Economic Model
 

You can't have a civilization without an economy! And Clash will provide an extremely detailed and realistic one. The powerful AI will handle all the details or it can be micromanaged to the nth degree. Your choice! 

• Team Lead: Mark Everson 
• Coder: Mark Everson
• Last Update: 01/21/2002 
• Current Forum Thread: Click here
• Related Model Thread: Merchant Agents
• Older Related Threads:Proposal for a Complete Infrastructure Model, Demo 5 Econ Model, Government - Economic models interconnections

12/17/00 Revision
Added Glossary & Production Function Sections, many small corrections
 

The economic progress of civilizations is a critical aspect in Clash.  It is perhaps the most important factor in the modern era. The most important things I'm trying to implement in Clash are:

Reduce Micro-Management by the Player  -  Allow the player to handle the economy in the game with as few actions as possible.  Implementing an economic strategy should Never require the player's attention at the level of a single map square.  Most Civilization-type games also require too much attention at the city or provincial level.  This typically involves the player commissioning the construction of a variety of "buildings" that affect the economy in various ways.  Clash won't have explicit buildings, except in the graphics.  Infrastructure for the economy will be handled in a continuous fashion in a way that is relatively hands-off for the player. Various levels of micro-management of the economy should be Available to the player If He or She Desires Them.

Get the correct flavor of growth rate and innovation effects of economies operating under the influences of Market, Traditional,  and Command forces (either Government Command like communist/totalitarian, or regulation)

Have a distinction between public and private sector; the government can influence private sector economic behavior but does not control it (having the public sector control itself will also reduce micro-management)

Use relatively realistic consumption patterns and growth rates for ancient and modern economies

Simulate the effects of rare or important goods and products - e.g. that of coal in the industrial revolution (special goods give the player new possibilities for cooperation and competition with other civs: enabling important trade pacts, struggles for trade dominance, and embargos in Clash)

Capture the importance of trade in stimulating economic growth.  Clash will use quasi-market treatment of pricing.  This supports relatively realistic functioning of merchant activities. Merchants will be agents that "make a living" buying commodities where they are priced cheaply and selling them or they are worth more.  Merchants will be intelligent agents, and not under player control.  Players can influence merchant activity thru taxation of trade.

The infrastructure system in Clash will be one of the most important ways that the player can guide their civilization.  It will allow the player to emphasize, above other priorities, religion, education, or a number of other things.  I think all this can be handled on a civilization-wide basis, avoiding the need, as in civ2, of every city/province explicitly having its own library, marketplace, temple, etc.  The Player will, if desired, also be able to "drill down" and  fine-tune infrastructure on a per-province basis.

Contributors:  Mark Everson, Hrafnkell Oskarsson, Dominique, Glak, Xiane, Druid2

CC (or C) -- Clash Cash, Monetary unit in Clash

PCI -- Per-Capita Income (Income per person)

Special(s) -- Specials are scarce "strategic" resources that are found locally (tin, oil), or specialty finished goods that can best be produced where the primary resources are, like wine

Before getting into the detailed treatment some particulars about the simple economic model used in the game need to be covered.

1)  Five sectors of the economy are handled explicitly in Clash:

Food (primarily agriculture, but also includes herding and fishing)

Resources (natural materials that are production inputs)

Production (finished goods created from input Resources)

Service Sector (Service Sector covers usual services and transportation)

Special (Specials are scarce "strategic" resources that are found locally, or specialty finished goods that can best be produced where the primary resources are, like wine)

3) The main economic action takes place in provinces, which consist of groups of map squares. [We may change this to the individual map square level if practical.]  A capital (usually city) square is the economic and administrative center of the province.  All the production that takes place in the province is pooled in the provincial capital to simplify the economic system in the game. The size of a province is somewhat fluid, and depends on the technology level, and the distance of squares from the capital.  The capital in a province is where military units are recruited, prices for goods are determined, and merchant transactions take place.

4)  I am assuming for the moment that productive capacities (and so real per-capita income, PCI) change by about a factor of 50 throughout the game.  PCI starts out at 5 units (using $100 per unit gives a $500 modern equivalent) which is basically subsistence.  The progress throughout history breaks down so that the levels of real per-capita income associated with different historical time periods in the developed world (of the time) are about:
 

Giving a rough PCI of $25,000 for developed countries now.

