The answer to such a question has several different aspects:

First, the very idea of representative democracy is based on the fact that the preferences of the individual constituents do not make for good decisions in complex questions. The Federalist Papers still provide one of the most readable general arguments on the matter, but modern research confirms that only 'elites' fulfill the information requirements of rational choice. In other words: In representative democracies, decisions are likely to differ from the preferences of constituents, because they supposed to be based on more information than most constituents bother to process. Differences do not indicate a systematic problem.

Secondly, voter preferences usually include impossible demands, such as high levels of service, no paying for it. Decades of studies in cognitive sciences have shown that human decision making is systematically biased, for example underestimating future cost of decisions with long-term impacts. This second aspect means that the decisions of representatives who actually consider the public good are, again, expected to differ from the preferences expressed by constituents. Differences between decisions and constituents' preferences might, therefore, actually indicate a good and desirable thing - responsible and sustainable decision making.

So in order to 'improve' the situation, one would have to first establish that the deviation is a form of corruption - meaning that the decision biases the result in favor of some constituents at the expense of overall social welfare. Here, many of the arguments in previous comments come to the front: The influence of lobbies where election campaigns need paying, the influence of unions and the public sector in countries that have party-based election system. In many developing countries, a fundamental mismatch exists between a minority of constituents who are integrated with the formal economy, pay taxes and interested in good policies, and a vast majority of poorer constituents that act as vote-banks for corrupt elites because they 'prefer' short-term handouts even though they are against their own interest.

In my opinion, important factors for the divergence are personal interests of the politicians and the party's sponsors. Industries with powerful and financially strong lobbys have a significant influence on the politicians' decisions because they are the ones who finance the electoral campaigns.

I think that threatening sanctions and implenting them, if a politician obviously diverges from his constituents' preferences and electoral promises. Furthermore the people should be given an opportunity to penalise a politician's misconduct themselves.

Politicians have their own preferences and cannot precommit to campaign promises. There is a whole literature on this called " partisan politicians". See the Persson Tabellini 2000 textbook.

More than anything, politics is a livelihood for many politicians. Divergence can also be explained by the desire to maximise returns on political involvement. If returns decline for whatever reasons and the politician can no longer see their way, they simply cross the floor. A constituency is important for as long as it sustains the politician's own interests.

The issue arises primarily in single-member constituency systems, but in all systems the voters have to make a choice from among those candidates that are available. Especially in closed list proportional systems one is typically able to vote for parties only, while the priority of the candidates elected from those lists is determined by parties. Moreover, the posed question rests on the assumption that the voters vote for candidates that have as similar opinions as possible with the voters. I believe many, many voters do not think like that, but vote for candidates that they find somehow more enlightened, sophisticated, knowledgeable or simply ”wiser” than themselves in political matters. Elitist? No doubt, but amazingly common among people who are not really all that interested in politics, but feel that they should vote anyway.

The parliamentary system depends on party discipline. In its absence, governments would be required to resign every time a significant government bill was defeated. That would mean that virtually no legislation would be passed.

By contrast, the US system permits Congress to defeat or the President to veto bills without the government collapsing.

It has virtually nothing to do with politicians' individual motives, personal interests or partisanship. It has nothing to do with industries, lobby groups or anything else mentioned above. It's "Political Science 101" ... a feature of the institutional arrangement called parliamentary democracy.

It would, of course, be possible to change these customs and traditions, but it would require the destruction of Political Parties as we know them, and the abolition of Cabinet Government as it exists in the "Westminster System." Even the most enthusiastic supporters of chaotic 18th-century British politics requiring the cobbling together of inherently unstable coalitions on the one hand or American-style political gridlock on the other should think twice about advocating such a "reform."

You might have a look at May's "Special law of curvilinear disparity". This is a classic theory to model the ideological divergence in political parties. He distinguishes between three strata in political parties, party elites, middle-level elites and rank-and-file members. These three groups have different motivations, with consequences for their political opinions. You could easily combine this with modern rational-choice attempts to capture politicans' motivations (Kaare Strom: office-seeking, vote-seeking and policy-seeking motivations).

Factors for the divergence are personal interests, lack of preparation in the resolution of problems, and the small ethics that this absurd social system generates, politicians and party's sponsors. Industries with powerful and financially strong lobbys have a significant influence on the politicians' decisions because they are the ones who finance the electoral campaigns.

I think that threatening sanctions and implenting them, if a politician obviously diverges from his constituents' preferences and electoral promises. Furthermore the people should be given an opportunity to penalise anpolitician's misconduct themselves.

Richard Fenno, in "Home Style," demonstrated the effort that Congressmen then took great trouble to find out their constituents' preferences. The preference that needs to be known is what voters will prefer near election day on their most important issues, and this is not and perhaps can not be known from a public opinion poll years before, when politicians have to make policy decisions whose results will be decisive at election time. So, according to Fenno, what Congresspeople needed was a deep knowledge of individual and group preferences, particularly for potential supporters in the general election. Of course, they also needed to be able to predict the impact of policies, in order to be able to predict constituents' results-based evaluations of policy outcomes.

