Description

How would you like an opportunity to invest $22.30 today, and to get $5.70 in cash dividends over the next 27 months, followed by growing cash dividends of $3.00+ annually thereafter? This should result in a total return over the next 12-27 months of 55-90%, assuming the stock trades at a dividend yield comparable to peers. CapitalSource (CSE) offers such an opportunity today.

This is a company that has excellent management, and its largest shareholders are two world-class investment firms, Farallon Capital and Madison Dearborn, who backed the company when it was founded and are also represented on the board. There is also a very clear catalyst in the near future, namely the conversion to the company to a REIT. This catalyst will cause us to realize the returns projected above.

Investment Summary
CapitalSource is a commercial finance company that commenced operations in 2000 with $511 million of equity capital, and with the backing of several well-regarded institutional investors. The company has grown rapidly and profitably over the last five years. The company recently announced that it will elect to be treated as a REIT starting in 2006. The implication of this is that it will be required to pay out most of its profits each year as dividends, and will not pay taxes on the bulk of its profits. Management has projected that the cash dividend will be $2.15 in 2006 and $2.90 in 2007. In addition, as part of the conversion to REIT status, the company will have to pay a one-time dividend of all the earnings it has retained over time. Management has estimated that this “E&P dividend” will amount to between $325 million and $375 million, or about $2.30-$2.70 per share, of which at least 20% has to be paid out in cash with the balance paid in stock (which for practical purposes will be the equivalent of a stock-split). The cash portion of the dividend is expected to be about $0.65, and has to be paid before the end of January, and the regular quarterly dividend will commence in the first quarter of 2006.

At its core, my thesis is very simple. If the dividends projected by management are delivered, the stock will trade at a price that reflects the yield, comparable to other REITs. If the annual dividend rate is $2.90 by the end of 2007, and the yield is between 8% and 10%, the stock price will be between $29 and $36.25. In addition you will receive cash dividends of $5.70 in 2006 and 2007 (the one-time “E&P” dividend of $0.65, plus $2.15 in 2006, plus $2.90 in 2007). From the current price of $22.30, this translates to a total return of 55% to 90% by end-2007, i.e. in 27 months.

Some of the key aspects of my analysis worth noting follow:
* Management is outstanding. The CEO built a similar business focusing on health-care financing in the 1990s, and sold it to Heller Financial in 1999 for $483 million. Most other key members of the management team have been at the company from the start, and have impressive credentials and experience.
* Expectations created by management in the past have tended to be conservative. In my view, management has a lot of credibility, and I am inclined to believe that they will achieve the earnings and dividend goals they have established. The achievement of these goals are critical to my thesis.
* Any lending business is inherently risky, and I get comfort from the company’s cautious approach to lending. Virtually all of the loans are senior secured loans. The few problem loans that have occurred have had very high rates of recovery of principal (averaging 89%).
* Related to the previous point: When the company is considering a loan, the borrower is simultaneously reviewed by two groups within the company, reporting to different bosses. The Business Lending Group is responsible for recommending the underwriting of a loan. Simultaneously, the CapitalAnalytics group (which includes 70+ professionals, mainly CPAs, and is in effect an in-house accounting firm) is reviewing the borrower’s business, historical financial statements, projections, etc. with a cynical eye, to assess the integrity of the financial statements. This dual-track approach to underwriting significantly reduces the likelihood of problem loans.
* Management is conservative in leveraging the company, with a debt to equity target of between 4:1 and 5:1. This is significantly less leverage than is possible, because management is determined to maintain a margin of safety on the balance sheet.
* The management team, and particularly the top two guys, has very large equity stakes in the company, and, along with the two largest institutional promoters, purchased 4.3 million shares at $22.30 in the recent secondary offering.
* Farallon Capital and Madison Dearborn own over 25% of the company, valued at over $750 million, and purchased additional shares in the secondary offering. Their presence as large shareholders and on the Board is increases the odds of a favorable outcome.


Ultimately, the success of this investment boils down to two questions:
1. Can the company achieve the levels of dividends projected for the next two years and convince investors that earnings and dividends will continue to grow from there?
2. Will the stock trade at the prices I’ve suggested above, assuming the dividends come through?

I’ll address the second question first: I think my 8% to 10% yield assumptions are reasonable, and perhaps conservative. The company has provided data on comparables in their presentation to analysts on September 21, which is available on their website (a 4-hour audio recording of the presentation and a 100+ page pdf document). CSE has the highest ROE, and is growing faster than its commercial lending REIT peers, which trade at yields of 7.5% to 10% next year’s dividends. With my 8% to 10% assumption, a stock price of $29 to $36 appears very plausible.

The rest of this write-up will be devoted to addressing the first question, beginning with some background on the company.

