Someone who reads my blog sent me this email:
Have you read "The Warren Buffett Portfolio" by Robert Hagstrom?
If you have what are your thoughts on Chapter 6, The Mathematics of Investing? In this chapter the author describes using probability theory based on Fermat/Pascal system along with the Bayesian Approach to determine if an investment has odds in your favor. In a nutshell it uses decision trees to help with the analysis.
Do you use probability theory and decision trees in your methodology for selecting stocks?
Yes. I've read "The Warren Buffett Portfolio." I just pulled it off my bookshelf and re-read Chapter 6 after getting your email.
No. I don't use probability theory and decision trees in my stock picking.
I don't think of investing in terms of probabilistic analysis. I think of investing in terms of a defensive analysis.
I look for value. Then I look for ways to defend that value. As long as the gap between price and value is large, you just buy the opportunity with the best defenses.
What are the best defenses?
There are a couple approaches that work. The strategy I tend to prefer — just because it seems to be undervalued by the market at all times — is defense in depth.
Defense in depth works by not depending on a single strong point to defeat an attack. Instead it uses multiple layers — or compartments — to defeat an attack. Ideally, an attack that could overcome one layer (breach one compartment) is incapable of overcoming the next layer, etc.
In this way, defense in depth preserves a defensible position at all times unless there are multiple, independent failures.
A very obvious example is when Ben Graham bought some bonds of a small railroad that was part of multiple larger rail networks. The company's bonds were guaranteed by about seven different railroads. The risk that all seven railroads would default was close to nil. People underestimated the safety provided by requiring seven defaults instead of one. If you are being guaranteed by seven good credits, you have just about the best credit imaginable. It isn’t the strength of any one defensive position that matters in that case — it’s the cumulative defense provided by having six more fallback positions.
If you look at most stocks, they depend entirely on one position for their defense.
If you asked: And what if earnings fail to ever grow? Or what if earnings evaporated?
That’s it.
There's no further defense.
No fallback position.
And if you’ve paid two or three times tangible book value for that stock — well, now you’re facing something like a 50% to 70% permanent loss.
And that’s ignoring the issue of insolvency entirely. Businesses do go belly up. And investors lose everything on some stocks.
So, for a great many stocks, your only protection against losses of 50% to 100% is the company’s earnings.
You have only one defense.
Now, sure it's a very big defense. Earnings are usually what investors spend the most time analyzing.
But unless the company is a regulated rail, utility, etc. or Coca-Cola (KO, Financial) or something equally impenetrable there's a real risk of maybe a 50% to 100% loss built into most stocks.
All these stocks depend on a single line of defense. They need to be able to make a profit on their product. They can't be pushed too hard for too long by their competitors, or customers, or suppliers, or employees, or by some tech change, or some taste change, or they will buckle and there will be no defensive position left from which to guard the value you originally bought the stock for.
Again, we’re talking about stocks generally here. If you buy a stock at a price-to-tangible book ratio of 1 or at less than its net current asset value or anything like that — the situation is different. You won’t necessarily lose money if competition destroys the special earning power of the business’s capital. If you never pay more than $1 of your money to buy every $1 of the business’s tangible capital — then you have a different kind of defense. You have the general earning potential of capital protecting you. The business can do badly in its current product lines and you might still be able to salvage your investment if they can eventually earn a mediocre return on capital someplace else.
Since most stocks in the U.S. sell for quite a bit more than their tangible book value — most stocks lack this defense. They are vulnerable to anything that reduces their special ability to earn good returns on tangible capital.
That special earning power is what needs to be preserved when you buy a stock at any price above tangible book value.
I'm not saying you can't depend on just one defense.
There are defenses — primarily spaces in the customer mindscape — where you can utterly dominate a position forever. Usually, these are products that consist of nothing but the position they hold in their customer's minds.
Businesses like FICO, Dun & Bradstreet, Moody's, S&P, etc.
These businesses are nothing but a name. An investment in a business like that is defended by a single strong point. But it's a very strong point.
