Govt Model Links

• Model Introduction

• Page 1 - Sec I - III

• Page 2 - Sec IV

• Page 3 - Sec V - VII
  (and Appendices 1-3)

IV.1 Computing De Facto Influences

In the next sections political power will be used by classes and the ruler to change government policies. That political power is real political power, which counts in the nominal or by law political power and de facto influences classes have leading to a higher political power than law says they should have. De facto influences are the way this model provides inter-classes interactions, allowing the UC to bribe high officers in the MC, FE. To compute real political power for each class we need the nominal values from the Government Profile and de facto influences. Here's how de facto influences are calculated:  

Ruler: He can bribe other "politicians" to gain their sympathy. De facto influence is computed as a scale parameter multiplying the total spending in bribes. 

Upper Class: UC de facto influence comes from its control over economy. They can bribe, they can endorse election campaigns and ask favors afterwards, use propaganda through the media or directly to their workers, etc. De facto influence is computed as:

0.3*PrivateProperty*(1-EconomicPlanning). 

Religious Class: RC de facto influence comes from the respect and worshiping they receive from the rest of society. The more intolerant the religion is, the more willing to use this influence is. That's because intolerant means religion sees itself as the only 
way to live, so it tries to impose that. De facto influence is computed as:

0.2*MCA_ImportanceOfReligion*(100-RCM_ReligiousTolerance)/10000. 

Military Class: MC de facto influence comes from its ability to militarily threat the government and politicians. De facto influence in this case is a constant around 0.15.  

Bureaucratic Elite: BE de facto influence comes, by its definition, from its influence on the government from the inside and its control over administration. BE's nominal political power is always zero, since BE's political power comes only from its de facto influence. BE's power is really unwanted. It's the unwanted result of a large bureaucratic apparatus. A measure of the size of bureaucracy is used to compute BE de facto influence:

0.5*Average(1-PrivateProperty;EconomicPlanning;SocialPolicies). 

Lower Class: LC de facto influence comes from labor unions and the ability to threat with labor strikes. This influence increases with higher Social Policies (which protect workers), with higher Civil Rights (which allow the formation of syndicates and allow strikes legally) and with the level of syndicating, seen as a social development in the tech tree:

0.3*(1/(1+exp(-2+10*SocialPolicies)))*(CivilRights/100)*Syndacting.

(assuming Syndicating Tech Development in 0-100% range)

For the ruler and all classes except BE, de facto power is really the capability of gaining other classes' power by some means. It represents how a class X can buy the sympathy of other classes to encourage them to use their legal (nominal) power in the benefit of class X. Because of this, all values described above should be multiplied by the total nominal political power other classes have. In other words, the MC cannot threat others encouraging them to do as the military want if those others cannot really help because they don't have any nominal power to please the military. That's why de facto influences are really those values given above, but multiplied by all possible nominal power the class can actually "buy" from others, which excludes the ruler (otherwise the player would be giving away power without wanting it) and excludes the power high officers loyal to the ruler in the MC have.

In the case of BE, de facto influence is its only source of power and it represents actual control over government policies and administration. It's not, as opposed to the other entities, power to buy other classes nominal power, but power it indeed possesses. So, even when the ruler thinks he has absolute power, the BE still can have some control.

Some of the sources for de facto influences can be undoubtedly recognized as corruption. The civ will be able to fight corruption decreasing de facto influences and forcing political entities to rely only on their nominal (legal) political powers. Although not yet determined, this will probably be done using the media and Civil Rights. In this way, civs with a free enough society (independent media) and having a decent tech level for media and communications, will be able to reduce the level of corruption.

Once we have de facto influences, we must compute real political powers. BE real political power is directly its de facto influence, as per definition. Political power the BE doesn't control is what all other classes and the ruler have left. This remaining power is distributed as:

Ruler: Nominal+DeFactoInfluence  

UC, LC, RC and MC: (1-Ruler_polpower)*(Class_NominalPolpower+Class_DeFactoInfluence)
(UC_nominalpp+UC_defactopp+ LC_nominalpp+LC_defactopp+ RC_nominalpp+RC_defactopp+ MC_nominalpp+MC_defactopp)

IV.2 People's Preferences

People have a clear opinion of what they'd like to see in the govt. As everything, people see things through the eyes of culture, so cultural attributes from the social model are taken in to drive people choosing what they'd want for the govt. What we need is an opinion from every person for each government policy that needs to be negotiated (i.e. all policies except tax rate) and for the political structure.

