Tuesday, March 14, 2006

How markets work with imperfect information

Our friend Quincy left a comment a while back that I didn't get around to: "The thing I don't get is the argument that, if the market depends on imperfect information to function, why it then follows that introducing more imperfect information, which benefits neither party no less, can be a good idea."

Well, markets have a great deal of imperfect information, but that doesn't mean, nor was I trying to imply, that they require imperfect information to function. Nevertheless, we all make a lot of economic calculations and decisions based on imperfect information (still correct but incomplete), and things turn out just fine. In fact, imperfect information can be cheaper overall. We often don't try to acquire complete information before making a decision, because of search costs. Time and emotional impacts can make a purchase more expensive than the monetary price would indicate. By the time we determine we can spend $x less by buying the item elsewhere, we might have spent more than $x in time.

Last September, I wrote about people who put up with long lines to save 30 cents per gallon of gasoline. Even if my car's 12.5-gallon tank were completely empty, I'd save only $3.75. I already dislike standing in line, and I get very aggravated waiting in my car in line. I'm willing to pay more elsewhere to save time, and because of my income versus limited free time, what I'm willing to pay is more than what my father was when he was retired (fixed income, lots of free time). Last New Year's Eve, I paid 20 more cents per gallon, knowing it would be cheaper down the street. That station was right off the Saw Mill Parkway, and since I was in a rush, its higher prices were worth the convenience. I'd never really considered the economic value of time until taught by my undergraduate microeconomics professor Dr. Ikeda, who as I've said is an Austrian economist's Austrian economist.

Here's another example of imperfect information and a willing acceptance of it. I bought a new raincoat last night. I spent 80% more than I originally intended, but by definition I did not spend 80% more than what I was willing to pay -- that is, for the coat I bought in the end. I wouldn't have paid what I did for the coats I initially considered, but I was willing to pay it for a much nicer coat of which I was previously unaware. The moment my preference changed, I moved to a completely different demand curve, and so instead of Calvin Klein, I came home with a Joseph Abboud coat. Remember, as I explained in my entry on how markets clear, markets are always in a state of flux, with supply and demand curves constantly shifting because of changes in information. Such changes include discovering something you were previously unaware of, something that, upon discovery, you prefer to what you originally planned to buy.

Is it possible I could have gone to another department store and paid less? Certainly possible. Might I have found a less expensive coat that may have been an inferior substitute, but whose lesser quality I would deem worth the reduced monetary cost? Also certainly possible, but again, how high would the search costs be? It was already well past 7 o'clock, and I wanted to get home with a new coat. So like consumers often do, I committed to the purchase then and there, because I judged that spending less money would not be worth rushing to another store. Also, my action demonstrated that I preferred the uncertainty of getting the best price to the risk of paying more elsewhere (because I'd get there, see the higher price, and pay it anyway instead of going back), or the risk of going home empty-handed. Another store might not have that coat at a sufficiently lower price, or even a coat I liked.

A third good example of imperfect information is something Alex Tabarrok noted a few weeks ago on Marginal Revolution. Two professors, John Morgan of Berkeley and Tanjim Hossain of the Hong Kong University of Science and Technology, held 80 auctions of their own on eBay. "[Their] findings suggest consumers pay less attention or even completely overlook shipping costs when making bids..." Bidders were attracted to lower prices and didn't seem to consider shipping costs as carefully as we'd think. Shipping costs were high enough that consumers could have spent less overall in other auctions, whose items sold for a higher price but had lower shipping costs (and less in the end).

I'm not surprised at the results. In fact, I maintain that Austrian economic thought would disagree with the professors that, "In theory, dividing a price into these two pieces should have little effect on overall demand for a good..." That theory has the big assumption of a perfectly informed consumer, and Austrian economics' theory of market forces is based on information being imperfect (not necessarily wrong, just incomplete). Consumers act upon the information at hand and what they perceive, not absolute truth. Besides, we again encounter the high cost of searching, even on something like eBay with lots of search tools. Someone might bid on 20 things he likes, unaware that the 21st item would be his best deal, but he might have to do something in real life that is more important than saving $5 or $10 -- especially if the potential savings is marginally insignificant to him. That brings me to a small problem I have with the experiment. I'm not saying at all that it isn't valuable empirically, but it's results are only with relatively low values. As prices increase, people tend to take greater care about their total spending.