Coca-Cola can make Pepsi (PEP, Financial) and Pepsi can make Coca-Cola. Neither company is stuck with just one formula. But they are stuck with just one brand name.
When Warren Buffett bought Coca-Cola stock he was betting on a single strong point. He was betting on the mental real estate the Coca-Cola company owned around the globe.
A single strong point isn't usually enough of a defense. By definition, most businesses are ordinary. There will never be more than a few extraordinary businesses. Most companies can never be Coke. So, extraordinary defenses will only rarely come from the business itself.
More frequently, you will find that price is your best defense when buying a stock. Especially a price that gives you two separate defensive compartments:
1. Earnings
2. Assets
Now, you don't want to double count.
But when you find a business like George Risk (RSKIA, Financial) — that struck me as a very strongly defended value investment. It was being offered — I think the stock was around $4.30 to $4.70 when I first came across it — at less than the cash & investments on its balance sheet (which were about $4.75). The company probably had "normal" cash earnings power — this was during the housing bust, so trailing earnings didn't reflect this — of maybe 40 cents a share.
If you bought the stock for $4.30 to $4.70 you were getting something which on an asset basis was worth $4.75 and on an earnings basis might be worth $6.00.
Now, you can certainly argue about any of these points. I’m not wedded to the exact assumptions. But the idea here is the two separate compartments of defense. You have one compartment with the $4.75 in cash and investments. And you have a separate compartment with the earnings that provide maybe $6.00 in value. I also thought the earnings compartment was pretty strong in the sense that the company had competitive advantages in its business. Basically it had (relative) pricing power because of its own reputation and the way its customers used its products.
Pricing power is the best way to defend your earnings value. And it was obvious George Risk had that defense. Because it had thrived for a long time at a severe cost disadvantage to its competitors. The kind of cost disadvantage that is very quickly lethal to most businesses in most industries. So, it was kind of like seeing someone drink poison first thing every morning and continue merrily on with their day.
The immunity got my attention.
So, I thought you were paying around $4.50 or whatever it was at the time for something worth more like $10. Let's call that half price. You were paying about 50 cents on the dollar. And that dollar of value was split into two separate compartments (that would have to fail independently). And compartment #2 had strong defenses.
So, no, I definitely didn't think probabilistically about the investment. I looked at the last 17 years of earnings. I looked at the exposure to new construction — especially residential — and saw that housing starts were at a low and unsustainable level. They were unsustainable in the sense that if we continue to have 2009-2010 type housing starts forever, we'd eventually live in households of five or six people, which I didn't think likely. So, eventually, housing starts were going to tick up. And the company wasn't losing money even in close to the worst construction environment around. So, I kind of asked how bad the situation could get.
I asked myself questions like — what if the company earns nothing for the next 10 years (because construction is much, much worse than I could ever imagine) and the market refuses to ever pay more than 10 times earnings for the company.
Okay.
If in 10 years’ time, the market valued the normal 40 cents of earnings at only 10 times, that would mean only a $4.00 earnings value. Add that to the $4.75 in cash and investments (which I figured would go nowhere for 10 years) and you get $8.75. If I buy the stock at $4.50 today and wait 10 years, I'll get to sell my stock for $8.75 at that time. My annual return will be 7%. That's about as much as I expect from the S&P 500.
And that's the downside.
You only hit that downside if the company actually earns nothing for a while. Even though it was still posting a profit in the worst housing bust in history.
So, I looked at it entirely in terms of defenses.
I kind of asked what "attacks" could breach my defenses. Now, you could obviously say well what if they misinvested the $4.75 and lost it all and then the core business actually starting losing money, couldn't that cause you to actually lose money on the investment.
Sure.
There’s always some way to overcome any defensive system.
But in this case it would take something extreme like that.
And that’s the best an investor can hope for.
To find a couple stocks with defenses like that.
So, I don't think probabilistically about stocks.
I think in terms of defenses.
