Bronte Capital Callisto Fund commentary for the third quarter ended September 30, 2024.
It has been a challenging time since our last substantive letter two months ago. The market kept making new highs with only a brief dip in late August and early September. For the quarter the fund was down 0.14% vs. +6.61% for the globally diverse MSCI ACWI and for September was down 3.36% vs. +2.32% for the index. We are not built for markets like this. We are a long-short fund, and we would expect to underperform a market regularly making new highs. Contra: we expect to outperform and historically have outperformed when markets are weak.
But we run at positive beta, and we would at least expect to run positive returns. In September we underperformed that expectation. This really was our first significantly under-model month for a couple of years.
Our short book has been very difficult. This is not a surprise. When the market makes new highs day after day we expect to lose money on the short book. We have a long book though – and that usually provides enough “hedge” so that we go up a little even when the short book is awful.
This did not happen in September.
First the short book
We have 700 shorts in our short book. Typical size of a short position is about 10 basis points – some shorts – particularly in biotech are run even smaller than this.
Over time, this short book chosen mainly by following bad people, has performed very well. But individually it is littered with little disasters. Many shorts have doubled. Some have doubled twice or more (and still gone to zero).
September was worse on our short book than we would expect – but not abnormally worse. This is not January 2021 when the question with the short book is what did not hurt.
A short that doubles or even triples does not tend to worry us. We keep positions tiny. A short that goes up tenfold (even if it eventually goes to zero) is however noteworthy. There was one of those. It was Summit Therapeutics Inc (NASDAQ:SMMT). We have put this in Appendix 2 if you are interested in process. Summit cost us well under 1 percent of the fund – but for a short that is still a very large loss.
The long book
The big problem in September was our long book. Many of our important longs simply did not participate in the rally.
Berkshire Hathaway Inc (NYSE:BRK.A) was down mid-single digits, Regeneron Pharmaceuticals Inc (NASDAQ:REGN) – an important long for us – was down double digits. Google (NASDAQ:GOOGL) did not participate in the rally in a meaningful way – and is currently subject to Justice Department monopoly issues. Visa Inc (NYSE:V) another long-standing long was down – and again the issue was Justice Department.
As a result, we have had to trim the book a little at least in part to deal with the lack of upside protection we are carrying.
This note goes through individual longs and writes up a new long in Appendix A.
Two longs that hurt – Regeneron and Visa
Regeneron – the biggest loser in the book by absolute dollars
Regeneron is a large position for us. We wrote the position up in June 2021 letter. The stock has almost doubled since the original write-up, but it is down on the month. Allow some time for an explanation
There are two ways of looking at Regeneron. The sum-of-the-parts way and the platform way.
Regeneron has 11 approved drugs but two comprise most of the cash flows. The two are Eylea and Dupixent.
Eylea is a VEGF drug that is injected into patients’ eyes and stops macular degeneration (the main cause of blindness in old people). The drug stops capillaries growing in the retina.
Eylea has been an enormous success because stopping old people going blind has huge economic (and social) benefits and until recently offered best-in-class efficacy, safety and convenience compared to other VEGF inhibitors. Even singlepayer systems (like Australia) are prepared to pay high prices for Eylea because stopping old people going blind saves the system a lot of money.
This is an old drug that should be protected from biosimilars for the next few years – but the company hopes that a long-life version will mean market share is maintained. A long-life version is attractive because it means fewer injections into grandma’s eyes.
Dupixent is a drug that addresses a range of inflammatory disease including atopic dermatitis and chronic obstructive pulmonary disease. The drug is highly effective against a wide range of indications. This is developing into one of the biggest drugs in the world.
The problem in September – causing a double-digit fall in the stock price were court cases that could mean that the loss of exclusivity for Eylea may (or may not) come sooner. (The court cases are yet undecided.)
The second way to look at the company (the way we look at it) is that it is a platform for developing new drugs. They have two key assets.
- A rapid and powerful method for genetically modifying mice, which allows them to produce mice that (a) faithfully recapitulate human diseases and (b) mice that can produce best-in-class fully human antibodies.