And if anything, there's an entrepreneurial opportunity here: sellers can advertise "zero shipping charges" and tout that as a selling point. "Don't be fooled by low prices with high shipping charges!" I don't do eBay, so I don't know how many sellers actually do that. If there aren't many, I'd say it's because the sellers judge that the increased sales (especially because information is imperfect and bidders may be unaware) are not certain enough to offset the lower profit. One shopkeeper could always undercut his nearest competitor by selling for 10% less, but he could also make hardly anything (or even go broke) doing that. There are limits to how much you can compete. The cost of increased advertisement is also a factor. So sometimes a seller will closely "follow the crowd" and adopt similar patterns, because he'd rather count on a share of the market than undertake the uncertainty of competition.

In Austrian thought, the fundamental role of the entrepreneur is to improve information. Because no single person can have complete information (which Hayek set forth in his essay "The Use of Knowledge in Society"), markets will have imperfections. The entrepreneur, in the course of calculating and bearing risk, and performing arbitrage, has a very real economic incentive to correct those imperfections: profit. By introducing better information, he will be more competitive than others and, in the absence of coercion (i.e. government regulations), be more successful. Should his information happen to be imperfect (like Coca-Cola's expectations of New Coke), then unless he commits coercion or gets others (like government) to do it for him, he will eventually fail. Failure is the free market's self-correcting "Darwinian" mechanism, which is why Austrian economics doesn't see recessions as inherently bad. They're necessary corrections as supply and demand curves realign, specifically because government policies skewed them from their natural tendency to gravitate toward equilibrium via entrepreneurs' actions.

I haven't read Morgan and Hossain's paper itself, so I can only go by the writeup in offering a couple of thoughts. "Morgan and his colleagues found that pricing on the Internet varies dramatically, contrary to predictions that the vast wealth of price information online would eat away at profit margins and result in one low price for consumers." But if consumers aren't saving that much by buying online, why are they so irrational to do it? It's because buying online still reduces their costs, just not always monetarily. They save time. One small example (both as an example and physically) is a small squeeze toy that I couldn't resist buying for my best friend. It cost $3 online and $5 to ship it directly to her. For me, it was like buying it for a fraction of the total price to buy it in person. I'd have spent so much time, not to mention on gas and wear on my car.

Morgan "compared auctions on eBay and rival Yahoo! with UC Berkeley Department of Agricultural & Resource Economics graduate student Jennifer Brown. After auctioning off identical Morgan Silver Dollars on both sites, they found that eBay auctions averaged almost 60 percent more bidders than Yahoo! and 30 percent higher sales prices on identical items." This shouldn't be surprising, either. eBay's larger population of bidders naturally means higher sale prices because there are more people to outbid each other, and eBay will have more bidders because of reputation. When eBay is mentioned, people immediately think "online auctions." When Yahoo is mentioned, people think "search engine" and "information portal." Both buyers and sellers will tend to use a marketplace that's popular for a specific service more than a general marketplace that offers the service but isn't well-known for it. I don't know if Morgan and Hossain examined it like I have (in their paper or elsewhere), but I think it's pretty evident with not much thought, even if it's not based in Austrian economics.


Anonymous Anonymous said...

That makes it simple. If I'm looking to buy something, I'll buy it through Yahoo! auctions. If I'm selling, eBay...

Honestly, it's quite possible that you could make a tidy profit buying things on Yahoo! auctions and reselling on eBay, no?

Wednesday, March 15, 2006 11:32:00 AM  
Blogger Perry Eidelbus said...

It's certainly possible to make a profit by engaging in that particular arbitrage, but could you make a greater profit doing something else?

I was still playing a little Ultima Online when the Southeast Asia tsunami hit. A few players had this idea to acquire gold in the game, sell it on eBay, then donate the money to relief efforts. Virtual items in online games are popular but not the most profitable if you have a decent job. Depending on the server, you could sell 1 million gold to another player, via eBay, for $9. But a normal player could take a few hours to earn that much. If players really wanted to maximize their donations, they'd have done better to flip burgers for a weekend, then donate that paycheck.

The idea was also self-defeating. It didn't take a genius to predict the drop in UO gold's eBay price, but I did pretty well to predict how much. I was correct that it would quickly drop by a dollar or two. Because players were earning much more virgual gold than normal, the market price fell to $7 per million in just a few weeks.

Wednesday, March 15, 2006 1:13:00 PM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home