Follow Geoff at Gannon On Investing
Have you read "The Warren Buffett Portfolio" by Robert Hagstrom?
If you have what are your thoughts on Chapter 6, The Mathematics of Investing? In this chapter the author describes using probability theory based on Fermat/Pascal system along with the Bayesian Approach to determine if an investment has odds in your favor. In a nutshell it uses decision trees to help with the analysis.
Do you use probability theory and decision trees in your methodology for selecting stocks?
Yes. I've read "The Warren Buffett Portfolio." I just pulled it off my bookshelf and re-read Chapter 6 after getting your email.
No. I don't use probability theory and decision trees in my stock picking.
I don't think of investing in terms of probabilistic analysis. I think of investing in terms of a defensive analysis.
I look for value. Then I look for ways to defend that value. As long as the gap between price and value is large, you just buy the opportunity with the best defenses.
What are the best defenses?
There are a couple approaches that work. The strategy I tend to prefer — just because it seems to be undervalued by the market at all times — is defense in depth.
Defense in depth works by not depending on a single strong point to defeat an attack. Instead it uses multiple layers — or compartments — to defeat an attack. Ideally, an attack that could overcome one layer (breach one compartment) is incapable of overcoming the next layer, etc.
In this way, defense in depth preserves a defensible position at all times unless there are multiple, independent failures.
A very obvious example is when Ben Graham bought some bonds of a small railroad that was part of multiple larger rail networks. The company's bonds were guaranteed by about seven different railroads. The risk that all seven railroads would default was close to nil. People underestimated the safety provided by requiring seven defaults instead of one. If you are being guaranteed by seven good credits, you have just about the best credit imaginable. It isn’t the strength of any one defensive position that matters in that case — it’s the cumulative defense provided by having six more fallback positions.
If you look at most stocks, they depend entirely on one position for their defense.
If you asked: And what if earnings fail to ever grow? Or what if earnings evaporated?
That’s it.
There's no further defense.
No fallback position.
And if you’ve paid two or three times tangible book value for that stock — well, now you’re facing something like a 50% to 70% permanent loss.
And that’s ignoring the issue of insolvency entirely. Businesses do go belly up. And investors lose everything on some stocks.
So, for a great many stocks, your only protection against losses of 50% to 100% is the company’s earnings.
You have only one defense.
Now, sure it's a very big defense. Earnings are usually what investors spend the most time analyzing.
But unless the company is a regulated rail, utility, etc. or Coca-Cola (KO, Financial) or something equally impenetrable there's a real risk of maybe a 50% to 100% loss built into most stocks.
All these stocks depend on a single line of defense. They need to be able to make a profit on their product. They can't be pushed too hard for too long by their competitors, or customers, or suppliers, or employees, or by some tech change, or some taste change, or they will buckle and there will be no defensive position left from which to guard the value you originally bought the stock for.
Again, we’re talking about stocks generally here. If you buy a stock at a price-to-tangible book ratio of 1 or at less than its net current asset value or anything like that — the situation is different. You won’t necessarily lose money if competition destroys the special earning power of the business’s capital. If you never pay more than $1 of your money to buy every $1 of the business’s tangible capital — then you have a different kind of defense. You have the general earning potential of capital protecting you. The business can do badly in its current product lines and you might still be able to salvage your investment if they can eventually earn a mediocre return on capital someplace else.
Since most stocks in the U.S. sell for quite a bit more than their tangible book value — most stocks lack this defense. They are vulnerable to anything that reduces their special ability to earn good returns on tangible capital.
That special earning power is what needs to be preserved when you buy a stock at any price above tangible book value.
I'm not saying you can't depend on just one defense.
There are defenses — primarily spaces in the customer mindscape — where you can utterly dominate a position forever. Usually, these are products that consist of nothing but the position they hold in their customer's minds.
Businesses like FICO, Dun & Bradstreet, Moody's, S&P, etc.
These businesses are nothing but a name. An investment in a business like that is defended by a single strong point. But it's a very strong point.