Given some opinion diversity and lack of strict voter ideology, the incumbents electoral majority will not be the districts opinion majority on every issue.

With increased party polarization, Congressfolk are increasingly focused on their primary constituency, voters who will support them against an intraparty challenge. Hence, they try to represent an intraparty, rather than inter party or constituency majority.

The answer to such a question has several different aspects:

First, the very idea of representative democracy is based on the fact that the preferences of the individual constituents do not make for good decisions in complex questions. The Federalist Papers still provide one of the most readable general arguments on the matter, but modern research confirms that only 'elites' fulfill the information requirements of rational choice. In other words: In representative democracies, decisions are likely to differ from the preferences of constituents, because they supposed to be based on more information than most constituents bother to process. Differences do not indicate a systematic problem.

Secondly, voter preferences usually include impossible demands, such as high levels of service, no paying for it. Decades of studies in cognitive sciences have shown that human decision making is systematically biased, for example underestimating future cost of decisions with long-term impacts. This second aspect means that the decisions of representatives who actually consider the public good are, again, expected to differ from the preferences expressed by constituents. Differences between decisions and constituents' preferences might, therefore, actually indicate a good and desirable thing - responsible and sustainable decision making.

So in order to 'improve' the situation, one would have to first establish that the deviation is a form of corruption - meaning that the decision biases the result in favor of some constituents at the expense of overall social welfare. Here, many of the arguments in previous comments come to the front: The influence of lobbies where election campaigns need paying, the influence of unions and the public sector in countries that have party-based election system. In many developing countries, a fundamental mismatch exists between a minority of constituents who are integrated with the formal economy, pay taxes and interested in good policies, and a vast majority of poorer constituents that act as vote-banks for corrupt elites because they 'prefer' short-term handouts even though they are against their own interest.

The reason for the divergence in the US is because legislators are able to foster a constituent-friendly view of themselves through their personal cultivation of the district and the use of perks of the office to do so. Consider the fact that legislators spend on average about 3-4 days a week in their home districts. Consequently, voters focus on what they know about their legislator when they vote and most of that is pleasing to ears and sight. Then, legislative incumbents run-up large reelection margins and use that cushion of votes to pursue their own interests rather than those of voters. Voters are not blind to such activities but they are so easily rationalized by legislators in their frequent visits to their districts and minimized by the truly excellent constituency service they provide voters. And the large vote margins discourage others from contesting the elections of incumbents; hence, incumbents face nominal competition.

To sum up a lot that has being said, I would say that one of the main reasons lies on what Glenn Parker has explained: people vote on a politician that seems to be closer to them. In fact, people in Brazil (a country where the Proportional Representation elections for the Lower Chamber of Congress is complex, with lots of candidates, and people have low levels of political information) people usually vote for the one that is not the closest to their preferences. I am willing to build a Political Website where this could be improved, while people could find whoever is closer to them in political attitudes (like a "political match"). And people might value more ONE characteristic than another: I might vote on somebody because I am more worried about the economics and think that this candidate will do better at economics than the other candidate, even though I identify more with that second candidate on a lot of issues.

Besides the lack of information of the constituents, the candidate's own preferences are not transparent, they advertise what seems to be good for election or reelection, and they avoid talking or declaring anything that could be unpopular.

A second dimension that is very important is the actual decisions made by the politician, just like Howard Doughty stresses. But Doughty seems to overweight the "institutional design" of countries to explain that. In fact, the influence of interest groups do matter (and interacts with this institutional designs), the influence of each politicians own preferences do matter, at least sometimes. And the politicians who seeks reelection or other political functions will calculate what voter's really value and try to fit this (in fact, they will try to fit a specific constituency desire in specific issues).

The choices voters make depend upon the choices available to them. In the U.S., the existence of party activists and party primaries tends to generate choices located near the median voters in each party which can be far from the overall median voter. As Dick Fenno outlined many decades ago, politicians have multiple constituencies, including the geographic constituency (all the (likey) voters), the electoral constitutency (those who voted for the candidate) and the personal constituency (those especially close to the candidate and/or providing special resources of time and money. Officials are responsive to one or more of these constituencies, but how much to each varies with electoral rules and even holding rules constant, can vary substantially over time with exogenous events/historical processes.

Fenno's construct of multiple constituencies suggests a tug-of-war between the geographic, electoral and personal constituencies, but we can observe as well that some politicians consider the news media to be a constituency. Phil Gramm, Newt Gingrich and currently Ted Cruz leap to mind. The relative importance of these constituencies shift with proximity to the re-election. I would propose as well that Jim Stimson also has something to say about why politicians and officeholders feel comfortable ignoring the public opinion as expressed through surveys. Stimson suggests that public opinion is like a sleeping bear, dangerous only when awake and moving (and I would add, that matters most when elections are at hand). If surveys indicate there is, for example, sustained significant opposition to a war, but that opposition did not prevent the election of both pro-war and anti-war representatives, an officeholder might find it easy to ignore public opinion. Finally, in my previous career as a journalist, I found that many representatives prefer the trustee model to the delegate model of representation. This provides ample wiggle-room for representatives determined to igonore public opinion on key issues and often provdies a representative with status among his colleagues.

