Friday, April 30, 2010

The Goldman indictment

I waited a couple of after the indictment of Goldman to post my thoughts on it, since I have mixed feelings on the topic. If you want to take a look at the indictment, you can go here:
As most of you who have followed my work know, I have been been more negative on the products and practices of investment banks than most of my academic brethren. I think that investment bankers promise more than they can deliver and that there is far less value in most of the products that they sell than they claim in their sales pitch. On the Goldman indictment, my sympathies lies with Goldman because I feel that this it is selective prosecution, based upon 20/20 hindsight, designed to advance a larger agenda of "financial regulation and oversight" by the Federal government and Goldman happens to be (at the moment) every one's favorite bogeyman.

First, the background. The deal that has Goldman in hot water is titled Abacus and was a multi-billion dollar Collateralized Debt Obligation (CDO), a fancy terms for a bond backed up by assets, and in this case, the backing came from real estate mortgages. On the face of it, the deal looks unremarkable. In fact, Business Insider was able to get its hand on the pitchbook used by Goldman for the deal:
The pitchbook has all the hallmarks of a standard sales presentation - the obligatory disclaimer that runs 3 pages, 63 more pages that reveal less than they should, uninformative (but colorful) graphs, and tables filled with enough numbers to numb the brain (which is the objective).

So, what made this deal stand out? Here are some of the factors for its being singled out:
1. A "big loser": The securities bundled in the Abacus deal were priced at the height of the housing bubble. Like other housing backed securities it did lose money. However, it was one of the "biggest losers", with losses exceeding in the billions.
2. The John Paulson connection: The seller of the securities in the Abacus deal was the hedge fund headed by John Paulson, one of the few winners in the housing bubble. His subsequent notoriety, chronicled in a book, has made him into a sage, at least in hindsight, about the housing bubble and its implosion.
3. Goldman was the intermediary: Among investment banks, Goldman Sachs was viewed as the only one that was able to cut its losses in the mortgage backed securities debacle and escape relatively unscathed. The fact that it was the broker in this transaction has evoked suspicion that is was partnering with Paulson to take advantage of the suckers on the other side.

The indictment of Goldman seems to rest of two claims:
1. According to the SEC, Goldman Sachs claimed wrongly that Paulson was buying the securities (packaged under Abacus), when it was the seller. I checked through the sales presentation that I linked to earlier to see if there was an explicit mention of Paulson but I did not find any. It is entirely possible that Goldman left the implicit understanding that it was a buyer (I assume that the SEC has something - emails, phone calls, phone video - to back up its claim).

2. Goldman had advance knowledge of the collapse of the housing market and took advantage of their clients: Even the SEC seems to recognize that this is a much weaker legal argument, but the Senate committee investigating Goldman Sachs had no qualms about making this the center piece of its accusations. Using emails from from Fabrice Tourre, who in addition to being an employee of Goldman seems to have forgotten that emails are not erased on the server when you delete them on your computer, senators accused Goldman of knowing that the housing market was going to collapse and actively exploiting investors by selling them securities that would be destroyed by this collapse.

I know that these are legal issues subject to the legal rules on what comprises reasonable. However, if this case were subject to what the rest of us (who are not lawyers and have the benefit of common sense) think as reasonable, it just does not stand up to scrutiny:

1. What if Paulson were the seller rather than the buyer and why should the buyer of these bonds (ACA) have cared? Implicit in the SEC's argument that holding back on the identity of the seller (Paulson) was somehow a deal breaker for the buyers of the securities involved in the Abacus deal. Notwithstanding the halo that Paulson might have acquired as a soothsayer in the housing bubble, he was a voice in the wilderness in 2007 on housing prices. I seriously doubt that ACA would have not bought these securities, even if they had known that Paulson was the seller. In addition, I don't think any intermediary in this market (securities) is required to reveal to the buyer the identity and motives of the seller.

