Why Groupon Is Poised For Collapse

Yes, This is one of the most dubious Groupon "sells".

There are a very small class of businesses that excess capacity which costs nothing to use. These mostly the type of business where you're just consuming an "experience" - adult classes, skydiving school, etc. Unfortunately, a big part of business, classes, is already competing with the Internet. And so the rest is skydiving schools, martial arts schools, and yoga schools, the few places where you just have to be there. But those are pretty marginal.

Restaurants certainly aren't in that class. Food is expensive. Food is a significant share of cost for anything but the most expensive restaurants. Sure, restaurants may throw food away each day BUT the only way to profit from that optimizing that is to sell food cheap on the contingency that its available. That's far from the Groupon model and clearly would "cheapen" the feel of any given restaurant.

I heard of a restaurant in France years ago that priced by the hour. That could actually cut waste to nearly zero - but it would destroy "the feeling of specialness" which many higher end restaurants cultivate.

(Micro)-Economics is much more like physics; here is how it tends to work, how it tends to optimize, how it tends to in practice actively subvert attempts to predict what it will do (and thereby actively resisting any attempt to create a macro-economic theory, or at least one that lasts for any period of time). It's a very physics-like definition of "efficiency". The fact that the dam broke in the most energy-efficient manner and the resulting water deluge can be modeled with a high degree of accuracy by assuming it will follow the least-energy course in an efficient manner doesn't make the resulting flood "good" or "desirable"... just... efficient.

On the other hand, start your economics model with anything else and it's like trying to rewrite physics starting with material ether or something. I have "faith in the markets" in much the same way I have faith in physics (which I will remind you contains a lot of similar "merely statistical" theories, what you learned in school is the aberration, not the rule, and also is really strong on the fundamentals and gets progressively weaker as the systems get larger, rather like economics), but that doesn't mean I blindly trust them. Physics will kill you without blinking. But if you want to prevent that, it's a lot better to come to understand the laws of physics than to try to create new rules for the universe by sheer force of denial. Trying to guide the physics while actively refusing to learn about it is doomed to failure, too. Same for microecon.

1.Mkt isn't like BM. BM permits -ve values & any market index say SPY must be non-negative.

2. We (quants) model market indices like SPY as a log-normal process like so : dS = rSdt + sigmaSdz where S = underlying ( or market index ) dS = change in S, dt = change in time r = risk free interest rate ( about 0.03 % currently ) dz = normally distributed rv with mean 0, std dev sigma

There is an embedded Brownian component in dz, so also in S. But that doesn't make S itself Brownian. More here: http://en.wikipedia.org/wiki/Wiener_process

"No observer could possibly predict which direction it will jump next"

Nobody is trying to do that in the very next instant. OTOH, you can certainly predict where the market will go, say, in a year. Well known econ result: DP ratio regresses positively with return. ie. Dividend-price ratio predicts stock return with high statistical significance. Source: Cochrance, French/Fama, et al. http://dash.harvard.edu/bitstream/handle/1/3122601/campbells...

"there are plenty of people making random guesses...there are also plenty of fools, and they are investing on all sorts of techniques which ignore any of the global considerations, and of course this is how bubbles expand."

Ha ha ha! Very amusing but very inaccurate. I think this is my general problem with geeks on the outside of finance. They happen to think markets are purely stochastic. X has randomness doesn't mean X is unpredictable. Otherwise why have APT, CAPM, covariance matrices, factor models, Heath Jarrow, Black Litterman...you've written off some 100+ years of economics in one breath !

To summarize, there are tons of observers, who predict where the overall market or some portfolio of equities ( or bonds or agencies or munis or treasuries or whatever else ) will go. To get data points for these predictions, you use a historical time window, compute coVar matrices, invert them, build a CAPM....etc. Anyways, its not rocket science. There are tons of CFAs & Finmath engg ( I'm one of them ) who do this sort of work day in and day out. Its our bread & butter. If I said something like "market is a lot like brownian motion & people are making random guesses", I'd be out of a job.