Government - Economic models interconnections

[axi]

Warning: To be able to fully understand the issues discussed here, one has to be familiar with both the govt and econ models and to posess both Mark's econ model spreadsheet and my govt-econ spreadsheet. Posession of Rodrigo's govt model spreadsheet is not necessary. If you do not have them, ask me via e-mail and I will send them to you. If there is no problem about this with Mark or Kull, all these model spreadsheets should be available at the website presently.

After comparing the procedures that take place inside the econ model (v.11, for demo 5) with what my computations do, some issues rise. I will adress them along with the explanations I give about my model.

a) Labor allocation: The econ model supposes that 1 head=1 labor, overlooking the fact that people can work longer or shorter hours, if they want to or if they are forced to. In my approach, Labor is still given, but it depends upon the conditions. I use Working Hours, which for each class are regulated by the govt and the Labor per capita depends from WH and the Active Population percentage, which still isn't equal for all classes. The reason for this differentiation is to give more realism and more emphasis in the class distribution, especially in what concerns minorities. Apart from complexity/speed issues and a small need for rescaling, I see no reason why we should not implement this feature, but the Team has still the liberty to decide. CtP uses WH regulation, why shouldn't we?

Nevertheless, even if we went for simplicity, because of the existance of the UC, the available labor would still vary according to economic conditions (more profits>more UC members>less labor). From what I saw in the econ model, once one labor unit is put on a sector, it can't be removed from it and only the addition of more units, due to pop growth, can be regulated. Can negative changes be handled? This reasoning is ok if you think that a farmer cannot be turned into a doctor at will, but what about things like the industrial revolution, when farmers left the country and went to town to become workers? Wouldn't it be better if all or part of the existing workforce was redistributed each turn?

b) Private Property: The govt model defines that PP% of all production (in all sectors) is made by private enterprise and the rest is state-owned. Furthermore the state can impose taxes on the private production (at a rate t%). I followed this approach in my computations, dividing everything into private and public, even investment, along this line. For the time being, the productive machine (the production function) is unique (or we have two machines with the same characteristics), which means A, a, b, c of the production function and the Product Distribution are the same. They could be differentiated to add some more features (f.e. difference in public vs. private wages or difference in productivity) but I do not think this is necessary.

The econ model on the other hand, does not use PP and tries to represent command economies with tax rate. But state-ownership means that the state, as a producer, has to provide the capital and pay the wages. This is entirely different in essence, so I think we can't just trash PP. It's just that Mark's market system becomes a little more complex. This is how I envision this:

1) Production phase: The production function is employed.
2) Tax payment and a first change in prices (as in the econ model).
3) Trading phase, where specials and external trades are computed and prices change again.
4) Distribution phase: Here comes my part (govt-econ) to determine everyone's income, investment and consumption (in CC).
5) Investment phase: Investment in productive infrastructure (kapital) and tech factor recalculation.
6) Consumption phase: The 3-4 agents of the matrix below bid for the FPS to be consumed and the prices get their final values. Parallel to this, investment is made in other infrastructure classes (every expenditure, even food consumption, will be regarded as such a class; that's why these will be parallel). Government intervention will occur at this time also, either by direct redistribution or by price subsidies on the desired infrastructure points.
7) Finally, demographics are computed (Birthrate-Deathrate-Active Population-Education Level-Class Distribution-Labor Reallocation).

Of course one could Merge all this procedure in one "big bazzar", where everything is bought and sold, as per the following matrix.

Agent_______Consumes___Buys_____Sells
External______N/A_______FRPS_____FRPS (optional)
State________N/A_______FRPSL____FRPS
UC__________FPS______FRPSL____FRPS
MC_________FPS_______FRPS_____FRPS (optional)
LC__________FPS_______N/A_______L
Minorities____FPS_______N/A_______L

Of course my understanding is not perfect, so if I am terribly wrong somewhere, please correct me.

c) Product Distribution: For my calculations, this is THE most essential variable. It determines how the product of the combination of Kapital and Labor is distributed among the contributors, who are no other than the classes and the state. The problem is how to define a proper and well-behaved function to do the job. I am still not satisfied with what I currently use, so any contributions would be welcome.

Mark is preoccupied about the UC starving the LC, but that won't be a problem. Because although dividing the product according to the marginal productivity of each factor is the most reasonable thing to do (I hadn't noticed this until now), sound reasoning has never been a priority in a society based on exploitation and expropriation. Until now, the function that I used for product distribution had nothing to do with marginal productivity. PD was determined by administrative power {EP*(PDs-0,5)}, legislative power {SP}, syndicalist power {CR*sy} and priviledge power {PP*20*pr}, in the margin between two starvation limits:
a) The maximum PD, above which the returns would be less than the capital invested, so the capital would actually be reduced (of course if PD=maxPD the UC would immediately starve to death, but then there wouldn't be any UC; the limit is to protect kapital belonging to the state rather than the people).
b) The minimum PD, below which the LC and the Minorities would physically starve, which is self explanatory.
Below is a copy-paste from my spreadsheet.

PD=min+(max-min)*(4+EP*(PDs-0,5)+SP+CR*sy-PP*20*pr)/8
minimum: PD=si*(X(LC)+X(Min))/YT/(1-t)
maximum: PD=1-Ko*D/YT/(1-t)

In extreme situations (extremely low tech, or severe taxation, or scarcity of food), it is possible that maxPD < minPD. Then the land gets so poor that starvation is unavoidable, but even then, the function has some purpose, to show us who starves less. In more prosperous times, the PD can be anything from 0 to 1, so then it is reasonable to include an optimum PD, deriving from marginal productivity and have this as a base for all the above deviations. How should we proceed with this?

d) Class demographics: This is another tricky (and probably buggy) part of my model: the functions that provide the social ascension and descension of people among the classes, according to their investment. Any suggestions are welcome.

