Vltava Fund letter for the third quarter ended September 30, 2017.
Three investors were sitting in a pub, arguing over how expensive the US equity market is. The first one asserted that the S&P 500 was trading at 24 times earnings per share – that is to say with a P/E of 24. The second claimed that the P/E was 21.6, and the third investor said the P/E was 17.3. All three had based their assessments on the same source – official data from Standard & Poor’s. How is it possible that their numbers were so different? It is because each was using a different definition of earnings per share. All three were correct, but each was speaking about something different.
This might have sounded strange to somebody casually passing by. After all, earnings are earnings, right? But it is not quite so straightforward in finance and investment. When we say earnings, this can be understood variously and so it is necessary to specify which earnings we have in mind. Earnings per share can essentially be categorised in two ways: first on a time basis into actual, historical earnings and future expected earnings and second according to how they are calculated.
The first investor, who stated that the market was trading at 24 times earnings per share (EPS), was referring to historical earnings – the actual earnings achieved over the past 12 months and calculated according to GAAP (Generally Accepted Accounting Principles). These earnings are of a high quality because they are real, and therefore not an estimate, and they are calculated in accordance with unified accounting rules. The second investor also was talking about historical earnings actually reported, but these had not been calculated in accordance with GAAP. Rather, they had been adjusted (one might say “enhanced”) by the individual companies’ managements. So, what does that mean?
In recent years, the managements of US companies have increasingly preferred to report their own, adjusted, non-GAAP earnings rather than those calculated according to GAAP. They use various terms for these earnings, such as adjusted earnings, operating earnings, core earnings, pro forma earnings, earnings ex certain items, etc. Such reported earnings have one thing in common: They exclude items prescribed by GAAP and include items prohibited by GAAP. The managements assert that such adjusted earnings more precisely reflect the actual profitability of their companies. It is interesting that these earnings almost always turn out to be much better than GAAP profits, and quite markedly so. Most probably, the motivation for these corporate managements is something other than an endeavour to provide a more perfect accounting.
According to various studies and statistical data, non-GAAP earnings reported by US companies in recent years have been fully 22% higher than their earnings calculated pursuant to GAAP. That is an immense difference. For individual companies, the differences are often even much greater. Every investor who wants to value publicly traded companies must know about this problem and know how to deal with it.
The most frequently encountered and largest items where GAAP and non-GAAP numbers differ are one-off depreciation of assets and one-off amortisation events. If a company suddenly writes off a chunk of its own assets, it either is admitting that its historical earnings were overstated (because its stated depreciation was too low in previous years) or it is promising that its future earnings will be overstated (because the large one-off write-down of assets will reduce ongoing depreciation long into the future).
Other items frequently tucked into non-GAAP statements are costs related to acquisitions and restructuring costs. Companies often present these as one-off items that should not be taken into consideration, even though they recur in the statements year after year.
My “favourite” item, which is widely used especially among companies in the technology sector, is remuneration to employees paid in company stock. The companies issue new shares and give them to employees in place of salaries. Because these are not cash payments, the companies imply that these are not costs at all, which is just plain ludicrous. And we are not talking about small change here, either, but rather large sums. Here is a beautiful example from the accounting of the company Splunk:
Last year, Splunk recorded revenues of USD 950 million. This may be the only number upon which the management and we agree. In its non-GAAP results, Splunk states that its net earnings were USD 55 million for last year. The company’s GAAP accounting, however, reports a loss of USD 355 million. The vast majority of the difference is due to salaries paid in the form of shares – a whopping USD 378 million. Now, losing USD 355 million on revenues of USD 950 million is a very bad result. The management, however, seems satisfied and is proud of its own tailor-made bottom line presenting Splunk as a profitable company.
