In late August, I wrote about Oshkosh’s upset win over BAE Systems in the FMTV recompetition. About ten days ago, Loren Thompson of the Lexington Institute wrote about the award as well, but with a very different view (“Army Truck Award Looks Very Suspicious,” Lexington Institute, 8 September 2009). He noted that while BAE Systems had bid 21 percent below its own current price for FMTV trucks, Oshkosh had bid fully 30 percent below—a figure he found suspiciously low. Noting that this was a fixed price contract, and complaining that Oshkosh was “teetering on the edge of bankruptcy,” Loren asked
since Oshkosh has already indicated that it plans to use the same suppliers used by BAE -- who contribute much of the value of the finished vehicles -- how is it possible for a relatively untested challenger to substantially underbid a company that has been building medium trucks for the Army for two decades?
The answer, he said, was that the Army’s source selection authority did not consider the risk of Oshkosh’s financial failure in the decision, and didn’t adequately question whether it could actually make money on its own aggressive bid. Awarding contracts on overly aggressive bids, he observed, was a “practice that is at the root of many acquisition disasters.”
We can generally agree on the point about overly aggressive bids. The problem for a governmental customer is more obvious when the contractor underbids on a pay-my-costs development contract, but it’s still present in fixed-price work. If Oshkosh couldn’t make money on this bid, then it could be in bigger financial trouble than it already is—if it is indeed in such trouble. Further, the article’s titular assertion that the award itself was “very suspicious” is itself very strong, perhaps even bearing the image of an envelope of cash changing hands. Such a strong accusation, and strong conviction about the relative merits of the bids, themselves merit some serious analysis. It seems that there are two assertions here:
- Oshkosh is too financially risky a company on which to rely, particularly if it cannot make money on this contract.
- Oshkosh cannot make money on this contract.
Let’s deal with these questions in order. If the Army should be worried about whether Oshkosh is a bankruptcy risk, then we should estimate that probability. The literature goes back to at least 1968, when Edward Altman published his Z-score methodology [1], which seems to remain the most popular approach. James Ohlson [2] and Mark Zmijewski [3] followed in the early 1980s with analogous methods based on different data sets. Each is an empirically-derived multivariate approach that utilizes a particular firm’s financial results from the past year to estimate the probability of bankruptcy in the following year. Each is a logit regression model, which provides an abstract number that must be transformed logistically into a probability. For reference, Altman’s method relies on seven variables from the preceding reporting year: working capital, total assets, retained earnings, earnings before interest and taxes (EBIT), total liabilities, revenues, and the market value of the equity.
The results would be alarming, if we could have confidence in them: Altman and Ohlson’s methods each estimate Oshkosh’s probability of bankruptcy in 2009 at about 80 percent. However, in the case of Oshkosh today, there is a big problem with using any of these backward-looking approaches: the company’s financial prospects seem to have improved considerably in 2009, so drawing on 2008’s figures provides an unwarrantedly negative view. (Zmijewski’s method here is even less helpful, as it is particularly sensitive to variations in net income.) Indeed, recent research [4] has indicated that the coefficients in all these models have degraded over time, and vary from industry to industry. As interpretations of accounting rules tend to vary with time and across industries, this is not surprising.
Fortunately, there is another option—literally, option valuation. Note that the last measure in Altman’s method, the value of the equity, is found in the market estimates of all investors, and not merely in the financial statements of the public auditors. That alone suggests that drawing on the collective wisdom of the market may provide better results. As Robert Merton observed in his seminal 1974 paper [5], owning equity is essentially holding a call option on the value of a company’s assets. That hits zero when the company is bankrupt. If we assume equity price movements to be a normally-distributed but lower-bounded (Brownian) process, we can use a transformed version of the Black-Scholes-Merton (BSM) options pricing method to calculate the probability of bankruptcy. In a recent and excellent book [6], Robert MacDonald works out the details, which use the market value of all the assets, the face value of the liabilities (analogous in the model to the option strike price), the expected rate of return of the assets, volatility of the asset value, and the dividend rate (the value of the dividends expected divided by the value of all the assets) as arguments in an application of the standard cumulative normal distribution function,
As appealing as this approach is theoretically, one could have reservations. As Hillegeist, Keating, Cram, and Lundstead noted in their paper [7] testing the utility of the BSM model,
the stock market may not efficiently impound all publicly-available information about PB [probability of bankruptcy] into prices. In particular, prior studies suggest that the market does not accurately reflect all of the information in the financial statements. Thus, whether a market-based PB measure derived from an option-pricing model or an accounting-based PB measure performs better is ultimately an empirical question.