History and Management
CapitalSource has an impressive management team and Board of Directors, and a blue-chip list of founding investors. The company was formed in 2000 with funding primarily by a handful of institutional investors. I estimate that the institutional investors paid about $6 per share. The company went public in August 2003 at $14.50. CSE sold 16 million shares in the IPO, and the investors sold total of 5.3 million shares, with each one selling about 6.7% of their stake. Management did not sell any shares in the IPO. Farallon Capital Management (with its affiliates) was the largest investor with 32 million shares, or 32.2% of the pre-IPO outstanding shares; immediately following the IPO it owned 25.8%, and as of March 2005 it owned 19.4%. Farallon is represented on CSE’s Board by Thomas Steyer and Andrew Fremder. Madison Dearborn Partners owned 18.4% pre-IPO, 14.8% post-IPO, and 11.3% as of March; Paul Wood and Timothy Hurd from MDB are on the Board. Some of the other initial investors were Highfields Capital Management, Wachovia Capital Partners, Och-Ziff, and Tully Friedman/Friedman Fleischer & Lowe. Tully Friedman, who co-founded Hellman & Friedman, is also on the Board.

The two key founders of CSE are John Delaney (CEO), and Jason Fish (President). Prior to founding CapitalSource, John Delaney was the founder/CEO of HealthCare Financial Partners, a provider of commercial financing to small and medium sized healthcare service companies, which was eventually sold to Heller Financial in 1999 for $483MM. Jason Fish was previously a partner at Farallon Capital, responsible for real estate activities and involved in both credit and private equity investing. CSE’s management team is made up of a dozen people, most of whom in one way or another worked with John Delaney at HealthCare Financial. Bryan Corsini, chief credit officer, is a former Fleet Capital Corp. lender who lent money to HealthCare Financial. Managing directors Dean Graham and Joseph Kenary and Chief Technology Officer Chris Woods all worked for HealthCare Financial under Delaney.

Members of the management team own a significant stake in the company. John Delaney owned 6.8 million shares and Jason Fish 6.3 million shares as of March 1, 2005. Other senior managers own smaller, but still significant, stakes. Delaney, Fish, Farallon, Madison Dearborn, and other Board members together purchased 4.3 million shares of the 17.5 million shares priced late last week at $22.30 in a secondary offering.

The Business
CapitalSource considers itself to be “the leading commercial finance company” in the country. It focuses on lending to small and medium-sized businesses, typically businesses with $5 million to $500 million in revenues. It currently has about 460 employees and $5.4 billion in loans outstanding. The company offers a variety of products, including senior secured asset based and cash flow loans, mortgage loans, term B/second-lien/mezzanine loans, and private equity co-investments. Their specialty is providing customized, negotiated debt financings, i.e. non-vanilla loans, where they can get premium pricing.

Loan size is generally $1 million to $50 million (though some loans are larger), with an average loan size of $6-$7 million, and a maturity of 2 to 5 years. Virtually all of the loans are variable-rate loans, generally tied to the prime rate or LIBOR. Most of the company’s borrowings are also at variable rates, so that the company believes it has minimal exposure to changes in interest rates.

Management attempts to identify specific markets which they feel are under-served, and where they can make high risk-adjusted returns by providing services that help the borrower get their transaction done. Many of the transactions are complicated, “high-touch” situations such as acquisitions or recapitalizations, which are high value-added financings that are less price-sensitive. The nature of these transactions also results in a fair amount of turnover in the portfolio, from loans being prepaid, and this tends to lead to acceleration of fees, resulting in a higher return for the company. For example, the average “yield” (or interest income plus fees) on the interest-earning assets was 12.04% in the first half of this year, indicating the premium pricing ability of the company.

[We pause here for a joke: A frog goes to his bank for a loan. The loan officer, whose name is Patricia Wack, asks him if he has any collateral to offer as security. From his pocket the frog pulls out an old tattered photograph of himself with his entire family: parents, siblings, cousins, uncles, aunts, etc. The loan officer looks at it, and then takes it to her boss. She explains the situation to him. The boss responds: “It’s a knick-knack, Patti Wack; give the frog a loan.”]

Happily, CapitalSource is much more careful than the frog’s bank in seeking security for the loans they make. Over 95% of their $5.1 billion loan portfolio at end-June was senior secured loans, with less than 5% being other types of loans. The senior secured loans are roughly equally divided into three types of loans:

(a) Senior secured asset-based loans. These are secured by specific assets of the borrower, such as accounts receivable, inventory, and machinery & equipment.

(b) Senior secured cash flow loans: These loans are based on CSE’s assessment of a borrower’s ability to generate cash flow sufficient to repay the loan and to maintain or increase its enterprise value during the term of the loan. These loans are generally also secured by a security interest in the client’s assets.

