Walking in the Dark: Labelism vs First Principles
Illustrated by destructuring the Price to Earnings Multiple
Labelism or (cluelessness) is the characteristic of uttering phrases that nicely fit within a context, but lack substance. We can see this clearly in GPT-3, as it provides output that fits a context but understands none of it.
People learn this by trial and error, and utilize it to signal status in current trends.
An example of this is pointing out that a stock is cheap given some standardized price ratio. While it may be true that a stock below 20x PE is cheap (in 2021), if one doesn’t know why, they may end up holding this belief in 2023, and possibly spreading it. Rather, if we know how the price of risk impacts the needed earnings in order for a stock to be considered cheap, we could use that to see at what standardized price a stock would need to be trading at in order to possibly be considered cheap.
We will use this example in order to understand and escape labelism, which sometimes yields valid rules of thumb, but can be dangerous if applied to the wrong context.
Destructuring the Price to Earnings Ratio
If we have a stock producing 10B in earnings trading at a 20x P/E, we can see that the stock trades at a 200B market cap.
However, hidden in that number, is the price of risk and an implication of growth.
In order to find our price of risk, we need to compute:
1/PE or 1/20 = 0.05 = 5%
This means that if we assume no growth, we can compute the value of a stock as:
Earnings / Price of Risk = 10/0.05 = 200B
Now comes the interesting part.
Five percent would have been an adequate price of risk in 2021, when the risk-free rate was around 1% and risk premiums were around 4%. However, in 2023 this has changed and riskfree rates are 3.5%, while the risk premium has shot up to about 5%.
This means that instead of the 5% cost of equity, we now need to consider an 8.5% cost of equity for a stock that produces the same 10B in earnings.
By applying our new price of risk, we can see what the stock is worth (assuming no growth):
10 / 0.085 = 118B in implied market cap.
Note that every stock has an individual price of risk, which we need to adjust the 8.5% for. For this example, I am assuming that the stock’s price to risk is equivalent to the S&P 500’s price for risk. In the real world, you need to adjust this - usually not too much.
Our new implied price of risk means that the 200B company may now be worth a good 80b less just because of the rise in the price of risk.
Flipping this to the implied earnings, the same company would now need to produce earnings of 200*0.085 = 17B in order to retain the same market value.
A Practical Example
We can apply this example to a company like MSFT 0.00%↑
We see that Microsoft is producing some adj. (capitalized R&D) TTM earnings of 76.1B, and has a price of risk of about 8.64%. This implies a market cap of 76.1 / 0.864 = 880B
, however the stock trades well above that at 1.85T indicating that investors are factoring-in growth, which is reasonable.
Flipping this around, 1850*0.0864 implies earnings of 160B, but if we assume perpetual growth we get something closer to practice:
1850 * (0.0864 - 0.035) = 95B
Note, the growth rate we use in this formula is a growth rate in perpetuity. If we want to model a number of years with a higher growth rate, we need to use something like a DCF model.
In essence, investors are assuming that Microsoft reaches 95B in earnings in the future. We can use this approach to see if we have reasonable expectations for a company.
What This Means for Us
The way to fight the labelist within us, is to continuously strive to understand the foundation behind a conclusion - to understand “Why”.
Today, we revisited the fundamentals of the PE, and attempted to analyze how they fit into the value of a stock. We are reminded that the building blocks of a PE are the cash flows to equity investors (earnings are just a proxy for these), the expected growth, and the risk we are taking on if we buy an asset.
A suitable PE can be inferred by analyzing how much a company needs to produce in earnings in order to justify the current price of risk.
In testing our thesis, we need to see if the target earnings are attainable for the stock in the future, else we are paying more for risk than the marginal investor is willing to pay.
Now, when we read a PE, we can see that it translates to:
How likely is a stock to produce the needed earnings in a reasonable time-frame.
If the expected earnings stretch too far into the future, we may need to reevaluate our thesis.
In the example, MSFT, as a single cash flow asset needs to produce 160B to justify its current 24x PE ratio, or viewed via the lens of a steady growing asset it needs to produce 95B.
My goal, wasn’t to value MSFT. It was to illustrate why thinking from first principles can help you walk in the dark. Anyone that wants to take on markets needs to be able to utilize first-principles, and be brave enough to act on it in key moments - I believe that knowledge is not enough and courage (BIG 5 trait “disagreeableness”) is a necessary ingredient in good decision-making.
Labelisem is fought by developing rules by which we base our thinking, and not differing to context. In practical terms, anytime we want to check ourselves, we need to try to explain an output before going along with the conclusion.
Thank you for reading my short exploration of labelism vs first-principles.
Next, we will discuss how labelism falls prey to marginal thinking, and what general strategies we can utilize to move away from its pull.
In the meantime, check out some of my past valuation models, and write down a stock that you would like me to analyze.