Is the price-to-book ratio too arbitrary to be useful to investors?

Question

Many financial portal websites show the price-to-book (P/B) ratio of companies. For example, Yahoo Finance shows this P/B ratio for TotalEnergies SE:

Under IFRS (IAS 16 — Property, Plant and Equipment), a company can choose to state the value of property, plant and equipment in the balance sheet using either the "cost model" (historical cost minus depreciation and impairment), or the "revaluation model" (fair value minus depreciation and impairment). It is allowed to account for one class of asset using the "cost model" and another class of asset using the "revaluation model".

Under US GAAP, companies use historical cost minus depreciation and impairment.

These accounting rules seem to affect the relevance of the book value and hence the price-to-book ratio. Historical cost often differs from fair value, so book value based on historical cost may not be relevant. When companies have the choice to use either historical cost or fair value at their own discretion, the book values are somewhat arbitrary ... There's also the issue of comparability across companies. For example, Company A and Company B each bought a piece of land for $100,000. The piece of land is their only asset. They have no liabilities. Each piece of land is now worth $200,000.

• Company A uses the "cost model". The land is carried on the balance sheet at $100,000. If the market capitalization of the company is $200,000, then the P/B ratio is 2.00.

• Company B uses the "revaluation model". The land is carried on the balance sheet at $200,000. If the market capitalization of the company is $200,000, then the P/B ratio is 1.00.

Given the problems I mentioned above, how is the price-to-book ratio relevant to the investor?

Answer 1

You are exactly right: price-to-book has plenty of limitations - the ones you mentioned and many more. For example: consumer company builds brands by marketing which is cost and do not increase book value. But strong brand can be of great value that is not present in book. Think of Coca Cola or Harley Davidson for instance.

So depending on company, book value can be a number more or less distant from reality. So will be the ratio.

Personally I look on book value when analysing banks and distressed companies. Sometimes it shows something when you analyze groups of stocks (indices) or long term value creation in single company.

But in most cases, other metrics are more useful.

Answer 2

The usefulness of price-to-book is extremely limited.

If company A has assets of $100 million, and P/B of 3, it has market value of $300 million (assuming no debt). If now company B buys company A, does the book value of company B increase by $100 million?

No, because that would require one-time expense of $200 million related to buying the company, so $200 million of "goodwill" is invented out of thin air. It is gradually amortized, and today the rules are rather strict: if the goodwill has lost its value, a gradual amortization schedule isn't enough but the value in balance sheet must immediately be reduced to the true value.

However, how is the situation before company B buying company A different from the situation after the purchase? The assets are the same. Yet book value is invented out of thin air.

Also, if a company creates software, investments into the software are classified as immediate expenses. Sometimes this is right, a bad software developer could create so much technical debt that what this developer created could be even worthless, or even have negative value, because if there are agreements with customers related to the bad software, it might be necessary to fix the software at the company's own cost or else every single customer will sue for breach of contract. However, a good developer could create something very valuable. This value is not reflected in the balance sheet. Thus, P/B of software companies is expected to be high.

I would only use P/B as comparison to other similar competitor companies in the same sector, and only as one metric: P/S (or better yet, EV/S) and P/E (or better yet, taxation-corrected EV/EBIT) are useful metrics too and you should make a good estimate using all of those metrics -- not just from one year, but in the case of E or S, using current P or EV but historical average of E and S from let's say 10 years.

Also, for true software companies you can totally forget P/B.

But for a company pretending to be a software company but which actually is a manufacturing company, like Tesla, at P/B 19 it's clear the decision should be "not to buy".