Hacker News Reader

Comment by allyt | Hacker News Reader

I think this can be generalized to say that start-ups have a different cost-benefit analysis - one where losses are capped at the (relatively small) value of the company. On the other hand, large companies have to be risk-averse because the worst-case is several magnitudes worse. When you're working at ConEd or AIG, the "tiny probability/worst-case-loss" factors start to matter, because "worst-case" can include investors losing life savings, federal investigations, and jail time for your boss's boss.

View on HN

I don't think greater risk accounts for all the slowness of big companies. Some (probably most) of it is simply breakage caused by their large size.

The major factor at work is simply a lack of personal responsibility towards the end product in a large company.

In a startup there's noone to blame but you, and thus no need for politics. As anyone who has worked in middle management or above in a large company can tell you half of the job is making sure you can always find someone else to blame, while taking credit for the successes of others.

http://en.wikipedia.org/wiki/Principal-agent_problem

Exactly! I didn't know it had a name.

Funny that some of those large companies you mention didn't seem to get the memo about being risk-averse... ;-)

During their executive retreats they decided to be risk seeking instead.

I wonder if the problem at the recently troubled companies (AIG, Merrill, Lehman etc) was not enough checks in place or not enough power allotted to the employees who knew something was wrong.

Same as always: greed

Ideally a system should use greed (and other emotions) instead of having to guard against it.

From memory: the Excel bug that caused Moody's to give high ratings of mortgage securities (one of the probable causes of the whole mess) was known, but management decided not to fix it. So, greed can override checks and balances.

The bug wasn't in Excel. It was Moody's own fault, not Microsoft's.

No, their version had a bugger where zero in the denominator came back as "infinity."

Sorry, I didn't express myself clearly. I meant a bug in an Excel spreadsheet made by Moody's.

The issue was not a bug, it was a fundamental misunderstanding of how real estate markets work. Moody's had a model for pricing mortgages which assumed real estate prices never went down. It was basically the real life equivalent of the famous "Assume a spherical cow..." problem.

Reply on news.ycombinator.com