Authors: David Einhorn
Tags: #General, #Investments & Securities, #Business & Economics
Triad Hospitals (a spin-off from Columbia/HCA) and MDC were two large long positions that each doubled during the year. Most of the rest of the shorts contributed to our returns, and we recovered from our tough February to finish 2000 up 13.6 percent. Again, we did not achieve our annual target, but demonstrated a lower risk profile than the market—where the S&P lost 9 percent and the Nasdaq imploded, falling 39 percent. No one was complaining. We finished the year with $440 million under management.
We closed the fund a second time. Except for a few formal capital-raising rounds when the portfolio was fully invested and we needed capital to pursue new opportunities, we have remained closed ever since. The market continued to return to rationality in 2001. Like 1997, we went the whole year without a serious setback. The large long positions all performed, as did our biggest short position, Conseco, an insurance and annuity company.
Conseco started as a “capital structure arbitrage,” an investment based on our assessment that one part of a firm’s capital structure is mispriced relative to another. We do not do a lot of arbitrage, but participate in extreme opportunities. Conseco had terrible news: It lost its A-rating from the A. M. Best rating agency. This made it difficult to compete to obtain and retain customers without making dramatic price concessions that eliminate the profit opportunity. Conseco bonds traded at sixty-five cents on the dollar, while the equity market had a different view and valued the company at $10 billion. The debt yielded over 20 percent, while the equity traded as if bankruptcy were improbable. We purchased the bonds and sold short the common stock.
At first, we lost more on the short sale of the stock than we made on the bonds. Conseco brought in a new CEO, Gary Wendt, who had a great record running GE Capital. Wendt signed for $45 million cash and 3.2 million shares of stock up front. He led an analyst day and promised an immediate turnaround. He would implement fancy-sounding GE management concepts like
Six Sigma,
where its people would be trained to become
Six Sigma Black Belts
and turn Conseco from its present weak state into a strong man, which he depicted with a cartoon of a powerful weightlifter. Wendt convinced the market that the problems would soon be fixed, and Conseco refinanced some of its debt at 11 percent.
I attended a meeting Wendt held in Conseco’s office in the GM building. About thirty investors and analysts stood around in a warm conference room until the group was fully assembled. When Wendt was ready to join us, we were asked to sit around a very large conference table. When everyone was seated, an assistant came in dragging a conspicuously fancy chair. Space was cleared at the middle of the table and when the throne was in place, Wendt joined the meeting. After a lengthy pitch, Wendt took some questions. To any question that involved the numbers, Wendt had no answer. Over and over, his response was, “Someone will get back to you.” I had seen enough. We sold our bonds and added to our short.
Conseco issued a series of “turnaround memos.” These self-congratulatory tomes appeared designed to provide good news to juice the stock at random times. It seemed to work, and the shares doubled. However, each time Conseco reported quarterly earnings, there were more questions without answers. For example, in one quarter, corporate overhead magically turned to a corporate profit. How do you turn overhead into a
profit
? No answer. The next quarter, the premiums and float fell in the insurance business, but capitalized customer acquisition costs and profit rose. How? No answer. Quite simply, the numbers did not add up, and Team Wendt was not interested in clarifying them. The stock continued to rise. Until it didn’t.
The next quarter, Conseco reported better-than-expected results. The results were, again, of low quality and raised many questions. The results were no worse than the previous batches. Nonetheless, this time the market did not buy it. The shares imploded. Eventually, Wendt resigned and the company went bankrupt. Subsequently, the company has reorganized, but is now much smaller. Its name lives on most conspicuously as the Indiana Pacers’ home court, the Conseco Fieldhouse. Some have observed that naming a sports arena is a good way to identify short-sale candidates.
Another troubled company with odd accounting was Orthodontic Centers of America (OCA), a rollup of orthodontist practices. The company accelerated revenue recognition and recorded more than all the profit from patients in the first months of a multiyear treatment cycle. As a result, toward the end of the treatment cycle, the average patient generated a reported loss. OCA had to rapidly grow the number of new patients to outnumber the old patients. Additionally, though OCA was 40–60 partners with the orthodontists, OCA back-loaded expenses by recognizing the orthodontists’ compensation expense on a “cash” basis rather than on an accrual basis. Based on our research, we discovered that OCA front-loaded revenues and back-loaded expenses to compound the impact of dual aggressive accounting practices on reported earnings.
