My only ever short thesis

A frothy market. A dodgy serial-acquirer. A wise analyst. A poetic justice ending. A real story of a short thesis, in which I played a bit role.

While this may sound surprising given my jibes, my first investing job was on the sell-side. In 2000, I was a lowly summer intern in Lehman Brothers’ equity research team. Somewhat ironically, I worked on a short-selling thesis. It was quite a learning experience, much like how I learnt that British summer was an oxymoron.

By way of context, mid-2000 still felt like boom times. Friday dressing was for all days. Summer interns flew business class. IPOs were plenty, as was booze at lavish office gatherings. It’s only with hindsight that bubble-tops get precisely dated. In real time, it took many months to realize that we had gone off the cliff, like those cartoon characters suspended in midair. A short thesis was quite unusual for those go-go times.

It was driven by an extraordinary analyst who I was lucky to work for. Brian Skiba was a senior analyst who headed the equity research team covering European Technology Sector, spanning software, internet, telecom etc. Uncharacteristically, he had not only held real jobs before working in markets but had founded two software businesses. He was a wise, wisecracking man who retained intellectual honesty and a sense of proportion, even as a bubble raged all around.

On Day One, he took me out to lunch. It turned out to be my longest meeting with him, as a hot market meant that he was constantly travelling for roadshows. He asked me “Why equity research”. In typical over-enthu mode, I started “Since I got centum in Maths in 2nd standard, I knew equity research was my calling” and blabbered non-stop while he finished his steak. Wiping his mouth, he asked “Don’t you want to ask me the same question”. So, I did. He replied “Money!”. He said there were far more fulfilling things to do in life than stock-market shit, including what he had done in his past life. But this was interesting work that paid for his home and kids’ college. He gave the job his best (he was #1 ranked) without taking it, or himself, too seriously.

Brian’s task for me was to study a company called Lernout & Hauspie, which claimed to be the leader in speech-recognition software. His sense was that there was something fishy about the company. But, he had too much stuff on his plate and hadn’t had time to flesh out all the details. For added joy, management refused access to an analyst who had downgraded them. My brief was simple: poke around Lernout &Hauspie to validate/negate a (short) thesis. With hindsight, there was no doubt about where I’d end up. Someone with his perspective recognized in 60 seconds what I took the next 6 weeks to validate.

So, I poked around. I read everything available and between the lines. I logged every number I found to get a more granular view of the business. As real-world jigsaw puzzles have many missing pieces, I did what I could to plug the gaps. But the secret didn’t emerge from a detailed grunt work or XL models. The secret lay in plain sight: Lernout & Hauspie was a serial acquirer!

Every cockroach in the kitchen was a descendant of this granddaddy. There was little organic growth or profits. Management had made 20+ acquisitions in short order to hide this, including acquiring their two largest competitors – Dragon & Dictaphone. They seemed keen to maintain a façade of an exciting growth story at any cost. They went to great lengths to hide like-to-like comparisons. They didn’t share splits across geographies, segments, organic vs inorganic etc. They were salesy but didn’t give one straight answer to any pointed question. It didn’t just smell funny, it stank.

So far, somewhat incriminating but still high-level. We needed something more specific. That turned out to be Korea which was, oddly, their largest geography. In certain quarters, over half their revenues were from Korea. Primary and secondary sources indicated that speech-recognition didn’t travel well across languages. Two decades back, it barely worked in English. Microsoft’s internal team had labelled their group “Wreck a nice beach” as that’s what the computer returned on hearing “Recognize speech”. Even stranger, their claimed use-case in Korea was registering stock-trades (customers simply spoke out their trade orders & Lernout & Hauspie software transcribed them for execution). It seemed impossible for unreliable software to be used for critical information on which real money was gambled.

Since I was less lazy back then, I cold-called Korea to try to talk to buyers of Lernout & Hauspie software. One kutti problem - no one who picked up the phone spoke English. Plus, I was randomly calling board lines of chaebols (which happened to also be in stock-broking) asking about speech-recognition software. Sometimes, I miss being young and stupid. After a few days without sub-titles, I stumbled on a strategy that led somewhere. To the person who answered my call, I only asked to be connected to anyone at their firm who spoke English. Since I was calling international, many were kind enough to oblige. Then, I’d work my way inward, getting to the right division and again asking for an English speaker. After a few weeks of misguided effort, I actually managed to talk to a few relevant people in the Korean stockbroking industry. Needless to add, they had never heard of voice-recognition software or Lernout & Hauspie. QED.

Whenever Brian managed to be in office, I was one of a dozen people who needed to talk to him. I usually got to walk with him from office to front-door. He digested what I found, synthesized it better than I could and nudged me towards which thread to tug at next, all in a few minutes. He never had me do unnecessary work. He nudged me away from dead ends. I was free enough to watch a movie on TV every weekday evening. He was first to realize when we had reached the point of diminishing returns. After 5-6 weeks, he felt we had enough to go on, and even let me to rotate onto a different desk for the rest of my stint to get a broader experience. I never asked him what he would do with this. He likely took it to clients looking for short-selling ideas. More than anything, he knew all along but just wanted to be sure.

Postscript: My work was done in April-May of 2000. In August 2000, business media started carrying articles about Lernout & Hauspie being a fraud. By September, stock tanked precipitously. Promoters and CEO stepped down. SEC initiated a formal probe. Stock essentially went to zero and company’s technology got sold for a modest sum. After a long trial, two eponymous promoters and salesy CEO were sentenced to prison. Korea business was revealed to be a complete fraud. I didn’t maintain contact with Brian subsequently, but he ensured that an uncharacteristic bonus check was sent at year-end to a summer intern. Money!

So, what did I learn that’s useful even now? At a subtle level, the idea of investing being an apprenticeship business got seeded in my head. It was invaluable to be guided in real-time by someone with wisdom and an ability to keep one’s head when others are losing theirs. Years later, after a disastrous start in investing, this idea was crucial to helping me see light.

At a more practical level, this was my first exposure to patterns common to bad businesses: serial acquisitions; egregious capital allocation; glib management; opacity; complexity; plain weirdness (Nasdaq-listed European company mainly selling in Korea? Exciting growth story can’t grow without M&A?); anomalous pieces of the puzzle that just didn’t fit together. While these elements are timeless, they also feel topical as new-age companies go nuts with easy money.

Apprenticeship is learning from better people. Pattern recognition is learning from history. These are pretty much all there is to progressing up investing’s learning curve, while minimizing expensive learning through own mistakes. Without my realizing it back then, Brian helped me take the first step on this journey. That experience helped me spot dodgy businesses. They just felt like Greek & Latin. Or Korean.

PS. This is the first essay of a series, tentatively titled “Connecting the dots” where I’ll talk a bit more about my journey. Hopefully, I’ll limit self-indulgence by making it about investment learnings.