File-Sharing and Double Moral Hazard

Every major label 's setting up an iTunes these days. They're all, in the immortal words of Johnny Cash, 'born to lose, and destined to fail'. Why? The music industry doesn't understand the microeconomics of it's own business. If it did, it would see that it's business model is not just misguided, but broken- because, DRM or not, the implicit contract it signs with listeners is being broken in both directions.

I reached this conclusion because, as I was scoping BoingBoing one day, I read Cory's statement, and it struck me as exactly right. For many people, digital music's more about risk than it is about music itself. Not legal risk - but transactional risk, the kind of risk you take when you buy a used car. Now, this statement has deep economic meaning. I'd like to explain why.

Fundamentally, I'm going to argue that consumers download music, as much to derive extra value from getting something for free, as they do because they want insurance against buying something they didn't want in the first place. File-sharing is as much about risk-sharing as it is about the 'theft' of value. Technological changes have made this possible - but the way the business model of the music industry is at odds with the implicit contract it signs with listeners is what makes it probable.

Here are the basic economics of the music industry: The major record labels assume market risk in exchange for value. They take on the risk of assuming search, development, and distribution costs, in exchange for uncertain profits.

We can also look at this through the lens of contract theory. Contract theory says that principals contract agents to do things they're unable - for whatever reason - to do. In every such transaction, we can say that there are extra costs incurred. Economists call these costs agency costs.

So we can say that labels are agents hired by music listeners - principals - to perform a function they don't have the time to do - find interesting and entertaining musical artists. The problem is that this simple transaction creates massive information asymmetries. There's no monitoring mechanism, so listeners can't see what the labels are doing; conversely, labels can't really tell what listeners' preferences are. Even worse, the principals can't influence the agent unless they can coordinate amongst themselves to do so.

Now, in most real-world markets, information is an issue. Neither side in a transaction is perfectly well-informed about costs and benefits. But in most markets, prices are considered the central economic mechanism of information transmission, because they convey information about future benefits and future risks. This point is intuitive if we think about it: prices reflect the scarcity of a good. Think of the price of blue-chip stock, for example.

But, partly because of massive buyer power (the influence the biggest retailers exert over the record labels), prices in the music business have long since failed to carry any pertinent information. Prices have become, if not fixed, as many suspect, certainly standardized. And this robs consumers of a vital means to gauge how much future value they derive and risk they take when purchasing different music goods. It also robs labels of the ability to really understand consumer preferences.

So this forces listeners to rely even more on the record industry's - the agent's - choices. In this case, the principals are kind of blindly reliant on the agent - they have no mechanism to monitor the agent.

So what if, under such a contract, the interests of the record labels - the agent - diverge from the interests of the listeners - the principal? What if, for business reasons, the labels are more interested in economies of scale, scope, and brand than providing music listeners with music they value?

In an extreme case, the labels might begin to impose agency costs beyond the search costs the listeners are exchanging value for - making transactions with record labels provide negative value for listeners. Conversely, we can say that listeners might find it more efficient to take on their own search costs. And this is what's happened today. Many people are more happy to spend time searching for new music on the net than they are simply buying the goods the industry selects and promotes.

It's traditionally argued that the web reduces search costs. But this argument helps explain a very curious phenomenon: why music today is one of the few markets in which people, are, curiously, willing to pay very high search costs.

So the net actually begins to make it possible for people to pay higher search costs at all. They do so because they replace the agency costs imposed by the music industry - which provide them little value - with their own search costs, which do result in a transaction that provides them value. Before the web, people had little option but to pay the agency costs the music industry demanded.

Economists have a name for problems like this: moral hazard. Moral hazard happens when the actions of an agent can be hidden from a principal, creating agency costs - because the agent is able to shirk, take additional risks, and generally not deliver on his end of the bargain. In this case, the moral hazard is that the record industry, because listeners can't monitor or influence it, can effectively shirk, and choose artists not based on listeners' preferences, but based on business efficiencies. This is effectively what the record industry has been doing - adding massive agency costs that replace the search value it is supposed to provide. It's compounded by the fact that music is an experience good, whose value is not directly knowable to buyers - another fact the music industry has been exploiting.

The way to change the incentives implicit in such a moral hazard-creating contract is straightforward in economic terms - insurance. Insurance provides an incentive for the recording industry to choose only acts listeners value. At the same time, insurance means that consumers don't have to pay agency costs - the costs of the music industry selecting acts no one wants to hear.

But doing so would create a double moral hazard. The second moral hazard is trickier: offering insurance to listeners provides listeners an opportunity to hide their actions from the recording industry. Listeners might take advantage of the insurance, and renege on buying music altogether. If the industry offered consumers the ability to simply return any music they didn't like, consumers might return all the music - even the music they did like, after having copied or consumed it.

But this is exactly what the internet has done - offered music listeners a second moral hazard, in opposition to the first. The net offers a kind of gigantic way to renege on buying music goods produced under moral hazard, and completely eliminate the risk listeners take in buying such risky experience goods.

The point is this: the net offers listeners insurance against the music industry itself. File-sharing isn't simply theft. Rather, file-sharing is risk-sharing - against an industry with the freedom to undertake hidden action in the extreme, and not live up to the contract it has written. Remember, the contract said that labels would assume the risk in exchange for dollars from listeners - so when moral hazard lets labels try and push risk to listeners, is it any surprise that listeners try and minimize it by parceling it out? In fact, we could go even further - saying that file-sharing is a way for principals to punish agents operating under extreme moral hazard, with the hope of bringing the agents incentives into line.

In this sense, we can see that the music industry has played a large part in creating it's own problems, which we can call a massive double moral hazard. Next time, we'll examine how it can begin to solve them.