For the second time, we outlined our concerns to the SEC. In March 2001, OCA announced that the SEC required it to change its revenue recognition to record patient revenues on a straight-line basis. The company delayed filing the annual report to restate results with 10 percent lower revenues and 25 percent lower profits than previously believed. Though the stock fell initially, the bulls believed OCA had put its accounting problems behind it. The restated results created easier future comparisons because OCA re-recognized the same revenues that it had improperly front-loaded and reversed in the restatement. By May, the shares recovered almost back to their previous highs. Partially changing the accounting, however, did not improve the overall bad economics of the business. By the following year, the cash flows badly lagged the earnings, and the shares collapsed. This created discontent among its orthodontists, who had bet their businesses on OCA stock. OCA eventually underwent a massive financial restatement and went bankrupt.
When the books closed in 2001, almost everything had worked. The fund returned 31.6 percent, and our assets under management reached $825 million. The bear market deepened, with the S&P 500 falling another 11 percent and the Nasdaq 20 percent.
Then, early in 2002, our two longest-standing and hardest-fought short sales finally paid off. As described above, after three years, Seitel finally imploded under the weight of its own bad business model and aggressive accounting.
Elán was a different story. Elán, an Irish specialty pharmaceutical company, had a small portfolio of branded drugs and a drug delivery technology applicable to a wide number of possible drugs. We sold Elán short in 1999. Elán entered into a series of licensing deals that appeared to be shams. Elán would invest $10 million in XYZ biotech company. XYZ, often a tiny company, would put out a press release trumpeting that Elán’s investment “validated” their technology. XYZ would use most of Elán’s investment to license Elán’s drug delivery technology to use on drugs in XYZ’s pipeline that were years away from commercialization. Elán recognized the license fee as revenue at 100 percent margins. Essentially, the money traveled a circle from Elán’s pocket to XYZ as an investment and back to Elán as a license. The market paid a rich multiple for this.
We also noticed that Elán’s line-by-line revenue and expense reports were usually nowhere near analyst models. And yet, the bottom line was always a penny or two ahead of estimates. A reported shortfall in one place would almost magically be made up in another. Additionally, Elán had a number of off-balance-sheet, special-purpose entities to hold various assets. These vehicles created an unexplained and growing benefit to Elán’s earnings. There were many other financial statement anomalies apparent almost every quarter. Like Conseco, Elán was not interested in clarifying the issues and, for a long time, the bulls didn’t care. In 2001, the SEC delayed approval of Elán’s financials. We thought the truth would come out, but it didn’t. The SEC completed its review, and Elán had to restate earnings—by one penny . . . one lousy penny! With the review behind them, Elán was “home free,” and its shares took off to new highs.
Over the years, I debated Elán with brokerage firm analysts. After discussing the numbers and the various problems with the earnings, they would conclude by asking, “So what? Why would you short Elán? They never miss earnings. They never will miss earnings. If they don’t miss, you are never going to win. How are you ever going to get paid on your short?” Elán was a rig job, but it was not up to the analysts to notice. Perhaps they noticed, but thought Elán should be rewarded for executing it so well.
In January 2002,
The Wall Street Journal
ran a lengthy story on the front page questioning many of the aspects of Elán’s accounting that we had observed for years. In the post-Enron environment, the reaction was quite different, and the shares cratered. Though most of Elán’s senior managers were accountants, the company began a serious accounting review. When it was over, the accounting fraud was even worse than we suspected. Elán had been selling off its drug portfolio to other manufacturers and booking the proceeds as “product revenue.” This explained some of the gaps in the financial reporting that we never understood. The shares that peaked at $65 upon completion of the SEC “review” in June 2001 were in the low teens in the spring of 2002 on their way to $1 in October of that year.
Fraud can persist for a long time, and investors, analysts, and the SEC miss things. But, sooner or later, the truth wins. If you
know
you are right, all you need is patience, persistence, and discipline to stay the course.
The year 2002 started nicely. We were up 12.9 percent by the end of April. Just around that time, we completed our research on Allied Capital, an investment that would require all of my patience, persistence and discipline. And more patience.
CHAPTER 5
Dissecting Allied Capital
In early 2002, the managers of a small hedge fund that specializes in financial institutions called to discuss Allied Capital. They came over and walked through their critical analysis, pointing out anomalies with Allied’s portfolio valuations. They wanted our opinion because they knew of our success shorting Sirrom Capital in 1998, a company with the identical business development company (BDC) structure and a similar strategy to Allied. The story was intriguing.
Allied Capital is the second-largest publicly traded BDC in the country (American Capital Strategies is the largest). Allied was founded in 1958 by George C. Williams as a small business investment company (SBIC) to take advantage of the Small Business Investment Act of 1958. Williams had worked for the FBI for much of the 1950s, and since the Small Business Administration (SBA) and all that new funding that was available from the SBA was also right there in Washington, Williams had the good sense to headquarter Allied in the city. The company went public in January 1960, selling 100,000 shares at $11 each. The company began making quarterly distributions to shareholders in 1963. Over the years, several affiliated companies were spun-out or created with similar mandates to make debt and equity investments in small, mostly private businesses that would provide recurring cash flows. Several of these companies would go public, but some remained private partnerships. Williams served as president, chairman, and CEO of Allied and its affiliated companies from 1964 until 1992, when he was named chairman emeritus.