- A database containing the genetic sequences of approximately three million peoples’, paired with deidentified electronic health records.
The combination is very powerful. The company can work out the genetic basis for many diseases and clone disease-susceptibility into the mice. This makes the mouse a great model for exploring the disease and its treatments. For instance, recently they developed a mouse that faithfully recapitulates amyotrophic lateral sclerosis (more commonly known as “ALS”). Maybe this will lead to treatment for that terrible and unfortunately not uncommon ailment.
We think the mouse and data technologies will allow the company to produce many effective therapies in the future. The main risk is that those therapies will not have the revenue of Eylea and Dupixent (which are amongst the biggest drugs in the world).
The loss this month, predicated on possible loss-of-exclusivity for Eylea does not much upset the reason we own the stock, and we do not think it is threatening. Maybe implicitly we are overstating the value of the mouse and data platform, but we do not think so. We are unworried about this one.
Visa
Visa is another stock we have held many years, and which has done well. The company, along with Mastercard Inc (NYSE:MA), has enormous powers in payments networks. Many businesses cannot survive without accepting Visa. If you want a history of the company detailing Visa’s exceptional power there is a good (though hagiographic) history on the Acquired podcast. There is another history of the beginning of Visa in Joe Nocera’s excellent (though now dated) book “A Piece of the Action”.
The stock has done well for us despite underperforming in recent years. Over the past few years the business has benefited from touch-to-pay and increased penetration during covid.
The problem is that globally dominant positions attract the antitrust scrutiny of regulators. The Justice Department sued Visa for illegally monopolizing the debit card market. The stock fell hard.
This is a risk that runs through our portfolio. We aspire to own businesses that become globally dominant and thus extremely profitable. Our portfolio companies have attracted regulatory attention in the past due to their exceptional profitability. It will likely happen again. We may be taking too little concern for that risk.
Generally, though we think it is okay. There is a countervailing force. America perceives itself to be in a cold war with China, and commercial prowess especially in technology is the means by which that war is fought. America does not want to hobble its champions - and this probably means that American champions will be less hobbled by antitrust than you might otherwise expect.
Alas the China protection does not apply to either Google or Visa. These are the two companies we have had that have been hit by antitrust issues and neither operates much in China.
Appendix 1 is a write-up of a new long this year – Capital One Financial Corp. (NYSE:COF). Capital One is – as the Appendix explains – a new competitive threat to Visa.
Reporting when we were wrong
In our December 2023 letter we outlined our reservations regarding Karuna Therapeutics.
We have been short Karuna Therapeutics for most of 2023. Karuna are developing a drug (which they have coined “KarXT”) for the treatment of schizophrenia. KarXT is a combination of a molecule abandoned by Eli Lilly And Co (NYSE:LLY) because of severe side-effects (xanomeline) and a generic used to treat urinary incontinence (trospium).
The bull case – simply stated – is that drugs which are effective in mental illnesses such as schizophrenia are hard to find. Moreover, they can’t be tested/explored in animals as there is no realistic mouse-model of schizophrenia.
The Lilly drug (xanomeline) was thought to work but with intolerable side-effects. What Karuna sought was a drug that did not cross the bloodbrain barrier and would neutralize the effect of xanomeline in the body but not in the brain. They chose trospium.
Karuna claim the drug has “compelling efficacy and a differentiated safety profile” compared to existing antipsychotics. When the company announced “positive” phase 3 results in August 2022, the company’s market cap rose to USD 9 billion. There are millions of people with schizophrenia in the US alone, so the opportunity for a better drug is massive.
We were not convinced by the phase 3 result. We thought that the trial design might have sharply understated the risk-profile of the combination (and/or overstated the performance of the drug), and that the FDA would identify these design issues in its review.
Alas, important people with a lot of money think we are wrong. Bristol-Myers Squibb Co (NYSE:BMY), a major global pharmaceutical company, has just bid 14 billion dollars in cash (substantially all borrowed) to buy Karuna prior to KarXT being approved. They believe the transaction brings “great science”, is a “clear strategic fit” and that KarXT will deliver “multibillion” dollars in sales at peak.
