A quick thought: William F. Sharpe said It's Dangerous to Think of Risk as a Number

During increased uncertainty in financial world, including the threatening higher inflation, what Bill Sharpe, the father of CAPM (pronounced: CAP-EM) and the receiver of the Nobel Prize in Economics in 1990, said in the title above, keeps ringing in my head.

I believe what he said is so true. Though mathematically, pretty much, we could model the risk, the big fat question is, does it work?

When we build our financial modeling for the next 3-5 years (personally, I think even after 3 years, nobody could really have a good crystal clear idea about what will happen). Of course, the business will still exist, but what happens to the environment, economy, politics, etc...I guess all we have just a good GUESS-ing. Even for some companies, the near-term is the only agenda item they are concerned about and the rest might be read as peering through the fog of uncertainty.

Normally, we have many cases or scenarios we put into our financial model, in general:

• Corporate scenario or base scenario
• A sunny or good scenario
• The bad or worst scenario

Then where is the Covid-19 scenario, the war, for example?

I did remember one book written by Enrique R. Arzac with the title : Valuation : Mergers, Buyouts and Restructuring (2nd Edition) which he said that it is necessary that in the valuation, we need to do some stress tests by assuming the revenue is BIG ZERO. Or zero growth in at the terminal period using Gordon Growth Model?

When I read that a couple of years ago, I didn't really believe I would like to use that suggestion, and somehow I feel it was too extreme an assumption to put into the financial model.

However, in one or two other books on financial modeling topics, the authors will suggest, that one of the testing the integrity of the model is to assume away extreme figures, such as BIG ZERO REVENUES, and see whether there is something wrong in the model. I've never entertained a thought when reading that such a "model test" could actually be put as one of the scenarios to be built. But for example, amidst the Covid-19 pandemic 1-2 years ago, all these suggestions are worth noting and used somehow in the model. A good model should also take into factor extreme conditions. How about if the company does not have the revenue at all or revenue drops off more than 50%. We could even say that Covid-19 is one of that "black swan" events. According to Nassim Nicholas Taleb in its beautifully-written book The Black Swan, The Impact of the Highly Improbable, a black swan is a highly improbable event with three principal characteristics: (i) It is unpredictable; (ii) it carries a massive impact; and, (iii) after the fact, we concoct an explanation that makes it appear less random, and more predictable, than it was. The question is, can we put the impact of the highly improbable into the Model?

By the way, talking about risk and CAPM, the insight I got is not all risks that other people are willing to pay. For example, if you are crossing a busy street recklessly, then I don't think there is somebody that wants to pay you for taking that risk. In the CAPM world, only systematic risks deserve an (expected) return to ask. Can we call this systematic risk a real risk that an investor is willing to pay?

Another thing to look at is investments that give us payoffs in bad times (for example, during this Covid-19 pandemic) then, logically, it should be valued higher or more expensive. Since we need to pay more expensive, then it means we are willing to accept the lower expected return compared to the investment that will give us payoffs in good times.

My simple question to you:

§?Loosing US$1,000 when you are in bad financial condition vs Gaining US$1,000 when you are in good financial condition, IS NOT THE SAME, though the money is the same, that is US$1,000.

§?Loosing US$1,000 when the money is scarce, is a lot painful, and we might call this 'REAL RISK'.

The CAPM is built under Markowitz's Modern Portfolio Theory and its main engine is diversification, which will quash away all firm-specific risks, and such risk will not be priced into the expected returns for common equity. Only the risk that can't be diversified than that investor or shareholders are willing to pay. However, this same rationale could not be brought into the case where the securities have limited upside potential and greater downside potential arising from firm-specific risks. For example, securities such as bonds are issued by the company. If the bond issuer is doing well, then the bondholders will only be limited to receiving the promised principal and interest coupons. However, if the company is doing badly, then the bondholders might not receive the payments of principal and/or interest coupon as promised. It might be in a lesser amount or even part of all principals are never recovered. So for such investment, the mean-variance framework will not work.