My reasoning is based in the assumption that although all members of a class invest part of their income, it is only a part of them (the "winners") that reap all of. So, if the investment is positive (investment exceeds depreciation, or the new kapital held by the class is more than the old; this always happens with the LC) the "winners" get to ascend one class. If it is negative, the "winners" stay in place and the "losers" descend one class. How many they are, it depends from the difference in the kapital held by the two classes in question and by the amount of kapital gained/lost. All these are valid for any class and any number of classes. If you add up all these moving populations, you can have a very nice indication of the inter-class mobility of the society (a high number indicates a land of opportunity; a low one indicates a caste system). This might be useful in some ways, f.e. in the Tech Model.

When I started with this, I had made the (false) assumption that kapital is not persistent (with some depreciation factor) but that it had to be reinvested as a whole each turn. So, if in one turn the UC would decide not to invest anything, it would lose all kapital and descend to join the LC. If the investment was less than the kapital, only part of the UC would do that, while the rest would keep their initial kapital. Now this can't happen suddenly (except if PP is lowered) and the reduction of the UC will follow at most the Depreciation rate.

In the extreme cases, where the income of the UC is smaller than that of the LC, to avoid the effect of the LC refreshing the, otherwise dwindling LC, the "winners" are kept still. That leads us to the question of what happens to the investment of the LC, since, according to the class definition, the LC can hold no kapital. For now and until we find a better way to do this, we suppose that it is handed out to the UC, free of charge. The same happens with the MC (if ther is one); they do not keep their excess investment, because that would break the law that defines them.

As for the Minorities, the do not participate in this. They cannot invest and they cannot move into any other class, except if they are accepted into the majorities by some change in the govt model. I think this is proper for them, otherwise, why discriminate at all?

e) The Middle Class: As I have already commented, it is relatively easy to insert any number of middle classes into the machine, IF you can provide a sound definition for them. The tricky part is that definition and that's what's causing all the bugs on the right side of the spreadsheet (as if the left side is bug-free! ). Currently, the definition of the MC that I am using is being economically independant, possesing the capital to employ their own labor. That means that they will lean ideologically from time to time towards the UC or the LC, while in an earlier approach, their income was strictly 50-50, and so would be their profile.

Because the labor is a datum (it is given), this definition determines the kapital and the profits the they MUST have. So if in the right side of the spreadsheet, you lower PP (say to 0,2), they have more kapital and profits than the whole private sector, forcing the UC into the negative. This happens of course because their demographic proportion doesn't correspond to the type of the economy. I hope that this anomaly can be overlooked if all policy changes are not abrupt but take some time, allowing for the population to readjust. Generally speaking, this is the most rational thing to do; if you want abrupt economic changes (vide revolution), expect at least some starvation to occur (which is the least of all evils right now, but what the heck!).

Another issue about the Middle Class is the type of labor it produces. Disregard my simplistic approach of WH and l, lc is most important: I use the formula lc(MC)=(1+2*E)*lc(mean) which, based on a comment from somebody that the MC represents skilled labor, gives more value to the labor of the MC (according to some education tech) and, eventually, increases the overall income of the MC and it's prevalence against the other two classes. Do you find this implementation good? I don't, I want to make things better. Any suggestions are (desperately) welcome.

f)What is not in focus: Some things in my spreadsheet are implemented for mere functionality and are to be provided more specifically elsewhere. So I use a primitive production function that disregards resources and lumps all sectors together - the econ model should do that. I use an investment consumption function which also disregards sectors and F,P,S, but this has yet to be resolved. I use a simple (also in bulk CC) redistribution function; this is going to be refined and micromanaged after we add infrastructure and consumption classes.
[This message has been edited by a gremlin (edited July 14, 2000).]
[This message has been edited by axi (edited July 15, 2000).]

[Mark_Everson]

Hi Axi:

As I explained to you in my email, I couldn't open your spreadsheet because my version of Excel is too old. But I think I can comment on at least some of your proposed changes...

On work hours...
This sounds reasonable to me. We will need to discuss exactly how it should work. The one point I would like to make before we get too far, is that there should also be diminishing returns for work hours above something like 10 per day.

Private property...
I was actually planning on just having the government be analogous to the UC to handle private property (or more property scratch that or more properly the government's share which is 1- private property). So some proportion of the land and capital would be owned by the government. It would then pay wages, just as the UC would do. Anyway, I would like to hear your and Rodrigo's opinions on just doing it this simple way. Normally the private property ratio would change slowly, but this would easily allow for things like "nationalization" of parts for all of the economy which would change private property discontinuously.

I'm not sure what I think about the "big Bazaar" part of your post. But we already have a moderately large bazaar between the trading the government does and what the merchants do. We will just have to see if having the extra things you propose would make the trading phase unduly complicated.

Product Distribution...
I don't understand the details here, and will wait for the spreadsheet. Given my limited understanding your proposal doesn't look too bad.

I am a little bit nervous that your particular formulas may well have bad behavior over certain ranges of the parameters. But as you say you aren't quite happy with them yet. I also need to think about the ramifications of your proposed changes with respect to writing the economic AI. In the model as I had it, the AI is actually fairly straightforward. My first impression of the sort of things you are trying to do is that it wouldn't be that much harder to do the AI, but I need to see some more details and think a bit more on it.

[axi]

A couple more things, before I'm outa here...

1) Don't let the figures frighten you. This is all the data that could be extracted from the model, not all is necessary for each turn of calc. Many calculations could also be truncated.

2) This model will probably work at a province level, but it can work in a province x sector level too, if we want to complicate things a bit (and add some cool features). This is a decision that must be taken soon though.

3) Private property is currently bound by the govt model. If the govt says 50%, you have as much as private capital has and you can't invest more or less than they do, because that would spoil the balance. This sounds bad, but then you can influence all investment if you want as well as PP itself. As I have said, we can always differentiate the two productive engines, but I am not inclined to do so.

4) Diminishing returns for WH: It can be implemented in the labor per capita formula. And we can have increasing returns in unhappiness for overworking.

5) If you dislike my whole take on the subject, be bold and say so; I won't be offended.