We think this is unlikely and Bristol Myers management will wind up regretting it badly.
But hey – we have lost money on the short – so for the moment we look like idiots.
The FDA is currently reviewing the company’s New Drug Application, with a decision expected in the second half of 2024. We will find out whether the “smart money” at Bristol Myers has torched 14 billion dollars or (more likely) we have fooled ourselves into believing we understand what an inadequate drug trial looks like.
When it happens, we will follow up and report the lessons from this experience.
Well, alas, it is time to follow up. The FDA approved KarXT on a label we thought was far broader than justified by their phase 3 studies. And it seems clear we were comprehensively wrong.
To make matters worse we had an esoteric financial bet that the drug would not be approved. This would have paid out enormously well if we were right. But we were not right, and this cost us a further fifth of a percent in what was already a poor month.
There has been some soul-searching to work out how we were so comprehensively wrong. Some lessons have been learned. We hope they are the right lessons.
Bronte reviewing hedging strategies
We have said this before – but it is worth repeating here. September – and the nasty period we had in Early 2021 – were characterised by sharp mismatches between our long book and our short book.
When implied vol is low (and options are cheap) there is an easy solution – which is to buy out-of-the-money calls on indices. We did this in December 2023 and it worked very well.
When options prices are high, we will need different techniques or – simply just shrink the book. For the moment we are just shrinking the book – but over the next few years we hope to do better. We will implement anything slowly and with extensive testing – and we will keep you informed as we go.
The Bronte Model
You do not buy a shorting-heavy fund like ours for months when the market is going up. You buy us for consistent alpha regardless of market moves.
We expect to underperform in a weak market (we run positive beta). Our results have historically been better than that. When markets are weak Bronte has mostly produced small positive returns (see chart below).
We expect to underperform rising markets, but we still expect to be positive. Most of the time we have been positive in rising markets. Alas September and a few other months produced worse than expected results.
And we like our book. We are finding some longs to buy (not many but some). The markets however are high and finding longs is difficult.
Finding shorts remains easy. It is easy to find companies with a market cap over a billion dollars that are worth zero. Shorting them however is difficult. This is a bull market and stocks that are worth zero and have a billion dollars market cap may soon have a two billion market cap.
We have explained all this before – but the possibility of shorts going from ten to zero via 50 or 100 is the reason we have 700 shorts – and is the reason we built Bronte the way it is.
Thanks again….
John, Simon and the Bronte Team
Appendix 1: Capital One and the Discovery merger
US Credit cards have become a concentrated business. The leading players in order are Chase, a part of JPMorgan (NYSE:JPM), American Express Company (NYSE:AXP), Citigroup Inc (NYSE:C), Capital One, Bank of America Corp (NYSE:BAC) and Discover Financial Services (NYSE:DFS). A combination of Capital One and Discover will become the number two player.
A book of old and stable credit cards is a good business. You almost certainly have a card in your wallet (or embedded in your phone) and you transact on it a lot. The bank makes good coin from that card – and given your financial circumstances and long history of paying – your bank can be reasonably assured that you will never default on that card.
Many Americans however have two cards. The number two card in the wallet is a problematic business. If you have two cards in your wallet there will typically be one on which you do most transactions on and which earns most the revenue.
The main reason for calling on the second card is that you are financially stressed, and the first card has gone beyond its limit. Being the “back of wallet card” is not a good business as you get a full share of the credit losses and little of the revenue.
Originating new cards is also difficult. The average American is inundated with offers for credit cards – and prime customers mostly file those invitations in a cylindrical filing cabinet. However, a financially stressed customer may accept a new card – and they will of course be far more likely to default. Originating cards is costly.
The difficulty (and loss expense) of stealing customers from other issuers (that you preferentially get customers more likely to default) makes market shares in this business hard to shift. It gives banks some pricing power.
Of direct consequence banks have pricing power – and credit cards have remained profitable for existing but not for new players.