That said, Hillegeist et alia rather convincingly demonstrate that the BSM model much more accurately predicts bankruptcy risk across industries than any of the older accounting-based measures (whose coefficients have drifted beyond recognition anyway), and thus, that it should be generally used in research and by practitioners. The math is a bit heavier, requiring application of the Newton-Raphson algorithm to accurately estimate the total value and volatility of all the assets, which cannot quite be directly observed in the market (not many people are placing calls on the value of Oshkosh’s receivables, for example). If we’re content with first-order approximations, we can skip that part, and just slightly underestimate the volatility and overestimate the value. Given the final results (below), this should hardly matter. However, partly to compensate, and partly to provide a conservative view of Oshkosh’s prospects, we can draw on Oshkosh’s 2008 share price gyrations to calculate volatility, and we can presume that Oshkosh’s asset value cannot be expected to appreciate faster than that of the risk-free rate of interest. Both these approaches should overstate the probability of bankruptcy.
In his essay, Loren noted that Oshkosh’s “stock price shot up 24 percent even before the [FMTV] truck award was disclosed.” The timing itself is not surprising, as information about contract awards and merger announcements has long been known to leak in advance. But more so, the increase in the stock price itself is evidence that investors believe that Oshkosh will succeed financially with the FMTV, and that this success will stave off bankruptcy—increasing the “distance to default” in the random walk of the stock price.
Indeed, running the numbers with the BSM method on Oshkosh’s equity (NYSE:OSK) provides a probability of bankruptcy today that is so close to zero as to be irrelevant. All the computation aside, this is easy to see by inspecting MacDonald’s equation: when the fraction of a company’s asset value that is found in its assets gets much above highly leveraged, the risk of bankruptcy—intuitively—drops precipitously. That’s where Oshkosh is today—as Charles Brady of BMO Capital Markets put it late last month, the FMTV award itself “basically takes bankruptcy risk off the table.” [8]
From the standpoint of the source selection, however, the most important figure could be OSK’s share price around the time at which the contract decision was being finalized: about $14 per share. Running this figure through the BSM suggests that Oshkosh’s risk of bankruptcy was probably not all that great to begin with: well below one chance in a thousand over the next five years. To save the $440 million said to be the difference between the bids [9], it’s hard to fault that anonymous Army officer for taking such a slight gamble. Could the counterbalancing cost really be claimed to exceed $44 billion? That’s inconceivable.
Finally, even if one is really uncomfortable with relying on the predictive power of financial markets, Oshkosh’s debt structure provides some serious comfort. Reviewing Oshkosh’s most recent SEC filing, it’s clear that its debt obligations don’t become a serious problem until after 2013:
| year
|
2009
|
2010
|
2011
|
2012
|
2013
|
thereafter
|
| long term debt
|
$25 MM
|
$50 MM
|
$50 MM
|
$262.5MM
|
—
|
$2314 MM
|
While that’s not necessarily good news for Oshkosh or its shareholders, the Army shouldn’t much care. The five-year contract will be expiring as the first big debt payments are coming due, and if Oshkosh does get into trouble, finding someone to service the trucks five years from now shouldn’t be difficult.
In short, the potential for financial distress—apart from the feasibility of its bid—was not a reason for choosing any supplier other than Oshkosh. Could the market, then, be overestimating Oshkosh’s operational capacity? I’ll deal with that question in a column to follow.
NOTES
- Edward I. Altman, “Financial Ratios, Discriminant Analysis and the Prediction of Corporate Bankruptcy,” Journal of Finance, 1968, pp.189-209.
- James A. Ohlson, “Financial ratios and the probabilistic prediction of bankruptcy,” Journal of Accounting Research, Vol. 18, No. 1 (1980), pp. 109–131.
- Mark E. Zmijewski, “Methodological Issues Related to the Estimation of Financial Distress Prediction Models.” Journal of Accounting Research, 1984, 22 (Studies on Current Econometric Issues in Accounting Research), pp. 59-82.
- J.S. Grice and M.T. Dugan, "The Limitations of Bankruptcy Prediction Models: Some Cautions for the Researcher," Review of Quantitative Finance and Accounting, Vol. 17, No. 2 (2001), pp. 151-166
- Robert Merton, “On the Pricing of Corporate Debt: The Risk Structure of Interest Rates,” Journal of Finance, Vol. 29 (1974), pp. 449–70. Merton built on the work of Fisher Black and Myron Scholes, “The Pricing of Options and Corporate Liabilities,” Journal of Political Economy, Vol. 7 (1973), pp. 637–54.
- Robert MacDonald, Derivatives Markets, Addison-Wesley, 2002, p. 604.
- Stephen A. Hillegeist, et al., “Assessing the Probability of Bankruptcy,” Review of Accounting Studies, Vol. 9, No. 1 (2004), pp. 5–34.
- Bob Tita, “Military Truck Contract Reduces Bankruptcy Risk for Oshkosh,” Dow Jones Newswires, 28 August 2009.
- Kris Osborn, “Sources: Cost Became Determining Factor in U.S. Army’s FMTV Decision,” Defense News, 14 September 2009.