Allied Capital Corporation I, Allied Capital Corporation II, and Allied Capital Lending were closed-end management companies that elected to be regulated as BDCs under the Investment Company Act of 1940. They made private-equity and mezzanine investments in small businesses. Allied Capital Lending made loans through the SBA’s 7(a) loan program. Allied Capital Commercial Corporation was a real estate investment trust (REIT) devoted to investing in small business mortgages sold by the Resolution Trust Corporation and the Federal Deposit Insurance Corporation. Allied Capital Advisers managed the assets of the four other Allied Capital companies. On December 31, 1997, these five publicly traded affiliated companies merged to form Allied Capital Corporation in a tax-free stock-for-stock exchange.
At the time of the 1997 merger, Bill Walton (no relation to the former basketball star) was chairman and CEO of all the merging companies. He assumed the roles from David Gladstone, who resigned as chairman and CEO of the Allied Capital companies in February 1997. Gladstone was a long-time Allied Capital executive, having served as an executive officer of the affiliated Allied Capital companies since 1974. (Gladstone would go on to co-found American Capital Strategies before starting his own publicly traded BDC, Gladstone Capital.) Prior to assuming these positions at the Allied Capital companies, Walton had been a director of Allied Capital Advisers and president of Allied Capital Corporation II.
The rationale for the merger of the separate Allied companies was to simplify Allied’s internal operations and create critical mass to raise the company’s profile with Wall Street and make it attractive to institutional investors. As of December 31, 1997, Allied reported $800 million in total assets, including a $200 million private finance portfolio with investments in eighty-nine portfolio companies.
I asked one of our analysts, James Lin, to replicate our Sirrom work on Allied. He built a large database of all of Allied’s loans showing the cost and value of every investment each quarter for several years. The database showed that Allied’s valuation patterns repeated Sirrom’s. Allied marked down the equity kickers of problem investments, while holding the related loan at cost. This was a good predictor of a future write-down of the loan. Small write-downs disproportionately preceded further write-downs. As in the Sirrom analysis, this indicated that Allied was slow to write-down troubled assets.
The pattern of loan and equity-kicker marks revealed the problem loans. Allied invested in a few public companies, where we analyzed the SEC filings and checked trading prices to see evidence of aggressive carrying values. To protect its existing investment and delay the day of reckoning, Allied often put more money into apparently troubled situations and/or restructurings without taking proportional markdowns.
From James’s database and Allied’s SEC filings, we assembled a list of questions to ask the company. We arranged a call with Suzanne Sparrow and Allison Beane of Allied’s Investor Relations department on April 25, 2002. This would be my first contact with Allied, and in many ways would reflect Allied’s general investor relations practice: Officials answer the easy questions and avoid the hard ones. During this call, and in a follow-up call the next week with Penni Roll, Allied’s CFO, we raised all of our issues and concerns and listened to the company’s responses. The first call, which we recorded in accordance with our standard practice, lasted about two hours.
Early in the conversation, I asked the key question of how Allied determines the value of its investments. “How do you . . . decide what to value the equity for? What do you need, like another financing round to come in that validates the value of the equity or do you do an appraisal? How do you do it?” I asked.
Sparrow described what she called Allied’s
Mosaic Theory
of valuation, “It is not quantitative definitively. Certainly, there are quantitative factors, but there are also qualitative factors,” she said. “And that’s where some of the BDCs, I think, diverge on methodology with respect to valuation. You see some others who treat it truly as a quantitative exercise.”
“That’s the beautiful thing about a BDC as a vehicle,” she said a moment later. “You don’t have, you know, the bank regulators leaning on you to say you must write-off this asset.”
I asked whether Allied began writing loans down when the risk increased so it would require a higher yield or whether it waited until it realized the investment was a certain loser. She responded that write-downs started “when we believed that we had permanent impairment of the asset.”
This was wrong. As a BDC, Allied has to use “fair-value” accounting, which requires them to value securities based on what they are worth today. An arm’s-length buyer would take into account higher risk and would demand a higher return on a loan that deteriorated. A higher return requirement translates to a lower value. It was aggressive and, in my opinion, improper to wait for an investment to be permanently impaired before writing down the value. My job during these calls was not to argue, but to hear their side of the story. I responded only, “I see.”
