Letter number 149 of March 2023
- QUESTIONS & COMMENTS
Of course it’s easier to predict the future when it’s behind us than when it still lies ahead and is only partly unknown.
It seems to us, however, that investors who had done their homework in the autumn of 2021 had every reason to sell any Orpéa shares and debts they may have had as quickly as possible, or even to sell them short.
We believe and will demonstrate in this two-part article that a properly conducted financial analysis inevitably showed that Orpéa was heading faster and faster towards a wall that was getting closer just as fast. Even without the publication on 26 January 2022 of the book Les fossoyeurs (The gravediggers) by the investigative journalist Victor Castanet, which brought to light maltreatment, embezzlement and shocking management methods within Orpéa, Orpéa was faced with profitability that was far too low to cope with the burden of its debt and a very overvalued share price. Sooner or later, this situation would have led to a major financial crisis at Orpéa.
On the equity side
Let's first take a look at what financial analysts focusing on the value of equity (sell-side analysts) had to say.
As the following graph illustrates, of the 13 analysts following Orpéa late 2021, only two did not issue Buy recommendations (Exane BNP Paribas and Oddo BHF). They were simply Hold. All the others recommended Buy, and some of them were quite pressing: "A rare entry point" for Berenberg with a price target of €110 on 13 December 2021 (price of €83). Kepler’s price target was €155 on 7 January 2022, when the share price was €86. Kepler titled its report "A wake up call", and not for the reasons it would do so today when the share price is €1.8, which we believe is still greatly overvalued.
How can such enthusiasm be explained? There is, of course, a herd effect that makes it very costly for an analyst to go against the grain of the market consensus, especially if he or she is wrong. This will be remembered for years, whereas analysts who are right are in such cases are forgotten in under than 3 days. Moreover, posting a "Sell" recommendation is a guarantee that the analyst’s to Orpéa's management will be cut off. Proximity to management is very useful for producing reports and reassuring investors as to the relevance of an analyst’s report.
At a time when the main investors and asset managers have developed their own financial analysis capabilities (the buy-side), when research is essentially intended for hedge funds which, for the most part, only see value in financial analysts’ reports as they allow them to position themselves just before the publication of a result, one would have thought that the analysis reports on Orpéa would be rather hollow and oriented towards the very short term. Well, no. Some of the reports we reviewed showed real analytical work, far from the superficiality of knowing whether or not the company was going to beat the EPS or EBITDA consensus when its results were published two days later.
Nevertheless, all financial analysts fell into the trap of IFRS 16, the standard that covers accounting for a company’s financial and operating leases, the pitfalls of which we have outlined on many occasions .
We know that this standard leads to treating operating leases and financial leases in the same way, which has always seemed to us to be a major contradiction .
As a result of IFRS 16, since the 2019 accounts, rent paid is split in the income statement between financial expenses and depreciation. It is therefore not included in the calculation of EBITDA since EBITDA is computed before depreciation and before financial expenses.
In addition, the company records the present value of future operating and finance lease payments on the asset side of its balance sheet as a right of use, and a similar amount on the liability side of its balance sheet. However, the crucial point is that these amounts are calculated over the residual term of the operating or finance leases.
When we calculate the net financial and bank debt/EBITDA ratio, the denominator is EBITDA which we think is more or less recurrent, depending on the economic situation and the profitability of current investments. This gives this ratio its relevance by making it possible to estimate after how many years the debt will be extinguished if all the EBITDA is devoted to this objective.
With IFRS 16, Orpéa adds to its traditional net bank and financial debts its lease debts over a period of 9.3 years as at 30 June 2021, for an amount of €3,020 million. Why 9.3 years? Because as at 30 June 2021, Orpéa's last financial publication before the scandal broke, the average maturity of operating and financial leases was 9.3 years.
The most likely outcome is of course that after 9.3 years, these leases will be renewed, because otherwise Orpéa's business and EBITDA would fall significantly. However, IFRS 16 does not take this into account. In fact, its interpretation is legal (what is the term of the lease? 9 years in France, for example, with 3-6-9-year leases), and not economic: for the business to continue, these leases will have to be renewed or Orpéa will have to sign new rentals for similar volumes of property in order to continue to house its customers and employ its staff. At constant activity, the volume of rentals is constant and not limited to the term of existing leases.