What the people want for some of the policies can be computed straight forward from cultural attributes and individually for each policy, while others need more sophistication. This sophistication arises from the need of coherency between variables. It's necessary to avoid people from choosing at the same time RC political power share equal to 80% and LC political power share equal to 50%, since all shares must sum 100%. Or, we must avoid people choosing a politically powerful aristocracy (high UC political power share) and at the same time choosing economic policies producing a communist system. So, economic variables and the political structure are all in "packages". These packages are built in the game guaranteeing internal consistency between variables, so people, through culture,  only needs to pick the package they see as the best. These packages are called ideologies.

Ideologies not only help doing a consistent modeling, but also add flavor to the game since now people with different ideologies can collide. For example, the French revolution was nothing but the battle of two ideologies: monarchy vs. democracy.

So, before people act in the political arena, we need to know which ideology they support and what they think about those other policies that can be analyzed individually outside the frame of ideologies (Directly Negotiated Policies). The following section IV.2.1 and IV.2.2 show how this is made. 

IV.2.1 People's Preferences on Directly Negotiated Policies

The policies in this category are:

-Slavery
-Ethnic Discrimination
-Religious Discrimination
-Civil Rights
-Foreign Policies

In the following lines it's shown what the people in UC and LC want for each policy, mostly given by cultural information stored in MCA (Majorities Cultural Attributes) and what the RC members want, mostly based on religion's doctrine, stored in RCM (Religious Class Mentality). BE wants to preserve whatever value the government currently has. What the MC wants is given by the relative influences it receives from UC, LC and the ruler, as explained earlier. So, for a given policy X, having what the UC and LC want (say, UC_X and LC_X) and what the ruler wants (from the Ruler's Government Profile, Ruler_X), then what MC wants for policy X is:

MCM*(MC_U*UC_X + MC_L*LC_X + MC_R*Ruler_X)

where MCM is a Military Class Modifier allowing MC members so slightly adapt their mentality depending on the policy and MC_U, MC_L and MC_R are the respective influences over the MC by UC, LC and the ruler.

Slavery (SL)

LC wants: SL=2*square_root((100-MCA_EthnicTolerance)*(100-MCA_Asceticism)/10000), rounded to the closest integer.

UCwants: SL=1.2*2*square_root((100-MCA_EthnicTolerance)*(100-MCA_Asceticism)/10000),   rounded to the closest integer.

RC wants: SL=2*square_root((100-RCM_EthnicTolerance)*(100-RCM_Asceticism)/10000), rounded to the closest integer.

MCM=1

Formulas say slavery will tend to be accepted if there's low respect for other tribes and a high desire for wealth. Because of the latter, the UC will be a little more inclined to slavery (20% more).

Ethnic Discrimination (ED)

UC and LC want ED=(1/(1+exp(5-MCA_Nationalism/10)))*(100-MCA_EthnicTolerance)/10, rounded to the closest integer.

RC wants ED=(1/(1+exp(5-RCM_Nationalism/10)))*(100-RCM_EthnicTolerance)/10, rounded to the closest integer.

MCM=1.1  

Religious Discrimination (RD)

RC wants RD=(100-RCM_ReligiousTolerance)*5/100, rounded to the closest integer.

Both the UC and LC want: RD=(100-MCA_ReligiousTolerance)*5/100, rounded to the closest integer.

MCM=1

Civil Rights (CR)  

RC wants: CR=RCM_Individualism

UC and LC want: CR=MCA_Individualism  

MCM=0.9  

Foreign Affairs (FA)

RC wants: FA=RCM_Aggressiveness  

UC wants: FA=MCA_Aggressiveness  

LC wants: FA=MCA_Aggressiveness*0.9  

MCM=1.2

In this case LC wants a FA policy a little less aggressive than the aggressiveness of its culture because, after all, it's them that are going to war. The military prefers a policy a little more aggressive. 

IV.2.2 People's Preferences: Choosing Ideologies

Each person will look at the available ideologies (discovered/invented ideologies) and will choose one depending mostly on his culture and how good it is to his class. To do this, the model computes an attractiveness for each ideology, on a class basis. That is, the same ideology has different levels of attractiveness depending on the class observing it. This will be made for UC, LC and RC but not for MC nor BE. As in IV.1.1, MC preferences depend on what UC, LC and the ruler want and BE wants always to preserve the current form of govt. 

