How political parties adjust to fixed voter opinions

Krzysztof Kulakowski

DOI: https://doi.org/10.1142/S0129183105008163

2005-04-13

Abstract:We propose a new version of the spatial model of voting. Platforms of five parties are evolving in a two-dimensional landscape of political issues so as to get maximal numbers of voters. For a Gaussian landscape the evolution leads to a spatially symmetric state, where the platform centers form a pentagon around the Gaussian peak. For a bimodal landscape the platforms located at different peaks get different numbers of voters.

Physics and Society

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **How do political parties adjust their political platforms according to the fixed opinions of voters to obtain the most votes?**

### Specific problem description

1. **How parties adjust their political platforms when voters' opinions are fixed**:
- The author assumes that voters' opinions are fixed in a short time, and parties will adjust their political platforms according to these fixed voters' opinions to maximize the number of votes.

2. **Evolution of party platforms in multi - dimensional political space**:
- The author proposes a new spatial model of voting, in which the political platforms of parties evolve in a two - dimensional or multi - dimensional political issue space to obtain as many votes as possible.

3. **Distribution of party platforms in different political landscapes**:
- The final distribution of party platforms in unimodal Gaussian - distributed and bimodal - distributed political landscapes has been studied. For example, in a unimodal Gaussian - distributed landscape, the party platforms will eventually form a pentagon around the peak; while in a bimodal - distributed landscape, the party platforms will be distributed near different peaks.

### Main research content

- **Model construction**:
- Computer simulations are used to study the evolution process of party platforms in multi - dimensional political space. The model takes into account the interactions between parties (such as repulsive interactions) and the distribution of voters' opinions.

- **Result analysis**:
- For the unimodal Gaussian - distributed political landscape, the party platforms will eventually form a symmetric distribution, similar to a pentagon around the peak.
- For the bimodal - distributed political landscape, the party platforms may be trapped near different peaks, resulting in an uneven distribution of votes.

- **Discussion and conclusion**:
- The relationship between the position of party platforms and their historical traditions has been discussed, and it has been pointed out that this relationship is not strong and is more based on the optimization of voters' opinions.
- The response mechanism of parties when the electoral landscape changes has been proposed. For example, voters' disappointment with the ruling party may lead to changes in the political pattern.

### Summary of mathematical formulas

- **Vote calculation formula**:
\[
w_i=\int dx\rho(x)g(y_i - x)\left[1-\frac{1}{N}\sum_jg(y_j - x)\right]
\]
where \(w_i\) is the number of votes obtained by the \(i\) - th party, \(\rho(x)\) is the distribution of voters' opinions, \(g(y_i - x)\) is the influence function of the party platform position \(y_i\) on votes, and \(N\) is the number of parties.

- **Platform position evolution equation**:
\[
\frac{dy_i}{dt}=-\frac{2y_i}{1 + 2\sigma^2I(i)}+\sum_j\frac{2y_i + 4\sigma^2(y_i - y_j)}{N(1 + 4\sigma^2)J(i,j)}
\]
where \(I(i)\) and \(J(i,j)\) are respectively:
\[
I(i)=\frac{1}{\pi(1 + 2\sigma^2)}\exp\left(-\frac{y_i^2}{1 + 2\sigma^2}\right)
\]
\[
J(i,j)=\frac{1}{\pi(1 + 4\sigma^2)}\exp\left(\frac{-y_i^2 + y_j^2 + 2\sigma^2(y_i - y_j)^2}{1 + 4\sigma^2}\right)
\]

Through these formulas, the author simulates the evolution process of party platforms in different political landscapes and draws corresponding conclusions.