Meson Builds Big Stake In “Revolutionary” EV Company; Holds 200 Shorts

• Venture Capital Investing Process Improvement Through “Machine Learning” 
• Kolanovic’s Deep Dive Into Artificial Intelligence Lacks Human Understanding
• Lakewood Capital Bets Against One Of China’s Richest Billionaires
• Great Investors Debate – Should You Be Investing In Family-Controlled Businesses?

Get The Full Warren Buffett Series in PDF

Get the entire 10-part series on Warren Buffett in PDF. Save it to your desktop, read it on your tablet, or email to your colleagues

Dear Partner,
For the quarter our performance was -7.4% vs. indices of 2.1% HFRI Hedge Fund Equity Index, 2.5% Russell 2000 and 3.1% S&P 500. Shorts broke even net of borrow fees while longs detracted 7.4% for the quarter. Although we do not normally provide intra-month updates, the recent public announcement of our largest position, Sevcon, being acquired at a 61% premium has made July an unusually successful month. As of this writing, we expect to be up approximately 20% Net YTD, well ahead of indices while maintaining a market neutral defensive positioning. We will provide up to date numbers when our administrator has completed the July close.

In our last two letters we describe our unique approach to long term fundamental investing using our machine learning, data-driven system we call Meson Gravity. In particular, we excel at short-selling where we have benefitted from a historical accuracy rate of roughly 80%. Short selling won’t make you rich in markets that do nothing but rise but we have still been profitable shorting on an absolute return basis since we positioned ourselves market neutral starting 2014. We hope for the best and certainly don’t pray for disaster but we are prepared to look unusually intelligent when a declining market finally arrives. Our penchant for short selling comes not from a place of cynicism but rather from a deep respect for how brutally competitive and hard business truly is, having built and run numerous private and public companies ourselves.
The ‘Buffett-metric’ US market cap to GDP stands at 135%, eclipsed only in history for 3 months at the 2000 bubble peak where it reached 145% briefly. The volatility index (VIX, aka the “fear index”) also reached an all-time low recently at 8.8% - the long term average is over 20%. Retail investor margin debt also has reached new highs. We wish long only investors, particularly those with margin debt, the best of wishes at their seemingly never-ending party that we have declined to attend. We have instead been busy working hard on investments that seem to us inevitable regardless of short term market gyrations: using data patterns to predict poorly performing businesses; improving efficiency of industrial equipment at lower cost; electrification and improving operations at decades old stalwart companies.
Portfolio Update - Machine Learning System Drives Diversified Positions
Our Gravity platform began to drive a portion of our diversified longs and shorts starting June 1 and we will have completely made the transition by mid-August. Our long portfolio consists primarily of our concentrated active positions where we have an ability to influence the future of the company in a positive way. Currently about 70% of our long exposure is in concentrated names like this. As we have built relationships with more private equity firms over time we will tend to be focused on opportunities where we can buy a meaningful toehold position and then work to acquire the company with a partner. We have two examples of this in the works at the moment in positions we have not publicly disclosed yet and I look forward to discussing them as they progress and we accumulate stock. This strategy is what we laid out as our goal as far back as 2013 when we made our first acquisition offer for InfuSystem but is very resource intensive and we have not had the resources or capital to execute on it consistently until recently.
Secondly, the opportunity set in the market leads me to conclude (perhaps 10% in jest) that all small caps are either activist opportunities or shorts. High quality management of good small cap companies realize they can personally make more money by going private so there has been a creaming effect over the last decade or so. The number of small caps that have become large companies is staggeringly small – either because they fall along the way or are taken private / acquired before they become large.
This skewed opportunity set is something we are well equipped to handle given our hands on approach where the undervalued small caps are ‘fixer uppers’. Our data-driven Gravity platform also recognizes this state of the world. This is unsurprising in that the system is designed to capture our fundamental investment approach and then implement it in a consistent systematic way, allowing us to reduce risk for our diversified positions.
The highest probability longs in our diversified portion of our portfolio are high quality larger companies that have a blend of good and consistent margins, ability to grow via acquisition (but not too much), and being moderately capital intensive (with some exceptions). Unlike ‘factor’ investors we are not looking for extreme top/bottom percentile values on any particular metrics – indeed these ‘red’ flags often indicate that the numbers are faulty rather than a true observation of interest. A recent academic paper for example showed that the lowest P/E stocks are not in fact the cheapest but are the most probable to restate “E” downwards! We aim for robust predictability in our approach, not extremes – 10 subtle yellow flags are better than 1 big red flag everyone else can obviously see. Example stocks in the portfolio include General Mills, Cisco, Church & Dwight, Paychex, and Comcast. Keep in mind these are <1% positions but the important idea to take away is that the best passive/diversified longs in this environment are boring, consistent, defensive businesses.
The short book by contrast is where we shine and the interesting opportunities are in this environment – regardless if the market continues its rise or not. Generally speaking they are companies that lose money, dilute their shareholders, are heavy on SG&A and low or declining revenue growth. You may recognize that these characteristics, while normally unappealing, describe some biotech homeruns. We do not currently short biotech or other similar industries where GAAP financials are essentially totally irrelevant to predicting success or failure. We have other data sets that we are working to incorporate over time for these more unique and subtle situations such as the track records of the individuals involved with the company, patents filed, etc. Keeping in mind that we have 200 shorts: examples of our highest probability shorts include alpha-En (which claims to have revolutionary Li+ battery tech), Pershing Gold, Yangtze River Development (Chinese real estate), VirnetX Holdings, and a couple hydrogen fuel cell players. We are highly confident in the long term prospects for these businesses and thanks to our system we can reduce the risk of short squeezes to a bare minimum due to the diversification and consistency of our portfolio management. As we say, economic gravity wins in the long run.
