Sorry, I don't buy it. For any signal to be accurate there has to be a cost attached to making it : http://octavia.zoology.washington.edu/handicap/handicap_principle.html

Prices are valid signals of something when applied to scarce resources. But when I'm trying to price an information good, where the marginal extra cost of the next copy is as near to zero as makes no difference; there's no cost difference to the seller between selling me a copy of the greatest band in the world or a file full of white-noise.

I just don't see prices can be anything other than totally arbitary.

The only reason music has a price is because copyright law allows the industry to pretend music is a scarce resource, and file-sharing still isn't good enough yet. How can you get an accurate signal out of a government mandated monopoly?

Attention *is* scarce, and can be priced. That's what AdSense succesfully does by varying the price of a click-through. But we'll never see accurate signalling of information quality by price. Price is just not the right way to guage it.

PageRank and social routing (http://zby.aster.net.pl/kwiki/index.cgi?SocialRouting) will help us identify information quality. I can't see that price ever will.
 

Hey Phil,

1) Dude, saying

"..But we'll never see accurate signalling of information quality by price."

is absurd.

Whether or not you "buy it" is, honestly, completely irrelevant.

2) ...because markets all over the world do exactly this every day; just walk into a Ferrari dealership.

In fact, that's what the price mechanism *is* - if prices didn't signal (aka "price in") aggregate expectations, markets wouldn't work.

3) Now, I have already explained why media goods are not strictly info goods and why the marginal cost is not zero. This is a specious line of argument.

Put another way, of course "music has a price". That's because making and selling music is not *just* about technology.

In the simplest case, you can just think about marketing costs.

4) Final note - anything can be "priced". Scarcity is not really a great determinant of price in real-world markets. Consider oil. It's nearly as scarce this year as it was last year; the difference is that demand has been vastly amplified by China. This means what matters is the expectation of scarcity.

5) If prices didn't signal those expectations, the oil market would fail, right...? So think about that in the context of points 1-3
 

Umair,

maybe I expressed myself badly. I won't spend much time arguing here but will read your other stuff and probably disagree again over there :-)

When I said "..But we'll never see accurate signalling of information quality by price." I didn't mean prices don't carry information. Of course they do. What I meant was, this didn't hold for information goods.

I'll accept that "media goods are not strictly info goods and why the marginal cost is not zero." but I simply don't believe that this makes enough difference to save the idea that prices can tell us much about information quality.

More damningly, I can't think of a single information product that I buy : eg. music, literature, technical information, art, software etc. where I regard price as a good predictor of value. Faced with two novels or two CDs, and knowing that one is twice the price of the other, I have no idea which is going to be better.

That's partly because quality is so subjective, but also because price is *all* about kinds of institution and business models that are used; and this *obscures* any other information.

One piece of software can be sold for $50,000, and from that I predict it's made by a large software company with an army of salesmen. Another piece of software is free from SourceForge.

Or two companies may have strategies for spreading the costs differently : X-corp spent 100,000 dollars developing their package. And plans to sell 100 copies at a $1000 each. Y-corp spent the same, but plans to sell 10,000 copies at $10 each.

The same development cost is spread differently. Sure price reflects market structure and company strategy, but in doing so, loses the other information it might contain.

Price certainly can't tell me anything about which of these pieces of software is likely to be more useful in my company, which is likely to have more bugs or the best architecture.