Waiting for a ride on the Phoenix

As a follow up to my previous post on Altman Z score, investors who use solvency analysis to avoid bankrupt companies should beware of the effects of an economic recovery. The other side of the coin of solvency analysis is the Phoenix effect.

When the economy comes out of recession, shares of near-bankrupt companies see eye-popping returns as they rise Phoenix-like from the ashes of near insolvency. Examples include Chrysler moving from $2 to over $30 in the 1982-3 recovery; Magna International from under $2 to over $80 in 1991-2; and Akamai Technologies from under $2 to over $18 in 2003-4.

Buying shares of near bankrupt companies is a dangerous but exciting game. To be successful, an investor needs to identify the Phoenix candidates and correctly time the turn in the market. The rewards are can be big. Buying a basket Phoenix stocks can yield returns of 100-200% over a 12-18 month period.


Phoenix is partly a small cap effect
The Phoenix effect can be characterized partly as a small cap effect. The chart below shows the relative returns of the small cap Russell 1000 relative to the large cap S&P 500. I indexed the start value of 100, at dates representing stock market lows coinciding with economic slowdowns since 1980. On average, the Russell 1000 outperformed the S&P 500 by about 17% one year after the market low. The initial upward thrust in the market has always been marked by large cap outperformance.

Interestingly, the recent March 2008 low was characterized by small cap outperformance which leads me to conclude that this rally is just a bear market rally and the March low was probably not THE BOTTOM in this bear.







Looking for Phoenix candidates
Phoenix candidates are not just small cap stocks, but shares of companies that are at risk of insolvency and benefit from the tremendous positive operating leverage from an improving economy and high financial leverage which put them at risk of bankruptcy. The obvious quantitative way of finding Phoenix candidates is to screen the market for shares of companies that are at risk of insolvency. However, there is a simpler heuristic: low-priced stocks.

Stock price is a factor that’s not in most equity quants’ factor lists. However, it is a deceptively simple way of screening for Phoenix recovery candidates. I remember that Jeff deGraaf, who was at Lehman Brothers at the time, reported in late 2003 that the return spread between the lowest and highest decile of stock price was about 70% - an astounding return to a factor for less than one year.

I roughly confirmed these results by running a backtest using the current components of the Russell 1000. Had you bought the lowest decile by stock price in December 2002 and held them for a year, the median outperformance compared to the top decile was about 110%. This simple study has problems, mainly in the form of a survivorship bias. The use of a median return instead of an average return does mitigate some of the survivorship bias issues. Nevertheless, it does illustrate the magnitude of the effect. Using a long-only approach, this study over the 2003 and previous recovery period suggest that a basket of Phoenix stocks has the potential to rise by a factor of between 2 and 3 over a 12-18 month period.


Phoenix candidate = low stock price + dramatic fall + insider activity
Just buying low priced stocks gets you partly there but we should eliminate stocks that have always traded at low prices. Phoenix candidates are stocks that have taken a pounding, or stocks that have fallen dramatically (70-90%) from the 52-week high. This is a likely indication that it is at risk of insolvency.

These companies are on the verge of Chapter 11 so buying their shares is highly risky. To mitigate downside risk of possible bankruptcy, add an additional insider activity screen. Ideally I would like to see recent insider buying in Phoenix candidates, which indicates that the fundamentals may be turning. At the very least, I would like to see the lack of insider selling, a sign that the worst is may over for the company under consideration.


Timing: Be patient, the Phoenix will rise
Right now, the weight of the evidence suggests that the turn has not occurred yet. I am preparing a list of Phoenix candidates for my portfolio but waiting for signals of a bottom before buying. In a future post I will write about how I would time the buy decision of these stocks.

The limitations of Altman Z

As s we go through a period of economic stress, I thought that it would be timely to review the Altman Z formula as a predictor of bankruptcy. The formula is a function of liquidity, balance sheet strength and earnings power:

Altman Z =
1.2 X Working capital/Total assets +
1.4 X Retained earnings/Total assets +
3.3 X EBIT/Total assets +
0.6 X Market value of equity/Book value of debt +
0.999 X Sales/Total assets

The original Altman Z score assigned fixed weights to each of the components. Different ranges for Altman Z score represented different levels of risk of bankruptcy. Subsequent versions of the formula varied the weights depending on whether the analyzed company is public or private and also varied the cutoff ranges for bankruptcy risk.


Altman Z was formulated for operating industrial companies
The main problem with this formulation of solvency risk is that the formula is not suited for many industries. As an example, when I first tried to apply Altman Z I found that many regulated utilities showed up as having high bankruptcy risk.

I found that Altman Z was not industry specific enough to my liking. For instance, low or negative working capital doesn’t score well on Altman Z but some industries can operate with zero or negative working capital. For example, a restaurant gets paid in cash, but their suppliers will generally give them net 30 on their payables and the inventory (food) turns over very quickly.

Another sector that the Altman Z doesn’t analyze is the financial sector. What does “sales” mean for a bank? Financials tend to be highly levered and their operating risks and exposures are not well disclosed.


A heuristic for solvency analysis of non-financials
In a recession, the combination of high operating risk and excessive leverage combine to produce insolvency for non-financial company. A better way of forecasting solvency risk is to look for companies that show:

• High operating risk: The top two deciles of standard deviation of EBIT or EBITDA margin over the last 5 or 10 years (pick your horizon)
• High financial leverage: The top two deciles of financial leverage, by total debt to market equity or interest coverage. Normalized decile scores by sector.

I have found that this heuristic, or simple rule of thumb, serves as a better forecaster of solvency risk as it neutralizes many of the industry specific effects that Altman Z failed to address.

Solvency analysis of Financials is difficult
The problems of creating a solvency test for financials that operate in real-time or relative real-time is not easy. It’s not hard to do after the fact, but at any one time, no one – not even the directors of the company really know what is embedded on the books at a financial. Société Générale, Barings, Northern Rock, Bear Stearns – the list of blowup surprises go on and on.

I have had some successes with a solvency risk test for lending institutions based on the following two characteristics:

• Excessive lending growth as a sign of lending portfolio quality: In good economic times, a bank can produce earnings growth by growing its assets, or loan book. In the long run, not all banks can grow their loan books significantly in excess of GDP. High asset growth comes at a cost of lower asset quality.
• Loan loss provisions as a measure of the current level of stress: Instead of the standard ratio of loan loss provisions to total assets, I like to use loan loss provisions to assets three years ago. It’s not the loans that you make today that go sour, it’s the ones that you made two or three years ago that tend to get into trouble.