Over the last 200+ years, there has been crudely a doubling in PCI every 50 years or so in leading-edge economies.

The people's relative desires for food, produced goods, and services will be governed by a simple utility function.  For low levels of PCI the people want mainly food for subsistence.  As PCI grows they will demand more goods (production) and some services.  As PCI levels approach modern values services will out-strip production as the most-desired sector.  The people per se have no demand for resources, but they are needed to produce the goods that they do want.

Special resources, up to a point, have a higher value than normal resources as production inputs.  The value of Special resources and the amount that can be used (if any) depends upon the technology being used in production.  Each milieu has several different sorts of Special associated with it.  Some specials, like gold, silver, and gemstones produce output directly in units of currency.
 

Although the province is the most important economic unit in Clash, a lot happens in individual map squares, which I will focus on first.  A map square in Clash is pretty large, about 60 miles (100 kilometers) on a side.  So even at relatively primitive levels of technology it can support a population of order tens of thousands of people.  In order to simplify the game graphically we'll treat 1,000 people as the base unit of population and call it a "head".  Each map square is endowed with a number of sites that economic activity can take place on.  For instance, farming takes place on food sites.  Exactly how much food is produced on those 10 sites depends on the people farming, the capital invested in the farms, and the technology level of the investments.  The exact amount is given by the "production function", which is discussed in detail in its own section below.  But the player doesn't need to understand the mechanics, since the results should be intuitive.  Those wanting more information can go to the section on the Production Function.

As mentioned above, sites come in three types, food (F), production (P), and special (Sp). Food, production and special commodity sites are generally limited.  There is no limit on the number of normal services sites in a square. At the beginning of the game, a square that has good farmland, a relatively small abundance of resources for production, and no special commodities that can be produced might have sites like 8F, 2R, 0Sp. This means there is enough arable land for 8,000 people to farm effectively (more are possible, but don't produce as much food per person), enough natural resources to keep 2,000 people busy, no special commodities available to produce, and (as always) an unlimited number of production and services sites.  The reason for the unlimited number of production sites is that, with some historical exceptions, production can take place anywhere that has the necessary infrastructure.  The origin of the unlimited number of services sites is that these occupations (doctors, housekeepers, clerical workers, etc.) don't require inputs of land or natural resources in the same way that farmers do.

At the subsistence level each head on a site produces about five units of output. One head on a farm site will produce five units of food.  The main reason I picked five as the base level of production is that if you multiply it by $100 (or 100 Euros or 10,000 yen) can get a feeling for the per capita income of your people in modern currency. Production sites produce a net of five units since they take in 5 units of resources and make 10 units of finished goods from the input.

As new technologies or techniques become available, either the people or the government can invest in improving the productivity of each site.  For instance if farmland is irrigable an investment can be made to irrigate it, improving its productivity to 6 (with one farmer per site), or even higher depending upon the climate.  With all the technology and investment the modern economy has its disposal, a single developed farm site might be able to yield 500 units (Don't quote me on the exact number, but this is the right order of magnitude).

To illustrate how things work let's look at the square mentioned above that has 8 farm sites, and 2 resource sites.  At subsistence level each head requires 4 units of food.  So if all 8 farm sites are worked, the maximum population sustainable is 10 (8 sites x 5 per site / 4 per head).  The peoples' relative needs and wants in terms of food, production, and services is described by a utility function.  Let's not go into that here; suffice it to say that what the people might want beyond the subsistence amount of food might be 1/2 unit of production, and 1/2 unit of services each.  The production might be used to create housing among other things, and the services might be those of village elders, priests, etc.  At this point the population is self sustaining since there is enough food for 10 heads, and there are 8 farmers, 1 on resources / production (technically there would be 1/2 head each on resources and production), and 1 on services.