------------------
"In a time of universal deceit, telling the truth is a revolutionary act."
George Orwell

[Mark_Everson]

Axi:

I finally had a chance to look over the spreadsheet. Wow, it's gorgeous! I'm going to give you my overall opinion first, and then I'll do the deeper dive into some of the points you raise. I don't claim to understand in totality your model yet, but I think I get the general idea of what it's doing.

I agree with the overall goals of what you're trying to achieve, but IMO it takes a lot more calculation than it should to get to final answers that are reasonable. Especially if we would break it down by both classes and sectors of the economy, there will be thousands of calculations necessary to run a model like this. Since we agree on overall aims (awaiting Rodrigo's word) I think we can work together on ways to quickly calculate approximate answers to the things that we really need so that we don't have computational overkill. I don't have anything specific right now, but my notion would be to take most of the information that you are inputting, in combination with the incremental return information for the sectors (which are needed anyway for the economic AI) and estimate the effects on each classes income and investments using a few relatively simple expressions. In a total "free market" society that is fair, each class would earn the correct incremental return on its part in the production process. In a biased market society certain classes would derive more or less than this "fair share". Please let me know what you think about this.

Some other points for discussion...

A) Labor Allocation

As I said before I agree this is a good direction in which to go...

"(more profits>more UC members>less labor)" this is assuming a constant social policy. As PCI increases, social policy will generally get more progressive, which can break this correlation.

You asked if workers can also be removed from a sector. Indeed they can, I was just keeping things simple in the demo 5 economic example.

B) Private Property

I'm still inclined to simply treat the government the same way we treat the upper class. An entity that owns things and employs workers. This is certainly the way that the government sector works in most market economies. If the government wants to own the means of production, and all the benefits of labor, then a 100% tax rate plus 100% ownership of other production resources does that. I'm not quite sure what your position is on this. I have reread this section several times, and see you saying we can get rid of private property but I'm not sure what you are replacing it with. Can you try explaining the distinction one more time?

Also, if we did use private property, the number that comes from the government model should just be a moderately strong guideline. I think it would be ridiculous to insist that the number be exactly 50%. So if we do use private property, that number should IMO just be a target. However, if things got too far from that target than the government could take action to bring things more into line. This could go as far as it nationalizing or privatizing an adequate proportion of the economy either gradually or in one big shot.

C) Product Distribution

As I stated above I agree with your general aims, but think the method is way too complicated. I also don't understand why maximum PD causes the "UC immediately starving". However, maybe this is one of the mathematical irregularities you speak of. It is just an artifact of your particular equations, then I don't think you need to explain it to me. I also don't quite get the distinction between social policies and redistribution. It seems to me that one number could adequately summarize the two.

D) Class Demographics

I basically agree with what you have written here except...

quote: As for the Minorities, they do not participate in this. They cannot invest and they cannot move into any other class, except if they are accepted into the majorities by some change in the govt model. I think this is proper for them, otherwise, why discriminate at all?

I very strongly object to this statement. This is what you would expect when the minorities have no rights! If minorities have full rights, then they should have UC and MC proportions fairly similar to the ruling majority. They will still not have political power. But they Will have economic power. It will require very little calculation power to put this extra feature in, and I feel it is absolutely necessary to model a moderately progressive multiethnic society (but one that still has a ruling culture). This means that when the player or their government discriminate against ethnic groups there is a wide range of economic outcomes as well as the wide range (0-10) of political outcomes.


Anyway, those are my ideas as of the moment. Do you happen to know if Rodrigo went on vacation or something? We really need is input before this discussion can go much further...

[roquijad]

I'm not on vacations, just stopped reading for the weekend. Wrong weekend to do so, however... everybody seemed to go posting lots of stuff!!!

I haven't read Axi's spreadsheet yet and I read the posts above paying poor attention. I promise I'll give better comments later. For now:

1) I think too that the State and the UC should behave the same regarding the productive market. It hires people and pays them wages and makes investments. BUT, as I have said many times in other threads in other times, the State has also a very important role producing investment in things the private sector is never interested in or has come to be part of only in the past 50 years or so. The so called "public goods", which include roads, aqueducts, sewer systems, military infrastructure, institutional infrastructure, etc. My biggest concern on the econ topics is having a clearly differenciated public and private behavior for these very important goods. If we achieve this, we'll have IMO a much better modeling for the difference between capitalistic and socialistic regimes and other important effects such as the absence of govt (anarchy/troubled times).

2) The govt model variables Private Property (PP), Economic Planning (EP) and Social Policies (SP) were meant to "inform" the econ model about the econ system the govt wants. In my mind, the main role of these three is to alter the econ model behavior so we can produce different econ systems without using arbitrary penalties or anything like that. I know Axi feels they're sufficient for what we need and so do I. I'd like to know if Mark sees them also as the sufficient set of variables to characterize any type of econ system. If so, this thread should also lead us to how and where exactly in the econ model they make a change.

3) I'd like to insist in the idea of not using the tax rate as a "major" variable to give shape to the economy. As it was planned in the econ model, communist regimes would have a tax rate of 100%, FE. Although this approach is feasible and can give some of the right outcomes, it's IMO more consistent with history and less bizarre to use PP, EP and SP as the major variables defining the econ system. Tax rate(s) should be kept IMO only to affect the private sector, so if PP=0%, tax rate(s) are irrelevant, as they actually are in communism.

4) Just as a bookmark, one of my greatest interests in the govt-econ interactions is how different regimes affect the govt's spending agenda, so FE in democracies the spending in religious infrastructure is much lesser (or null) than in fundamentalism. This was something I tried to solve in my very first version of the govt model, but I couldn't do it in a good way. I hope we can eventually do something on this matter in this thread.

5) Minorities should IMO be treated as LC or, at most, LC and MC. With slavery they should consume less and produce more. I don't see any other problem for this.