As it has been indicated by some fellow scholars,on the one hand unrealistic expectations of the constituents make it seem as if their representatives are digressing from what their focus should be.On the other hand, strong lobbies that fund the election campaigns influence the representatives to such an extent that the wishes, problems and preferences of the constituents become a secondary concern for the representatives.

More specifically, in developing countries where low literacy rate,women's marginal status,little or no accountability and rampant corruption are common features, it is more likely for the representatives to ignore their constituents' needs either due to obligations in favour of the influential groups or personal gains.

May be their personal or party preferences are different from constituents' preferences.

While the sentiments of those who argue that elected representatives should reflect their constitutents preferences is very touching, the fact remains that, in most representative democracies (except those with simplistic two-party systems such as the USA), the real choices are not among individuals, but among candidates for political parties. So, in any election I am going to vote for the candidate of the party which best reflects my views. I frankly wouldn't care if the other parties ran candidates who merit Nobel Prizes in Literature, Medicine and the Peace Prize as well and would (if Roman Catholic be "fast-tracked" to sainthood by the Vatican authorities) and if the candidate for my preferred party was an ignorant buffoon; I'd vote for the ignorant buffoon against the stellar statesmen (or women) for the simple reason that in a parliamentary system, people vote according to party affiliation and not personal choice. That may be hard to accept, but it's Political Science 101.

As for the United States, my New York friend Kurt Vonnegut (1922-2007) clarified the situation nicely in an article he published in Harper's Magazine in November, 1972. He said that there are two imaginary parties in the USA - the Republicans and the Democrats. He added that there are also two real parties in the USA - the Winners and the Losers. So, since both the imaginary parties are controlled by the Winners, this much is certain: in every election, the Winners will win.

This is not to say that there are no differences between the imaginary parties. The Winners who run the Democratic Party, for example, occasionally allow some scraps to fall from the elegant tables of the elites and regularly express sympathy for the Losers. The Winners who run the Republican Party, are cruel Social Darwinists who openly despise the Losers. This explains why Democrats have been slightly more successful than Republicans in the past century or so. Republicans only win when they are able to feed the fear of "terrorism" or "communism" and whip up racial hatred or religious fundamentalism - all of which are unfortunately easy to do.

In any case, the main difference is this: Republicans do evil things with glee; Democrats do evil things with regret. So, Americans are invited to make their choices accordingly.

Another perspective is available from that crafty old statesman Edmund Burke who is often called the "father of modern conservatism." In his Letter to the Electors of Bristol, he famously castigated his constituents for having the temerity to express their views about how he should vote in Parliament. He would not, he insisted, take instruction from the "swinish multitude" (which is pretty funny in light of the fact that he was writing about half a century before the Great Reform Bill of 1832 which extended the franchise to a portion of the male members of the middle and working classes).

My own preference is for the approach of G. D. H. Cole over a century later (ca. 1920), when he stressed that a robust democracy depended on a population that was experienced in self-governance in domains other than the casting of a ballot in a parliamentary election. Servility in the workplace makes it almost impossible to be other than servile in formal politics. Unless, therefore, we devise systems for workplace democracy, elections every four years or so are mainly empty rituals and symbolic exercises in "show business."

All these suggestions constitute very interesting research avenues: As Gérard Roland points out, it is clear that politicians have their own preferences and cannot precommit to campaign promises (partisan politics). This is the basic problem and we may explore which variables empirically explain congruence. Ellen Pfeiffer also highlights the point that lobbies may explain observed divergence of politicians from constituents' preferences. And as Bernard Grofman suggests, politicians have many "principals" or "constituencies".

The influence of parties on congruence is a further interesting aspect. There I can report some interesting results: Politicians’ party affiliations should influence political representation when they are elected under a proportional system. In contrast, majoritarian systems force politicians to converge to the median position of their constituents, thus, muting the role of party affiliations. We can confirm these predictions with unique quasi-experimental data within a common party system by matching referenda decisions of constituents with voting behavior of their representatives under the two different electoral systems (see video by following the link; the working-paper is available on my RG-Profile).

Do you agree with our identification strategy? Are we missing an important point?

http://www.youtube.com/watch?v=m6mAggnzc0A

Very nicely executed research design with highly important results. Well done!

if the representative appointed with cabinet post, he/she should serve all citizens, not peoples in his/ her constituency only. if there's preferences, it'll become bias already (next u'll hear cronyism and nepotism).

2. in today globalize world, a MP activities will be known by everyone. they've responsibility is not only to themselves but their party & coalitions. thus, normally he/ she have to do beyond his/ her constituent scope.

3. have to look at roles play by media too. the politicians might do more for his/ her constituency in general, at the end media will only reports what they see has 'news values', beside influences by agenda setting, etc. so the questions might be bias also based on what only seen in the news.

Respresentative parliamentary system is only one of the means of political ambitions. Divergence of constituencies preferences are often due to elected member of parliament are not in the framework of their contituency alone, but national interest, party interest and so do their personal interest. Jayum & Agus (2008) analyses from 2008 Malaysian General Elections and By-Election, where empirical explanation shows that 1. Politicians are not sincere in solving society issues, but rather sweep them under the carpet, and hope the society will forget, 2. Ethnic distance is a political project for the benefit of politicians, 3. By Election process are more crucial than General Election 4. Often, constituency preferences are just pieces of national issues.

