Dragon Systems

I find the story of Dragon Systems hilarious and horrifying so I’m never going to miss an opportunity to tell you about it and one occurred today. The story, quickly, is (1) Dragon Systems, a closely held speech-recognition company, hired Goldman to advise it on a merger with Lernout & Hauspie, (2) Goldman assigned Dragon an extremely JV team of bankers, (3) Dragon sold itself to L&H in June 2000 in an all-stock deal, (4) L&H soon turned out to be a massive fraud, and (5) L&H filed for bankruptcy in November 2000 and Dragon’s shareholders lost $300 million. Dragon sued, Goldman won a trial in January, and today they won some more, for reasons I’m not clear on. That is, I’m not sure why they had to win again – the judge issued an opinion finding in favor of Goldman, even though a jury did the same thing in January. Also it’s not much of a stirring win: Read more »

There’s a lot going on in today’s Wall Street Journal story about how Hewlett-Packard “missed a chance to back away” from its acquisition of Autonomy – which H-P now thinks did a lot of revenue recognition fraud and on which it has taken zillions of dollars of writedowns – but this description of the board’s approval process for the deal is the only thing you really need to know:1

H-P directors and bankers calculated how much revenue Autonomy would have to add over 10 years to justify such a price. Autonomy’s trajectory alone wouldn’t get there. The deal required assuming more revenue growth as a result of the tie-up than H-P usually assumed in acquisitions, said people familiar with the matter. But the directors believed they could make the numbers.

Those calculations were done without knowledge of the alleged fraud but who cares? Here is, roughly, H-P’s thought process:

  • The price that Autonomy demands is X2
  • Autonomy’s DCF value is Y1 based on expected revenue growth but no synergies
  • Y1 < X
  • Well but the DCF value is Y2 with regular-to-aggressive synergies
  • Y2 < X
  • Well but … well we could make up another number Y3
  • Y3 >= X
  • LOOKS GOOD.

If you notice that current revenue numbers are made up, then that changes Y1, but Y1 isn’t an input into Y3 – the actual value that H-P put on Autonomy – because that number was also just made up. Read more »

This Dragon Systems lawsuit that’s bopping along in a Boston court gives me the absolute heebie-jeebies; it is the investment banking version of the dream where you show up to school and there’s a test that you forgot to study for. One day the youngish team of Goldman bankers is patting themselves on the back for guiding their client, family-owned speech-recognition software maker Dragon Systems, to a $580 million sale to a Belgian acquirer. The next:

the 2000 all-stock deal was quickly followed by an accounting scandal that led to the collapse of suitor Lernout & Hauspie Speech Products NV. Goldman Sachs, which advised Dragon on the deal, called witnesses to counter allegations by Dragon founders Jim and Janet Baker that its negligence cost them their life’s work. … Dragon’s founders, who are seeking hundreds of millions of dollars in damages, claim that four Goldman Sachs bankers assigned to the transaction committed gross negligence by failing to pursue questions about Belgium-based Lernout & Hauspie’s finances that should have led them to avoid the deal.

Oops! I really do feel for the bankers here. It’s Friday afternoon and nobody’s reading so I feel comfortable making this confession: when I was a banker I once underwrote some convertible bonds for a company that we’ll call Company X, and later when I was a blogger Company X went bankrupt and those bonds are … those bonds are not doing so hot, these days. That keeps me up at night, sometimes.1 But in my defense it took like a year! And Company X wasn’t a giant fraud, either, it just turned out to suck at making money. The Dragon team doesn’t have those excuses: Read more »