6) IMO the econ model must be in essence a free-market system, as it is now. Govt-econ interactions should only be the way in which the govt forces the market to behave "unnaturaly". FE, I think is right to set prices by the marginal productivity, because it reflects the natural tendency of the market. Then the govt affects this as a modifier, using FE the SP variable increasing wages, as if there were minimum wages by law. So, in essence, govt variables should be used as modifiers. This, along with a good modeling for the State's econ behavior, will produce IMO a good modeling for all econ issues.

7) I'm inclined to the idea of having the demographic shares out of the govt setting. My reasoning is given PP, EP and SP (and perhaps UC privileges derived from the political structure), then you should be able to predict how many people will a)live completely from kapital returns (UC); b)live autonomously from their own work and kapital (MC); c)live from labour only (LC). IMO demographic shares cannot be derived from the econ variables such as wages, because most of the most important econ variables will be given by the govt econ setting. After all, FE, we know that if PP=0, EP=100 and SP=100, then you have a pure communist system and therefore UC and MC demographic shares must be zero and LC=100%. So, I don't know how to do it exactly, but I have a strong feeling that demog.shares can be computed only from PP, SP and EP without bizarre effects. This would also solve this part of the problem easily.

[Laurent Nal]

Hi Roquijad,
Sorry I try to participate in that discussions not having read social and govt models. I don't have time and so I concentrate myself only to economical coding.
But I would like to say something about your point 4) here. I lived in France 25 years and for 4 years I'm in Poland so I know two very different ways to considerate religion in Democracies. In France effectivelly the govt doesn't "invest" in religious structures for many time except to promove atheism and religion indifference (which I think are one important form of answer about the great questions of humanity ! In fact for me atheism is not only the decline of religion but another form of spirituality - but it would be better to speak about that in a religion's threat). We call that Laicism, but it allways was a war again religion. Real laicism would be the effort to respect all religions. And it necessite a very big investment from part of the govt.
The Poland example show that democracy can coexist with a dominant religion but only in respect to freedom of consciouses. And in fact very more money is Poland is give FE to religion education than in France. The idea is to help people to live their own convictions in peace, not to make a Theocracy.
I'm not saying that the Polish model is the best or even good, it is building itself. And it is very hard to do it.
What I wan't to say is that you can't simplify in that way: Fundamentalism (Theocracy) -> interest for religion ; Democracy -> no interest for religion.
Perhaps my ideas would complicate a lot but I think it is closer to the reality (past and even more future !).
One more thing very interesting for what you are looking for: In poland where the govt was making war to all of the great religions, people itself invested a lot for religion infrastructures.
The last point (very not in place here, sorry) is very actual : when the great religions are very weak in a civ, even educated people go to sects.
regards
Laurent

[axi]

Points, points, points...

1) Labor allocation: If you want me to explain away the formulas I use, please ask. If you think they are too complex, we cold try to simplify things.

2) Private Property: NO Mark, you have completely misunderstood me. I am not the one who proposes to trash PP, the current econ model does that.

quote: I'm still inclined to simply treat the government the same way we treat the upper class. An entity that owns things and employs workers.

This is what I have in mind too.

As for having PP fixed or floating, considering the scale of the game, having true PP deviate from policy PP would be irrelevant, because nationalisation or privatisation is a much faster process than the differential economic growth responsible for the deviation. In other words we assume all adjustments are immediate, which is much simpler to do too. The question here is if we will let investment be disproportionate to the corresponding property shares (let the player invest more or less than the UC at will). If it is, we have to add a simple privatisation/nationalisation mechanism each turn, so that the Kapital remains balanced. This is irrelevant with the question whether we should differentiate the two productive engines (A, a, b, c, PD), which must still be considered (my vote is negative on this one).

3) Tax Rate and Redistribution: You can have an economy with 0% PP and 10% tax rate but no Redistribution, which means that all are state employees with equal income, but get to decide how to spend their money by themselves. You can also have 0% PP and 100% tax rate, but very high Redistribution, which means that the State decides how the money is spent. If EP=0, the two systems coincide, since the ruler cannot control how Redistribution is done. If the ruler invests according to the people's needs, there will be no discontent. Right now the only difference is investment, since it is based in the personal income, in which redistribution is currently not icluded (Since this sort of income is out of the control of the people, it is added directly to consumption. I must find a way to extract investment for redistributive regimes though; maybe by applying the investment function to the public (per capita) income).

All such regimes can be described as Communism, the question is which one is pure? As far as I know, this has been THE biggest issue of controversy among the various leftist and communist organisations around the world since 1917. I don't think WE can give a definite answer.

4) Social Policies and Redistribution: They are completely different things. SP is a policy, R is an economic figure. If you're suggesting that the formula for R should read R=SP*(PS+TT) then it's understandable, but a) I wanted to leave a margin of pocket money for the player (ok 0,5 is too much; maybe it should be replaced by something more appropriate) and b) I wanted to let the player redistribute more than he has too if he feels like it.

5) Product Distribution: For max PD the Kapital will survive (We may want to lift this limit, to make possible actual Kapital shrinkage; this is more of a leftover from my false assumpion that Kapital is not permanent), but with no profits, the UC will have no income and so if they don't go to work, they will starve to death. Simple, isn't it?

6) Demographic Shares: Rodrigo wrote his post without having examined my take on solving the demographic problem (not the REAL one, the one in the model of course! ), so I expect him to study the spreadsheet before making further contemplations.

As for the Minorities, it seems that we're having a disagreement. Mark claims that not all minorities are laborers and historically he is correct (f.e. some of the richest people in ancient Athens were not citizens, but "metoikoi", immigrants from other cities). Rodrigo claims that in the model they should be thus discriminated and a reason why is IMHO, that otherwise they would act exactly as the majorities (politically and economically) and so there would be no reason for the discrimination other than cultural attributes. This means that keeping a large EG into the minorities is a policy aiming solely to the manipulation of the civ's cultural attributes and has nothing to do with the economic motive of providing the economy with cheap labor, which is incorrect.