Now let's suppose that an irrigation technique is discovered that allows the productivity of this square's farmland to be increased to 6 per site after the appropriate investments are made.  If the people and/or the government decides to make these investments the 8 farm sites could yield 48 (8 sites x 6 per site) units of food.  This means it would be possible to support 12 heads on the land at subsistence amounts of food.  So one solution might be to keep food at subsistence, and increase workers on resources, production and services.  Alternatively, the peoples' utility function might indicate that they would rather have somewhat more than 4 units of food per head and less production and services and try to maintain an equilibrium population of 11 heads.  (The alternative with more food per head would also give a healthier population due to better nutrition.)  The choice between the two alternatives is fairly complicated, but it will not be the player's decision.  (Unless the player has dictatorial powers to order the people on what exactly to do!)   The outcome will generally depend on the desires of the people.  The player can also influence this decision through, for instance, tax policy.  I will discuss taxes later in the context of the provincial economy.

Improvements in the productive capacity of sites will be done on a square-by-square basis.  The reason I want to keep track of production capacity, for instance, on a square-by-square basis is for military purposes. The presence of enemy troops on a province's soil gives them a way to hurt the province's economy through extensive pillaging of developed sites.  This will also be the case for things like roads and fortifications.  All other kinds of physical and institutional infrastructure will be handled at the provincial level or higher.  [We are considering doing this at the square level and will decide what we like best from playtesting.]  Physical infrastructure consists of things like healthcare infrastructure (including aqueducts and sewer systems...), transportation infrastructure, housing, and defense/military infrastructure.  Institutional infrastructure covers educational and religious institutions, as well as banking systems etc.

 Provinces in Clash will consist of a group of map squares and one or more cities.  One aim I have for Clash is to limit the number of sub-divisions of the civilization that the players must deal with.  The goal is to have the player deal with 10 or less economic sub-divisions (although whether we can meet this goal remains to be seen).  Additionally, the interface should allow one to change the characteristics of a number of sub-divisions at once, even when the player is in micro-management mode.  Provinces group a large number of squares, usually 20 or more, into a single economic unit.  Production from all the squares in a province is pooled, and the economic development of the province is handled in a uniform fashion.  Note that to be historically accurate, the size of economic units in the game should be of order one or a few squares for times prior to the modern era.  My view is that this piece of historical accuracy carries too large a price in terms of the playability, so I'm inclined to try the size above and see how it works.

Cities are the administrative and trade centers of provinces.  Because of these important functions, the population density in cities will tend to be much larger than that in the surrounding countryside.  Cities present both opportunities and problems of their own, due primarily to their increased population density.  Squares designated by the player (or the people themselves, TBD) as cities grow faster (~50%) in population and development than the countryside, and the administrative center of the province (capital) grows 2x as fast.  This factor represents some migration within the province.  However city growth will slow, or even stop when the city/urban squares have insufficient  water / health infrastructure to support their population.  For now, I will focus on how of the province as a whole functions, omitting those features which are specific to cities.  Also, the handle city is used loosely here.  It means a place that is an administrative center, or has significantly larger population density than average, or a key trading location.  In the modern era there will really be cities all over the place, but only those that are especially large will be treated as cities for game purposes.
 One element I want to add at the square level that relates to cities, is a bonus for the amount of production that takes place in a square.  In the transition during and after the industrial revolution, when supply chains in production became possible, concentrating production into a small area like a city could provide big benefits.  This type of effect was generally only very small before the industrial revolution.  To capture the importance of cities as production centers during this period, concentrations of productive capacity will get a bonus.  For a really big city around 1900 my guess is the production bonus should be of order 25%.  Anyone with a better guess is encouraged to state it ;-).  I don't have the details yet, but this effect will push city growth in the way that happened in the real world.  As the economy beomes even more developed, the bonus will gradually shift to being derived from the production in the whole province.  Suggestions are welcome as to how to fit cities into the rest of the world model structure in Clash in this or other ways.

