Seth Klarman talked with HBS professors recently, and the video was just released.
He described his approach to finding investments as looking a “mile wide” as the first step:
Most investors I come across don’t have the army, technology, and specialties that a shop like Klarman’s have, a fixed cost that almost no investors can afford. Going a mile wide requires that fixed cost as table stakes.
However, the interesting thing about financial markets, is the price of admission is zero for many instruments, but principally publicly traded equities. Ignoring investors who can’t read a financial statement and play markets based on price / sales or price /earnings ratios, too often I see investors play across many disparate industries. How can a one person band know enough auto parts, consumer internet, financial services, payments, etc. – and realistically believe they can construct a portfolio in which they have enough of an edge to pick outperforms in so many different industries?
Most often I see investors consume all the company’s filings and presentations – and go on to write a long investment memo with an associated model that is based on everything the company publicly files. The output is essentially a restatement of management guidance, a nod saying “I believe you, company A.” There is no original research, no competitive analysis, no value change diligence, no scuttlebutt of former employees, no industry contacts, etc.
Going a mile wide without doing the hard work is dangerous, and virtually impossible if playing in the pools of investments that the big guns play in. Skill is often mistaken for luck.
To be sure, I am guilty of the above at times. But for the most part, I’ve sworn off of single company investing (with effectively monoline products), and mostly allocate to investments that are either explicitly or implicitly the equivalent of investing in an investment fund, a fund of funds approach.
There’s no perfect way to invest, but one person playing a mile wide seems like playing a physical game with hands tied.