.0B). It is difficult to break out the profitability of the branches (neither company provides numbers), but recent M&A activity across the banking industry has indicated there is a lot of value placed on core deposit franchises, particularly now that short term interest rates are rising and the spread vs. funding in the wholesale market is much greater. It is something to think about, particularly given that DSL’s chairman, Maurice McAlister, is 80 years old and owns 20% of the stock. Here is a second way to think about it: because the business of retaining adjustable mortgages purchased from mortgage brokers appears so profitable currently (aided by negative amortization) and requires so little infrastructure, FED will come across as a far more efficient operating company. In the first half of 2005, FED’s 630 employees produced $40M of net income. By contrast, if we exclude DSL’s $46M in gain-on-sale income, its 2,500 employees yielded only $70M of net income. The difference is due almost entirely to DSL’s more extensive branch network. Viewed this way, FED looks increasingly like a mortgage REIT such as Thornburg Mortgage (TMA) or Annaly Mortgage (NLY), which trade at 1.3 and 1.2 times book, respectively.
3) FED’s Negative Amortization
FED’s negative amortization balance was $20.4M at the end of Q2/05, up from$10.3M at Q1/05 and $3.2M a year ago. Clearly, as the table shows, the growth is accelerating as interest rates rise.
FED’s Negative Amortization
Q2/04 Q3/04 Q4/04 Q1/05 Q2/05
Negative Amortization ($M) 3.2 4.0(est.) 5.6 10.3 20.4
Quarterly Increase in NegAm NA 0.8 1.6 4.7 10.1
After-tax Increase in NegAm NA 0.5 0.9 2.7 5.9
Net Income ($M) 16.2 18.2 16.4 18.5 21.7
NegAm as % of Net Income NA 3% 6% 15% 27%
Net Inc. - Increase in NegAm NA 17.7 15.5 15.8 15.8
Equity ($) 440 458 477 502 516
Tax Affected NegAm (at 42%) 1.9 2.3 3.2 6.0 11.8
NegAm as % of Equity 0.4% 0.5% 0.7% 1.2% 2.3%
One sell side analyst dismisses this, noting that the $20.4M constitutes only 23 basis points of FED’s $8.7B in net loans. What she does not mention is that at least 27% of earnings in the second quarter was from negative amortization. ($10.1M *(1 - 42% tax rate)) / net income of $22M.) It was probably higher. Unlike Downey, FED does not provide data on the amount of loans with negative amortization that were prepaid in the quarter. (Prepayments have the effect of reducing the existing negative amortization, thus masking how much negative amortization is flowing through the income statement.) About 14% of Downey’s Q1/05 negative amortization balance prepaid in the second quarter of 2005. If we assume the same rate for FED, that means there was an additional $1.4M of negative amortization that flowed through interest income ($10.3M * 14%), which means the total contribution to net income from negative amortization was actually 30%. This is a large figure that is comparable to Downey, and like Downey, will be even larger next quarter. (GDW, by the way, derived about 15% of its net income from negative amortization.)
4) FED’s Rapidly Deteriorating Underwriting Standards
There are four quality standards of underwriting that FED breaks out, with “verified income/verified asset” being the best and “no income/no asset” being the worst. Five years ago, almost everything was underwritten to the verified income/verified asset standard—the highest standard. As you can see in the table below, in the first six months of 2005, of the $2.3B in single family mortgages that were originated, 14% were originated with no income/no asset verification—the lowest standard—, up sharply from 6% in 2004. Management is not happy about it, but reports that the competitive environment requires them to underwrite this way. Comments from Golden West and others confirm this. This is clearly not a positive development for FED or the industry.
Residential Mortgage Origination According to Quality of Underwriting
Q1& Q2 Q1& Q2 Q1& Q2 Q1& Q2
2005($M) 2004($M) 2005(%) 2004(%)
Verified Income/Verified Asset 396 148 17% 15%
Stated Income/Verified Asset 758 264 33% 27%
Stated Income/Stated Asset 809 508 36% 52%
No Income/No Asset 314 60 14% 6%
Total 2,278 981 100% 100%
---FED vs. DSL Valuation---
The market has focused too much on FED’s 2006 projected earnings growth over 2005 (14%) vs. DSL’s negative 17%. The only reason DSL will have down earnings is because 2005 will be up an estimated 94% from 2004. (because of the gain-on-sale bonanza). However, looking at the two-year growth rate from 2004, DSL’s growth rate is projected to be notable higher than FED’s (66% vs. only 41%). (See the table below.) The premium doesn’t make sense because FED does not have any sort of sustainable competitive advantage. Let me stress again what the basic business of DSL and FED is—they call up independent mortgage brokers (or vice-versa) and buy variable rate mortgages that the brokers have originated through their contact with a person looking to refinance or purchase a home. FED and DSL rarely have any contact with the home-owner. They are almost purely a financing device which means the whole concept of brand and customer loyalty is non-existent. It is also the reason why FED has been able to grow its balance sheet by almost 70% in the last year while only adding only 28 employees (to 634 currently) and zero branches.