Provinces are the primary economic unit in Clash. The main criterion for whether a square can be part of province or not, is the ease of transportation from that square to the central city of the province.  Right now I'm working with the notion that any square that is less than one movement point from the capital city (or could be with road construction) qualifies for membership in the province. Using a flat-terrain movement rate with roads of approximately 1/4 movement point, a province could be as many as 40+ squares once the appropriate roads have been put in.  If the provincial capital were on a navigable river, or sea, this number could be even larger.  In addition, a limited number of " inconvenient" squares can be added to province, just to make the game play smoother.  By inconvenient squares, I mean ones that are just outside the provincial boundaries and don't fit naturally into another province either.  As technology progresses opportunities will exist for making provinces larger than at the start of the game.  In contemporary times I think a province could be very large.  For instance the entire continental United States might be spanned by as few as six provinces.  The player will be the judge of whether he or she wants to increase the province size as the game progresses.  Squares with better transportation connections (roads, canals, etc) to the administrative city of the province will get small bonuses to their economic productivity in all the sectors.  The quality of roads affects the bonus.  The center gets the bonus for the type of roads it has.  The bonuses I have in mind is to average the road quality to the adminstrative city using the following values ( track, +1%; dirt road, +3%; paved road, +5%; modern highway or RR, +10%).  [Note that it would be better to use the effective movement rate between a square and the admin. center to figure the bonus.  This would give larger bonuses to closer-in, which is realistic.  However since we do not want to encourage players to have large numbers of small provinces I don't want to use the more realistic method.]
 One important goal of the province system I've sketched out above is to support a pricing mechanism for goods and services.  The people, the government, and merchants will form a simple market in each province. Running a model market in each province has some serious computational overhead associated with it, but it also has some advantages.  A market with reasonable pricing provides "free" AI with respect to the people's economic desires.  When there is a choice of which sites to upgrade, or where to put a new head to work, a correct market price gives the model economy all the information needed to make the right decision for the people. Correct pricing also tells the government the cheapest place to make, for example, weapons or munitions.  The market also allows for including in Clash important, and fun to play, historical occurrences such as economic competition between states in the area of strategic goods and embargoes.

 For most purposes, provinces function as a kind of "super" map square with a huge number of economic sites.  The squares in the province will make food, resources, production, services, and specials independently, but all the produced quantities will be pooled together to become the provincial output [may not be the case if we handle econ at a square level].  Any specials that can be used locally have their effects on the economy immediately (see below).  Any surplus specials are available for trading. Once this gross production takes place taxes are levied by the government.  The government takes a percentage, the tax rate, of the generated food, resources, production, and services.  Any surplus of each of these commodities that the government does not spend locally can be converted into money for the treasury.  If the government does use the commodities locally, they can be given directly back to the people (but perhaps redistributed, as in a socialist system), or used by the government to develop sites or build infrastructure. Specials are taxed later, using a different system.

After taxes are collected the people know what they have left, and what they would prefer to have, from the utility function.  At this point merchants and government agents propose trades to the people.  Production sites also bid for resources at this stage.  For instance, if the province produces extra tin (an important input to an ancient economy, since it is needed to make bronze) merchants serving areas without the important metal will try to obtain it here where it's available, and bring it to where it is in demand.  Most often the merchant would offer some other special, for example wine, in trade.  With good transportation, trade in bulk commodities like resources, food and goods also is feasible.  Once all the trades are proposed by all the merchants and agents present, the people consummate whichever deals they like, starting with the best first.  Trades that happen are taxed by the government.  There can be different tax rates for internal vs. external trade, etc.  The government can also, if it has sufficient power, declare various commodities to be state monopolies.  At the end of this activity the people and government build whatever infrastructure etc. they would like, the people consume the remainder, and the economic happiness for the people is determined.  The economic happiness is essentially the ratio of what they were left with to the value of what they originally produced modified for merchant transactions and government givebacks.  When the government gives back to the people, they only count the contribution at 80% of its face value, because what the government gives them is usually not exactly what they want.  This connection will help to reproduce the slower growth rates in socialist economies.

