The counterpart to this ("why bother?") would logically be some sort of "hassle and conversion loss" penalty. Official same-civ coinage would have a value multiplier of 1.00x for all trading purposes. Other good coins perhaps 0.98x, precious metal bars and gems 0.95x, and most trade goods a lesser fraction; one interesting option would be to have it based on a table or equation keyed off of (value / weight), starting at say 0.90x and going down. This would simulate the transport hassle with a simple mechanic, and encourage forts to trade smaller, more valuable items naturally.
And by your proposal, like an officer appointing in mintage, I believed you are talking about a top-down approach in order to make the standardized and generalized coinage across an entire civilization possible. Hence I believe there must be other monetary and financial structures working behind it. I think invoking the infrastructures from the past history, indeed may help the design of economy system in DF. but it needs to be more concrete in order to translate it into DF and function properly.
However I don't get the part of multiplier. Are you talking about the property of transaction cost during exchange using commodity money? It's exactly the difference of transaction cost make precious metal a good candidate for the role of money (one of the other money characteristic). And there's many economics theories and ACE models built to test the emergence of money through this property. So, on top of that, you are trying to make even more difference in transaction cost into different coinages(money)? I guess I can make a model and test this theory. I imagine that it won't work as good as it hope for. A 2% or 5% difference in transaction cost isn't going to be large enough to make the separation in medium of exchange on different markets trading with each other. (Common medium doesn't need to be the best kind, but one of the better kinds which exchanged with largest quantity. Hence it's likely the one with largest quantity will eventually dominate the market as in history). Still, we only see the role of money, and lack the financial infrastructure that will allow money to flow, and controlling the speed of transactions. (If it's a top-down approach, and you know MV=PQ, right?)
I was attempting to come up with an easy to compute simulation mechanism to represent a wide variety of "messy" real-world considerations that are probably not going to be practical to model in detail; that could additionally be "patched in" to the existing system with an easy transition mechanic.
Right now, all items are treated as having an intrinsic "worth", and almost all trade interactions are carried out using that worth without any consideration for actual value or, particularly, overhead. (I will attempt to use "worth" for the game-computed internal abstract price of an object or service in this discussion, before any external modifiers; "value" would be after multipliers for being requested, etc.) The only limiting factor is that in certain versions, there is a maximum weight of goods that the traders can carry away with them, which assigns in effect a very weak and single-stepped value/weight filter. (If your primary trade good for the first trade caravan is ordinary obsidian mechanisms, you may find that the traders cannot carry away enough of your worth for you to buy everything you want and could theoretically afford of theirs, for example.) There is little need for coinage, as almost all items are traded with equal ease and at their full worth.
Realistically, a supply caravan would want to come away primarily with easily-spendable money, and other items in decreasing order of some combination of fungibility and liquidity. Recognized foreign nation coins would probably be the next tier, followed by bullion bars and gemstones still considered at some level as "money", just less efficient forms in small quantities. (Historically larger amounts were simply by weight; if you have to pay 10,000 Roman pounds of silver as
Danegeld, it probably doesn't matter nearly as much what shape it's in.)
There's a fuzzy boundary, but somewhere you probably cross from "money or close equivalents" to "trade good". The level of additional hassle necessary to convert a trade good into money when the caravan returns home (or gets to its next destination) is difficult to model. In the interests of having a first-cut approximation, I suggest looking at what you might call "specific worth", the worth of the item divided by its weight. The heavier an item is, the more hassle it is, and the caravan should consider it having a value which is less and less of its worth as it gets heavier. If set with care, this single, easily computable number could represent a wide variety of real-world economic factors and generate both a better feeling and better operating simulation for very little computational cost.
It is easily extendable by biasing the specific worth tables based on the distance from the caravan civ to the fortress, to represent the additional overhead of long journeys; further refinement might be by assigning a variable weight to tiles crossed based on their terrain and savagry, so a caravan that has to cross many squares of haunted jungle and a difficult mountain pass would have a stronger preference for easily-portable, high-valueable, non-spoiling items (and once this has been calculated, you could also use it to increase the expected profit).
Another related approximation would be to introduce a (fractional) multiplier to represent spoilage / upkeep; for foods, live animals, etc. This might help offset the "prepared meals" export problem somewhat, and would have useful synergy with other effects.
The current equation for trading with a caravan, abstracted, looks something like this (trader on the left side, fortress on the right):
worth * request * profit = worth * request
What I'm proposing is something along the lines of the following:
worth * request * efficiency * upkeep * profit = worth * request * efficiency * 1/upkeep
where:
worth = current internally computed DF "true" worth, or its eventual replacement
request = multiplier based on whether an item has been specially requested and at what priority, otherwise 1
efficiency = multiplier based on how easily traded, carried, and exchanged the good is; modifiers close to 1 for money, and dropping with increasing weight for non-monetary trade goods; may be itself modified by distance or weighted distance
upkeep = multiplier based on an abstracted idea of how perishable a trade good is, or how much upkeep is required e.g. for live animals. Inert goods at 1, other items less; a simple table might start with preserved food at 1 (reflecting that the caravan can eat it themselves in an emergency), perishable food at 1.2,x live herbivores & omnivores at 1.5x, carnivores at 2.0x, and fish at 3.0x (you have to carry all the water as well); may be itself modified by distance or weighted distance. This could also be used to handle fragile items which require extra care in transit and extra packaging; for a somewhat arbitrary example, you could assign green glass a 1.2 modifier and crystal glass a 1.5, for instance. From the caravan side, these additional costs / inefficiencies / overhead have *already* been expended, so they effectively make things more expensive; for the fortress side, they *will have to be* expended, so they effectively reduce the value of relevant trade goods.
profit = the profit ratio the caravan is expecting