In other words, in the calculation of the net debt/EBITDA ratio, only the lease debt over 9.3 years is taken into account in the numerator, post IFRS 16, whereas it is implicitly assumed that the EBITDA in the denominator continues to infinity as a going concern. This is a serious error in reasoning, which we warned our readers against . The only correct reasoning at this level, short of completely unravelling the effects of IFRS 16 on the accounts, would be to take the lease debt not over 9.3 years, but to infinity, to be consistent with the EBITDA reasoning. A quick estimate shows that we would then have to add about €12,600m (sic) and not the €3,020m actually added as at 30 June 2021 under IFRS.
As all of the equity analysts whose work we were able to consult reasoned post IFRS 16 in their calculations of the level of debt, they did not think it was unsustainable for a company with a very high property component (47% of assets are freehold). The ratio they calculate, using the accounting data as it is, is in fact 10.0 (based on EBITDA for the last 12 months of €986m). With this approach of considering leases as debts, the economic and financial reality is that this ratio is 19.7 , i.e. double that calculated by equity analysts, which is a massive difference!
Furthermore, continuing to reason post IFRS 16, as the same causes produce the same effects, in when calculating the transition from enterprise value to equity value, most equity analysts took a figure for IFRS 16 debt that is largely undervalued (€3,020m, as opposed to €12,600m), leading to an overvaluation of their estimate of the share value. The perpetually discounted EBITDA in their DCF calculations was based on post-IFRS 16 EBITDA, i.e. without rents, while the debt drawn down only took into account the debt of the rents over 9.3 years, and not to infinity.
The same was true for the multiples method, with an enterprise value on the basis of an EBITDA before rent, multiplied by an EBITDA multiple, less net debt including IFRS 16 debt limited to 9.3 years.
It should be noted that the loss in value borne by investors in the announced rescue plan is around €8.8bn, not including the fall in the value of secured debt that will earn less interest, which is not very different from the additional IFRS 16 debt not seen by equity analysts (€9.5bn).
* * *
If equity analysts had unravelled IFRS 16, reverting to the situation where rent on an operating lease is considered an operating expense, not an EBITDA inflator, would they have been more effective?
We have reason to believe this, as this work would have led them to restate the accounts instead of taking them at face value, and in so doing to realise that Orpéa's EBITDA was maintained by a significant change in the group's property policy which led to it being boosted in the short term, but reduced in the future. Unfortunately, none of the equity analysts seem to have realised this fact.
Orpéa had announced a goal of increasing the proportion of its establishments that it owned to 50%, with the balance being rented. Accordingly, this proportion had risen from 47% in 2018 to 49% in 2019. But by 30 June 2021, the ratio had fallen back to 47%. Why?
Because Orpéa wanted to generate capital gains in order to maintain its EBITDA margin, which was undermined by the increase in post-covid personnel expense.
We can see very clearly in the income statement, cleaned up from the effects of IFRS 16, that if the added value (after the payment of rents) is maintained at around 73% of sales, and that the EBITDA is stable at around 17-18% of sales, despite staff costs jumping from 53 to 57% between 2018 and the first half of 2021, it is because the "Other" item jumps from 0.4% to 3.9%. This item includes capital gains on the sale of property, as Orpéa becomes the tenant of buildings that the group used to own.
We have two comments in this regard, the effects of which are cumulative. Firstly, selling property to generate capital gains, while becoming a tenant, means increasing the proportion of rent, which leads to a reduction in future EBITDA, in absolute terms and as a percentage of sales. Secondly, while we have no problem considering, for example, that the capital gains made by car rental companies when they sell them after 6 months are part of their operations, the case is quite different for the buildings housing Orpéa's nursing home and clinics business. The sale of buildings cannot be recurrent because, if they are sold regularly to maintain the level of EBITDA, there will come a time when there will be no more buildings to sell.
All this leads us to believe that the quality of Orpéa's EBITDA was clearly deteriorating, which should have led any analyst who was thinking of abandoning this criterion, which is only a practical approximation of the operating cash flow, to concentrate on the original, the operating cash flow.