To simplify the economic system I propose that the price of services should always be set at 1 currency unit. (contest: who can think of a good name for the currency unit in Clash?) For now I will simply refer to the currency unit as the CC for Clash Cash or simply C to save typing.  Among the three sectors in Clash's simple economy, services are the one that is least tied to the land.  Also, some services, like artworks are much easier to transport per unit value than goods in the other sectors.  I could raise several other arguments, but I would rather hear what people have to say about this.  If I don't use services as the unit of exchange, I'm not sure how to set the unit of exchange for a simple model economy like exists in Clash.  Opinions and suggestions welcome.
 The equivalence between services and money allows the government to take its taxes out of the province in units of currency.  All that it needs to do is to trade the food and production part of its take for services.  This trading will have consequences in terms of structure of the economy.  If "cash taxation" goes on for a while, eventually the people will make more services than they themselves would want in order to have enough excess to pass onto the government.

 Before moving onto the provincial economy, I need to mention how special commodities work.  Specials are scarce "strategic" resources that are found locally (tin, iron, coal), or specialty finished goods that can best be produced where the primary resources are (ceramics, wine, cloth).  I think that the way I've come up with for handling specials will introduce important real-world effects of "strategic" resources into Clash, without the awesome amounts of micro-management needed to achieve these effects in games like Colonization.  Adding a special commodity to an economy that is without it will make it significantly more productive.  This productivity increase applies until a point of diminishing returns is reached.  The point of diminishing returns usually depends on the size of the provincial economy.  The fact that specials will usually be produced in amounts that are much greater than the local economy can use provides a stimulus for merchant activity.  When it is worth their while, merchants will deliver specials to places were they are desired.  I'll discuss this further in the section on province economy.

A square that has a special commodity associated with it will have a number of sites for it. If the local economy can use the special workers will generally go to work the special site before going to others because of the value of its products. If  an excess of the special is produced, it gets recycled back into the fundamental commodity type it represents.  So for instance if excess wine were produced, it would just count as services (= entertainment), since this is the production type wine turns into.  ( This game mechanism will cover the real-world case that the people would only produce as much of the special as they could either use themselves, or sell to others. The technology level for the special's extraction depends on the particular special.

As mentioned above, special commodities are things that are important economic inputs.  Without these the economy will perform under its potential, and grow slowly.  With a good supply of the critical strategic commodities of a given technological level, the provincial economy will perform significantly better than expected for that age.  All specials increase the production of one sector of the economy.  For instance, iron and coal enhance the production sector when the technology to use them properly is available.  Other examples include the spice or salt specials, which augment the food portion of the economy, and wine that enhances the service sector. The number of units of a special that can be used in a province's economy is 5% of that sector's production. (This number will probably vary by special, this is just for purposes of example)  So if the production sector output for a province is 100, up to five units of the special iron could be used.  Each special has a multiplier associated with it to convert it into output units.  For instance, iron might have a multiplier of 4.  If our province with production output of 100 can get its hands on 5 units of iron, the production output bonus would be 20 (5 units x 4 per unit).  So the final production output with the iron would be 120 units.  The final thing that specifies a special is how hard it is to transport.  The merchants trying to make money by delivering specials will know this number, the player won't be directly involved.  The function of specials in Clash will require some fine-tuning, but I think the mechanism I've outlined here will introduce some fun game-play in terms of trade between civs that requires little effort to manage on the player's part.

 Development takes place in one of two ways.  Either the people do it, or the government does. If either party wants to invest in a particular sector, for instance food, they can do so provided that at least some sites are not "maxed-out" with respect to the current technology.  If improvement is allowed, whoever wants to do it must pay an upgrade price to improve the site.  For instance, consider a food site that is producing at the subsistence level of 5.  Suppose the land in that square is in principle irrigable to a food production level of 7.  The cost to irrigate one site one step might be 4P/16S meaning that 4 points of production (tools) and 16 points of services (labor) are needed for the job.  Once the resources are committed the upgrade, in this case to 6 food, happens.  For this particular site the upgrading process could be done again to get up to a production level of 7.  During the industrial revolution some advances might allow the productivity of production sites to go from 20 up to 40.  The improvements would be done in steps similar to the irrigation example above.  So going from productivity 20 to 21 might require 10P/10S, as would all subsequent upgrades until getting to 40.
 Some technologies also allow for an increase in number of sites in a square.  One example of this is the technology of drainage, which allows re-claiming of swamp land.  In such a case there is a cost to bring a site online if it is one of the new ones available with that technology.  Once the new site is activated in that particular square it behaves just like any other.  Improvements paid for by the people will be distributed, prorated by population, among the squares in a province.  I don't know exactly how to handle this yet, but don't think the distribution of improvements among the map squares of a province is much of a problem.