IMHO it depends from society to society and that it is strongly related to the definition of slavery. If it is totally EG related (All Jews in ancient Egypt were minority and ALL were slaves), then it's like in the spreadsheet. If it is partially EG related (ALL slaves in ancient Athens were minorities (foreigners) but NOT ALL minorities were slaves), then slavery is a condition a person is reduced into for some economic reasons along with some ethnic ones. There must be a class distribution among minorities too and we have 3 kinds of people: citizens, minorities and slaves. But things would be kept simple if we did not distinguish between majorities and minorities in the class distribution. We would still need a function to determine how many slaves we have (none if SL=0). If slavery is totally unrelated to EG, then this case would be similar to the previous one, if not simpler. All this is IMHO too much trouble and I don't know if it's worth it.

7) Please let's decide pretty soon (but after seriously pondering) if we're going to include the Middle Class. From a computational point of view I find it pretty troublesome and, from a theoretical one, I don't think this should be our priority; I'd rather try including some more flexibility to the model.

8) We must also decide if we're going to employ this model at a civwide, province, sector, or province x sector level, or even parts of it here and there. The decision is essential for the complexity of both the methods used and of the game itself. I would say province for all the classes stuff, sector (optionally province x sector) for investment, civwide (optionally province) for private/public stuff and ruler's policies. The optional implementation should be for micromanagement.

9) Rodrigo, could you make that turn evolution macro? It is required, in order to test the model dynamically. Mind you, Mark has an older version of MS Excel. While making it, you will have a chance to explore the mechanism and point out the facts that I have overlooked.

10) Calculations: As I have already pointed out, the spreadsheet demonstrates all variables that can be derived from the model, not only those that are needed for one iteration. It is possible that in some places I have not chosen the best way to make calculations, so somebody more experienced should bother to check the calculations flow (A salesman would say that the product has no guarantee ). We can compress all the detail into small solid functions, but this is a programming issue, so I'm not the expert. We'd better ask FSmith about an OOA for this pretty soon, or we're in danger of making bad design decisions.

11) AI: I'm afraid that the more liberties we provide the player with (green-coded slider variables), the more it will be difficult both for the AI and the human player to optimise their play. Who will be more badly dazzled by the complexity is still an open question...

[Mark_Everson]

My comp at home was dead this morning. If I can't fix it, you will just get very brief things like this from work

I am only responding to Rodrigo here, since I started this before Axi's post...

1. Agree completely!

2. I think PP and SP are reasonable minimal inputs for the econ model (btw what does 100% SP Mean? 100% even incomes/benefits, or 100% of GNP spent on social needs?). I believe EP is best Instituted by govt Actions. Since EP can go in Many different directions. Then the sum of these actions (in CC) is compared to EP given by the govt. model x GNP. So although the govt model specifies an EP number, it would only be a target for achievement by the AI or player. I don't like the tax rate being meaningless as PP -> 0. Perhaps we should use effective tax rate being an average between govt-controlled and private sectors. So a Bad communist government that doesn't return much in value to its citizens would have a high effective tax rate, whereas one that truly gave everything back to the people would have a low one...

3. Have to hold this discussion at a later time...

4. Definately. Political power should feed back into what the govt builds. I think this is relatively stratightforward with infrastructure types I have.

5. Well at least minorities being MC as well as LC is an improvement (MC/LC ratio affected by treatment of minorities in the civ). I guess that compromise would make me mostly happy.

6 & 7. Agree.

[Mark_Everson]

I had one thought about how we might handle the private property issue. I'm not sure if I even like it yet, but I thought I would put it out here to see what people think.

My proposal allows the player or AI to essentially do whatever they want over the three choices for PP that I elaborated above. The idea is that the government can take its share of the property in Any of the three buckets which consist of: a) land and capital; b) infrastructure of the kind that people might invest in in a capitalist society; and c) labor. At the two extremes of PP there is no choice. PP = 100% means the government controls nothing of the above (a-c). PP = 0% means the government controls All the above. For values in between, the government gets its choice of what it owns, but the value of it must equal 100 - PP fraction of the GNP (which for now I am taking to be the sum of a through c). Doing it this way PP = 50% can be handled in a variety of ways, but always indicates a very substantial part played by the government in the economy.

Here's an example. In this example I will take the parts of the economy to be worth a (land and capital) = 30, b (infrastructure that people use directly like healthcare or housing) = 20 , and c (labor) = 50. The total value of GNP is 100 here. So let's say PP from the government model = 50%. In this case the government could control directly everything except labor giving PP = 30 + 20 = 50% (and it would probably then tax the labor moderately heavily since the people have their infrastructure needs already provided for them). Or the government could take 50% of each of the three things, or it could even control all the labor, but none of the other components. For the intermediate cases of PP I think we would need some sort of limitation on how fast the government could change between different versions of the economy with the same PP value. The people would get used to it being one way, and the government changing things drastically would certainly cause large economic dislocations.

I could elaborate further, but I'm not sure this is a good idea, so I will just leave it here and see if either of you are interested in the concept... One attractive aspect of handling things this way is that it makes it clear that with PP = 0% there is nothing to tax.

[Mark_Everson]

Laurent:

Definitely give us input. You will need to code some of this stuff, you should Like it That said, I think the govt model can already handle the points you raise. Hopefully Rodrigo will correct me if I'm wrong.


Hi Gentlemen:

BTW my comp turned out to be fine, at least for the short term.

Some general things we need to firm up at the base of the system IMO...

Does PP 0% mean the state controls:

1. 100% of capital and land but nothing else...
(this is like what a mixed-market econ does when the govt owns some of the productive apparatus, but the people still buy infrastructure with the wages paid to them) I think this is Axi's viewpoint based on his point (3) above (at least with low taxes and no redistribution).

or

2. 100% of the Stuff (kapital, land, And various infrastructure like housing -- socialist tendencies but not to communist)

or

3. 100% of Everything incl labor. making it communist in my book, though my definition may be off. (This is my best guess as to what Rodrigo is thinking.)

