Risk & Probability In Investment Decision Making

You have likely heard the phrase “more risk, more reward”. But, this is the wrong way to think about an investment or capital allocation decision.

The financial definition of risk should be:

The probability of loss X amount of possible loss from the probability of the gain X amount of possible gain.

Warren Buffett BRK meeting 1989

Why should the notion of risk be married to probability?

The answer is simple. Because the mind does not automatically think in terms of probability.

Think about how considering risk in a probabilistic way changes investing. Before making a decision to invest the question should be asked:

Will the investment produce aggregate after tax receipts including those at the time of sale, over the prospective holding period, grant at least as much purchasing power as when the investment was made plus a modest rate of interest on the original investment?

To answer this question with accuracy (but not precision), the investor would need to know:

1. Certainty of the long term economic characteristics of the business being maintained
2. Certainty with which management can be evaluated, both to their ability to realize the full potential of the business and to wisely employ its cashflows ( and consider what may prevent their past success from continuing )
3. Certainty with which management can be counted on to channel the rewards from the business to investors rather than to themselves
4. A correct purchase price determined for those future cashflows
5. Levels of taxation and inflation affecting investors pricing power over time

None of the above variables are knowable with absolute certainty. Many are subjective in nature and not something that can be defined to a decimal. Without a doubt, all of those variables need to be considered in the riskiness of the cashflows. Therefore, investors and manages should use the tools of probability in their capital allocation process.

The right way to think about risk is the way Zeckhauser plays bridge, it’s just that simple.

Charlie Munger

Richard Zeckhauser is an Economist and Professor of Political Economy at the Harvard Kennedy School. When he plays bridge he thinks via decision trees and attaches probabilities to each branch. When the facts presented by the cards in play change, he updates the probabilities of each branch.

The role of a CEO of large organization is that of capital allocator. The evidence that this is the correct way of thinking is found in the writing and examples of the best capital allocators on the planet. Consider the following.

Michael Mauboussin at the investment firm Legg Mason has pointed out some of the similarities between the best investors including:

1. That they are probabilistic players that focus on processes over outcome, and exhibit a constant search for favorable odds and understanding of the role of time (having patience).
   
2. He noted that success in a probabilistic field that requires weighting probability and outcomes is expected in a value investment mindset.
   
3. Success drivers include a high degree of awareness of the factors that distort judgment.

Charlie Munger prefers not to do deals that require accountants or brokers. He views them as too complicated and adding to the risk of the transaction. ‘With a preference to multiply by 3 rather than pi.’ He and Warren would rather go to the seller's house and work it out on a piece of paper.

Similarly, he prefers investments into companies with superior economic characteristics (in line with the thinking of Phil Fisher) as opposed to statistically cheap bargains (Graham or Klarman). Reasons for this approach include that it radically reduces the number of variables to be considered to arrive at a better judgment of value in the short term and long term.

After considerable growth, the team at Blackstone private equity group eventually adopted an underwriting process that is probabilistic in nature.

The partners at Blackstone analyze deals based on changes imposed by Hamilton E. James following from his time at DLJ Merchant Banking, later know as  Credit Suisse First Boston following their acquisition of the firm.

When James joined Blackstone in the early 2000s he made changes to help the growing organization follow better risk analysis processes. Retrospectives of past deals showed that deals where the partners overweight management performed worse and vice versa.

Blackstone already conducted a detailed analysis before these changes, such as producing 150-page detailed reports “forecasting everything from every division down to how many coca-colas they are drinking in their conference rooms and the price of coke”.  An initial change was to require that for partners to continue researching a potential transaction, they must submit an outline on their thesis for others to weigh in.

The partners were required to add factors to the deal they were evaluating that included the possibility of fluke events that could sink a company or turn the investment into a success. Consider the below examples of events used in the book King of Capital:

• Terrorism
  • Event once 20 years – don’t affect the base case because its 1:20
• Oil goes from $30 to $140 in a year
  • Never happened before, the most oil has gone up is 20 bucks; how could it go to $100?
  • Add a probability to this happening.
• Probability of labor problems

A probability is then assigned to each of these potential events. They assume the probabilities are 1/10, 1/20, 1/50 and would generally not be put in the base case because they were individual very unlikely.

The chance any single one of them happening is tiny, but the chance none of them happening is also tiny. If you multiply out the combined percentages you get a 55% chance one of them will happen and it kills the investment. They conduct the same analysis on the upside and aim to be lucky.

For Blackstone, the objective is to be rigorous and consistent in the analysis of potential transactions. Prior to this change in process, the careful observer of the growth of Blackstone will note that transactions occurred roughly by the seat of their pants. Only as Blackstone scaled did they add a chief investment officer, Mossman, to be a filter for each of the deals, removing himself from any conversations with management to provide objectivity much like a value investor would.

Comparing and contrasting Warren and Munger against Blackstone point to a similar conclusion.

Risk can’t be calculated precisely. And the superior investors in markets are not acting as if this is the case.

In stock markets, heuristics just don’t cut it because the odds change when new bests are placed.

The variables that go into investment decision-making are the same types of variables that go into decisions in the executive chair. They have subjective elements and require the decisionmaker to predict the future. Therefore in assessing the risks of a given investment or course of action, probabilistic thinking is the foundational tool that should be used.