For this note we need to explain Capital One’s business and then why Discover’s business is different – and what makes the merger with Discover transformative.
Capital One – the math/computer science led giant
Capital One is not like the other companies on this list. Capital One was a scrappy startup which originated cards by scoring people with computers and accurately offering pricing to good customers largely to get them to switch. The company is still in the hands of its founder (Richard Fairbanks) who has made dozens of major and correct calls over the history of the firm. The most important of these was buying regional banks and ING Direct (AMS:INGA) to get an online and small customer banking presence to make themselves largely free from wholesale funding markets.
The early part of John’s career involved much shorting of subprime credit card companies who were intent on stealing the low-end of Capital One’s credit card base. Most of these failed in some way and Capital One emerged the unequivocal winner.
If it were not for the Discover merger we would not be buying Capital One because – frankly – even the best run subprime credit card issuer in the world is a subprime credit card issuer. This is a difficult business for which bad outcomes are possible.
Capital One’s business is two-peaked with respect to the cards they issue. They are the super-champion at “good subprime”. These are people who have difficult finances and may have defaulted in the past but have realised that you need a credit card to survive in American society (to rent a car, to buy things online for example) and are determined, within their means, to try and fix their credit. These people still have a high default rate, are economically sensitive and will have rising losses when unemployment goes up. Issuing cards to them can be a difficult business.
They are also champions – along with American Express – in super-prime cards used by people who do all their transactions on the card, have a high spend including on travel but pay their balance every month. They are usually interested in the airline points. Think of an upper-income family that travels a lot but has and intends to maintain pristine credit.
These card holders are hard to find (you have to offer them very good terms and lots of points to get them to change) but they are very profitable once found. Many of our readers fall into this bucket.
What Capital One has mostly avoided is issuing to the vast collection of what America calls middle class and the rest of the world calls working class families. Think a family with household income of 60-80 thousand and a couple of kids, but with a credit record that is mostly clean but with some inevitable live-frompaycheck to paycheck financial stress. Their view – these customers want pricing that looks close to super-prime but with a credit profile that can (and unfortunately often does) go pear-shaped when unemployment or other stress such as a medical emergency hit. Their view is that business is mostly mispriced (at least for them).
The great era of profitable growth for Capital One ended at the financial crisis – and the company has – with a banking licence – has had to build up a large capital buffer. The stock is about 50 percent higher than its pre-financial crisis peak – but traded sideways with volatility for about a decade. The derating back to a low-teens price to earnings ratio has been sharp.
We think the capital buffer problems have largely been solved the right way (by earning the money needed to meet larger capital requirements).
American Express and Discover
Most credit cards transact on the Visa/Mastercard networks. As noted in the main letter Visa is an enormously powerful company. There are however two independently owned networks. One is owned by American Express and the other by Discover.
American Express have competed with Visa by charging more than Visa. The idea was that by charging more than Visa you could offer more services (such as very good service if your card is lost overseas) or more airline points or both to customers. The cost for these were carried by the merchants who accepted the card. This made merchants reluctant to accept the card – with the reluctance offset by high value customers American Express brings.
For years we thought this was a losing strategy. We could not imagine many customers who thought it prestigious to pull a “Centurion Card” from their wallet – as if carrying a black American Express card was a symbol of their prestige and status. We noted that there was at one point a Harvard Alumni Association co-branded VISA which seemed to have more status, or at least loudly announced you went to Harvard.
We were wrong. American Express has grown for decades with a charge-morefor-more-services and prestige pricing strategy.
Discover competed with American Express the more logical way – by charging less. This makes merchants very willing to accept the Discover Card and Discover has near universal acceptance in the United States.
The problem is the lower take means that Discover cannot compete for high value customers on airline points. Instead, they took a different tack.
Discover is the king of the American lower-middle class. The typical customer is precisely the customer that Capital One avoids – the family with 60-80 thousand household income that mostly lives from paycheck to paycheck, but which has and tries to maintain good credit.