(c) First mortgage loans: which are secured by first mortgages on the real estate of the borrower. Clients consist of owners and operators of hospitals, senior housing and skilled nursing facilities; owners and operators of office, industrial, hospitality, multi-family and residential properties; resort and residential developers; and companies backed by private equity firms that frequently take out mortgages in connection with buyout transactions.

Less than 5% of the loans are not senior secured. CSE makes Term B, second lien and, to a lesser extent, mezzanine loans. Many of the company’s loans are to businesses being acquired by a private equity firm, and the company sometimes purchases equity in a borrower at the same time and on substantially the same terms as one of their private equity sponsor clients. These equity purchases generally range from $250,000 to $2 million for any client. They sometimes also obtain warrants to purchase equity in their borrowers, usually exercisable at a nominal price. These equity investments together make up less than 1% of the assets on the balance sheet.

Note that CSE is very different from Allied Capital and American Capital Strategies, in that it is almost exclusively a lender, with minimal equity investments, and is therefore not subject to management’s discretion in valuing investments.

One good indicator of CSE’s caution in lending is the recovery rates on problem loans. Loans are written to a zero loss tolerance, but some problems are inevitable in this business. In these cases, the security behind the loans is key. From inception through August 2005, $7.3 billion in loans had been originated. As of June 30th, the company had reported 21 troubled loans, i.e. delinquent/non-accrual, totaling $255 million. Of these 21 loans, 13 (totaling $132 million) had been resolved, with a net recovery of principal of 89%. Of the 13 resolved loans, the first mortgage loans had the highest net recovery (96%), followed by senior secured asset-based (93%), and senior secured cash flow loans (80%). These are very high recovery rates. Management has said that they expect the remaining 8 unresolved loans to have similar recovery rates.

All loan proposals are simultaneously reviewed by two independent groups within the company. While the Business Lending Group involved (such as Corporate Finance or Structured Finance) is analyzing the borrower’s business with a view to originating the loan, a separate group called CapitalAnalytics (reporting to an independent Chief Credit Officer) is taking a hard cynical look at the financial statements of the borrower, and at the company’s projections. Bear in mind that most of the borrowers are small and medium sized businesses, without the kind of controls in place at larger companies. CapitalAnalytics has over 70 professionals, primarily CPAs, and serves as CSE’s in-house accounting firm, with audit capabilities. Its purpose is to provide forensic due diligence designed to minimize losses because of fraud and aggressive accounting. After a loan has been made, CapitalAnalytics also monitors the borrower on an ongoing basis, to catch problems early. The group is managed as a separate P&L center within CSE, and its costs are passed on to clients.

The Loan Portfolio & Credit Risk
From the founding of the company in September 2000, the loan portfolio had grown to $1.99 billion at the end of 3Q03 (right after the IPO), and $5.07 billion at the end of June 2005. Management sees significant opportunities in its current markets and in other specialized areas, and expects the portfolio to grow significantly in coming years, to over $8 billion by the end of 2006, and about $11 billion in 2007. This rapid growth clearly has risks, but everything I see suggests prudent lending standards that have not deteriorated with time.

Clearly, the biggest risk in a business of this nature is credit risk. A significant increase in bad loans could wipe out the profitability of any lender. But CapitalSource has demonstrated that it is a cautious lender. The discussion on the CapitalAnalytics group above is one indication of the care taken in originating loans. Management’s significant equity stake in the business also decreases the likelihood of imprudent lending.

But a few troubled loans are inevitable in this business. As CSE’s portfolio has matured, there has been a gradual increase in non-accrual loans, from 1.16% in mid-2004 to 2.20% at August 2005. Management has increased loss reserve levels to recognize this increase in problem loans. They forecast charge-offs of about 50 bps of loans annually, and have adequately provided for this. Again, the recovery rates of problem loans has been very high, because of the senior secured nature of the lending.

Interest Rate Risk
The vast majority of both CapitalSource’s loans and its borrowings are variable rate. As of June 30th, 94% of the company’s loans to its customers were tied to prime or LIBOR. In addition, 61% of loans have interest rate floors, so that the company should benefit from rising short-term rates. Of the $4.34 billion in debt at end-June, $3.79 billion is variable rate, tied to LIBOR or CP rates ($1.32 billion in credit facilities from banks and $2.47 billion of term debt raised by securitizing packages of loans). The remaining $555 million of borrowings is from two issues of convertible debt issued in 2004, $225 million at 1.25% and $330 million at 3.5%. As a result of both assets and liabilities being predominantly adjustable rate, management believes that the company has little interest rate risk, and should benefit modestly from an increase in short-term rates.