2. Goldman knew the housing market was going to collapse and took advantage of its clients: I find this argument to be beyond absurd, especially given the evidence to back it up. In fact, let's take this argument at face value. If Goldman were that prescient about the housing market in 2007, there was a dozen other ways (most of them more profitable and less work than Abacus) that they could have made money on this belief. So, why construct this convoluted way to make money? Furthermore, investment banks are not monolithic when it comes to views about markets. Having worked with investment banks for almost 30 years, I can guarantee you that at any point in time, views about whether a particular market is under or over priced (equity, bonds, real estate) diverge across an investment bank. For every strategist/analyst at the bank who is bullish, there is one who is just as strongly bearish. Thus, I find Tourre's emails (about what he thinks about the market) to be sensational but completely irrelevant to this discussion. (As an analogy, think of the following: If you were a real estate broker who believes that houses are over priced, should you stop selling houses to clients who want to buy houses?)

Did Goldman take advantage of "naive" clients"? Probably, but that is the nature of trading. All trading is predicated on exploiting the lack of information or good sense on the part of the the investor on the other side of the trade. I don't like what they did because it is bad business practice, in general, to take advantage of your customers. However, it is not illegal. If it were, home buyers should be suing brokers who sold them houses in 2007 and 2008 while secretly believing that these houses were overpriced, customers should be suing electronics salesmen who sold them video disc players, knowing that DVD players were the standard of the future, and voters should be suing politicians who told them that their pension and health care benefits were secure, while undercutting the basis for these benefits.

I know that a lot of people would like to see Goldman fall, and that some of them work at Goldman's competitors. While I understand the urge to bring the mighty back to earth, I think that failing to support Goldman at this time is a huge mistake. To me, this case reveals everything that is wrong with both politics and law - the use of ex-post evidence to back up a case (Paulson made money of the housing crash.. so, he must have known that the crash was coming), suspicious timing (just in time for the new law on regulating bank) and scapegoating.

Friday, April 16, 2010

Currency Choices in Valuation

I am currently in Bogota, Colombia, doing a seminar in risk. One of the topics that came up yesterday was about the choice of currency to do a valuation in, and how it affects your inputs. In particular, the question that I was asked was whether an analyst should value a Colombian company in Colombian pesos or US dollars, and the implications of this choice. Here is how I responded:

Which currency should I do my valuation in?
If you do your valuation right, it should not matter. Your value for a company should be the same, no matter what currency you choose to value it in. Thus, a company that is under valued by 20%, when you do your valuation in pesos, should remain under valued by 20%, when you do your valuation in US dollars.
Given this proposition, you should pick the currency with which you are most comfortable with and where it is easiest to get the financial information. My instinct, given the latter requirement, is to do valuations in the local currency since most financial statements are richer and more detailed in the local currency. Your choice of currency should not be a function of the investor for whom you are doing the valuation. Thus, you should not try to value a Colombian company in US dollars, just because the investor for whom you are doing the valuation is dollar based.
How is my discount rate affected by my currency choice?
In the context of discount rates, the input that is most influenced by the currency choice is the riskfree rate. If you work with a higher inflation currency, the riskfree rate will be higher. In the Colombian context, the Colombian peso riskfree rate was 6.5% and the US dollar riskfree rate was 4% last week. The difference of 2.5% is entirely attributable to differences in expected inflation.

Just as a side note, while getting a US dollar riskfree rate is easy (I used the T.Bond rate), I had to work a little harder to get the riskfree rate in pesos, since the peso-denominated Colombian government bond does have some default risk embedded in it. In particular, I subtracted out the default spread for the Colombian government (about 2%) from the bond rate (8.5%) to get to the riskfree rate.

The other inputs remain pretty stable. Betas should measure the business risk of the company. I have never understood the rationale of a widely used practice of using betas against the S&P 500, when doing dollar based analysis, and switching to betas against local indices, for local currency analysis. Those of you who follow my work know that I am firm believer in using sector or bottom up betas. For Ecopetrol, the Colombian company, I estimated a beta of about 0.80, based on the fact that it was an oil company, and used that beta for both US dollar and Colombian peso analysis. Even more dangerous is the practice of using the US equity risk premium, for US dollar analysis, and the much larger Colombian equity risk premium, for peso analysis. The company is a Colombian company and you cannot make the country risk go away by switching currencies. Both the dollar and the peso analysis therefore should use the higher Colombian risk premium.

As a final note, the cost of debt should be in the same currency that you estimate the cost of equity in and this is true no matter what currency the company actually borrows in. Therefore, if the company borrows in US dollars but you are doing your analysis in pesos, you will have to restate the cost of debt in peso terms.