I think that difference between elected politicians´ decisions and constituencies´ interests depends on a complex set of reasons. This includes the information gap, as well as a wide arrange of compromises arising from the existence and behavior of political actors. In a typical parliamentary system, like western Europe´s, there is also a mediation due to party discipline. In many cases, mostly when the electoral system is proportional or hybrid, legislators represent party policies instead of constituencies´ preferences. Also, we most consider individual and collective subjectivity. Only considering all these aspects, and maybe others, from a complex perspective, can give us an answer.

Interesting question. This problem's prevalence could be attributed to the fact that the electorate has no role once the politician is elected, and the suspended animation of electorate is brought back to life only when the next elections are due. It is then hoped that the electorate takes care to elect a better politician to represent them.

1) Unless we address the electorate's participation at more frequent intervals, the politician is bound to have an agenda that is different from the constituents.

2) Large electorates can also be a reason for making the voice of the electorate of negligible value.

3) Often in India we have seen that election issues, regional or national, have not been the issues that are directly affecting the people. No political party has a plan of action on issues affecting the daily life of people, for example, security measures in the geographical region, are issues outside the direct powers and action of political class. They ar emore administrative in nature with only the important decisions being taken by the political class. So they would not be interested in such issues except for the shouting brigade activity during the polls.

I find it very interesting that the arguments here are essentially social or political, and not economic. I would be very interested to see an economic analysis. I do not have one ready.

However, I have a ready answer from game theory, which I studied many years ago at Rice, and this very question occupied an important place in the course. And there is a correspondence to certain market arrangements which I will explain.

Political parties are a means to control the outcome of a political process (i.e. to control governmental policy) with less than a majority. Membership in a party is nearly compulsory, otherwise the representative has no voice at all. And we have seen (and it is mentioned in one response above) that party discipline is enforced by various means. The Tea Party is using primaries. So party membership is necessary, and members are coerced to all vote together on important issues. So a majority of the party members controls the entire party's vote. By various strategies such as consolidated campaign contributions, an even smaller minority can exercise control. There may be no relation whatever of this minority's preferences to the majority constituent preferences.

A similar situation in corporate governance is holding companies. Rockefeller used a holding company structure to control a large number of oil companies with very little ownership. At only two levels, it works like the political parties. 51% of the holding company controls it. The holding company owns 51% of the base company. This is the reason that, importantly, S&P does not allow companies on its indices unless at least 51% of the stock is publicly owned so that at least in theory they can be responsive to the public. It also does not allow limited partnerships or trusts for the same reason. I was until recently a shareholder in Uranium One, which was 51% owned by a Russian company. Well, they sold out entirely at what most shareholders thought was an unfairly low price. But they had no say.

A related finding from game theory is that the preferences of a group among three or more alternatives, whether it be the constituents or the parliament, depend on the order of voting on pairs of propositions. A single person will generally not exhibit such inconsistency. With such effects, it is unsurprising to find less than perfect correlation among preferences any way you cut it. The Republicans, for example, generally prefer spending on the military over not. If they had voted on that first, they would have appropriated full spending for the military, then let the sequester go into effect. But the Democrats were able to prevent such action until the sequester issue came up, so by controlling the order of voting the Democrats got what they wanted. Now the Republican constituency is undoubtedly unhappy with their representatives.

The Democrats continue to be able to control the general order of voting in important issues like the continuing resolution and debt ceiling. The media widely accuses these politicians of incompetence. But actually I think they are simply at war, deeply ideologically divided, and each side is more or less as effective as it can be on the parliamentary moves. Public relations is another matter.

In an article that appeared in Harper's Magazine in November 1972, Kurt Vonnegut commented on the Republican Convention that had recently nominated Richard M. Nixon for a second term as US president.

Vonnegut said that, in the United States of America, there are two imaginary political parties - The Republicans and the Democrats.

He added that there are two real political parties - the Winners and the Losers.

And, since both the imaginary parties are run by Winners, this much is certain: in every election, the Winners will win.

There are, of course, minor differences of personality and style. There are also slight differences in policy. Republicans worship a crude form of Social Darwinism. Democrats do as well, but they also occasionally allow some crumbs to fall from the High Table.

I would say that Republicans do evil things with glee, whereas Democrats do evil things with apparent regret or, as President Johnson was wont to say, "with a heavy heart."

There has been some change since, 1972, of course. Now, the religious right, which even Barry Goldwater considered a threat to sanity back in 1964, has high-jacked the Republican Party and let the likes of Michele Bachmann and Sarah bin-Palin loose upon the land. The tunes of Ted Nugent now trump the strains of Mitch Miller, Lawrence Welk and Debbie Boone at Republican fund-raisers and a screeching paranoia about Muslims and anti-colonials (like George Washington?) is heard from coast to coast.

Somehow, some Losers seem to have captured the role of Social Director at the Republican Country Club but, apart from the fact that the ladies of the Junior League now feel distinctly uncomfortable in the company of uppity trailer trash, the situation remains basically the same.