Coca-Cola can make Pepsi (PEP, Financial) and Pepsi can make Coca-Cola. Neither company is stuck with just one formula. But they are stuck with just one brand name.
When Warren Buffett bought Coca-Cola stock he was betting on a single strong point. He was betting on the mental real estate the Coca-Cola company owned around the globe.
A single strong point isn't usually enough of a defense. By definition, most businesses are ordinary. There will never be more than a few extraordinary businesses. Most companies can never be Coke. So, extraordinary defenses will only rarely come from the business itself.
More frequently, you will find that price is your best defense when buying a stock. Especially a price that gives you two separate defensive compartments:
1. Earnings
2. Assets
Now, you don't want to double count.
But when you find a business like George Risk (RSKIA, Financial) — that struck me as a very strongly defended value investment. It was being offered — I think the stock was around $4.30 to $4.70 when I first came across it — at less than the cash & investments on its balance sheet (which were about $4.75). The company probably had "normal" cash earnings power — this was during the housing bust, so trailing earnings didn't reflect this — of maybe 40 cents a share.
If you bought the stock for $4.30 to $4.70 you were getting something which on an asset basis was worth $4.75 and on an earnings basis might be worth $6.00.
Now, you can certainly argue about any of these points. I’m not wedded to the exact assumptions. But the idea here is the two separate compartments of defense. You have one compartment with the $4.75 in cash and investments. And you have a separate compartment with the earnings that provide maybe $6.00 in value. I also thought the earnings compartment was pretty strong in the sense that the company had competitive advantages in its business. Basically it had (relative) pricing power because of its own reputation and the way its customers used its products.
Pricing power is the best way to defend your earnings value. And it was obvious George Risk had that defense. Because it had thrived for a long time at a severe cost disadvantage to its competitors. The kind of cost disadvantage that is very quickly lethal to most businesses in most industries. So, it was kind of like seeing someone drink poison first thing every morning and continue merrily on with their day.
The immunity got my attention.
So, I thought you were paying around $4.50 or whatever it was at the time for something worth more like $10. Let's call that half price. You were paying about 50 cents on the dollar. And that dollar of value was split into two separate compartments (that would have to fail independently). And compartment #2 had strong defenses.
So, no, I definitely didn't think probabilistically about the investment. I looked at the last 17 years of earnings. I looked at the exposure to new construction — especially residential — and saw that housing starts were at a low and unsustainable level. They were unsustainable in the sense that if we continue to have 2009-2010 type housing starts forever, we'd eventually live in households of five or six people, which I didn't think likely. So, eventually, housing starts were going to tick up. And the company wasn't losing money even in close to the worst construction environment around. So, I kind of asked how bad the situation could get.
I asked myself questions like — what if the company earns nothing for the next 10 years (because construction is much, much worse than I could ever imagine) and the market refuses to ever pay more than 10 times earnings for the company.
Okay.
If in 10 years’ time, the market valued the normal 40 cents of earnings at only 10 times, that would mean only a $4.00 earnings value. Add that to the $4.75 in cash and investments (which I figured would go nowhere for 10 years) and you get $8.75. If I buy the stock at $4.50 today and wait 10 years, I'll get to sell my stock for $8.75 at that time. My annual return will be 7%. That's about as much as I expect from the S&P 500.
And that's the downside.
You only hit that downside if the company actually earns nothing for a while. Even though it was still posting a profit in the worst housing bust in history.
So, I looked at it entirely in terms of defenses.
I kind of asked what "attacks" could breach my defenses. Now, you could obviously say well what if they misinvested the $4.75 and lost it all and then the core business actually starting losing money, couldn't that cause you to actually lose money on the investment.
Sure.
There’s always some way to overcome any defensive system.
But in this case it would take something extreme like that.
And that’s the best an investor can hope for.
To find a couple stocks with defenses like that.
So, I don't think probabilistically about stocks.
I think in terms of defenses.
Follow Geoff at Gannon On Investing
