15,000% increase in value per employee: 1990 to 2017 - investing implications

It is no secret that it has been extremely difficult for most investors to find appealing value investments in the current environment. The ‘Buffett-metric’ US market cap to GDP stands at 131%, eclipsed only in history for 6 months at the 2000 bubble peak where it reached 145% briefly. In order to stay fully invested, many investors have to reach outside of their value discipline and some, like ValueAct Capital have been returning capital to investors. 

Since the industrial revolution, technology has been a wedge between those that can adaptively use tools to amplify their capabilities and those that are displaced by it. A map of the 2016 US election outcome approximates the boundary of the amplify vs displace effect. Today, automation and AI are accelerating this trend to unprecedented impact. In 1990 the ‘Big 3’ in Detroit (GM, Ford, Chrysler) had a market capitalization (inflation adjusted) of $65 billion with 1.2 million employees. Today – the ‘Big 3’ in Silicon Valley (Apple, Google, Facebook) have a market cap of over $1.9 Trillion with 210,000 people: a > 15,000% increase in value per employee. If you thought you could avoid this by excluding tech from your investment universe, good luck: in 2016 for the first time , there were more $1B+ acquisitions of tech startups by NON-tech companies. After a century of dominance, Buffett favorite Gillette lost 16% of its US market share in just 4 years to startup Dollar Shave Club thanks to YouTube and inexpensive overseas manufacturing. It now seems clear that no business will be spared the impact of these exponential technology forces.

Our efforts to use data as a core part of the investment process started in 2009 with screening the universe one metric at a time and using Excel. If only Buffett had even this kind of basic search capability instead of manually flipping pages in the Moody’s manual in 1960, who knows what would have happened? By 2014, we had built a database driven system to take in hundreds of screens and integrate them together. In 2017 our new system takes all those perspectives and captures history going back decades so we can rigorously and scientifically test all of our assumptions with hard data. Our goal is two pronged: A) To generate multiples of our capital in long positions in companies that we can influence and improve their strategy and operations via entrepreneurship and targeted activism and B) To hedge external / macro factors and generate absolute return by shorting low quality companies. The goal is to make consistent profits while being positioned strategically to benefit disproportionately when there is a market dislocation. 

As Buffett says in his recent Berkshire annual letter:

Charlie and I have no magic plan to add earnings except to dream big and to be prepared mentally and financially to act fast when opportunities present themselves. Every decade or so, dark clouds will fill the economic skies, and they will briefly rain gold. When downpours of that sort occur, it’s imperative that we rush outdoors carrying washtubs, not teaspoons. And that we will do.

Today is not a storm to wait out but the best opportunity set with the right skillset

Times like 2009 are not just to be weathered through, they are to be taken advantage of aggressively. Anyone can pay for insurance, but what if you could create insurance that pays you a premium and then offers purchasing power exactly when the opportunity set is most interesting? We believe in the importance of preparing for the once-a-decade ‘downpour’ so strongly that we have invested considerable effort and expense in our ‘ark building’. The upshot is that in preparing for the worst, we end up with a powerful toolkit that can provide a significant competitive advantage across all periods. The machine learning capabilities we have developed are not ‘black boxes’ but rather systems to enhance our capabilities as long term fundamental investors. With these tools we can apply our principles more consistently for our more diversified passive positions while continuing to use our entrepreneurial efforts to grow select businesses.

The number of investors who outperform with the ‘traditional’ value-focused stock-picking methodology has dwindled as the world has changed in deeply structural ways.  The fundamental change driver is technological and to a lesser degree, demographic – the monetary policy changes are mostly symptomatic, rather than causal as is commonly misunderstood. 

It was impressed upon me by a mentor, who is one of the most successful investors of recent times, to always be on the ‘right side’ of larger trends. In my view the clearest and most predictable trend in capitalism are the exponential price/performance curves in technology: Moore’s Law for integrated circuits has now spilled over into the price/performance of solar panels, batteries, sensors, drones, robotics, etc. Consumers of these types of products are delighted by the massive increase in capability at a given price point. Producers, on the other hand must keep pace with an exponentially deflating revenue per unit of capability! 

A ‘traditional’ stock picking approach is firmly on the wrong side of this trend which is why we have sought to do more and physically transform companies we invest in as entrepreneurial activists. We concentrate our long investments on businesses positioned strategically to benefit from declining input costs and capability advancements. With respect to stock picking - humans are simply bandwidth constrained, emotionally biased, and can rarely appreciate the compounding effects of change.

To be perfectly clear: we 100% subscribe to the idea that we seek to buy companies for less than their intrinsic value and short companies at prices significantly above their intrinsic value. My argument is that 1) the market has become more competitive as more parties analyze and make investment decisions this way and 2) the business environment itself changes faster and in a more exponential way making ‘intrinsic value’ much harder to assess. To understand the ‘dynamic intrinsic value’ of a Company requires real business depth of understanding how things evolve in a market and also within a Company. Some types of businesses can scale and quickly saturate their market while others may have positive feedback loops where they get even stronger as they grow larger. A copper miner expanding its production will cause prices to decrease and mean revert its revenue growth, Amazon builds more datacenters and expands its lead as it furthers its per-unit cost advantage over rivals.

This is Part 1 of 3 of our open discussion on the machine learning implications for fundamental investing - click below for more.