I've been thinking about game economics, specifically in massively multiplayer games. They suffer from a host of problems, but the situation gets infinitely more confusing and muddy if you begin to talk about large numbers of unique items, like we will certainly see in the relatively near future.
The whole point of MMORPG economics is that everything is fungible. That is, there is no difference between this gold piece and that gold piece, and there is no difference between this Codpiece of Flugorth and the other Codpiece of Flugorth. Demand varies, of course. Over time, space, and between items, demand varies wildly. Today the Codpiece of Flugorth is worth 10,000 gold, tomorrow only 150. When combined with the various other bits of Flugorth, the price becomes ten times that of the pieces individually, because they form a set, but even that set can be considered perfectly identical and interchangeable with other sets of the same components.
This interchangeability is the cornerstone of MMORPG economics, both the good and bad elements. However, it can't last. As time marches on, more and more massively multiplayer games are going to feature more and more customized content. Even if they take the cheap route and allow for only small amounts of customization on top of solid fundamental blueprints, that small customization will change the value in chaotic ways. Even just letting you die your gear various colors will make the price vary wildly, often with the expensive variants costing more than double the cheap variants.
With true variation between objects you can end up with something more like SecondLife. One detail worth noting is that SecondLife allows for very easy mass production, however, so it is possible for many people to have the exact same product. To a large extent this focuses and centralizes the production of goods so that many of the people in SecondLife own the same sets of goods, even though there are technically an unlimited number of unique variations they could use instead.
However, SecondLife's economy is not suitable for other massively multiplayer games (and is arguably not suitable for SecondLife), so their method of mass production and weekly dole isn't one that should be adopted willy-nilly. I personally would prefer to think of a system where all the players tended to create uniques, rather than a few players creating uniques and everyone else buying them.
In such a system, the economics would be very different and would need to be carefully planned. For example, what purpose is there behind trading? Are some uniques flat-out better than others (IE a unique wooden sword versus a unique flaming sword of badassium)?
An economy of uniqueness seems to require a few rather unusual attributes in the rest of the game.
First, there seems to be a need for an unlimited number of "slots". If someone can only equip one weapon, one hat, one suit, then there are only so many unique things they'll want. They may end up with ten or fifteen hats, if they're obsessive, but they'll usually only wear one of perhaps two favorites. To encourage people to gather uniques, it is important that a large number of uniques be simultaneously viable for play without sorting through them every time.
One example of this would be a wardrobe that the character would automatically dress from, picking random (perhaps themed or otherwise fashionable complements) pieces. This would allow you to see a variety of uniques over time, meaning that none of your uniques get forgotten in the back of some closet somewhere.
However, this is only one step in the right general direction. You're still limited by the number of clothes your character can wear. equally important is increasing the number of slots. For example, your players may want to collect houses, NPCs, dance moves, poetry - things we can't even think of. It's important to allow this to (A) exist in multitudes during gameplay and (B) vary such that no uniques get lost in the closet.
Once this basis is set up, we can talk about the actual economy of uniques. Without this kind of revolving, wide-spectrum use, an economy of uniques would simply be a stilted and clumsy normal economy.
One aspect of any economy is how difficult it is to manufacture wealth. Most MMORPGs have various means of manufacturing wealth, but the biggest is through killing monsters. This automatically scales with your level, meaning that you generate much more wealth if you are higher level and killing nastier monsters.
A uniques economy could have the same basic philosophy - perhaps the components of the uniques are collected from the corpses of monsters - but it doesn't really fit the needs very well. The reason for the monster farming is to create a treadmill, but when uniques get involved there are a lot of other ways to suck down player time, and level-based treadmills should be easy to play down without losing players.
A uniques economy could also have the opposite philosophy, where you can build anything you want whenever you want, like in SecondLife. A newbie can theoretically build anything that an experienced player can. However, this also has problems, largely in the proliferation of non-unique uniques (I call them "hello world" uniques).
A middle ground can be reached by allowing users to literally grow their content. Real-world time is spent while their character manufactures or tends a given product. The next iteration(s) of the product naturally descend from that, allowing the users to tweak their products to their own desires, rather than programming them from the ground up.
This middle ground has a lot of advantages, but the biggest is that the amount of uniqueness between player content will be much, much higher. Even just a few days in, newbies will have manufactured suitably unique newbie gear. After a few months, two players will have developed such radically different equipment that trade becomes useful. Crossing two "lines" of unique content could also be fun...
This method of growing content does result in a huge number of "spares", so you need to take some pains to eliminate them. This can be done through making things wear out, or through making them break when you descend from them, or from selling them to NPCs, or any number of other means. However, at some point some players will be producing literally hundreds of times more stuff than they can use, and this is a serious threat to the economy because they will flood the market.
Therefore two things must happen. One is that those players should have some recourse for all their spares (perhaps donating them for prestige). The second is that there must be a difficulty in marketing. It must be hard for a player to mass market goods (or even give goods away to many people).
There are a lot of ways of doing these things, but thiey all have ramifications on the overall nature of the economy. For example, if it is difficult to mass market goods, then the economy is hugely fragmented, which will result in a large number of players who specialize in moving goods from one fragment to another. These brokers will probably use out-of-game channels to organize and, in time, they will flat-out replace your implemented market with third-party market(s) that unite the fragments into one, more smoothly-operating economy. You can inhibit them to some extent, but doing much inhibiting will make the players angry.
It is still important to prevent an in-game smorgasborg of uniques, though. Players routinely encountering lists of hundreds or thousands of unique items will cause problems with swamping. There are various means of dealing with this, too, such as a central market that randomly segments, or having players rigidly separated into shards.
These are some of the difficulties a uniques economy will cause, I think. But I also think it's inevitable. How about you?