The government may use interest groups to find out what people want, and use favor-buyers to create public support for their policies. Therefore, because these three groups are regarded more highly than others, it is a violation of the "equality of influence" goal of democracy.

         If the distribution of voter preferences on a continuous spectrum is a bell-shaped curve, with the mean located at "50" and the positions of political parties at 25 and 75, these parties will shift their policy standpoint closer to the center. They always gain more voters than they lose. Extremists are forced to vote for that party closest to them no matter how far this party is from their ideals. An orientation to the future changes the logic. In this case, they use the threat of abstention to keep the closest party from moving to the center. For extremists, small differences between parties near the center become irrelevant. With a bimodal distribution democracy will produce chaos, with radical policy shifts when parties in power alternate. Centrist parties must either move to one extreme, or can only govern in a coalition with an extremist party. When a centralized distribution becomes polarized, this can lead to revolution. When one extreme imposes its policies by force, warfare breaks out, leading to rule by a clique of under-dogs that eliminate their predecessors. In countries with no middle class, the development of democracy will bring into power a leftist government representing the lower, working classes.

         According to this model, the economic and social condition of society will be reflected in the form of political system. Although new parties may emerge either between or outside of existing parties, principles called "integrity" and "responsibility," create relative immobility ensuring stable equilibrium. Those who form new parties, -- in the US, "third" parties -- may aim to gain office, but in practical effect they only shape the platform of existing parties. Ideological immobility may prevent existing parties from capturing the votes of any substantial block of newly enfranchised voters. Therefore a new party will arise to exploit this political potential on one extreme and a party that then finds itself in the middle may wither.

        Factors that change the distribution of voter preferences include important events such as WW II or a social revolution like industrialization. If a party is formed on one extreme to force an existing party to shift away from the center, that shift represents an ideological victory such that the disappearance of the party just formed is only part of the plan.

        Therefore, what we observe in the American political system can be explained based on logic. Our two parties tend to converge toward the center, but fear of losing extremist voters keeps them from becoming identical. They tend to have a wide spread of policies and a deliberate equivocation about each particular issue.

         Downs uses the distribution of voters on a political scale to determine:

        Under "proportional representation" (not the US "plurality" or "winner take all" system) no single party gets more than 50% of votes, so voters are selecting parties based on the estimated probable coalition. It can be so difficult to rationally select a government that voters treat elections as merely expressions of preference. Governments are forced to adopt non-integrated sets of policies. During the election, parties sharply differentiate their ideologies, but try to be as ambiguous as possible regarding how they will compromise in different possible coalitions. If there is sufficient consensus, even though individual rationality is difficult, democracy can still work efficiently.

         Pareto optimality -- when no person can improve their position without making someone else worse off. This doesn't happen, even in a democratic society except by accident, and democratic governments don't even try. They generally act against high-income citizens and in favor of low-income people.

         Politics is based on decision-making, which requires information. A bad decision is very expensive. Therefore, voters must purchase (with time, money, effort and emotion) enough information to avoid a bad decision. Because we live in a world with an excess availability of information, the greatest cost-saving issue is to collect only that information which will save the cost of a bad decision. Voters need enough variety in their information sources to eliminate false reports, while at the same time matching their principle of selection to their biases, needs and values.

         Some citizens receive more politically useful information for free than others do. Decision-makers want to focus the attention of citizens on areas that will lead to support for their policies. When decisions are delegated to others, the "rate of return" for gathering information is reduced, making lack of information rational. When variations in the amount of information received, or the ability to assimilate it, become systematic, this influences the distribution of political power. Therefore, although democratic societies are rooted in political equality as a basic ethical premise, inequalities of power are inherent because of the costs of information and inefficiencies in information processing.

         The expense of being continuously well-informed about events in one issue area, which is a pre-condition for political influence, means that citizens must limit their awareness to only those areas where their information costs are lowest and political influence is the greatest. These areas are directly connected to employment and the flow of income. Because it is irrational to be politically well-informed, voters don't know their views before voting. And most citizens do not influence the policies that affect them. Therefore, true political equality is impossible even in democracies because of 1) uncertainty; 2) the division of labor; and 3) people act rationally.

         People vote when the value of their vote is greater than the cost. The value of a vote is not just the party differential, or strength of preference. There is also the long-term value of even voting incorrectly, because the act itself makes democracy possible. If no one votes, the system collapses, so citizens derive a benefit from the system continuing. (Therefore, society gives a small measure of prestige or esteem to those who vote, and the value of that esteem can outweigh the cost of voting.) Small changes in the cost of information or other costs of voting can have a large effect on the distribution of political power. It is these costs which disenfranchise low-income citizens relative to wealthy citizens. When voters cannot establish a party differential based on policies, they will use individually rational, but socially and politically irrational criteria, and vote based on personality of leaders, the appeal of historic leaders, or family heritage.