I would suggest a historical perspective on this question. First, party remains the dominant force in both Congress and the state legislatures because it has been integrated into the fabric of legislative organization through the committee system. Picking up on the two-Congress idea (representatives are both delegates of a number of constituencies and legislator colleagues in an institution where power is the currency), we seem to be in an era where the status within the institution is more important than serving the constituency.

Second, where policy was the focal point of party action in the past, the rise of media politics has weakened the link between policy and party during elections. Candidate and party images are regarded by most candidates and consultants as more important than party positions on policy. But even among voters who educate themselves on policy questions, Most do not choose candidates on the divisive issues like abortion and gun control where parties define their differences. Rather, voters parse candidates on approaches to valence issues like the economy and security, where goals are shared (who wants a bad economy or more crime?) but the means of obtaining those goals differ.

I do not want to excuse the parties for their failures to recognize and convert societal preferences into policy. Part of this failure is surely the myopic self-serving nature of information scavenging in the cable tv/internet age. We are free to choose our sources of information and ignore those that disparage our views, and my experience with candidates and officeholders is that they are more selective than most of us in that behavior. They watch/listen/read the sources that support their positions and do not explore the range of opinion unless forced to. Officeholders often seem to believe that their position ins the favored position, and they are provided evidence to support that view by staff, consultants and party whips.

But I also think that the advent of the permanent campaign in the early 1990s altered the political landscape. Campaigns are inherently tribal, and the focus of parties become one of gaining and maintaining power to the exclusion of other goals, notably governance. We have seen this for the last 20 years in Texas, and we have been treated to a seminal example in Congress since 2010. Parties conceive their prupose as taking control of the government and keeping control, while policy and governance are tools to be used to achieve power. So who cares what the population thinks?

Hi Walter, I agree but three important things should be added to the picture:

1. The electoral college, codified in the constitution, has a winner take all approach to allocating state delegate votes. This makes it darn near impossible for any party which becomes 3rd in size to have any impact on presidential elections. There is no negotiation and alliance forming as in parliamentary democracies.

2. Congress passed legislation encouraging state primaries by direct vote rather than caucus, to which some authors have attributed primary battles and radicalization, since party bosses do not control the direct primaries.

3. Unlike a parliamentary democracy, there is no way to call for early elections if the government becomes gridlocked. This possibly contributes to politicians drifting out of sync? Because there is no reset by a new vote when a critical issue comes up and the views of the citizens evolve in response to it.

After getting a note from David, I realized many reference points were in confusion due to a question being asked by Europeans about European election methods (majoritarian and proportional), and my answer was for the American primary & Electoral College systems (with which you may not be familiar). Also, the goal of representation may be different in different countries, and I'll start there.

GOAL OF REPRESENTATION

It is NOT assumed by many people in the United States, and certainly not our founders, that the GOAL of representative governent is to re-create the preferences of voters in a parliament. Indeed, this often results in deadlock, and the U.S. constitution provides NO mechanism for recognizing deadlock and calling new elections. This has been painfully obvious recently as due to strongly opposing parties being in control of each of our two Houses of Congress, our government has deadlocked and nearly failed several times in the past year.

The House of Representatives most closely approximates the ideal expressed in David's research of replicating the views of constituents. To this end, districts are small with only one representative from each, making their offices and attentions accessible to constituents. The Senate was deliberately constructed otherwise so that it gives representation more or less by region (state) quite independent of population.

Unlike European governments, the President, whose consent is necessary for most kinds of legislation, does not come from the Congress but is separately chosen by an Electoral College process, in which there is proportionality to population but a winner takes all votes methodology in each region (state).

The idea is this collection of representatives will be able to take into account the needs not only of the majority, but also the minority, and various regions, and not led one completely override another. Americans are very concerned about a possible dictatorship of the majority.

There is also the negative example of ancient Greece, which could never unify its city states. They destroyed themselves with inter-city warfare, in fact. Europe has twice in the 20th century nearly destroyed itself that way, and before that was continuously at war since the fall of the Roman Empire. By comparison the U.S. history was marred by only one violent internal conflict, severe in magnitude but restricted to the years 1861-1865. I believe our differences are just as strong as anywhere in the world, or stronger, and we cannot be said to have achieved this record for lack of internal division or even hostility. So the goal of our governmental structure is to achieve election of officials who can mediate and compromise rather than just carry forward the irreconcilable preferences of the population.

DYNAMICS THAT RESTRICT POLITICAL PARTIES TO TWO

Most of my comment above was dependent on the election dynamics that have resulted in no more than two influential political parties at any time in the history of the United States. This is quite different from European and other so-called parlimentary democracies. A 3rd party would by definition be "smaller." A smaller party is unlikely to win more than one or two Senate seats or the presidency since these are conducted along majoritarian lines, with in fact two levels of majoritarianism in the election of the president. Without either Senators or the Presidency, a 3rd smaller party cannot influence the legislative agenda.

What has happened is that a faction within a major party can gradually take over that party, as currently seems in progress with the Tea Party within the Republian party. Or, a large party can fail and a 3rd party or a newly formed party can suddenly find itself in the number two position.

