Why is Warren Buffett selling his Apple shares? As Monsieur de la Palice would say, for one simple reason: he must think the share price is too high.

His stake peaked at 5.9%, making him the second-largest shareholder behind mainly passive fund manager Vanguard (8.9%). With a value of $178 billion, his stake in Apple represented just under 20% of Berkshire Hathaway's market capitalization.

Warren Buffett began investing in Apple in 2016, when the stock's P/E ratio was 13. Since then, earnings per share have almost tripled and the share price has risen 9-fold, giving today a P/E ratio of 37. As net cash is fairly negligible in relation to Apple's market capitalization (1.5%), reasoning in terms of P/E ratio rather than operating income multiple is not biased[1].

According to the current consensus of analysts, Apple's sales should grow by 10% per year until 2029, from $361 billion in 2024 to $572 billion, while net income should increase by 14% per year, from $92 billion in 2024 to $179 billion in 2029.

Achieving these forecasts implies that :

• operating margin to rise from 31.7% of sales in 2024 to 36.7% in 2029; Apple has publicly warned that the margin of its future products will struggle to match that of its flagship iPhone;
• pre-tax ROCE, up from 41% in 2016 to 128% (sic) in 2024, still capable of rising to around 150%.

One could be forgiven for not believing it. Apples may fall from apple trees, but trees don't climb to heaven.

Some will argue that the market is efficient in large stocks like Apple, by which they mean that Apple's share price always corresponds to its intrinsic value, insofar as this can be measured, because it faithfully reflects the market consensus. Its acquisition would therefore generate no more and no less than the required rate of return, taking into account its market risk. They'll be amazed that Warren Buffett was able to turn his $39bn investment into €178bn in less than 8 years, earning his shareholders far more (around 25% a year) than the required rate of return on Apple shares (around 10%).

But this is a misunderstanding of the notion of efficiency, fed by abuses of language.

Eugene Fama[2] (Nobel Prize 2013), who did the fundamental work in this field, distinguishes 3 forms of market efficiency:

- the weak form, which states that today's price reflects all the information contained in past prices and volumes. This denies the relevance of chartist analysis, especially if transaction costs are taken into account. This is a form of efficiency on which there is virtual consensus among researchers;

- the semi-strong form, according to which prices reflect all available information, and new information arriving at the market causes prices to vary accordingly (as the election of Donald Trump has caused the prices of certain stocks to vary);

- the strong form, which states that prices even include private information held by those with privileged information. Virtually no researcher believes this. Otherwise, there would be no need for regulations on insiders to preserve market integrity, which would be useless since, on average, insider transactions would be unprofitable. This is the opposite of what we see, hence the very low popularity of this form of efficiency.

 These 3 levels of efficiency do not imply that a share price corresponds to its intrinsic value, as is often said in common parlance. Indeed :

- The fact that new information is incorporated into the share price does not guarantee that this information is correctly integrated, given our biases, emotions and other beliefs.

- On the other hand, at most, it will be the average interpretation (the market consensus) that will be incorporated, and it may be wrong, especially when events with a low probability of occurrence occur.

Indeed, the fact that financial markets can be efficient in the sense of weak and semi-strong forms does not prevent us from being wrong about the current or future value of a share. Who would have thought that Boeing, the world's leading aircraft manufacturer, would be piling up endless operational difficulties from 2019 onwards, culminating (we hope for their sake) in an aircraft door coming off in mid-air? Admittedly, the share price may have incorporated this possibility with a low probability, but when the series of problems arrives and the low probability becomes a fact, the share price naturally adjusts. Similarly, who could have predicted that Tesla would go from nothing to a major carmaker? Here again, the small probability of this truly astounding development may have been in the price, but it's only logical that prices should adjust sharply when this chimera becomes a reality.

Admittedly, we can say in everyday language that the market has not been efficient for Tesla and Boeing, even though it has been efficient in the scientific sense for these two stocks taken as examples. But the market can never be efficient in the conventional sense, with a share price that always gives the investor a return corresponding to the required risk-adjusted rate, making the current price an excellent predictor of future prices.

Let's reiterate that the fact that the current price faithfully reflects the market consensus does not guarantee that in a few years' time, the results expected today will occur exactly, especially for stocks that are not among the least risky on the market.

This is why we believe it would be appropriate to banish the word “efficient” in its current meaning, which corresponds to an illusion. As Warren Buffett reminds us, he has just made the largest and most profitable financial investment in the history of mankind.

It's not because markets are inefficient that Warren Buffett has been able to make his shareholders a lot of money. It's because he saw, earlier than the market average, Apple's ability to raise its prices without discouraging purchases from its aficionados, its capacity to expand into services, and so on. But is this surprising for someone whose first task, as described by his long-time co-investor and Berkshire Hathaway vice-chairman Charlie Munger, who recently passed away, is: “to set aside plenty of time to read and think in peace”?

As to whether the development of passive funds and other ETFs, not to mention the reduction in the number of people who take time out to read and reflect, will reduce the speed of incorporation of new information, or even the incorporation of new information into share prices, we're not worried.

Indeed, in a competitive world, if such an evolution were to occur, it would open up a boulevard to those who read and reflect, boosting their performance and, in a logical pendulum swing, restoring interest and outperformance to active management. This would not go unnoticed for long, and would lead to the opposite of the current trend, with an influx of funds into active management and a smaller share reserved for index funds.

Thus, it is not impossible to think that equity markets are structurally inefficient in the common sense of the term, but are efficient in the scientific sense of the term, in its weak and semi-strong acceptation.

This chart produced using data from Robert Shiller, shows that neither deflation nor inflation is good for equity markets.

Deflation, because it induces a decline in activity, which reduces the free cash flow generated by companies.

Inflation, because it also reduces free cash flow by inflating working capital requirements.

With Simon Gueguen, lecturer-researcher at CY Cergy Paris University

The damage caused by global warming today, and especially in the long term, is a legitimate cause for concern that calls for political decisions. Nearly 200 countries signed the Paris Agreements in 2015, and more than 100 of them have pledged to aim for carbon neutrality in the coming decades. The speed of the transition and the nature of the decisions taken are difficult to anticipate. In economic terms, however, the trend is towards the gradual introduction of rules penalizing the use of fossil fuels. As finance anticipates the economy, these prospects are already affecting polluting companies. They will have to bear the cost of the transition, and this anticipated cost is reflected in a discount on the price of their equity. Equally, these companies must provide their shareholders with a higher immediate return, known as the carbon premium, to compensate for the uncertain consequences of the transition. This premium has already been measured in academic studies. The study we present here[1] has one particularity: its sample size. It comprises 14,400 listed companies in 77 countries.

The first result of the study is that the carbon premium is present and significant in all the countries in the sample. The order of magnitude is easy to remember: around one percentage point of premium for one standard deviation of variation in the group's emission rate. Differences between countries are relatively small. Comparing the two biggest carbon emitters, China and the United States, the authors measure 1.18% for the former and 0.95% for the latter. The fact that the premium is slightly higher in China may come as a surprise; this result shows that institutional concern for carbon emissions (higher in the USA) does not translate into a higher premium. Similarly, energy consumption does not affect the level of the premium (only production counts).

The study verifies that this higher profitability linked to the carbon premium does indeed correspond to a lower valuation. For the sample as a whole, and after taking into account differences between countries and sectors, the ratio of balance sheet equity to equity value (market capitalization) is 13.2% higher for one standard deviation of additional emissions. If we prefer, the inverse ratio, i.e. the PBR, is lower, all other things being equal, for companies that issue more.

Finally, the bonus affects all sectors of the economy, not just those associated with the highest emission levels. It takes into account both direct and indirect emissions, i.e. those of customers and suppliers. The authors thus emphasize that the market is capable of identifying carbon emissions across the entire value chain, and is not content with company-specific data. They also show that the variation in the emission rate (i.e. the variation in the variation) is also reflected in the level of the premium. This is why the emission premium can be low for companies that are still polluting in relation to their sector, but are playing the transition game.

According to the authors, the results suggest that the market plays an important role in transition incentives. If governments find it difficult to coordinate to implement a carbon tax, due to divergent interests and competition problems, the market integrates at least part of the cost of transition into security prices (and therefore into the cost of equity), and does so by taking the entire value chain into account. The difficulties encountered in meeting commitments to converge towards carbon neutrality since the Paris agreements suggest that transition costs will increase. For this reason, the authors believe that the carbon premium observed on equity markets should increase in the coming years.

Finally, it should be noted that, like much of the work on this subject in finance, only carbon-related pollution is taken into account. Other sources of pollution would merit equivalent work, but they are generally more difficult to measure by the market... and by researchers.

In response to a shareholder's question about his assumptions concerning Esker's working capital[1], in the context of a discounted free cash flow valuation, we read from the independent expert Finexsi:

“Any improvement in working capital does not mechanically translate into an increment in the intrinsic value of the Company, as the cash generated is not necessarily available (at least not in full) to be passed on to shareholders.”

The first part of the sentence is quite relevant, as an improvement in working capital can be bought, as it were, by paying suppliers later, who will, for example, remove their discounts for prompt payment, or increase their sales prices. Or by making customers pay more promptly, but who may ask for discounts or lower prices. So a variation in working capital doesn't necessarily have the same impact on value as a simplistic Excel model might lead finance beginners to believe[2], since it can negatively affect other valuation parameters, such as margins in this case.

The second part of the sentence seems to us to be wrong. A discounted free cash flow, or DCF for short, simply assumes that free cash flows are generated. And nothing else. And particularly not that they are available to be passed on to shareholders. They simply have to be available, i.e. after the company has paid all its overdue suppliers, taxes and personnel, and made the investments it deems necessary or advisable. This free cash flow is then available for :

- either reduce the company's debt by repaying debts, or be temporarily invested in cash pending future use,

- finance investments not included in the business plan used as the basis for the DCF calculation, in particular acquisitions,

- or to be distributed to shareholders in the form of dividends or share buy-backs.

However, this last allocation is only one of many. It is not required for free cash flow to be included in the DCF calculation.

There are rare cases, due to poor governance caused by incompetence, ignorance, dishonesty or abuse on the part of a majority shareholder who wants to disgust and discourage minority shareholders, where we can observe situations where the cash generated, and correlatively the equity capital, accumulates within a company without any need and constitutes a microeconomic, or even macroeconomic, mess. It's a mess because the equity generated, which could be used to finance investments and risk-taking - which is what it's there for - is then neutralized by being invested in treasury bills or money-market mutual funds. This was the case at Apple for tax reasons, until activist shareholders forced the company's management to back down in 2012, forcing it to initiate dividend payments and share buybacks so that Apple's free cash flow did not accumulate endlessly in its treasury, and correlatively in equity.

And in such cases, which again are statistically rare, investors are entitled to apply a discount to the valuation of this excess cash. This is what Apple was doing at the time. But this discount comes after the calculation of discounted free cash flows in a DCF, not in this calculation.

It should be noted that in the case of the valuation of a company entering an LBO, a case dealt with by the independent expert, we can trust the LBO funds, who are no choirboys, to ensure that the free cash flow generated by the company is used correctly. They are not passed on to shareholders, but used to repay the LBO debt. Whether or not they are passed on to shareholders, they are taken into account when calculating a value by discounting free cash flow.

Regularly on the Vernimmen.com Facebook page^[1] we publish comments on financial news that we deem to be of interest, publish a question and its answer or quote of financial interest. Here are some of our recent comments.

Robertet or the value of a voting right _{(November 23)}

There are times when we need to consider the value of a non-voting share, and the subject is not a simple one.

Last week's exit of Firmenich from the capital of Robertet, world leader in natural raw materials for perfumes and aromas, is illuminating. Robertet is the only one of the 818 companies currently listed in Paris to have both listed voting and non-voting shares.

In 1987, because the family's control over Robertet was fragile, the company had stripped ordinary shares into investment certificates (IC) and voting rights certificates (VRC): an IC is a share deprived of its voting rights.

ICs were created in 1983 by the French government to finance nationalized companies. BNP, Saint-Gobain and others issued CIs to the public, without diluting the State's stake in their capital, as the State kept the associated VRCs. Private firms such as Bouygues and L'Oréal followed suit. But the innovation, at least for listed companies, didn't catch on, and soon all these ICs disappeared through redemption or conversion into ordinary shares with voting rights.

All, that is, except those of Robertet, which now has 3 stock market quotation lines: ordinary shares, ICs and VRCs. The latter are very rarely traded, as they are mainly held by the family. The ICs are much more widely traded, but their liquidity is reduced by the fact that Firmenich held 85% of them.

Firmenich sold most of its shares and ICs to FSP and Peugeot Invest. The latter negotiated with Robertet's founding family to convert their ICs into ordinary shares. The parity retained values the VRCs at 10% of the value of the shares, which corresponds to the average discount observed in 2024 between the value of the IC and that of the share.

As the founding family has consolidated its power over Robertet (it now holds 63% of the voting rights with 38% of the share capital), the IC discount has logically narrowed from 30% in 2010 to 20% in 2019, and 10% now, as voting rights have lost much of their stakes. Thus the value of a voting right depends very much on the configuration of the shareholder base. Other factors come into play, such as the ability of a third-party purchaser of control to improve margins, or possible competition between several third-party purchasers.

It's easy to see why it's hard to see a discount on Hermès shares, where the limited partnership structure leads to solid family control and it's hard to see how one can do better than the current management. Or why, in agricultural cooperatives, the principle of one vote per cooperative member, irrespective of the shares held, means that voting rights lose all value. As a result, all shares in a cooperative have the same value, whether or not they carry voting rights, since a voting right is worthless. In this case, the absence of voting rights does not justify any discount.

