Please bear with my slowness. Iâm still on Buffettâs 2010 letter. Each of his letters is a rich mine of wisdom, and I donât want to miss a single piece.
In his 2010 letter, Buffett has a section titled âOn Reporting and Misreporting: The Numbers That Count and Those That Donât.â He explained in that section why net income is almost always irrelevant when valuing Berkshire. Buffett said â[âŚ] Charlie and I could â quite legally â cause net income in any given period to be almost any number we would like.â This is because net income counts only realized gains or losses on investments, but not unrealized ones. They can legally realize some investment gains if they ever want to juice up Berkshireâs earnings reports, though they never play this type of trick. Nonetheless, any of their decisions to realize investment gains or losses will affect Berkshireâs net income numbers. Considering the giant investment portfolio they carried and their extremely infrequent investment transactions, Berkshireâs net income numbers will be thrown out of the track every once in a while.
If net income is irrelevant, so is cash flow. Like net income, cash flow counts only realized investment gains or losses. And so is discounted cash flow model, which is cherished by many investors, especially those who have academic backgrounds. Academic researchers would spend respectful effort to evaluate the effectiveness of discounted cash flow model. Nice, but irrelevant. I wouldnât be surprised if they find that sometimes it works while sometimes it doesnât.
Buffettâs wisdom is, allow me to repeat, investors shall understand the business first and then choose appropriate financial ratios or numbers when valuing it. Diving into numbers before developing a thorough understanding is often dangerous. Any valuation model, in a nutshell, is a mechanical number cruncher. They bear no knowledge of the business and they simply marshal input numbers to produce an output. If investors didnât do a good job to feed in sensible numbers, they shouldnât expect the output to make any sense. There is a concise term for this phenomenon: garbage in, garbage out.
Fluent investors may have noticed an underlying connection from this to Buffettâs investment practice: Buffett only invests on businesses he could understand, though this is frequently cited as the reason why Buffett didnât invest on hi-tech names. Businesses are like people, each has its unique strength and weakness. An investorâs job is to understand the uniqueness and pick representative numbers, so he knows what numbers to look at when evaluating a business among its peers. It is said that investment is half art and half science. I think to understand the business is the art half.
Back to Buffettâs letter, he mentioned many different numbers when discussing different business segments or individual companies under Berkshireâs umbrella. Let me list a few.
Even the same number may have different meaning to different companies. For example, when discussing insurance float, Buffett said: âThis float is âfreeâ as long as insurance underwriting breaks even. [âŚ] I also expect us to average breakeven results or better in the future. If we do that, [âŚ] float [âŚ] can be viewed as an element of value for Berkshire shareholders.â So for a lousy insurer that always incurs underwriting losses, investors may want to mark down the value of the float it carries.
Lastly, I would like to say something about discounted cash flow model. Admittedly this is not from Buffett so I will take all the blame if what I say is inappropriate. Advocators of DCF model may agree that for a specific company one wants a suitable model to be chosen, but still argue that when applied to a collection of many companies, DCF model will stand out because personalities of the companies will be averaged away. I agree that DCF model, as well as other sophisticated models, has merit. But if I want to evaluate a collection of many companies, Iâd rather use P/E ratio. Yes personalities are averaged away, but then sociality emerges. We have to consider social behaviors when looking at a collection of many companies. Because popularity is an important character of social behavior, Iâd choose P/E for its superior popularity.Also check out:
In his 2010 letter, Buffett has a section titled âOn Reporting and Misreporting: The Numbers That Count and Those That Donât.â He explained in that section why net income is almost always irrelevant when valuing Berkshire. Buffett said â[âŚ] Charlie and I could â quite legally â cause net income in any given period to be almost any number we would like.â This is because net income counts only realized gains or losses on investments, but not unrealized ones. They can legally realize some investment gains if they ever want to juice up Berkshireâs earnings reports, though they never play this type of trick. Nonetheless, any of their decisions to realize investment gains or losses will affect Berkshireâs net income numbers. Considering the giant investment portfolio they carried and their extremely infrequent investment transactions, Berkshireâs net income numbers will be thrown out of the track every once in a while.
If net income is irrelevant, so is cash flow. Like net income, cash flow counts only realized investment gains or losses. And so is discounted cash flow model, which is cherished by many investors, especially those who have academic backgrounds. Academic researchers would spend respectful effort to evaluate the effectiveness of discounted cash flow model. Nice, but irrelevant. I wouldnât be surprised if they find that sometimes it works while sometimes it doesnât.
Buffettâs wisdom is, allow me to repeat, investors shall understand the business first and then choose appropriate financial ratios or numbers when valuing it. Diving into numbers before developing a thorough understanding is often dangerous. Any valuation model, in a nutshell, is a mechanical number cruncher. They bear no knowledge of the business and they simply marshal input numbers to produce an output. If investors didnât do a good job to feed in sensible numbers, they shouldnât expect the output to make any sense. There is a concise term for this phenomenon: garbage in, garbage out.
Fluent investors may have noticed an underlying connection from this to Buffettâs investment practice: Buffett only invests on businesses he could understand, though this is frequently cited as the reason why Buffett didnât invest on hi-tech names. Businesses are like people, each has its unique strength and weakness. An investorâs job is to understand the uniqueness and pick representative numbers, so he knows what numbers to look at when evaluating a business among its peers. It is said that investment is half art and half science. I think to understand the business is the art half.
Back to Buffettâs letter, he mentioned many different numbers when discussing different business segments or individual companies under Berkshireâs umbrella. Let me list a few.
- Insurance float, when discussing insurers on page on page 6,
- Earnings on unleveraged net tangible asset, when discussing manufacturing, service and retailing operations on page 12,
- Sales per employee, when discussing CTB on page 13, and
- Equipment utilization, when discussing XTRA on page 16.
Even the same number may have different meaning to different companies. For example, when discussing insurance float, Buffett said: âThis float is âfreeâ as long as insurance underwriting breaks even. [âŚ] I also expect us to average breakeven results or better in the future. If we do that, [âŚ] float [âŚ] can be viewed as an element of value for Berkshire shareholders.â So for a lousy insurer that always incurs underwriting losses, investors may want to mark down the value of the float it carries.
Lastly, I would like to say something about discounted cash flow model. Admittedly this is not from Buffett so I will take all the blame if what I say is inappropriate. Advocators of DCF model may agree that for a specific company one wants a suitable model to be chosen, but still argue that when applied to a collection of many companies, DCF model will stand out because personalities of the companies will be averaged away. I agree that DCF model, as well as other sophisticated models, has merit. But if I want to evaluate a collection of many companies, Iâd rather use P/E ratio. Yes personalities are averaged away, but then sociality emerges. We have to consider social behaviors when looking at a collection of many companies. Because popularity is an important character of social behavior, Iâd choose P/E for its superior popularity.Also check out:
