So Christmas is coming up, and I think I did pretty well for travel this year. Counting actual travel (not just visiting friends and family), I had 2 days in Toronto, 2 in Montreal, 11 in Africa, 13 in SE Asia and 30 driving across the US, for 58 total. Better than one in seven days on the road - not too bad! Last year, I only managed 28. Now that I’m doing the workaday, though, I don’t think I’ll come close this year. If I don’t take any days off except for vacations and Christmas week, I’ll sneak in 28. What can ya do? Ni modo.

No worries, though; 19 more months of work and I get some time for a real trip.

(Does it make me a geography geek that I got 35/50 on this geography quiz in the SF Gate this weekend? What’s worse is that I found two mistakes! Bahrain is also connected by a bridge to Asia, and Montenegro is the newest UN member. Hey, I used to read the atlas growing up, not Tolstoy.)

Before I go, here’s an interesting econ puzzle. A guy is selling his company; he knows what he values it at, but you only know that his value is drawn from (0,100), uniformly distributed. You value the company 50% more than he does. How much do you offer?

75 seems reasonable, because the expected value of the company for the seller is 50, and 50 + 50(.5) = 75. The answer, though, is zero. Once you make an offer, he either rejects (signaling that his value was higher than your offer) or accepts (signalling that his value is lower than your offer). If he accepts your offer of $b, you now know he values the company at some x drawn from a uniform distribution (0, b). Therefore, the expected value of the company is b/2, and your value of the company is b/2 + (.5)(b/2) = 3b/4. Since your value for the company is 3b/4 and the company is costing you b, your profit from the deal is 3b/4-b = -b/4, which is negative, and thus you should offer b=$0.

This is a mindblowing result, isn’t it? (Well, perhaps mindblowing for the 1% of people who really like math, finance or econ, but the rest of you will have to trust me). You can think that, owning the company, you will make 50% more profit that the current owners, but nonetheless you won’t offer to buy the company. Private information is a bitch. This example is directly applicable to real life as well - convert the scenario above to buying a used car…

Maybe it’s my complete ignorance of economics and finance, but I’m skeptical of the claim that there exist situations in which you could be completely ignorant of the seller’s value of an item, yet know that your is 50% more. Specifically, your value depends on his. Could you enlighten me with a more concrete example?

Sure: you’re a private equity firm considering the purchase of a privately-held company. You think they’re incompetently run, and placing your CEO in control would expand revenue by 50%. You’re not exactly sure, however, what the current profits are.

You’re also not completely ignorant of the seller’s value: you know it’s between 0 and 100. You can write similar problems with, say, 50 and 100, or problems where you have to pay some amount of money to determine a better lower bound (”costly verification” models).

Here’s another example: Your oil company has refinery and distribution networks that allow you to sell oil cheaper than any other producer. A company, which has done some private internal exploration, is considering selling their concessions to an oil field. You don’t know exactly how much oil is there, but you know that you make more profit with the oil there than the current owner can.