FED vs. DSL Earnings Metrics
2005 2006 Earnings
2004 2005 2006 2005 2006 Earnings Earnings Growth
EPS EPS EPS PE PE Growth Growth Since 2004
FED $3.85 $4.76 $5.43 12.4 10.9 24% 14% 41%
DSL $3.85 $7.45 $6.38 8.5 9.9 94% -17% 66%
On a price-to-tangible book ratio, FED trades at a 20% premium to DSL (1.9 vs. 1.6). The premiums make little sense in light of FED’s lower ROE in the first two quarters of 2005 (16% vs. 22%). Critics will point out that DSL’s 23% ROE is not sustainable because 2Q/05 had such a large amount of GOS. They are correct, but even if we exclude the entire $48.8M in GOS for the second quarter and assume it does not portfolio the loans (a rather severe assumption), DSL’s ROE would still have been a fairly respectable 13.6%. If we ease some of the GOS assumptions, it is easy to get to a 17-18% sustainable ROE, in line with FED. (FED also employs more leverage to get its 16% ROE—5.5% equity to assets vs. 6.7% for DSL (and 6.8% for GDW). FED just raised $50M in senior notes (which counts as Tier 1 capital) so that it would not invite the scrutiny of the regulators who get nervous with equity/asset ratios close to 5.0%.) FED should trade more or less in line with DSL.
FED vs. DSL Book Value and Return Metrics
Tang. Bk. Price/ Tangible First Half First Half
Per Tangible Equity/ 2005 2005
Share Book Assets ROA ROE
FED $30.94 1.9 5.50% 0.96% 16%
DSL $39.96 1.6 6.70% 1.38% 22%
Gain-on-Sale First Half Sustainable
as % of H1/05 2005 ROE ROE
Net Income (exclude GOS) (exclude GOS)
FED 0% 16% 17%
DSL 40% 13% 17%+
---More Reasons to Be Short the Group In General---
A flat or inverted yield curve is horrible for ARM lenders. The most common index off which option ARMS are priced is the average rate of the 1-month Treasury bill for the last twelve months (12MAT). This is currently 2.85% and will eventually catch up with the Fed Funds rate (currently 3.50% and likely to get to 4.00% at least). The spread that lenders charge on these loans is typically 200 to 280 basis points over the index, or 4.85% to 5.65% currently. 30-year fixed rate mortgages can be had at 5.80% currently, with 15-, 7-, and 5-year hybrids well below that. Over the next year, most rational consumers will refinance into the lower cost fixed rate mortgages. The consumers that do not are probably the ones making the minimum payment who cannot afford to make a fully amortizing payment even if the interest rate is lower. What many of the ARM lenders will be left with is a shrinking pool of mortgages that are worse and worse credit risks— a classic case of adverse selection.
Competition is increasing and irrational. The recent growth and success of FED, DSL, and GDW has not gone unnoticed by other players in the mortgage market. Countrywide is moving into the ARM market in a major way. IndyMac (NDE) is now aggressively offering a 12MAT loan with a 200 basis point spread vs. the 260+ spread that FED is currently booking. Because Washington Mutual sells everything off into the secondary market (i.e. retains no risk), it can afford to price irrationally (ask any competitor and they will confirm this). It is happy just to generate the gain-on-sale. The entire ARM market is very transparent with very low barriers to entry. Variable rate mortgages as a percentage of total mortgage originations are also close to a peak (30%), and with the flattening yield curve, this percentage (and the entire ARM market) will shrink fast. When volumes dry up, so will margins.
The regulators hate negative amortization. They banned it from the credit card industry several years ago, and it is only a matter of time before they do the same with mortgages. Herb Sandler, the influential CEO of Golden West, has the regulators’ ear and will tell you that he himself is asking for intervention. Greenspan and company also realize that jacking up short term interest rates too much (more than 4%) merely to cool off the California and Florida property markets would be comparable to administering chemotherapy for a common cold. With contained inflation and low long term rates, the rest of the economy and world does not need higher short-term rates. Regulatory fiat is a far more precise and effective way to target the real estate speculation occurring on the coasts. Expect some pronouncements this fall.
---Other Notes---
Golden West actually trades at a higher valuation than FED (2.5 times book and 11.8 times 2006 EPS), but Herb and Marion Sandler have justifiably earned reputations as legends, having successfully guided Golden West through many different economic, real estate, and interest rate environments over the last 40 years. I think the stock will also come under pressure for the remainder of the year, but I believe GDW is a best-in-class operator and will largely avoid the pitfalls that DSL and FED face (i.e. because of poor underwriting).
Most of these ARM lenders say that these minimum payment option ARMs are great for that BMW salesman with volatile month-to-month income or the consultant who gets paid a big bonus at the end of the year. We all know that line is complete bulls--t. Borrowers are taking out option ARMs because they want to speculate on real estate appreciation.
Catalyst
DSL’s sell-off and increased scrutiny of its peers
Regulatory restrictions on option ARMS
Slowing growth of variable rate mortgage originations due to flat yield curve