The crucial problem with non-GAAP earnings is that to a large degree they are a product of the respective management’s own creativity. Then too, of course, they also are unaudited, provide investors with distorted information, and make comparing individual companies much more difficult. Such earnings data are therefore of low quality and unsatisfactory in their representativeness. The second of our three investors in the pub was using this type of earnings. He is free do so, of course, but he should be aware of what he is doing. The third investor took an even more aggressive approach and used non-GAAP expected earnings for the coming year. Here, we are very much skating on thin ice. Working with earnings data that is of low quality to begin with and then made even more imprecise by the element of future forecasting is quite foolhardy on the part of the investor. Nevertheless, such numbers are actually being used. When people take such enhanced non-GAAP earnings still further overstated by typically overly optimistic future projections, then they come to a high EPS number which, being the denominator in the P/E calculation, makes the overall market appear to be much cheaper.
Beware of comparing the incomparable
If we consider the dearness of the market as a whole – for example, relative to historical averages – we must always be mindful of which earnings we are using in our calculations to avoid comparing apples and oranges. The GAAP EPS figure for all companies in the S&P 500 as of this past June was $104. The index’s P/E is therefore 24.1. The average for this indicator over the past 70 years was 17.5. By this comparative measure, the US market is expensive. Non-GAAP earnings (operating profit) achieved over the past 12 months were $115.90. This would mean a P/E of 21.6. Now, this P/E must only be compared with the historical average based also upon non-GAAP earnings. In recent years, non-GAAP earnings have been approximately 15% higher than GAAP earnings, which would put that average historical P/E around 15. By this measure, too, the market currently looks expensive.
If analysts or commentators want to persuade others that the market is cheap, they often take forecast non-GAAP earnings for the coming year and compare them to the average calculated on the basis of actual GAAP earnings. Taking today’s situation as a concrete example, it would look like the following: The 2018 forecast non-GAAP EPS for companies in the S&P 500 is presently $144.7. This gives a P/E of 17.3, which is slightly lower than the historical average P/E for the market (17.5). Thus (one might argue), the market is not expensive at all. In fact, however, if the market were to be trading somewhere around its long-term average P/E as calculated according to estimated non-GAAP earnings for the next year, that P/E would have to be approximately 12.5.
So, if someone says that the market is on the whole expensive, that person should clearly state with which earnings he or she is working and with which long-term average the comparison is being made. It is not difficult to orient oneself with these numbers; a person just needs to be attentive to the details. It is much more difficult, however, to make sense of the numbers reported by individual companies. Our investing is founded upon selecting individual stocks and in analysing them. To achieve that in practice, over time we have established the following internal rules:
- We never automatically accept numbers as they are presented to us but instead approach them with a healthy dose of scepticism.
- We are aware that company managements have various motivations and that we must consider any given company’s overall culture. If we fail to understand the presented financial statements of a certain company, it is often because the management does not want anyone to understand them.
- We unequivocally prefer GAAP numbers to various non-GAAP adjustments. Even though GAAP is an imperfect set of accounting standards, it is universally known and is the same for all companies. If we feel that GAAP statements require some adjustment in order to provide a more accurate picture of a company’s operations, then we make such adjustment ourselves. This is also very important when comparing individual companies to one another.
- We place greater importance on free cash flow than on earnings, because this indicator is harder to manipulate and is crucial for a company’s valuation.
All this is rather labour-intensive, but no one can expect to make big money in the equity markets without putting in the work.
The managements of a number of companies try to divert attention away from the fact that their companies are doing poorly in terms of profitability and to persuade investors (and bankers) that other lines higher up in the income statement are more important than is the bottom line. At some point in the past, the concept and abbreviation EBIT (earnings before interest and tax) first came into use. It was supposed to demonstrate better a company’s ability to generate the cash flow necessary to service its own debt. It has become apparent, though, that this figure does not fulfil its purpose very well because both interest payments and taxes are real and often very substantial costs.
Nevertheless, EBIT is under normal circumstances higher than is net profit and so it has a certain enticing appeal. This indicator is not only still in use but it has been augmented with two additional items: depreciation and amortisation. The argument was made that depreciation and amortisation are noncash items and therefore can be ignored. Even though anyone who has ever managed a company knows that depreciation is a very real business expense, such protests were swept aside. The new indicator thus became known as EBITDA (earnings before interest, tax, depreciation and amortisation). Buffett’s partner Charlie Munger, who never shies away from using colourful language when it is properly descriptive, refers to EBITDA as “bullshit earnings”. And he is right.
