Shamus Young, over at the Escapist, has written about the general failings of game economies. It's a short and slightly tongue-in-cheek look at why economies fail: i.e. it's the loot generated through a player's actions. Reading over at his blog and through the comments on the article there are a few camps:
These people want to include aspects of reality such as stricter carry limits and smaller, less frequent combats. They also want to devalue products which are more numerous - which would lead them to become ridiculously cheap and would thus mean that blacksmiths etc are run out of business by the player.
These people want to reduce the impact of selling loot by reducing the amount that vendors are willing to pay and/or increasing the cost of new items sold by vendors.
My opinion lies with a (seemingly small) third camp who are technically ultra-realists in the economic sense. i.e. The games would work as they do currently, with unrealistically large encumbrance levels and item storage etc. However, what would happen is that the world would act in a realistic way. This is also something i've thought about before, specifically regarding to MMO gameplay.
First off, there's a limited number of people but new people can be born and thus grow into adults and populate areas. These people will be free to join whatever faction they wish, whether they be bandits, bad guy minions, traders, townspeople or guards etc. Though this would have to come with some limits such as "tendencies" to go into a particular profession/lifestyle (e.g. gentry would be more likely to remain gentry or become knights in a medieval setting). To compensate for unlimited population grow you could, upon birth assign a person fertility % (i.e. their chance to be able to procreate) and, on top of this, you would have a population vs food check with each area having an upper limit of food production. If the population rose above this limit food would have to either be bought in from another area (trade) or people would starve and die (you could choose whether children or old adults would be more likely to die first with people in their teens and twenties the least likely to suffer from starvation).
Fewer farmers/hunters in the area would mean less food and thus a smaller population and if the player or bandits rode into down and killed all the food-makers then the rest of the population would suffer unless some people switched roles to start producing food.
This style of thinking would also be reflected in the items of the world. Each item would require the base materials, the right profession (or professions) to be present and time to produce. A sword, for example, might require 8kg iron ore, 5kg ashes, 1kg leather, a blacksmith, a tanner and two days to make. A better sword might require multiple blacksmiths or more of one material or additional materials. This allows the player to be more involved in the game world because they can commission, bring materials or products to different places to enable the creation of items and help with the local economy. Of course, the initial number of each item in the world would be limited and, to stop items becoming limitless and thus devalued, they would degrade with use and time, depending on the item.
For example, a cloth tunic might cost 13 coins, would take 2 weeks to produce and degrade within a year of use. If it was not in use then it would last 5 years (damn those moths!) and to account for use and disuse some simple formula could be implemented to say 1 day use = x% of 1 day non-use (and then divide that by the item's total "lifespan"). This mechanic would also allow a repair trade/economy to flourish though you'd also have to limit the effectiveness of each time the item is repaired, depending on the item type. e.g. Repair 1 would refresh 50% of the item's lifespan, repair 2 would refresh 30% etc. Until the cost to benefit ratio decreased to the point where it would be more cost effective to buy a new item. Item effectiveness could also be tied into it's repair/lifespan stats so that an old rusty/heavily used sword would be less likely to do heavy cutting damage than a brand new shiny sword.
The end result of these mechanics is that the player is unlikely to completely destroy any economy because any influx of items into the market would require them to be removed from the market elsewhere first. Plus, no vendor would be likely to have that much money to buy them - or want to buy them. Why would a blacksmith want to buy a sword for very much money when they've already got 20 in the shop? He might pay for the raw materials but that's not likely to be a large sum.
Of course, i've got my own ideas for magic items and telling that here would just be giving all of my secrets away :p