The Bermuda Triangle of Valuation (Learning from Damodaran)

As I have mentioned in previous posts, I did a course about Valuation over the summer. This post follows a similar post about the dangers of valuation models. In my previous post - also inspired by an Ashwath Damodaran talk - I explored the importance of 'stories', which necessarily drive every valuation. I would like for this post to serve as a cautionary note for myself - and any readers - about the common errors we might make when valuing a company. I have attempted to capture the main points Damodaran makes in his talk.

The Bermuda Triangle of Valuation is:

• Bias and Preconceptions
• Uncertainty
• Complexity

Bias and Preconceptions

• It is important to distinguish between Pricing and valuation. Most people think of a number first and build a valuation to fit the number. Value cannot precede valuation, because models can only tell you what you want to here. If you want to buy, you will keep fiddling with the inputs till the model says it is a buy. Thus, it is probably useful to question how each input number was derived.

• You start out with an opinion, and then your valuation simply confirms what you learned. The more you know about a company the more biases there are in your valuation. That being said you need to know the business to develop a story.

• If you know the managers of the company, you know too much to value the company effectively. If you get too close to the company, you are less effective.

• When everyone is saying something in the market, it is hard to step back and objectively disagree. The less confident you are the closer your valuation gets to the market. 

• Suggestion bias. If I say I think its worth 15, but value it anyways, you will probably arrive at a similar price.

• When you see decimals in a valuation, the valuer is using decimals to intimidate you. Decimals mean nothing, because you cannot be that precise about the value of a company.

• Small changes have significant effects. Thus, it is important to question each input.

• Don’t let comparable multiples tell you what the price is. If you pick the comparable, you are making several assumptions!

• Several people deny bias. Denial is no good (Stop lying to yourself). Instead, you must think of ways to structure processes that help reduce bias. 

• Ask people to be transparent about where they come from. What was there hypothesis? What were their findings (attempt to disprove the hypothesis?) The hypothesis is full of bias, so you can then take the findings with a pinch of salt. That being said, its much easier to point out the (often small) issues with a hypothesis than it is to create a hypothesis yourself.

• Where M&A is concerned, you don’t ask the deal maker if the deal makes sense; they have bias!

Uncertainty - We often ignore this as we don’t know how to deal with it.

• Every input involves assumption. We make estimates. The difficulty of making estimates is variable. We want companies where it is possible to make estimates. The harder it is to make an estimate, the greater the uncertainty is. Uncertainty is an opportunity, and so the pay-off to doing valuation is greatest when there is greater uncertainty. However, we must be aware that our numbers are only one possibility.

• Estimation uncertainty (Micro uncertainty) can be refined through cashflows and KPI estimates.

• Economic uncertainty (Macro uncertainty) can be accounted for through the discount rate.

• Spending more time doing valuation doesn’t make uncertainty go away.

• Continuous uncertainty is much easier to deal with than discrete uncertainty (bankruptcy, nationalisation). We must deal with discrete uncertainty. This is because these improbable - but not impossible - occurrences, can result in insurmountable difficulties for companies.

• Things we can do to deal with uncertainty:
• Less is more: Aggregate things instead of breaking them down, when you make forecasts and assumptions. That being said, looking at the lines for historical data allows you to spot significant changes, and can help guide your research.
• Make sure your valuation is not at war with itself; be clear what your assumptions are.
• Make sure your assumptions affect all the relevant inputs, (eg. inflation affects interest rates, discount rates, and growth rates)
• Be realistic.
• Even if you disagree with the market, try to understand what the market is telling you.
• Law of large numbers: Use averages to compare. But, blind averages may conceal significant anomalies.
• Don’t use the discount rate to account for fear.
• Use distributions instead of fixed figures. Considers the different possible outcomes. Draw a bell-curve?
• Don’t look for precision. You will be wrong 100% of the time! You are only looking for a ball-park figure to confirm what your story will look like in numbers.

Complexity - As valuations and models are complex, we may lose sight of what exactly it is we set out to achieve.

• Companies are 'tentacled octopuses'. This complexity may make it difficult to understand a company's business.

• Data accessibility means we have more data. However, this means we must filter the useless noise, and hunt for the jewels.

• Legal/accounting complexity also raises issues.

• Build simple models.
• In complex models with lots of variables, input fatigue sets in. (You may start putting random numbers in.)
• So, the model becomes a blackbox. You don't know what went in and don't know why a given output came out.