Over the last 12 months , it was €294 million. Faced with a net bank and financial debt of €7,435 million. That is 25 times more. How can anyone imagine that operating cash flow of €294 million, which is before any maintenance or modernisation investment , could reduce net debt? The level of debt was clearly unsustainable, and Orpéa had rushed headlong into major investments in organic and external growth (€2,950m between 2018 and 2020), with only €650m in operating cash flow over the same period, requiring increasing recourse to debt. This was at a time when its ROCE had declined from 4% to 2%, far from a cost of capital of 6-7%, and at the level of its cost of debt. A reverse leverage effect was very close.
* * *
In the second part of this article, to be published in the April Vernimmen Newsletter, we’ll analyse other minor mistakes made by equity analysts - so that you can avoid making them too - and look at how debt analysts may have gone wrong, even though they had avoided the IFRS 16 trap by excluding the effects from the company’s accounts as we recommend.
 Bank of America, Berenberg, Bryan Garnier, CIC, Exane BNP Paribas, Gilbert Dupont, HSBC, Jefferies, Kepler, Oddo BHF, Portzamparc, Société Générale, and Stifel. We did not have access to the reports published by HSBC and Stifel.
 A company enters into an operating lease when it simply wants to have an asset at its disposal, without wanting to acquire it. It enters into a financial lease when it wishes, at the end of the lease, to acquire the asset in question for a modest price, taking into account the rental payments already made, which amounts to spreading the payment of the acquisition over time, as in a leasing agreement. It is therefore a method of financing the asset. To account in the same way, by adding the value of the leased asset to the asset side and the present value of the lease commitments to the liability side, seems to us to fly in the face of common sense, since the company's intention is radically different in the two cases: to have the flexibility of leasing without wanting to acquire the asset, versus acquiring the asset through leasing.
 (6.841 + 3.020) / 986.
 (6.841 + 12.600) / 986.
 99.6% of the market capitalisation, i.e. approximately €5.8bn based on a price of €90 and 70% of the unsecured debt of €3.8bn.
 Second half of 2020 and first half of 2021.
 Depreciation and amortisation are in the order of €300 million, which gives an idea of the amount of maintenance investments.
If the average late payment in relation to negotiated payment terms is down in Europe to 13 days, this average hides divergences as illustrated by this table produced with data from the Altares surveys:
This dispersion of behaviour increases when the horizon extends beyond Europe, with payments on time in only 45% of cases, and delays of more than one month compared to the agreed deadlines in 13% of cases:
With Simon Gueguen, lecturer-researcher at CY Cergy Paris University
Announcements of so-called "horizontal" acquisitions (in the same sector and at the same level of the value chain) are events that provide information on the valuation of listed companies in the sector. The acquirer's managers have access to information on the fundamental value of their company and therefore on potential market over- or under-valuations. This information content is well known theoretically, but difficult to measure empirically. The problem is that the revaluation of the target (and even the acquirer) at the time of the announcement is based on multiple sources. In addition to the market value correction, movements are linked to anticipated synergies and the specific price proposed for the transaction. This month's article avoids this problem by looking at sectoral revaluations during acquisitions. If management is well informed, then the choice to favour the acquisition of listed or unlisted companies should lead to a revaluation, upwards or downwards, of listed companies in the same sector.
The intuition of the article is as follows: let us consider the case of an acquirer having the choice between two targets that are equivalent in terms of their fundamental value and anticipated synergies, and that differ only in the fact that one is listed and the other is not. In this case, the price of the former is a market price that is sensitive to fluctuations from various sources. If the listed target is undervalued, the buyer chooses it. If the listed target is overvalued, the buyer prefers to buy the unlisted target. Thus, all other things being equal, the acquisition of a listed target sends a signal of undervaluation to the market, which should lead to an upward revaluation of companies in the same sector. Conversely, the acquisition of an unlisted target should lead to a fall in the price of companies in the sector. By looking at the stock market reactions on companies not directly involved in the transaction, we measure a revaluation linked to the information of managers on the fundamental value in the sector independently of the transaction itself.
Derrien et al base their empirical study on a sample of nearly 8,000 horizontal acquisitions in the US market between 1990 and 2015. The sample includes 88% unlisted acquisitions, which represent 50% of the sample by value.