UNEXPECTED REVERSAL OF REPRESENTATIVENESS

A few decades ago, Congress passed laws intended to increase the degree to which representatives reflected the views of their constituents, by encouraging states to have direct primaries, that is, the candidates from each party were chosen by voters rather than in caucuses by party bosses. This has meant that to get onto the general ballot, candidates had to appeal to a majority of their party, which may not be a centrist position at all. Whereas the old party bosses would be thinking about who would be able to win in the general election. So there has been an increase in divisiveness over the last few decades. In the general elections, voters are faces with candidates chosen in primaries such that none of them represent the majority of all voters, and so no choice is possible which represents a majority of voters.

This goes back to what I was saying about game theory, and controlling the order of voting and other details of the voting process. It is possible to manipulate the results that way. Should we go back to the caucus system? Maybe we should look at that. Should the U.S. adopt a European style parliamentary system? I don't think it could be done so the question is a bit moot. But I have not seen anything coming out of Europe that even argues for it. In particular, I think the idea of a unified monetary policy between such diverse countries is a move back toward a gold-like standard, where the Euro replaces gold, and the 20th century and all its bloody wars was basically a process of learning economies could not be managed that way. And the European parliaments voted for it. Except Britain of course.

This discussion turns very interesting.

I suppose most people will agree that political representation is about making constituents’ preferences present in politics. This is not directly the same as saying that politicians should represent constituents’ preferences. However, it is definitely interesting to know whether politicians actually do what constituents want. In a number of research articles mentioned in this discussion (and on my RG site) my coauthors and I observe significant divergence of politicians from the majority’s preferences. The reasons for this divergences are not fully clear to us yet, but we are working hard to identify some of them: party affiliations and interest groups will certainly play an important role.

Regarding party affiliations we think that we have some interesting results which do not only matter for European countries but also for the United States as well as other democracies. Electoral systems determine the role party affiliations play in political representation. Congruence between politicians and the majority of their constituents should be relatively independent of individual party affiliations under majority rule while congruence should be influenced by individual party affiliations under proportional representation. Employing quite unique Swiss data where we can directly match constituents’ preferences with politicians’ decisions in parliament, we find that these hypotheses broadly hold (see also my earlier post; here is again the video link http://www.youtube.com/watch?v=m6mAggnzc0A)

From a policy perspective, such results do not imply that parties do no matter under majority rule while they would only matter under proportional representation. Our results rather show that even politicians elected under majority rule diverge from the majority of their constituents but their individual distance (either to the left or the right) with respect to the preferences of the majority of constituents does not systematically depend on their party affiliation as opposed to politicians elected under a proportional system. Put differently, politicians from both majoritarian systems (e.g. United States) and proportional systems (e.g. many European countries if we include also mixed electoral systems) diverge from their constituents’ preferences but absolute (!) divergence in a majoritarian system does not strongly depend on party affiliations.

Conference Paper Preference Representation and the Influence of Political Par...

It would certainly be interesting to come up with a theoretical framework to explain the observations, as it sounds like that's where David et. al. are headed. I would suggest one place to look is the change in party candidate selection in the U.S. and the hypothesis that moving to direct voting by party in primaries caused a greater disparity in preference representation. Of course there will not be the same crisp data as in the Swiss case.

Additionally, I would just caution that while a researcher may have clearly in mind the limitations of the scope of his or her research, it may be misapplied by others. It's almost certain to be in fact and we see this in all fields, but in politics it can have great consequences. In other words, be sure others don't assume as I incorrectly did that you were meaning to imply alignment of preferences was a principle measurement of the quality of governance. Only one factor. Constituent preferences can change quickly, and be based on incomplete information and uninformed feelings, and often do not consider long term consequences of a particular preference.

WORKING HYPOTHESIS?

Here is one of those ideas that comes up while jogging at the gym. It may be more in line with what you are asking. Taking a cue from David's results and my own anecdotal evidence from over the big pond, I see two general factors that could be supposed to influence directly the likelihood a representative agrees with a majority of constituents on a particular issue:

1. The fraction of the constituency that is required to elect a representative.

2. The number of uncorrelated preferences.

Examples:

Majoritarian - if there is one principal issue, presumably 50% of constituents will agree. If two uncorrelated issues on which the candidates take positions at random, that could be cut in half, and so forth as issues are added.

Direct primary (American) - as long as party voters mostly vote as a block, then on just one issue there could be as little as 25% matching, though likely it would be more. For uncorrelated issues it should divide down as above. (liberal vs. conservative issues are correlated in most cases, but some issues are not)

Proportional - I'm not very familiar with this since it is not used here, but if I understand you correctly, the matching could be as small as half the percentage representation of the smallest party, and again would divide down for multiple uncorrelated issues.

If you figure out a way to test this, keep me posted. Interesting question.

QUALITY OF GOVERNANCE - I wanted to keep this a separate post since it is less crisp, but I don't see how you can consider preference representation without considering quality of governance since they are probably correlated in unexpected ways!