Sparrow continued to defend carrying loans that deteriorated at cost. “Grade 3 tends to be carried at cost because nothing has been lost yet; we don’t believe there is permanent impairment there yet,” she said. “So, it’s only when we believe that truly it’s gone and once there is a write-down we take the position that it is permanent. We’re not taking it down because we think it’s going to come back. I mean, obviously we’re going to work real hard to make it come back if we possibly can, but we don’t want to tell our shareholders, ‘Oh, it’s down today, but it’s back tomorrow.’ You know, if we write it down, we think it is gone and so it’s permanent.”
It was plain she was openly admitting improper accounting. “Right, I understand,” I said. I wondered whether Allied realized that what it was doing was wrong or whether the company was simply unsophisticated.
Then I asked her about writing down the loans when the equity kicker has been written down. Allied doesn’t need to, she said, because it believed at that point it was not losing principal or interest on the loan.
The market values debt based on Treasury bonds, which are presumed to have no risk, and then adds a spread representing the credit risk of the particular debt instrument. So, I asked whether Allied valued its loans based on spreads to Treasuries.
“No,” she said, “long term it’s tended to show fair-value over what a willing buyer and willing seller over a reasonable period of time would be willing to exchange assets. So it’s not supposed to be a fire-sale or a liquidation kind of valuation.”
This seemed like a non sequitur. I hadn’t referred to a fire-sale or liquidation values. I knew her answer was wrong, and thought she was plainly avoiding my question. “Sure, I understand,” I said.
We discussed several winning investments in the portfolio. She volunteered Business Loan Express (BLX), Hillman, and the Color Factory. We discussed Allied’s rapid growth in fee income. This came from Allied’s strategy to have more “controlled” companies, meaning Allied owns the majority of the equity. By controlling companies, Allied can charge various fees for services. Allied was principally a mezzanine lender until around 2000, when it shifted strategy to add controlled companies to the portfolio. According to Sparrow, almost everyone working at Allied helped provide services to controlled companies. Allied even billed Sparrow’s time. We reviewed the real estate portfolio and then asked to discuss the specific private finance loans. Sparrow suggested a follow-up call with Allied’s CFO, Penni Roll, to cover those.
So we reassembled on May 1 for a lengthy call with Roll. I wanted to create the same kind of static pool data we created with Sirrom. A static pool analysis looks at loans in groups based on when they were originated. This analysis is particularly helpful in analyzing growing portfolios, where new loan growth can sometimes mask the developing losses in earlier loans. We hadn’t been able to do this for Allied because it did not disclose the loan maturity dates, even though this is required by SEC Regulation S-X. (Allied began disclosing individual loan maturity dates in its 2004 annual report.) We had trouble tracking loans by year of origination because Allied’s corporate restructuring in 1997 made data older than that difficult to compile.
Lacking the data, I asked Roll about the historical credit loss rate. She indicated it was less than 1 percent of principal per year. (Later, Allied showed this figure in its SEC filings, but stopped after we questioned its accuracy.) That figure seemed absurd to me. In a low-interest-rate environment, Allied charged interest rates in the teens for mezzanine loans to middle-market companies. No one achieves a credit loss under 1 percent a year over time on these types of risky loans. An excellent average annual loss rate would be 3 percent to 4 percent. Allied’s loans had to be riskier than high-yield loans, and much riskier than bank loans, which recently experienced loss rates much higher than Allied claimed for its portfolio of mezzanine loans. Loss rates on risky corporate debt instruments spiked in the bear market. Apparently, none of this hit Allied’s books. Was it truly better investing or simply an accounting regime that delayed losses until they were deemed
permanent
?
I asked her whether Allied’s loans were more or less risky than an index of publicly traded high-yield bonds. Roll said, “We think what we do from a structural perspective of the instrument itself is less risky. If you look at a high-yield bond portfolio, a high-yield issuance typically has very little teeth in the financial instrument itself. And very little covenants, and payment default is always the biggest thing that can put you in default versus a lot of financial ratios.
“So Bill Walton, our chairman, kind of equates it to, if you’re going to a basketball game, if you’re the owner of a high-yield instrument, you have a ticket to watch the game. If you have a subordinated debt instrument like ours, you’re on the court playing the game. So, you don’t have a lot of rights as a high-yield bondholder. As the holder of a highly structured privately placed subordinated debt instrument like we would have, you have a lot of teeth in your document. You have tight financial covenants, you have covenants with respect to what they can and can’t do with the company, assets they can or can’t sell, people that have remained in the company to ensure its success. You have rights to review any acquisitions, rights to look at corporate structure, change in corporate structure, and a lot of teeth in your document.”
This was not true. While investment-grade bonds often have skimpy covenant packages, it is standard for high-yield bonds to have exactly these types of covenants and restrictions—that’s teeth. People with hands-on experience in bankruptcies and financial restructurings know the absurdity of Walton’s in-the-game-vs.-watching-the-game analogy and Roll’s clueless parroting of it.