Takeover bid for Esker: the watchdog market authority (AMF) and the independent expert can do better _{(December 7)}

The current takeover bid for Esker, a specialist in the digitization of finance department processes, led by the 2 investment funds Bridgepoint and General Atlantic in alliance with the company's management, highlights a number of points that could be significantly improved.

The lowering in 2020 of the threshold for expropriating minority shareholders from 95% to 90%, a reform favorable to issuers wishing to delist, has been accompanied by new rights for minority shareholders. Minority shareholders can now interact with the independent expert during the course of her/his mission, commenting on his/her preliminary report and making suggestions. It is now up to the expert to include these suggestions in the final version of his/her work, giving his/her point of view on these remarks and what he/she has done with them in his/her final assessment of the conditions of the offer.

Gilles Chaufaud, an equity investor specializing in small and mid cap companies, who has been following Esker since the late 1990s and therefore knows the company very well, wrote to the independent expert (Finexsi) to underline 2 points, which prompted 3 comments from us.

1. Anonymizing comments weakens their impact

His comments have been anonymized in the report. An AMF-vetted prospectus is not an on-line stock market discussion forum where anonymity reigns. The expert should give the names of the investors writing to him/her, as anonymizing them reduces the scope of their remarks, as a reader would assume that they were ordinary people. However, this is not a question of an ordinary person, but of a professional who is just as qualified in his field as the independent expert.

2. Inconsistent growth and working capital assumptions

Beyond the projection horizon, the appraiser forecasts zero variation in working capital, whereas he anticipates perpetual sales growth of 3.5% a year, and has assumed a constant working capital of -4.2% of sales over the projection period. This assumption implies that working capital would become progressively less negative.

Finexsi's response is surprising, to say the least: “Any improvement in working capital does not automatically translate into an increment in the intrinsic value of the Company, as the cash generated is not necessarily available (at least not in its entirety) to be passed on to shareholders”. While we agree with the first part of this sentence, we disagree with the second. In a DCF, the main methodology used by the expert, it is not assumed that cash will be returned to shareholders, but that it has been generated. It is then used by management to make acquisitions not modelled in the projections, or to repay the debts of the planned LBO, or to pay dividends, or to leave cash on hand. However, the fact that it is passed on to shareholders is not a prerequisite for the validity of a DCF.

3. Gilles Chaufaud underlined a 2nd point to the independent expert (Finexsi): he was surprised that the Kyriba transaction was not included in the comparables. In fact, General Atlantic (GA) has announced that it is acquiring a minority stake in Kyriba, another specialist in software for financial management, controlled by... Bridgepoint. And GA is paying a higher multiple for a minority stake in Kyriba than Bridgepoint and GA are proposing to pay together to acquire control of Esker. ... It is likely that Esker and Kyriba will be merged in the future. Why else, in the takeover bid memorandum, would GA and Bridgepoint have given themselves the option of abandoning their bid if they did not reach the 60% threshold, in addition to the legal threshold of 50%? Indeed, if the 90% threshold were not reached at the end of the takeover bid, Esker would remain listed on the stock exchange. With a minimum of 60% of voting rights and share capital, Bridgepoint and GA could have more than 67% of voting rights at the AGM, enabling them to have a merger between Kyriba and Esker approved, due to the abstention of a number of shareholders. In other words, the reference to this transaction, omitted by the expert, deserved better than to be hushed up.

The AMF and the expert's response on this point were particularly disappointing.The AMF could have demanded that a paragraph be devoted to this post-takeover merger project, which Esker's management mentioned at an investor meeting, and that it be included in the offer document. However, this is not the case, contrary to the principle of good information for shareholders, who must decide whether or not to tender their shares to the takeover bid.

As for the expert, he states that his policy is to only take into account completed acquisitions when calculating transaction multiples, and not those that have only been announced (which we have not verified; we have seen experts change their methodology according to their files). The simultaneity of the 2 transactions, Bridgepoint having brought GA into Kyriba on October 16, after announcing their joint intention to take control of Esker on September 19, means that we find this answer light.

Finexsi points out that the higher multiple paid by GA for Kyriba can be explained by the latter's higher growth rate. This seems a good argument, but does not stand up to analysis, since the expert has not examined the growth rate of the comparables retained elsewhere, and averages their multiples without considering their growth rate.

These non-answers can only sow doubts, especially when, on the one hand, you have a transaction between 2 funds that are no choirboys, to which the company's management has rallied, and the public on the other. All the more reason to be perfectly transparent and not to consider the shareholder who questions you as a dog in a bowling alley. 

^[1]  Like it here