Due to demand from certain investors that are unable to invest in any securities with reduced liquidity (which occurs when we take board positions), we are launching Meson Gravity LP August 1. This fund is exclusively the diversified machine learning driven positions and does not include any concentrated positions that result in reduced liquidity or require lock ups. Our main fund, Meson Capital LP will continue to represent the full spectrum of how we invest and make money and I expect returns to be higher in Meson Capital LP than in Meson Gravity LP thanks to the benefit of concentrated active investments. The tradeoff is a 2-year rolling lockup (of 1/8th of capital per quarter) which eliminates any withdrawal cliffs for us and higher volatility that is unavoidable with concentrated positions.
Electrification Sequel to Sevcon
As much as can be said that our >2X return in Sevcon is a ‘mistake’, it was certainly not the optimal outcome and I believe that had we had more control and/or fewer historical factors to navigate, we could have done much better. The market for electrification is being transformed by a number of different dynamics currently. The three key components of an electric drive system are 1) the energy source (battery in mobile applications, grid for stationary) 2) the power electronics and 3) the electric motor.
The cost of batteries is exponentially deflating at roughly 15-20% per year which by most estimates will make electric vehicles cheaper on initial purchase than an equivalent gas-powered car by 2021 in Europe and perhaps by 2023 in the USA. Consider that a) the total cost of ownership tipping point is probably 2 years sooner given lower operating costs and b) EVs intrinsically perform better than gas-cars: how long will it take until the sales of 1) more expensive, and 2) lower performing, gas vehicles drops to nearly zero, totally replaced by EVs?
When compared to the battery, the power electronics and electric motor are ‘good enough’ and more mature on their cost curves. Sevcon’s strength is its expertise in power electronics where a customized subsystem is required for any new vehicle design. Modern electric motors are little changed in literally a century. Nikola Tesla could look at a top of the line P100D motor and immediately recognize its operation. It was only 40 years ago when power electronics and microprocessor control became feasible to operate an electric motor at a variable speed without physical brushes (that wear out and are expensive). Prior to this only “on/off” control was possible. This enabling technology allowed for a total paradigm shift from DC-brush motors to AC-induction motors and certain players such as Danaher became enormously profitable industrial giants on the back of it.
Today, I believe there is another paradigm shift in electric motor technology that is about to take place and I have been working to position us over the last two years for the best way to be a driving force in this. Fundamentally all electric motors today operate in an analog way – sine waves of electricity drive them (to oversimplify). There is a class of motor that operates in a fundamental digital way called switched reluctance “SR”. The basic idea has been around for over a century and there are very niche commercial applications requiring extreme reliability and durability, such as in the core of a nuclear reactor. There was a brief period about 20 years ago where an attempt was made to make this the standard of motor technology but, like electric cars, the other enabling technology was not ready yet.
In the case of EVs, 20 years ago the ‘false start’ was that batteries were nowhere near acceptable cost or performance and in SR motors the power electronics and computational requirements were not good enough then. The control strategy of an SR motor is orders of magnitude more computationally intensive than a conventional motor and it requires nonstandard power electronics circuity. Both of these problems were very expensive to solve – until very recently. Our favorite class of problems is the substitution of material cost/complexity with information/compute complexity because the latter is highly predictably improving thanks to Moore’s law. This hardware improvement/commoditization transforms them into software problems.
There is one company in the world that has successfully commercialized the new generation of SR motors and I have been working with them for the past year to expand the scope of their efforts and create a strategy to capitalize optimally on this changing technology landscape. They have a commercial product, it works, and it can be produced in volume. It is the most efficient electric motor in the world, is more compact, more robust, and cheaper to produce than the current motors out there. Over 90% of an electric motor’s total cost is the electricity it uses, efficiency is a big deal – the government is mandating improvements. The key is software excellence in the control, electronics, and design. This is not a dimension that incumbent motor companies have a hope of competing on as it is a fundamentally different skillset than shaping metal and winding copper wire that they are used to competing on. It is also protected by numerous patents and I have been working on building relationships with the world leading engineers in this extremely niche discipline over the last two years.
In the interest of full disclosure, the company referenced above is a private company and like our other concentrated active positions, we will be involved hands-on to create value. Our main fund Meson Capital LP is a hedge fund and has a ‘side pocket’ described in its documents that handle the potential situation that there is a mismatch in liquidity due to private or restricted positions. This is a standard provision and ensures that no individual investor is disadvantaged by a private investment becoming disproportionately large. This could happen either because a private investment is wildly successful or due to the situation where redemptions require the private investment be distributed in-kind. We are restricted to keep the side-pocket of private investments below 25% of the fund and we are currently below 10%. Depending on the opportunities available we may use this but certainly set the bar much higher on return expectations for illiquid investments. We are well aware of the potential issues that can result in a long term illiquid position and do our best to avoid them. For example, we do not invest in early stage companies that are still in the technology risk phase before commercial products.
To be clear we have never encountered an issue with the side pocket during 5+ years of experience with positions being illiquid due to our board seats so I do not anticipate any issues but wanted to be forthright with this provision. We recommend investors note that potentially 25% of their investment in Meson Capital LP could become illiquid in the worst case scenario and include this in their portfolio allocation decisions. For investors where liquidity is a requirement, Meson Gravity LP has no side pocket and is liquid on a monthly basis. As a reminder: the General Partner has the vast majority of their net worth in the main fund Meson Capital LP and is subject to the same liquidity restrictions and side pocket as LPs.