How are my cash flows affected by my choice of currency?
The key rule here is that your cash flows have to be in the same currency as your discount rate. Thus, if you decide to do your analysis in pesos, you cash flows have to be in nominal pesos. If you decide to do your analysis in dollars, your cash flows have to be in nominal dollars. If it is a company with Colombian operations, this will often mean that you have estimate the cash flows in pesos and convert them into dollars. You have to use forward or expected exchange rates (and not the current spot rate) to make the conversion. In fact, if you want to preserve consistency, your expected exchange rate has to be computed from either interest rate or purchasing power parity. In the context of Colombia, for instance, the 2.5% higher inflation in Colombia that I have built into the riskfree rate will translate into an expected devaluation in the peso of about 2.5% a year.

Can I avoid this currency choice altogether?
You could, if you do your analysis in real terms. Thus, your discount rate has to be a real discount rate; the real riskfree rate is about 2% (I used the inflation-indexed US treasury to get this) and you can build the rest of the inputs on top of this rate. Your expected cash flows should be real cash flows; thus, you cannot count the inflation component of growth. Again, if you do it right, you should get the same value.

The bottom line: Make your choice of currencies at the start of the process and stay consistent with that choice all the way through. If you are wrong about expected inflation, it will cancel out - both your discount rates and cash flows will change. If you are inconsistent about inflation, applying one rate to cash flows and another to discount rates, your valuation cannot be salvaged.

Friday, April 9, 2010

Stock versus Flow: My thoughts

When analyzing companies, the three financial statements that we primarily use are the income statement, the balance sheet and the statement of cash flows. We obtain the inputs for earnings and cash flows from the income and cash flow statements and the numbers for debt, cash and working capital from the balance sheet.

While all three statements are governed by accounting standards and are audited, there is a key difference between them. Income and cash flow statements represent flow statements: they measure how much the company earned and spent over the period. Balance sheets capture the values of assets and liabilities at a point in time and thus represent "stock" statements.

So what? Stock statements are inherently less trustworthy than flow statements, because the numbers may not be representative of what the company did over the course of the year. This can be manifested in almost every number extracted from a balance sheet:

a. Debt: The debt that is reported on December 31 of a fiscal year will reflect what was owed on that day. A company can therefore pay down debt on December 30 and borrow again early the next year, manipulating its debt figures. In fact, there is a story in the Wall Street Journal about banks doing exactly this to make themselves look less indebted and thus safer. 
Another common way in which debt can be kept off the books is by using lines of credit or seasonal financing during the course of the year but to pay them down by the end of the year.

There a couple of clues that we can use to detect this practice. One is to look at quarterly balance sheets, in additional to the year-end balance sheet, with the intent of finding big changes in debt from quarter to quarter. While enterprising companies may still be able to hide debt, it is much more difficult to do so on a quarterly basis. The other is to look at interest expenses as a percentage of the year-end debt. If a firm has debt for the bulk of the year, it has to pay interest expenses on that debt, even if it retires the debt towards the end of the year. As a result, the book interest rate (interest expense/ book debt) will be disproportionately high (relative to what you would expect the company to pay.

b. Cash: The cash on a balance enters intrinsic valuations as an add-on to the estimated value of the operating assets and relative valuations when we use enterprise value multiples (where cash is netted out of debt). However, the cash balance on the balance sheet may bear little or no resemblance to the actual cash balance today (which is really the number we should be using in intrinsic and relative valuation). This is one reason why some firms actually trade at negative enterprise values, where market equity is updated to reflected today's value but debt and cash remain frozen at year-end values.

c. Working capital: Net working capital is the difference between non-cash current assets (inventory and receivables) and non-debt current liabilities (payables and other accrued liabilities). Any of these numbers can be altered over short periods. For instance, receivable collections can be stepped up and inventory cleared (how about those year end clearance sales) just before balance sheet dates to make working capital look smaller.

In closing, I am not suggesting dynamic balance sheets. That would be too expensive and not quite practical. However, I am suggesting that we be more careful about balance sheet based analysis. Never trust a single balance sheet; if you can get quarterly balance sheets, do so; if you have access to the current numbers (on cash, debt and working capital), even better. The last may seem unrealistic but if you are the acquirer in a friendly merger, you should be able to demand and get this information from the target company.