The results confirm the theoretical intuition. After an acquisition of a listed company, the revaluation (measured by the abnormal profitability observed at the time of the announcement) is positive, between 0.02% and 0.34% depending on the specifications of the econometric test. When the target is unlisted, the revaluation is negative, at around -0.2% irrespective of the specifications.
These amounts may seem small compared to the stock market effect of other market announcements, and event studies often uncover more dramatic movements. However, the large sample size makes these results statistically significant. Above all, the idea of the paper is not so much to explain market movements as to prove the existence of information content on the fundamental value of targets in a sector during acquisitions. In this sense, the exercise is successful thanks to an original idea.
In the rest of the paper, the authors conduct various tests that confirm that the results obtained are attributable to the effect tested. For example, they show that revaluations are higher when analysts have diverging predictions about the future performance of the stock. More importantly, positive revaluations are followed by increases in sectoral returns (and vice versa for declines). This is consistent with the idea of a gradual correction of valuation differences, and especially with the fact that these revaluations are financially justified.
Finally, the paper empirically confirms that an active acquisition market contributes to the efficiency (in the informational sense) of financial markets, as does a liquid secondary market.
 F. Derrien, L. Frésard, V. Slabik and P. Valta (2022), "Industry asset revaluations around public and private acquisitions", Journal of Financial Economics, 147 (1), 243-269, January 2023.
In Chapter 40 of the Vernimmen , devoted to start-ups, we emphasise that, with some exceptions (mainly real estate and receivables financing), start-ups are meant to be financed by equity and not by debt. In good times (as 2021 was for venture capital), financiers are becoming more inventive in pushing the limits of this assertion. Thus, some lenders have ventured into the field of start-ups that are not yet profitable, but promising, and whose business model provides some form of visibility. This led to the development of loans to companies with recurring revenues (secured by subscriptions or long-term contracts) and close to financial balance.
Initially (in the 1990s-2010s), in the United States, these were loans of a few hundred thousand or a few million dollars to small start-ups, amortisable and offering returns of 30 to 40%. Now, both in the US and in Europe, Annual Recurring Revenue based financings are larger, bullet loans with a term of 5-7 years, quite similar to the unitranche loans set up in the context of LBOs. Their cost is also more reasonable, but still between 5 and 10% above the risk-free rate!
They are based on a covenant relating debt to recurring annual income (secured contractually), a ratio that is intended in the medium term (2 to 3 years) to be transformed into a classic debt/EBITDA ratio when the company becomes profitable, the debt then taking the form of a classic unitranche debt.
During the first phase, the ratio covenant is accompanied by a minimum liquidity clause to ensure that the company has enough cash to continue its operations. Also, during this first phase, the company benefits from a deferred interest payment (PIK). The financing is secured by a commitment from the shareholders (usually a venture capital fund) to reinject capital in the event of a breach of covenants - the equity cure. Without this clause, very few lenders would venture into such financing. However, these transactions remain very risky and are reserved for direct lenders, not banks.
In the early 2020s, the amounts lent were between 1 and 3 times annual recurring revenues (1-2x in Europe and 2-3x in the US). The return to reason for technology stocks in mid-2022 has put a stop to these deals... for now (as have most of the highly leveraged deals).
Regularly on the Vernimmen.com Facebook page 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.
Should Berkshire-Hathaway pay dividends?
Berkshire Hathaway is the investment company of Warren Buffett (92 years old) and Charlie Munger (99 years old), whose share price has risen since 1964 (when Warren Buffet took control) by 3,787,484% (sic) compared with 'only' 24,708% for the S&P 500 index, dividends reinvested, i.e. an average rise of 19.8% per year for 58 years, i.e. exactly twice the return of the S&P 500 (9.9% per year) over the same period.
Berkshire Hathaway has never paid a dividend, and has therefore always reinvested its earnings, with the success mentioned above, again in 2022, when its share price rose by +4% compared to -18% for the S&P 500 index with reinvested dividends. At 31 December 2022, cash net of debt was $30bn, and $153bn gross of debt.
A shareholder tabled the following resolution for a shareholder vote: "Whereas the corporation has more money than it needs and since the owners unlike Warren are not multi billionaires, the board shall consider paying a meaningful annual dividend on the shares."