The way Discover attracts these customers is with a cash-back card. They give back to the card holder 50 cents per hundred dollars spent. The families that hold this card really value that 50 cents, and they know that if they default on Discover this perk will go forever. They do not value fancy airline points or travel benefits because mostly they are not rich enough to travel except by car.
Capital One has experimented with marketing to Discover’s customers and has found that Discover customers will default preferentially on the Capital One card. They like that 50 cents.
Capital One is buying Discover
The news that made us interested in Capital One is that Capital One is buying Discover.
This is a transformative acquisition. What it gives Capital One is a network with near universal North American coverage. These are truly rare assets – and it will allow Capital One to compete with Visa rather than be just another captive of Visa’s near-monopoly network.
Capital One are appropriately blunt about Discover’s card business. They do not love it. They would not be buying Discover for their book of business. But it is profitable, and they will leave it alone. They will not be changing the branding or anything like that. Discover customers will – if all is right – not realise that anything has changed. (Alas Discover Cards will need to say that they are issued by Capital One Bank in small print.)
There are however enormous synergies from moving Capital One cards to the Discover network. Capital One will look far more like American Express, except that it will be larger and with wider acceptance.
The guidance given by the company for post-merger earnings do not include all of these synergies. The only cards that will – at least in the guidance – be moved to the Discover network are debit cards. This should be easy to do. As noted in the letter Visa is currently being sued by the Justice Department for using their network power to illegally monopolize debit cards. Under these circumstances it is highly unlikely that Visa can or would offer any obstruction to moving the business.
That limited amount of synergy gets Capital One earnings in 2026 to about $18 a share. (Current share price about $150.)
The big driver however is getting the vast book of high-end transactors off Visa’s network and into Discover’s cheaper network. The savings from that will make earnings grow sharply for many years without requiring more regulatory capital or without accepting more credit risk.
The company however cautions against shareholders putting this in any shortterm company model. The problem is that Discover’s network is under-invested. When I touch my phone on a Visa network the payment typically clears in under a second. The network rarely (but still occasionally) rejects the payment not for credit reasons but simply because network connectivity is not perfect.
Discover’s network by contrast is far clunkier. Payments clear slower and network problems blocking payment are more frequent. Capital One will have to spend a few hundred million bringing that network up to scratch, decentralising data to solve connectivity problems and just making it faster.
Capital One will not transfer high value transacting customers to the Discover network until the network quality is the match of Visa. This will happen – it will just take time. But when it happens earnings should rise fast – and Capital One will be able to compete for these high value customers with a competitive advantage over their most important competitors – the competitive advantage will be that they are less beholden to Visa’s near-monopoly.
Our holding
We hold Discover Financial Stock rather than Capital One because it trades at a discount to the Capital One stock it will convert into. However, what we are really doing is holding Capital One stock for what we think will be almost a decade of earnings growth. It is also the best kind of earnings growth – it comes without requiring substantial additional regulatory capital and without taking substantial new credit risk. It comes from disintermediating Visa.
The problem with this position is that it is levered to a large riskier subprime business. This means that the position has been started around 3 percent rather than the 6-8 percent we are comfortable with when holding robust and unlevered business models.
Appendix 2: Summit Therapeutics
The short book hurt us in many ways. Several deeply junky stocks went up – and one – Summit Therapeutics – in which we may have made a mistake – stood out above all. Even excluding Summit, our short book would have faced challenges in September, but this position exacerbated the difficulties.
A little background is necessary. Summit is (or was) a small drug company that we scored as having some suspect associations and hence a short candidate. Our database of bad people in financial markets is very broad.
Summit was not such a stock. Our short signals were triggered light amber, not bright red. This meant the company required a bit more analysis.
The company was associated with Bob Duggan who spearheaded the highly successful pharmaceutical company - Pharmacyclics. For readers who are interested in the history of Pharmacyclics and Duggan, it is written up in Nathan Vardi’s “For Blood and Money”.