The problem is that these three have different effects, so we Must agree before trying to put in mechanisms to use them or it will be madness

IMO we need Something (perhaps not PP) that allows 1. to happen in the economy.

I personally prefer something similar to 1. augmented with tax rate and redistribution as in Axi's (3) in his post above. However I'm not exactly clear on how the redistribution works with the complex (post demo 5) economy with different infrastructure types etc. FE (exaggerated) the govt may decide the people all need lumps of Steel more than anything else as in the 'Great Leap Forward' in China. Although this might be technically a redistribution, the people have no real use for steel. But I guess maybe happiness could take care of this. so the people would recognize the right of the govt to redistribute and give them what it thinks they need best, but without much housing or clothing they will be very unhappy with the govt for practical reasons.


On the middle class: I prefer using it. If you insist on LC having no capital then I think we really need it.


I think Axi's proposals on what level things should happen at (FE PP civwide (optionally province)) are reasonable guidelines.


General notes on the econ/gov model intersections. I want to keep out as much as possible Abrupt jumps in response versus parameter changes. Such things make the AI much much more difficult. That is because then small changes in parameters (player or AI decisions) can make a big difference, meaning that Many more possible strategies need to be searched. This is a Big problem in Civ, too many abrupt things.

One such thing is rigidly enforcing PP every turn. The UC won't invest if it is just going to get nationalized anyway, giving them no gain. So to make it feel like the real world we then need to add another check to the AI... Is it worth investing, and if worth it, will the govt take it away, making it not worth it? In this case the fix is fairly simple. But I really don't want 20 such things coming out. That is why I think we need just a few simple equations for income distribution and investment among the classes. A long calculation chain makes determining the impact of a decision much more time consuming...

Sorry to end this here, but I need to get back to work!

[roquijad]

OK. There's a lot to talk about, but I'll refrain myself to avoid a mega-super-huge post... so I'll just make a huge one

Hi Laurent and welcome to this very theoretical thread!

Mark's right. What you're looking for will in essence exist in Clash. Since the govt model doesn't have fix/pre-set regimes as in, for instance, Civ2, then there's room for slight variations for any given "typical" regime. In the case you mention, french democracy would be somewhat more divorced from religion than Poland's. It'll be possible for the govt model to produce small differences like this, mostly based on differences in culture, which I guess is the main reason for the french-polish case. Although we're still discussing on how the govt decides what to invest in, I'm pretty sure the flexibility the govt model has in the types of regimes will be "passed" to the econ model in the final solution we make for investment.

On another comment, "people themselves invest in religious infrastructure", as I understand the econ model, this was planned long ago, so it'll also be there for you.

You don't know how happy I'm to see we can fulfil a potential player's aspirations!
Please insist in whatever you think should be there in the game.


Alright now...
First, 4 short things:
1) Working hours: I agree too. They should be used.

2) Middle Class: Let's use it. I'm working in the flexibility issue for the number of classes and although I don't have the thing solved, the approach I'm leaning to currently needs these 3 classes (UC, MC, LC). I know it sounds strange to speak about flexibility in the number of classes and at the same time encourage the usage of specificly 3, but I'll ask you to trust me for the moment.
BTW, I like Axi's definition for MC: "a class that has enough kapital to employ their own labor".

3) I saw the spreadsheet. Complex, but useful to play with.

4) I sent you Mark a copy of the new spreadsheet with a Macro in it. I also asked you in that mail a copy of your econ model spreadsheet. I haven't recieved anything from you and I'm wondering if the mail I sent was lost during you computer bug.
-------------------


There's a lot to say about, but I rather prefer to invite you to organize things a little bit. First, I know things tend to get mixed, but some things in previous posts belong IMO strictly to the econ model and not to this thread, like "should things be managed at the province x sector level?". In order to have a lesser mess, I believe we should keep things separated and let this thread only for what it was created for. OK?

Second, there're several topics I found where it seems we're not working exactly with the same definitions or, alternatively, things where we don't have an agreement about and IMO must be solved before getting to specifics. My proposal is to discuss those things first, so we can have the clear concept for each thing and the right interpretation for each variable. Done that, we can go and interconnect both models. Alright?

So, I'm gonna leave this post only to name the points I find ness. to discuss first and write my stand for each. They are:
-Minorities
-Player/Govt
-Tax Rate
-PP, EP, SP
-Govt & goods

Before exploring each, I want to say that it's only my perception that we've a disagreement on those, so please just tell if I got things wrong.

Minorities: Two comments confused me. First, one by Mark, "If minorities have full rights, then ..." and second, Axi's reasoning on the expotation of minorities and the relation with slavery.
My position and the guideline I took while developing the govt model, is that people discriminate others for a cultural reason, which can be a)low tolerance to other tribes; and b)low tolerance to other religions. b) is self explanatory. a) means the discriminating tribe feels in someway superior to the rest. This, of course, doesn't imply the discriminating tribe wants to economically take advantaje of the other tribes (slavery in some form). Slavery would be the result of a) plus a high desire for wealth and that's btw the reason I manage them as two separated variables in the model.

If we treat minorities this way, I think we'll have the right outcomes and I guess this is why Axi, being although inclined to the idea of discrimination inherently attached to slavery, agreed with Mark: "Mark claims that not all minorities are laborers and historically he is correct". Another one by Axi: "NOT ALL minorities were slaves [in Athens]". Since we all want to be historically accurate, what we'd want is the govt model to produce a discrimination, but not all discriminated people is slave and where all those who aren't slaves can behave "almost" like normal citizens. I believe I achieved that:

The govt model was meant to produce these possible scenarios:
1>Minorities and slavery don't exist. FE, US today.
2>Minorities exist, slavery doesn't. FE, black population in the US in the early 20th century.
3>Minorities exist, "soft" slavery exists. FE, roman empire. Here the term "soft" means not all minorities for a given EG are slavered. Not all slaves were forced to be so. Not all slaves are so for a lifetime. In other words, the widely most common form of slavery in history.
4>Minorities exist, "hard" slavery exists. This is exactly like the above, but some specific ethnic groups can be "brutally" slavered, that is, they're forced to be slaves, so are their children and they cannot get their freedom back. FE, slavering of africans. The EGs not brutally slavered are assumed to live like in case 3.