All persons measure the attractiveness for a given ideology with these four questions:

§         How much power the ideology offers to the class I belong to?  (Desire for Power Attractiveness-DPA)

§         How much the ideology reflects my culture? (Cultural Attractiveness-CA)

§         How good the ideology is to solve my economic aspirations? (Economic Attractiveness-EA)

§         How good the ideology is to solve other problems my civ is going through? (Current Circumstances Attractiveness-CCA)

Each question leads to an attractiveness level (DPA, CA, EA and CCA). Summing all four we get the total attractiveness the ideology has for someone in a given class. The following shows how to compute the four effects for a given ideology seen by class C. Variables in brackets [] represent info taken from the ideology being processed:

1) DPA=K1*[C_pol.power]

i.e., the greater the political power the ideology offers to class C, the greater the attractiveness. 

2) CA=K2*exp(-(ABS([RCpolpower]-exp(-0.04*(105-ImportanceReligion)))+ ABS([MCpolpower]-exp(-0.04*(105-Aggressiveness)))+ ABS((Individualism/100)-(0.2*[PP]+0.8*(1-[SP]))) 

i.e., the following cultural effects are counted in:

§         The higher the cultural attribute "Importance of Religion" is, the greater the attractiveness ideologies offering high RCpol.power have, and vise versa.

§         The higher the cultural attribute "Individualism" is, the greater the attractiveness ideologies offering capitalistic economic systems have, and vise versa.

§         The higher the cultural attribute "Aggressiveness" is, the greater the attractiveness ideologies offering high MCpol.power have, and vise versa.

3) EA=K3*(A1*[PP] + A2*[EP] + (A3+A4*(ID/3)+A5*(100-RCM_Individualism)/100)*[SP])

where:

In essence this says the UC likes an economic system with high private property, low economic planning and no too high social policies. The LC doesn't care about private property, cares something about economic planning (because this means some level of control on abusing employers) and cares very much about social policies. Indeed, LC has growing care with the more differences between rich and poor (ID=Income Distribution). RC only cares about social policies and it's a growing care the less individualistic religion is.

4) CCA=K4*[Ruler_pol.power]/ES

What's said here is people will find more attractive those ideologies with high ruler political power when the empire is seen unstable (ES=Empire's Stability) as if they were "looking for the ruler's leadership".

Having total attractiveness for each ideology seen by each class, each ideology is multiplied by its corresponding "Knowledge Level". 

Knowledge Level (KL): How known an ideology is. It goes in the range 0-100%. If 100% it means the ideology is perfectly known by the population. When the ideology tech has not been discovered yet, KL=0%. At the moment of discovery it becomes 10%. From that point and ahead, KL increases its magnitude (up to 100%) every turn based on communications techs available.

What KL tries to do is simulate how ideologies spread through the people slowly. When an  ideology exist (the tech has been discovered), it's known slowly by people at a rate given by the com techs available. This way when ideologies are discovered, no sudden changes occur. Multiplying attractiveness by KL with a low KL will turn ideology's attractiveness really small, so people won't be very enthusiastic about them. When KL=100%, people can see all the pros and cons of the ideology and it appears with its whole attractiveness.

Since the current Government Profile can also be seen as an ideology and since the same happens with the Ruler's Government Profile, attractiveness are computed for them too. This is like people measuring how attractive is their current government and how attractive ruler's ideas are.

From this point ahead, only 5 ideologies are kept. The Ruler's Government Profile, the Government Profile and the 3 highest built-in-the-game ideologies processed. So what we have up to now is a matrix of attractiveness. Something like this:

These numbers are used to determine what share of the population in each class support each ideology. This new matrix of support shares will be called SSM-Support Shares Matrix. We'll say the support share for ideology I in class C is:

exp(W*TotalAttract_I)/sum_over_i(exp(W*TotalAttract_i))

where W is a scale parameter. Using this exponential formula, differences in attractiveness are exaggerated and classes tend to concentrate in the most attractive ideologies. For the matrix of total attractiveness of the example above, the SSM results in (with W=0.05):

So, FE, 15% of the LC likes the current state of things. 90% of the RC supports ideology3.

In this matrix MC and BE are missing. In the case of BE, we can add a row with zeroes in all cells except Government Profile having 100%. This comes from BE definition.

For the Military Class, as said earlier, high officers mentality is given by relative influences of LC, UC and the ruler. The support share the Ruler's Government Profile has in MC is: 

MC_R+(1- MC_R)*(UC_X*MC_U+LC_X*MC_L)

where UC_X and LC_X are the support shares for the Ruler's Government Profile in the UC and LC respectively. For any other ideology I, the MC support shares are computed as: 

LC_I*MC_L + UC_I*MC_U

where UC_I and LC_I are the support shares for ideology I in UC and LC respectively.