But the accounting creativity does not stop there. You may sometimes come across the abbreviation EBITDAR. The R at the end stands for rent, and EBITDAR is often used in relation to retail chains where it entails costs for leasing retailing premises. Some take this even further and present their results as REBITDAR, which stands for recurring EBITDAR. But the gold medal must go to an accounting line item termed Adjusted REBITDAR, which is adjusted recurring earnings before interest, tax, depreciation, amortization and rent. Such item tells us essentially nothing about the company and its value. In terms of cooking the accounting numbers, it comes in just a smidgeon short of what we might call EBE (earnings before expenses). I think it is just a matter of time before we will see that, too.
So how expensive are the markets?
I very often hear that the markets are expensive. The word “markets” nearly always is intended to mean the US market, as if other markets did not even exist. The US market truly is expensive. According to most basic indicators, it is above the 90th percentile of its expensiveness, meaning that in less than 10% of its history has it been pricier than it is right now. Its fair value could be some 20–25% lower. But this in and of itself says nothing about how the market will develop going forward. It can continue to rise and remain expensive for a number of years to come. It may move sidewise and gradually become less expensive as the earnings of US companies grow and catch up to the overstated price multiples. And, of course, its expensiveness might be corrected more quickly by a dramatic drop. We cannot say which of these is the most likely, but we certainly can say that neither is the market a bubble waiting to burst. In any case, though, the situation urges caution, and therefore our net exposure to the US market is only 28% of the entire portfolio.
There are nevertheless many more markets on our planet that do offer good investment opportunities. Europe, the second-largest market, is substantially less expensive than is the US market, and we also regard the Japanese market to be inexpensive. An investor is currently much safer to be active in these markets. It is interesting that in Europe, and particularly so in Japan, the manipulation of financial statements is not as widespread as in the US. I would say this relates to those markets being less expensive and so managements of companies in those environments are under much less pressure to demonstrate good short-term results.
The dearness of entire markets affects us as active managers only indirectly. One can always find attractive opportunities; it just demands more effort in expensive markets. Our current portfolio is valued at approximately 11 times actual, historical earnings (as we ourselves have adjusted them). This means that the multiples of earnings we pay for our equities are about half those of the US market. But that’s not all. The companies we hold are above average in terms of the quality of their businesses. That quality is reflected in the combination that is the types of business, market positions, financial soundness (meaning low debt), and managerial and business cultures of those companies. The portfolio is not just inexpensive as a whole, but so, too, is each individual share. All the companies in our portfolio are highly and regularly profitable over the long term.
Such companies are not easy to find. They are not offered on every street corner, but they do exist and the role of active investors is to select the best among them. Investors most frequently talk about the returns from individual stocks (or even entire markets), but they forget about the other side of the same coin, which is risk. Our portfolio is not only lower-cost than the market but also much less risky. The low level of risk is due to both the valuations of the individual stocks (their price-to-value ratio) and the quality of the underlying businesses.
Changes in the portfolio
We sold Irish Continental Group. We had bought this shipping company last year in July, just after Brexit. At the end of June last year, the UK equities market and a number of stocks of other companies doing business in the UK had gone through a period of panic selling by investors who had considered Brexit to be just about the end of the world. I remember well how at the time we were looking with disbelief at the prices of some stocks and contemplating which to buy first. One of our purchases was ICG. About a year later, we decided to sell the shares with a gain of 35% because it seemed to us that our investment thesis had been fulfilled and the shares were no longer priced attractively.
We put the funds that were freed up into another company, coincidentally also involved in shipping – albeit on the other side of the world and with different types of cargo. We have had this company in our portfolio previously, back in 2004–2007, so it is no mystery to us. This business is outside the main interest of investors and the main indices. Furthermore, this company constitutes a special situation that we think the market is evaluating entirely incorrectly. Every time I have explained this investment to one of our shareholders, the reaction has always been the same: buy as much as you can get! I can now report: mission accomplished.
Daniel Gladis, October 2017
See the full PDF below.