We tend to make the assumption that when considering an investment opportunity, executives seek to fully understand its financial attractiveness. Likewise, we expect that politicians would be interested in some quantitative measure of the impact of an intended policy. However it is more common for decision makers to rely on an overly simple single number rather than analysing and understanding the opportunity and risks in an investment. For example, in older days, but perhaps no longer, anecdotes were doing the rounds at my former employer Shell about senior managers showing little interest in deeper insights in a potential investment by declaring “Just give me the number”.
As a result, more often than not, information on the attractiveness of investments is provided to decision makers as a single number, along with a few narratives alluding to risks involved. The perceived advantage of a single number is that it gives clarity, that it is unambiguous and that it would seem to be a solid basis to make a decision. But trying to reduce complex risks to a single number sows the seeds of poor decisions.
This is also illustrated by Patrick Leach in his book “Why can’t you just give me the number”. In the first chapter it is argued that many executives are highly skilled in ‘the game’ of their business, but that they do not know the odds. To win you need both, like in poker. It will then be argued that in poker one can calculate the odds precisely, but in business that is impossible. How do we mathematically calculate the chance of a positive return over say a ten year period of some investment with all sorts of potential issues looming on the horizon. So then we might as well just stick to some reasonable number on the basis of equally reasonable assumptions, add some gut feel and go with it (or not).
Of course, many poor decisions are made in this way. Do not get me wrong, gut feel, or judgment, is very important. It is the raison d’etre of executives. If we would not need judgment, we could automate the decision making process. Enter the inputs into the model and the right decision follows in a split second. Actually, for many processes that is how things work or will work in the future (we call that Artificial Intelligence). But for big bets, investment decisions, strategies and policies we are not there yet, if ever. We need human judgment. But this judgment should be supported by an appropriate level of good analysis providing a reasonable sense of the odds.
But still, why can’t executives not just rely on the analysts to provide a single but well considered sort of representative base case quantification of an investment opportunity?. Why would they want to be bothered with more sophisticated representations?
The main reason is that single number representations tend to be biased. Many books are written about this, for example Kahneman’s ‘Thinking, fast and slow’. Executives should be aware that their project managers and analysts are humans, too, and may be influenced by some or all of the following types of biases: overconfidence, confirmation, recency, availability, anchoring and vividness (and other). These are all mechanisms in the human mind that influence estimates and assumptions that are being made in the analysis process. There is hence no doubt that a single number representation of some investment opportunity will be biased. The same applies to the quantitative impact of a policy measure. Add to that the executive’s own biases with regard to the future of the business environment and we have a great mix for inferior decision making.
One way to at least attempt to de-bias the analysis is by using probabilistic approaches (probabilistic investment analysis). This implies that for the inputs (costs, timelines, production, sales) we consider ranges rather than rely on single number estimates. There are various ways to translate these ranges into the uncertainty ranges around key decision metrics, such as return on investment or NPV. The mechanics of these methods do not need to be complicated. Such a probabilistic approach, rather than a deterministic one, gives executives a much better sense of the uncertainty and risk associated with the decision to be made. If executives ask for such information, to be provided in suitable diagrams, it not only gives them a better understanding but it also encourages their project analysts to more systematically think about the uncertainty in the inputs and more carefully consider (and quantify) the risks. This thus helps in making the quantitative foundation for decision making more objective and less biased. At the same time I am the first to recognize that not all uncertainty and risk can be captured in numbers and that supplementary methods of analysis and communication can be needed.
Nevertheless, whilst the adage ‘Garbage In Garbage Out’ always applies, I am convinced that probabilistic investment analysis improves the odds of making good decisions…
If you would like to learn in about two hours what probabilistic investment analysis is about you can sign up for our Virtual Learning Bite ‘Why probabilistic investment analysis’.