A Wikipedia article lists several definitions of governance, NONE of them quantifiable. An identical summary is present on the World Bank site. I found a 2012 paper by Rothstein & Teorell http://iis-db.stanford.edu/docs/623/Rothstein%26Teorell2012.pdf that quotes an IMF 1996 quasi-definition: "promoting good governance in all its aspects, including by ensuring the rule of law, improving the efficiency and accountability of the public sector, and tackling corruption, as essential elements of a framework within which economies can prosper." I'll return to this in a minute, but since there is not a consensus I'd like to start from scratch and enumerate some alternatives:

We can take a "natural" approach to QOG as preference representation, but then if divisive preferences exist, the government may get nothing done which endangers economic prosperity and possibly the mutual defense. Europeans seem to take a "dissolve the government and hope the people change their preferences" approach, while Americans have no provision except to stand and fight. And on one occasion in 1861 it did come to fighting.

We can take a "get things done" approach, how long do issues remain before the legislature (Americans flunk that one), how many unresolved issues, how often do policies have to be revised, etc. During the Civil War, with the South not represented, the North got a lot done. A path was selected for a transcontinental railway and construction begun. The disastrous Era of Free Banking was ended and groundwork laid for what later became very effective central banking. And slavery was abolished. While the South would have followed different policies (choosing a more southern path for the railroad, keeping slavery), one has to admit these three things brought the country up to then-modern European standards and led to prosperity. In the ancient Near East, when wars developed, democratic legislatures were abolished and a strongman elected absolute ruler in order to make quick decisions on defense. (sometimes it was hard for the legislature to re-assert itself afterward)

Notice that in the get-things-done approach I drifted into an economic evaluation. Perhaps it says something about my character that I didn't drift into a moral evaluation. But countries in good financial shape are able to defend themselves, and have the leisure to explore moral issues.

The IMF definition's mention of accountability could be construed as related to the topic of representation of preferences. It's not exactly what they meant. But really that's what the question of this thread gets at ... how and if representatives are accountable to the preferences of constituents.

The IMF's efficiency could boil down to "getting things done" or to economic prosperity.

The IMF's "rule of law" and "corruption" points could be taken as pointing toward fairness and protection of minorities, which I mentioned in one of my posts earlier. I think it should be a point in quality of governance, but perhaps not the main and certainly not the only point. If the country is not prosperous or capable of defending itself, nothing else matters.

Well then, if nothing else matters, then perhaps quality of governance should boil down to is the country prosperous and able to defend itself? And then add the fairness point about corruption (which will benefit prosperity) and rule of law and protection of minorities (which I suspect but cannot prove will also benefit prosperity, and almost certainly make citizens more willing to defend the country).

The field seems wide open for defining and quantifying quality of governance. If it can be done, then studies on things like representation of preferences could be evaluated quantitatively against quality of governance.

There may not be absolute answers on every issue. For example, liberty vs. prosperity. How much liberty is one willing to give up for prosperity? Certain countries, like China, seem to be actively pondering this trade off.

@Robert Shuler: In addition to direct primaries, gerrymandering of districts has amplified the distortion of "majority preference" and the move away from centrist candidates ... or at least away from sitting centrist politicians from casting centrist votes. Add to that the as yet uncertain impact of Citizens United unleashing even more money from both the left and right from *outside* the represented district into the politics of primaries and quantifying the "why" becomes even trickier!

May be their party preferences are dominat over their own personal promises to people and their preferences

@Dean, yes I agree.

I'll point you to the public choice literature.

Above all, you should look to Downs' An Economic Theory of Democracy (1957). There, he points out the rational ignorance hypothesis. Rational ignorance occurs when the cost of educating oneself on an issue exceeds the potential benefit that the knowledge would provide.

So, it's very hard to know all about the opinions and actions of a single politician. Imagine being informed about all the issues.

It's probably one source of divergence between politicians and constituents' preferences.

Felipe, thanks for the interesting reference. I found it, and also a paper which is available for download. Don't know if it includes all the details of the book of course.

http://www.hec.unil.ch/ocadot/ECOPOdocs/cadot2.pdf

Hi.

As long as you stay in a Public Choice frame you might continue wondering about the lack of constituent's representation in the political arena (problem of preference aggregation, aka Social Welfare Function). As soon as you broaden the scope and look at Public Policy research three problems arise: the policy-making process, the interests of collective political actors, and a double principal-agent problem between "the people" politicians, and the administration.

The policy-making process is no straight route from idea to regulation (Howlett, Ramesh, Perl 2009). Besides the individual citizens you have corporate actors you also try to see their preferences "represented". Political actors furthermore enter the policy-making process in two arenas: they have to be responsive to the public and they enter bargaining and negotiation processes in political arenas (so called nested games, Tsebelis 1992). Policies and regulations are mostly designed in groups like committees. Here all kinds of "games" are being played between the actors (Scharpf 1997).

An additional constraint is the fact that politicians are usually professionals who are in 99% of all case (my estimate) organized in political parties. The parties are not just in charge of aggregating interests, but - and a similar remark has been mentioned before - follow their own agendas. Or better, they try to influence public agendas. And, to pick up Felipes reference on Downs, they try to maximize votes. (Ergo focus on median voter and so on.)