Future Capital Requirements
Historically, CapitalSource has had to raise money in the debt markets on an ongoing basis to finance its lending, and will continue to have to do so. One of the implications of the REIT election is that the company will now also have to raise equity capital, if it wants to keep growing, because most of its profits will have to be paid out as dividends. In order to maintain a debt to equity ratio of about 4.5:1, which is the target, the company will have to sell new stock regularly, perhaps annually, if it wants to grow the balance sheet.

The company last week sold 17.5 million shares in a secondary offering at $22.30/share, raising about $377 million (before the greenshoe option). Management has also stated that they expect to raise $630 million in equity capital next year, and perhaps $550 million in 2007.

This necessary ongoing dilution of the shareholders of growing REITs has discouraged me from investing in REITs in the past. But in analyzing CSE, I realized that under certain circumstances they can be very good investments. Here’s what I mean: If a REIT can earn a high ROE, its stock is likely to trade well above book value. In that case, increasing shareholder’s equity by some percentage requires less than proportionate dilution. For example: Say a company has 100 shares outstanding with a BV/share of $10, for a total of $1000 in shareholders’ equity. Let’s assume this company is earning a 20% ROE, or $200, or EPS of $2.00. The stock is likely to trade well above its $10 BV. Now assume the entire earnings are paid out as dividends, and the stock trades at $25, reflecting an 8% dividend yield.

If the company wants to grow its business by 20% next year without increasing its leverage ratio, it will have to raise $200 in equity capital. At the $25 stock price, it will have to sell 8 shares, for 8% dilution in shares outstanding. If it can earn the 20% ROE again next year, total earnings will be $240, and with 108 shares now outstanding, EPS will be $2.22. So a shareholder in this company benefits from the 8% cash dividend, and also from the 11% EPS growth, which will presumably translate into an 11% growth in dividends and a higher stock price.

The key, of course, is the high ROE, which translates into higher earnings, a higher dividend, and a higher stock price (relative to BV), thus requiring less dilution to grow the company. CSE’s ROE has been growing steadily since inception, as it has gradually put its capital to use. I’m focusing here on pre-tax ROE because the REIT election will cause the company’s tax rate to decline sharply over the next few years as it takes advantage of the REIT structure:

Pre-tax ROE
2Q05 29.3%
1Q05 27.2%

2004 23.3%
2003 19.6%
2002 12.2%
2001 4.9%

So can the company earn high returns on equity (20%+) over time? The company’s stated goal is to earn pre-tax ROE of at least 20%. I think that the company should be able to earn 20%+ ROE for some years. I’ve attempted to create a basic model for the business, using assumptions I consider realistic, not overly optimistic or pessimistic.

100 = Equity
450 = Debt (assuming a debt to equity ratio of 4.5:1)
550 = Total Assets
523 = Loans (assumes 95% of the assets are loans, the rest is cash)

60.1 = Interest & fee income (assuming an average yield of 11.5%)
21.6 = Interest expense (assumes average 4.8% interest cost)
38.5 = Net interest & fee income
5.2 = Provision for loan losses (PLL) (assumes provision each year of 1% of loans)
33.3 = Net interest & fee income after PLL
3.2 = Total other income (0.6% of loans)
11.0 = Operating expenses (assuming expense ratio of 2% of assets)
25.5 = Net income before taxes

i.e. 25.5% pre-tax ROE

A few comments on the assumptions: Management’s debt to equity target is between 4:1 and 5:1. The 11.5% yield (which includes interest and fees) and 4.8% interest cost are management’s guidance for 2006; both of these are more conservative than achieved in recent years. The annual loss provision of 1% is higher than management’s expectations. Operating expenses as a percent of assets have been declining, and are expected to be below 2% annually by the end of next year.

One can debate these assumptions. The net interest spread (yield minus interest cost) and the loss provision are the two key factors that can affect the ROE calculation. I’ll let you be the judge of what assumptions you think are appropriate.

Risks:
* Rapid loan growth could lead to deteriorating loan quality, resulting in greater troubled loans.
* Shrinking spreads due to increased competition in the business.
* Extended periods of disruption in the capital markets, rendering it difficult for the company to issue debt and/or equity.
* A significant economic or real-estate recession.

Catalyst

*Conversion to REIT, and initiation of dividends, starting with the one-time “E&P” cash dividend of approximately $0.65 in January 2006, followed by regular quarterly dividends commencing in the first quarter of 2006. The details of the one-time dividend will likely be announced by early- or mid-November.
*Shareholder “rotation” as growth investors sell, and value-oriented and income-oriented investors purchase the stock. This process should be largely completed by the end of the first quarter of 2006.