This resolution failed to pass with 98% voting against, and only 2% for, among the B shares held by hundreds of thousands of small holders, the score being 99% against for the A shares.
A lesson to ponder for all those, many among politicians, trade unionists and journalists, who think that shareholders are like dividend-hungry leeches and ignore or forget that a policy of return to shareholders is judged by the marginal rate of return that managers are able to find and achieve. As soon as profitable investment opportunities (earning more than the cost of capital) exist, and management has demonstrated its ability to carry them out, shareholders are prepared to do without dividends altogether. It is a very rational decision, without any passion behind it, as illustrated by the shareholders of Berkshire Hathaway who have no regrets about not having received any dividend since 1964.
Are large groups pushing up inflation?
We don't think so in Europe, but we wouldn't say the same thing in the US.
If the large groups, which in France are summarised as the CAC 40, were to feed inflation by increasing their selling prices more strongly than is justified by the rise in their production costs and the maintenance of their margins, we should see their results rise faster than their turnover, thus showing an increase in margins.
However, when we look at the 2022 results of the 38 CAC40 companies (which close on 31 December), two thirds of them have a growth in results that is lower than the growth in turnover, implying a fall in their margins: Thus, by way of illustration for products that contribute to the daily life of consumers who may have this suspicion: Carrefour: + 16% for turnover and + 8% for results, Danone + 14% and + 1%, Crédit Agricole: + 3% and -7%, Michelin: + 20% and + 9%.
Among the 13 groups whose results grew faster than sales, we find, for example, Hermès: +29% and +38%, Safran +25% and +55%, Dassault Systèmes: +17% and +20%, whose activities are not remotely or closely related to the household basket and are therefore not likely to inflate inflation significantly.
In the United States, where competitive pressure has diminished since the 2000s, and where, for example, a telephone package costs around €60 per month compared with €20 in France, with 3 national operators compared with 4 in France, the situation is likely to be different.
Greed, short-sightedness and incompetence
In a recent article entitled How start-ups should manage their finances, we read the testimony of two confident entrepreneurs: "When you have at least 1 or 2 million in cash, you can invest it in US Treasury Bonds. At the moment they are paying 5%! This will give you a free runaway. "We bought US Treasuries a while ago. There is little chance that the US Treasury will default. If not, the world will be in big trouble."
We don't dispute that the world would be in big trouble if the US Treasury defaulted, and that this is very unlikely.
We find, in our dual role as educators and investors, including in start-ups, that it is a very bad idea for a European start-up to invest its cash in dollars. The only case where this might make sense is if the start-up has debt in dollars or makes significant purchases denominated in dollars, so as to constitute a natural hedge against currency risk, which should not concern many European start-ups since they rarely have debt, and even less in dollars!
Why is this a bad idea? Because in order to get 5% on US Treasury bonds, you have to buy a two-year paper, which only yields 4.4%, whereas French Treasury bonds currently yield 2.7%. So for a rate differential of 1.7%, the company takes a currency risk on the dollar. If the dollar falls by 1.7% against the euro in one year, the gain in the return on the dollar Treasury bill against the same maturity in euro is wiped out. A fall of 1.7% means that the euro falls from its current rate of $1.06 to $1.08. When you consider that the dollar has fluctuated between 0.96 and 1.4 over the last 10 years, you can see that a variation of at least 1.7% in one year is a possibility ... if not a certainty.
But there is not only currency risk, there is also interest rate risk. If in a year's time US one-year interest rates have risen to 6.2%, because inflation is finally higher than expected and our start-up needs the funds, it will sell its treasury bills at a loss, at a constant exchange rate, erasing the difference in yield.
We wish good luck and a lot of courage to these apprentice treasurers to go and announce their brilliant idea to their governance committee, or even, when the time comes, the possible disaster of an investment that would have lost 10% of its value with a dollar at 1.17, especially in the context of a rising tug-of-war between the Democrats and the Republicans on raising the ceiling of the US government debt. ... In any case, we have advised start-ups that have put their trust in us by welcoming us to their capital and their governance committee, as we are not only teachers, to invest wisely in euros and to focus on operations.