The cultural profile of the people includes how investment-oriented they are.  Among their desires is that to invest to improve their, and their children's lives.  The people will invest in a variety of things, enhancing sites as above, some infrastructure like housing and roads, and at higher levels of PCI things like education and health-care.  Depending on the cultural profile a certain portion of their consumption is reserved for these investments.  Let's take as an example the square being irrigated in the example above.  Suppose that the people in the province would most like to increase the food supply, and that the square in question is the only one that is irrigable.  At this near-subsistence level the people's capacity to invest might be 1% of their net income.  Further, let's assume their net income's value is 400C.  This amount of net income might come from 100 heads.  So the people might save 4C per turn to purchase the improvement.  Once they have saved the upgrade price of 4P (Say 4C, although it might be different) + 16S (16C) = 20C they will have bought the improvement.
 Another aspect of development is where to put new heads.  As the population grows new people will sensibly elect to go into a line of work that makes them the most money!  The profession (sector) making the most money is usually the one with output having the highest price.  If all those sites are taken, the new head can either settle for second best, or try to migrate to someplace they think is better.  Sometimes the local economy is so screwed up that even the best choice is lousy.  This will drive emigration if the people know about someplace better.  A corollary to this is that a ruler should not jack up the tax rate too high, or people will spontaneously try to "run for the hills".
 If the people are not moving fast enough in terms of development as far as the government is concerned, the government can expend resources to accelerate things.  I think it will be common in the game to subsidize, to some extent, the people's improvement of sites.  For instance a 20% subsidy on irrigation would cause the people (or local authorities) to irrigate more and faster than they otherwise would.  The government can also directly build site improvements itself by paying for the whole thing.  This is the way Clash handles things like five-year plans in planned economies.  My impression at the moment is that this is a very expensive way to do things, and it will be a lot of fun to see how it works out in the game.  About the only case where I can see government purchases of production capacity paying off is where the government has declared a monopoly on a special good.  Government monopoly of things like iron and salt production were common in China at least throughout the first millennium, if I'm remembering correctly.  One thing to keep in mind, is that if the government helps the people to build more production capacity than they want, the people will Never build any production capacity of their own.  Only if the government can effectively use the surplus production will be economy be "balanced".

 Finally, I'm sure the military types are asking "But how do I build an armored division?"  Well, I think an armored division, as well as many other goodies, will simply have a cost in terms of the three sector outputs, just like irrigating a farm site did.  For all the hardware in a modern armored division you'd probably need something like 2 million production points.  I don't know off-hand what the hardware requirements for a Roman Legion are, but they might be something like 50P and 10S. I Really don't want to get into a production system where you need x units of iron and y units of leather to "build" a legion.  My experience with systems of that type is that they become an  "accounting game" fairly quickly.  We'll probably have some spirited discussions on this topic ;-)
 

Note:  This writeup just contains the outlines of how infrastructure works in Clash.  There is a more detailed proposal by Axi that we haven't finished discussing, but that goes beyond what is written here.  If you are interested, consult the more detailed infrastructure proposal.

The infrastructure system in Clash will be one of the most important ways that the player can guide their civilization.  It will allow the player to emphasize, above other priorities, religion, education, or a number of other things.  I think all this can be handled generally on a civilization-wide basis, avoiding the need, as in civ2, of every city/province explicitly having its own library, marketplace, temple, etc.  The Player will, if desired, also be able to "drill down" and  fine-tune infrastructure on a per-province basis.  Making changes in the infrastructure priorities of the civilization should in fact be so easy that our main problem may be keeping the player actively engaged in this aspect of the game.  This will probably require some sort of feedback to the player about how their particular infrastructure strategy is playing out fairly often.