There're of course more possible cases, but I felt that was good enough. Now, in cases 2, 3 and 4, there're minorities which are "free" (not slaves). They can live "almost" like citizens (majorities). This "almost" is managed by the govt variable "Ethnic Discrimination (ED)". And here's where Mark's comment is in conflict with my mental idea. Using ED, free minorities can have about no rights at all and then, economically, they'd unlikely control kapital. Or, they can have many rights, making possible some kapital control. But, I feel they never reach to the point of "full rights" or "full rights except participating in politics". After all, they're intentionally discriminated. This is why I believe that within free minorities you could find at most UC and MC, but never UC.

Of course, I don't want to split minorities into classes for simplicity, although I know F_Smith would love to! So, my proposal is treat each minority EG as being partly slaves (with a share) and having some kapital, which will be interpreted as the kapital free minorities in the EG have. The share of slaves should be given by govt variable Slavery and maybe from some economic data, representing the need of some of people to become slaves. The amount of kapital they control should be given mostly by ED.

The only problem I foresee is in the riots model. Here we'd have an EG which is partly free and partly slavered and so, I'd expect them to behave differently in the riots model. If felt ness., I believe we can refine a little that model to produce two different Pro-Action Feelings from the same EG instead of the unique we have now for minorities.

Opinions?


-Player/Govt: In some comments in previous posts it's said that the player would have to choose values and take decisions for economy, like when Axi says "let the player invest more or less than the UC at will". Since the player is just one more actor defining govt policies, there's no direct control by the player on economy, at least for PP, SP and EP. So, we should only look at the govt setting. Now, there're things unsolved, like how much the govt wants to invest in each area, which are not defined by SP/EP/PP. But that, again, shouldn't be something the player'd control at will. As we have said, the type of regime should be THE tool saying how the govt manages investment and that sort of things. At least, govt type must give the guideline. Anyway, the important thing here is we don't have a clear picture of how the govt decides investment and salaries for public employees. There's a mechanism missing I believe at least Axi is leaving for player control alone.

Another thing is AI. What an AI-player would choose, knowing he can't affect directly economy but through the final govt setting? I feel it's not too bad. After we finish govt-econ connections, we'll have a good idea on where the optimums are for a given criteria. If the criteria is "maximize total wealth", probably this leads to a type of crude free-market that we'll know how to generate. Given that, we can store "strategies" as the values the ruler wants to apply. The rest is simply interaction with classes to get the final govt setting. AI will choose strategies, probably mostly based on the character's attributes.


-Tax Rate: Why you really want tax rate when PP=0%, Mark? If PP=0% then everybody is an employee of the State. If so, a person gets a wage from the State. Now the State ask him to give money back? Why not just pay him less? I really don't see the sense of using tax rate in this situation.


-PP, EP, SP: What each means, exactly? This was supposed to be solved already in the govt model, but it's obvious now it's not that clear after all. Here's my view:

PP=Refers to the posession of infrastructure, equipment and land (kapital). As a corolary, PP=0% means all factories, all farms, all equipment, all companies belong to the State. As a corolary of the latter, all people is an employee of the State.

SP=Redistributive practices. The intention to homogenizate living standards (or effective income). Includes things like free education, free health, specific taxes to rich people, etc. When SP=100% wages are all the same or effective/real wages are the same. Everybody has the same access to all available goods. When SP=0% then wages tend to be equal to marginal productivity and nominal wages are equal to effective/real wages, that is, you have access only to what your wage can buy (no free things, no subsidies in health, educ, etc).

EP=This is the toughest. Axi defined it as the control level of the productive machine, both public and private. I initially bought it, but here's the thing: All public econ activities are by definition controlled a 100% by the State. If so, EP should only be useful for the private sector. In that sector it has sense: Controling monopolies bahavior, mitigating externalities, setting standards for industries, defining rights of use for public goods, etc.
It seems like the case of tax rate: EP is useful only when PP=0%. That's unless Axi has a good interpretation for what Economic Control means in the public sector.

A complex question is whether the govt can change these policies regardless of the money it has... Increasing SP implies higher spending. What if the govt doesn't have it? Maybe here's the key to what Mark suggested when nominal values in the govt shouldn't be taken exactly in the econ model... I'm not sure....

-Govt & goods: I have the feeling that you, Mark, see the govt giving goods to people in socialist/communist regimes when it redistributes. I ask this because I remember when I read the econ model a loooong time ago I found something like "the govt gives back to the people but the things it gives are not ness the ones the people want" and a phrase in one of the posts above reminded me of this. Do you see it like that? In those regimes the govt at most guratees you an amount of certain "basic" goods. You, as a citizen, may or not use that, but also you have money (or something equivalent) to get whatever you want. So, it's the same like in free-market, at least in this regard. You get paid, but by the govt, and you go and buy the things you want. The problem has been that communist regimes have suffered great shortage of goods, so in fact you can get very little with your money. Sometimes even those guaranteed goods aren't even present.
I'm saying this only to avoid a bias in happiness due to a socialist system. Like in free-market, economic happiness should mostly be given by income and/or consumption.


Long post.... My clock says 4:36 AM...
[This message has been edited by roquijad (edited July 20, 2000).]

[Mark_Everson]

Rodrigo: Haven't read the whole post but got to the part of you not having gotten the spreadsheet. I have re-sent it... (I sent it first immediately after receiving your email last nite) If you still don't get it, maybe try sending me an alternate email address.

[F_Smith]

Hey --

One thought: The problem with the 'social classes' and 'uc'/'mc'/'lc' divisions is that they are seperate things!