We regarded Duggan’s past (major) success with Pharmacyclics as a positive sign from the company’s perspective. This is one of the reasons the stock was scored as light-amber and not sufficient to short without the presence of other red flags. However, in this case, we had good reasons to be sceptical of the drug beyond the people.
The concept Summit are trying is a bispecific molecule targeting PD1 and VEGF to treat cancer, and the company is claiming that they are on the cusp of dethroning Keytruda which is currently the world’s top-selling drug.
To provide context, it's important to understand the science behind these therapies.
Cancer cells often evade the human immune system through various mechanisms, including the PD-1/PD-L1 pathway.
This pathway acts as a "brake" on the immune system, suppressing its ability to destroy tumor cells. Drugs that block this interaction have revolutionized cancer treatment, generating billions in sales. Keytruda, the dominant player in this category, highlights the immense potential of these therapies and generated USD 25 billion of sales in 2023.Unfortunately, not every patient responds to PD1/PD-L1 inhibition, and even those that do are generally not cured. Hence the need for better drugs. Many companies are developing drugs to directly promote the immune system killing of tumor cells and many are developing drugs that take the “brakes” off the immune system in a similar fashion to the PD-1/PD-L1 drugs. Many of these companies are developing their drugs in combination with a PD-1/PD-L1 inhibitor and are hoping that the effects will be additive or synergistic. Some companies are even combining these mechanisms into a single molecule.
VEGF inhibitors are another major class of cancer therapy. They work essentially by controlling the rate at which your body produces capillaries. The idea is that you zap a cancer using radiotherapy or similar and that wounds the cancer. A VEGF drug will stop capillaries forming thus limiting the ability of the cancer to feed itself and grow back. These were amongst the biggest drugs in the world 15 years ago. The best-known drug is Avastin.
Summit’s experimental drug (Ivonescimab) basically combines both types of drugs into one molecule. Our belief was that:
- Both types of drugs are already approved for use in non-small lung cancer – the main indication Summit are targeting – and there is basically no reason to expect that a single molecule combining these two mechanisms would perform better than two molecules doing the exact same thing. In fact, there may even be commercial incentives in the US favouring the use of two infusions over one.
- PD-1/PD-L1 and VEGF inhibitors have already been tested in combination in real clinical practice by individual oncologists and by a number of companies in preclinical and clinical research. So far, these studies suggest that combining a PD-1/PD-L1 inhibitor with a VEGF inhibitor is no better than just taking a PD-1/PD-L1 inhibitor.
Earlier this month, Summit released phase 3 results claiming that Ivonescimab performed twice as well as Keytruda alone in the context of patients with locally advanced or metastatic non-small cell lung cancer. This is a very big indication. To say the stock was strong is understating it. It climbed about 10-fold from where we first looked at it.
There are however significant caveats.
- The trial was conducted in China alone and is non-admissible for getting the drug approved in the USA.
- The trial used a weaker control arm than the current standard of care in the US.
- The company claimed a benefit on “progression-free survival”, which is merely a surrogate endpoint for (and far less important than) “overall survival” (lifespan). Improvements on progression-free survival often do not translate to overall survival. A drug that improves progression-free survival but not overall survival is useless. Ivonescimab’s effect (or lack thereof) on overall survival has not been determined.
- Since US patients are inherently different than Chinese patients and Chinese clinical trial standards are different to the US, any “benefit” seen in China may not translate to the US. (Several Chinese trial results have not been reproduced outside China.)
- There are a few reasons to believe the trial may have suffered from selection bias. For example, the control arm performed worse than the control arm in other trials – raising the possibility that the trial was biased by over-weighting very sick patients into the control arm.
We largely had to cover the stock because it got big the wrong way. Our loss was well under a percent of the fund but only because our starting position was tiny. We hope the drug works. If it does the survival rate for many cancers will be dramatically improved. Moreover, Summit could be a great stock from here.
However, many competent people we know (including research executives at companies that have trialed similar combinations) doubt it. But we are mostly out of the stock – and when it passes or fails its phase 3 trial in the US we will report back. If it fails we will have intellectual satisfaction – but would prefer a refund.