Depending upon their culture, the people in the civ will desire a variety of goods and services. Among them are housing, entertainment, healthcare, religious institutions, education, and other things.  So it will be possible to have a culture that invests more in education that usual, at the expense of, for instance, buying consumer goods. The player can also influence the people's desires through either subsidizing or taxing some of the people's spending on infrastructure.  In addition, the government, if it has enough power, can simply mandate the amount of spending in a given area.  The people will then receive the government-mandated level of services regardless of what their desire for them is.  However, if done in more than a few selected areas, this is generally a lousy way to run an economy IMO.

In the same way that the people or the government can invest in production capacity for the various economic sectors (food, resources, produced goods, and services), they can also invest in infrastructure of different types.  The investments can be either in physical infrastructure, or institutional infrastructure.  Physical infrastructure consists of things like transportation infrastructure, housing, healthcare facilities, and defense/military infrastructure.  Institutional infrastructure covers educational and religious institutions, as well as banking systems etc.  A list of the infrastructure categories that I've got so far are in the table below.

Each infrastructure class is handled at the provincial level for bookkeeping purposes.  However, typically the player will only have to deal with infrastructure on a civ-wide level.  Each infrastructure class is modeled crudely like a storage container.  A resource is put into the container, and the benefits therein can be drawn upon either immediately or at a later date. The amount of resource in the container also suffers shrinkage for a variety of reasons.  In other words, some amount of constant investment is needed to keep the infrastructure at the same level.  For instance, using education as an example, services (of teachers, and time spent by students) are invested to give an education level.  This education level will provide a variety of economic and other benefits to the civ. However, without future investment this education level will slowly shrink as those educated previously grow older and die, or as what they've learned becomes outdated.  So, in the ancient world where average life expectancy might be 40 years, giving a useful life of education investment of about 30 years, the shrinkage rate for education would be about 3%.  I call education a persistent investment, because the use of its benefits to society does not reduce the education level. The only way the education level is reduced is by shrinkage.

Not all infrastructure classes are persistent.  The opposite type I call consumed.  An example of a consumed type is the "social safety net" class.  This represents charity, like food for the hungry, services for the destitute, etc. The food aspects would be similar to the function of a granary.  When the stored food is used, that much of the level is used up.  These social safety net functions can be fulfilled by the people themselves, governments, or religious institutions, or some mixture of these.

So, to sum up, each infrastructure class has an input, a shrinkage rate, and an output that is either persistent or consumed.  Changes in technology can have effects on this system into different ways.  Better technology can leverage the value of the input, or change the shrinkage rate.  So, for instance, if a civ discovers the alphabet, the services put into the education area would get a bonus, perhaps a 50% bonus.  The "social safety net" shrinkage rate might change due to the discovery of canning, which would make stored food much less susceptible to spoilage.  As healthcare improves, and expected life span rises, the decay rate for education would also, as a consequence, fall.

If we assume that the government isn't intervening in the investment decisions of the people for now, how do the people know what kind of infrastructure they should buy to make them happiest?  The table below summarizes the information I've given so far for the infrastructure classes that I have at the moment (suggestions welcome).  Things aren't completely worked out here, but at least this can serve as a basis for discussion.  The second column shows roughly on average how the people spend money not allocated for food.  So, other things being equal, they would spend 10% of the non-food amount on durable goods.  I've adapted these figures from the research paper, "Fundamental Similarities in Consumer Behavior" by K. Clements and D. Chen, Applied Economics 28(6) 1996 p747. The figures here are obviously somewhat skewed toward representing modern economies.  If we go with this basic system we will also need to figure out the numbers that might apply for an ancient or medieval economy.  Again, the "average non-food % used is Only an average over all cultures and all times.  For a given civ these values will be modified by the technological regieme and also by the culture.  The input column shows how much of which of the basic commodities is necessary to purchase a single unit of the infrastructure class. So, for durable goods, each unit costs one production point.  For religious infrastructure the unit cost is two-thirds services and one-third production point. When you purchase a unit of the infrastructure class its level would increase by one.  When the prices in the province are known you can figure a purchase price for each of the infrastructure types. The column that says "example price" uses one illustrative example from a province that has low availability of finished goods. The remainder of the columns just summarizes information about typical shrinkage rates etc. as discussed above. Don't worry, the player will never have to deal directly with these numbers.