The 'uc'/'mc'/'lc' are 'Economic Classes', organized along specifically economic lines, with economic purposes. A 'social class' isn't necessarily organized along economic lines, (for example, a 'noble' class).

These are two seperate concepts.

[Mark_Everson]

Rodrigo:

Yep, we need to agree on the foundation first. Your comments in these areas are certainly helpful to me...

Minorities vs social class: I was thinking of ethnic discrimiation of 0 for my comment above "If minorities have full rights, then ...". Perhaps you consider technically that there Are no minorities for ED =0. If so I hope that eliminates the confusion caused by my statement. I still think that for only small discrimination, say ED=1 there should be UC of other EGs than the majority. they would just be smaller in proportion to those of the majority culture. So in going from discrimination of 9, up to 0, there would first be a small amount of MC in the EG around say ED = 7 and that would grow... then around ED = 3 you might start getting some UC also.

Splitting minorities into classes: This would only be worth it if we use a simple mechanism. I'm with you that we don't want 20 classes... I just think if you insist on minorities always being LC essentially then you lose a lot of the benefits minorities can bring to a culture in terms of investment etc. in the real world. If we can come up with a simple game mechanism to do this I think it would be valuable. Your compromise sounds reasonable to me!

BTW when you use the word "slavered" (no such word) it should be "enslaved". Not meant as a criticism, your English is Much better than my Spanish (basically nonexistant)


Player/Govt: My view is that the player Must have real economic choices to make, but that these should be constrained by the type of government. The govt system can limit the player's economic choices, but must not eliminate them, or much of the economic system is a waste for the player... My thought is that government spending on, say, infrastructure is determined by who holds the power in the govt. This would crudely mimic a real political process. Take, for example the govt budget for infrastructure as 100k CC. (This high-level number would be set using PP and SP inputs.) Lets suppose we have the following political power:
Class___Political Power
Ruler_____25%
UC_______8%
MC______10%
LC_______25%
RC_______16%
WC_______6%
BE_______10%

Now, the UC, through their govt reps, would want different infrastructure (FE more entertainment, less social safety net perhaps) than the other classes. We would take the UC power of 8% and thus spend 8k CC of the infra budget on what they want. (to do this we may need to add to the ideology info to get what infra they want) And so on for the other classes except the Ruler. The ruler gets to spend his/her share of 25k CC on anything within the infrastructure category.

If the Ruler wants to push education it could all go there, but the people might be made unhappy economically if they feel their basic needs are not being filled at the expense of education.

So an autocratic ruler has great scope to do what he will. Whereas a democratic ruler has a much smaller amount he can move the levers of the economy, but its never zero.


AI: I agree about what you say being true for the govt AI, but the econ AI must have much larger scope...


Tax Rate: Sorry, I just see it differently. for PP=0% the workers still have a Fair Share value for their labor (whether they get it or not they can sense it IMO). This should be their nominal wage. If the govt takes more in taxes then the people will receive less than their fair share, and will be less happy the more of this is taken. Doing it your way, the government can set the people's wages anywhere it wants, say 0, and it is still effectively a 100% tax. Why Not use tax rate, which we already Need in the system anyway for capitalist systems to also handle this one peculiar case?


PP, EP, SP: Thanks the added info helps a lot.

I still think EP is not needed. I think govt influence can cover the EP factors... As I picture the infra model and the capability of the player to subsidize or tax individual commodities, I think the govt can, from tax revenues, duplicate a very large proportion of EP-type activity. Otherwise we will need a whole set of new knobs to handle EP at the detailed level. (Since minimum wages are clearly different from food price supports) Also, a govt that owns nothing, but regulates everything economic very strongly is not different in effect from a government that owns a substantial chunk of the economy and uses that to further its aims IMO.

Money limitations vs SP: Good Point! I think a give SP level implies a tax rate >=SP value, but we need to discuss further.

Govt & goods: Yes, you got my viewpoint correctly. Your statement of how it works, appears to me a good ideal. But in practice, in places like the Soviet Union what happened was more like my view. The govt gives the people what It Wants, and then they can squabble amongst themselves as to who gets which pair of lousy shoes, etc. Or in another case the government can give the people quality education, but they may not value it as highly as the govt does because they would prefer entertainment... But I agree with you that mostly econ happiness is a factor of what you recieve. But that is Not its value in money, but rather its utility to you. So if the govt distributes glorious housing to starving people, and gives nobody food, the people value this very low by their utility. Only when people have enough of other things do they value magnificent housing appropriately.


F_Smith:

Economic and social classes are indeed not the same. It is a simplification we make in the model to keep the model more compact.

[This message has been edited by Mark_Everson (edited July 20, 2000).]

[F_Smith]

Personally, I would like to request ya'll change your wording to 'Economic Classes' then, because altho ya'll know what you mean, others won't (I didn't).

A lot of ya'lls discussion sounds way off base, until the reader realizes the topic is 'economic classes' and not 'social classes'.

I think this is causing some very wierd problems for the model design, too.

For one thing, talking about 'discrimination' of/against an 'economic class' basis doesn't make much sense, really. Am I wrong?

And rolling the 'Govt Class', 'Religious Class', 'Military Class', etc, in with the UC/MC/LC is not going to work, once you consider we're talking about 'economic classes' . . . is it? And you shouldn't ahve a seperate 'working class' -- that doesn't make any sense, from an economic viewpoint. From a social class viewpoint, yes -- but that's an entirely different concept. The two types of classes will have to have different behaviors -- one based on economics, the other based on social class backgroud and ethnicity. In fact, all members of the govt of social classes will fall into the uc/mc/lc somewhere anyway, won't they?

That's where we had problems, after all. The only solution I see is to seperate the social and economic variables. Because you can guarantee a given number of 'economic classes', since the definition of the economic classes can be designed to include 100% of the population.

So 3 fixed 'economic classes' is fine, already scalable. But there should be no limit on the 'social classes' -- whatever those will be.
[This message has been edited by F_Smith (edited July 20, 2000).]