Having the SSM matrix, we have what we needed in section IV.1, that is, what the people want for every single policy. 

IV.3 Setting Government Policies

Knowing what people want, now we can model how they'll use their real political power to actually change government policies to their convenience.

IV.3.1 Directly Negotiated Policies

Using the procedures stated in section IV.1, suppose we have each class preferences for each Directly Negotiated Policies and through a proper interface the player putted the values he wants for each, so we have something like this:

Suppose the actual real political power shares are:

To compute the final value the government will take for the Foreign Affairs policy, we make a weighted sum of the values each entity wants, where weights are the respective political power shares:

FA=37*12% + 33*25% + 21*26% + 43*7% + 32*11% + 40*18% = 32

The civ's government will have a Foreign Affairs policy of 32. The same is done for each policy. This mechanism allows each policy to take a value which is more sensitive to the desires of those with larger political power. In other words, the more power an entity has, the more successful it is imposing its view on each policy. The mechanism was named "Negotiation Procedure" because it simulates how different entities, having each its own view on some issue, produce a single output. This output is not one of the originals in dispute, but a new "negotiated" one, reflecting the fact that no entity was able to fully impose its opinion. The procedure is general enough to cover a scenario in which all actors have relevant political power shares, like in the above example, and also situations where a despotic ruler holds all power. In this latter case the procedure, without any change, makes the final government policy value equal to what the ruler wanted. 

IV.3.2 Ideologically Negotiated Policies

Again we'll use the Negotiation Procedure, but in this case classes don't have a unique value they want. A class, as shown in section IV.1.2, may have its population divided between several ideologies, so there's no way to determine specifically what a class wants, as a whole, for any of the Ideologically Negotiated Policies. That's why we'll use the Negotiation Procedure on a "party" basis. It'll be like each ideology has a party in the government trying to impose its ideology and each party has a political power to do it.

To compute political power each party has is easy. If, for example, LC has 30% political power and 12% of the people in LC supports ideologyX, then 12% of LC political power is held by ideologyX supporters. Doing this analysis for each class and ideology we can compute all political power each ideology gets from different entities and then sum up. Taking the Support Shares Matrix example in section IV.1.2 and the political power shares in the example right above, the ideologies political powers would be:

So, people wanting to preserve the government as it is (supporting the Government Profile) have 21% of total political power. This one is like the "Conservative Party".

You can see that the ruler, although having only 18% of total political power, will be able to effectively impose his position on Ideologically Negotiated Values at 27% because, as can be seen in the Support Shares Matrix, the values he proposes found followers in classes, like in the UC where 20% of them support him and therefore will use their political power to back ruler's view. Also, since he has influence over the military, he can also get some support from there.

The Negotiation Procedure is ready to be applied. Suppose ideologies look like these: 

Using this info and the political powers each ideology has, the Negotiation Procedure would give us for Social Policies:

SP=5%*8% + 35%*8% + 55%*36% + 40%*27% + 43%*21% = 43%

Again, it can be seen how the Negotiation Procedure's final value tends to be closer to what ideologies with higher political power encourage.

It was assumed here that all ideologies were allowed to participate in the political process. This won't be always the case because through Special Actions the ruler will be allowed to ban ideologies forbidding them to participate. That's why instead of using the SSM directly to compute ideologies political powers, we'll use a "Representation Matrix" which is equal to SSM when no ideology is banned and it's an altered SSM when banning exists. The procedure to compute the Representation Matrix is explained in Appendix 3. 

IV.4 Circular Effects

Although the Negotiation Procedure is simple to apply, we'll need to apply it several times in order to reach final values. This happens because there're circular effects involved. FE, when Private Property is changed, it is changed using, among other things, real UC political power. But real UC political power depends on Private Property through de facto influences. So a change in PP leads to a change in UC political power, which leads to a change in the entire Political Structure, which leads to more changes, like the relative influences UC, LC and the ruler have over the military, leading to a change in MC preferences. This might look like a mess, but it's not too bad. Some of the calculations described in previous sections must be carried on several times to achieve an equilibrium point, but there're two good things: 1) The process is needed only when negotiations are called, not every turn; 2) It takes only about 10 iterations to reach equilibrium, which is quite fast.

You can see more of this in the provided Excel Workbook (click to download)

 
We Want Your Feedback! 
Contact us today or join in on the forum!
Project Lead: Mark Everson,  Web Master: Dominic
Copyright © 99-2001 The Clash of Civilizations Team