The principal-agent theory claims that tasks are delegated when this can only be performed by the principal at prohibitively high costs. This in turn means that the principal can never fully control the commissioned agent, because it would take him or her as much time and resources to undertake the control as it would do to the job him- or herself. Thus, there cannot be full control over the agents actions. (Trust plays an important role here (Easton 1975). But to implement policies policy makers become the principal vis a vis the administration (Howlett, Ramesh, Perl 2009:167-168). Thus, they are in "double trouble".

And what is "representation" anyway? To "stand for" or to "act for" (Pitkin 1967)?

Another good reference in line with the public choice approach is Olson's Logic of the collective action.

Olson looks why small groups are able to make their preferences prevail over the preferences of the majority. It happens when we have benefits concentrated in a minority with the costs diffused.

So, we have constituents working for interest groups instead for the interests of the majority.

http://en.wikipedia.org/wiki/The_Logic_of_Collective_Action

http://economics.about.com/cs/macroeconomics/a/logic_of_action.htm

@Felipe, thanks again for a link to interesting literature. While I find it "interesting," I do not find this argument compelling as I did Downs'. And the fact that huge groups like "all wage earners" and "farmers" have achieved Congressional price floors seems to rebut the examples.

Politicians diverges from their constituents' preferences because they "can." They have the power to diverge when divergence does them no harm due to the political system or the political climate of the times.

The issue has been studied extensively in the research area of politcal economy or public choice respectively.

http://en.wikipedia.org/wiki/Political_economy

http://en.wikipedia.org/wiki/Public_choice

Single members elected as representatives of their electorate are influenced by many other constituencies. Their party they belong to, political lobbies of any kind. Money transfers to parties, to single representatives by lobbies influence their decision up to corruption. So the ideal world of a representative of their electorate is a weak representation of the decision making process.

The legal principle that representatives are only responsible to their individual personal conscience makes them unaccountable for following some other influencess instead of their constituency. There is no imperative mandate binding them to their constituency interests.

To overcome these weaknesses political scientist recomment more direct decisions through plebiscites. However, public opinion can be manipulated as well by media compaigns in particular, when the impacts of a decision is complex and not easily understandably for not well informed people. This leads to a dilemma of decision making. Are representatives bias by outside influeces, are they better informed to make a decision based on expertise or is a plebiscite more efficient due to the principle of the wisom of the crowds? There is no easy general answer for this.

As someone who has actually run for political office (DISCLAIMER: I have never won an election), I think the disconnect between idealism and realism. Trickle down economics is a perfect example- ideally, if you give more money to the wealthy, they take the extra money and invest it in things that will create jobs and improve the quality of life for all, with a net benefit to all. The reality is that people are greedy and selfish, and many wealthy people will just keep the extra money.. (Pope Francis had a great analogy- instead of the cup running over, the cup just gets bigger.) So, you are an idealist that gets elected on such an idealistic platform, and to find out, once you are on the inside, that your idealistic assumptions were dead wrong. You then have one of two choices- pursue the flawed idealistic policy that your constituents elected you on, even though you know it will fail, or pursue a realistic policy that will be a success that is different than the idealism your constituents expect. As a longtime government employee, I would spend countless hours explaining to citizens that solutions are not necessarily as simple as they may appear from the outside, as there is information that the voters do not have access to that the policy makers do.

Dear David,

It is impossible to imagine democracy (representative democracy, if we want to use the term) without political parties. Politicians and political parties are therefore essential 'evils'. The most important purpose of a political party, however, is to serve as a link between the people and the state. They are to take people's dreams to government machinery to help the government formulate programs to realise them. Some of the dreams may be outlandish and impossible. But most of them relate to improving the conditions for better living standards and these are achievable.

However, in practice we see the politicians and parties becoming 'power' hungry and forgetting the very purpose for which they came into being in a democracy. The dereliction of their duty to their constituency becomes more pronounced due to reasons some of which I would like to list below:

1. Illiteracy in the constituency

2. lack of enlightened community leadership

3. Apathy in government machinery

4. Ineffective or corrupt Judiciary

5. politicised police force

The above mentioned do not come into existence by them selves. The citizens also have a huge responsibility in making these organs and their politicians defunct. Sometime back, while interacting with a friend who had some political connections, he mentioned a politician as having said: "well, if you believe that I am bad and corrupt, why do you vote for me? There are over 40% educated and working class people in my constituency. Why not put up a good, clean candidate from amongst them?"

The above dialogue has an obvious answer to your query. I hope this helps!! My apologies, I cannot quote any studies on the subject yet.

Here is a new article highlighting the importance of electoral systems for congruence: https://www.researchgate.net/publication/311422322_Preference_Representation_and_the_Influence_of_Political_Parties_in_Majoritarian_vs_Proportional_Systems_An_Empirical_Test

Article Preference Representation and the Influence of Political Par...

Campaign promises are aspirational. They are designed to lure votes from as many constituents as possible. The campaigner knows that he/she will be competing against the interests of 535 members of Congress if he/she is President or 534 members of Congress and a Presidential veto if he/she is a member. Odds favor the failure to be able to satisfy the promises made to most constituets.