Divided we stand: The Polarizing of American Politics

by Steven Hill, National Civic Review, Winter 2005

Editor’s note: this article was first published in January 2006, but it’s sharp and incisive analysis of what is causing political polarization in Congress and in the US is more relevant than ever. In fact, it’s as if nothing has changed over the last eight years.


A recent Zogby poll found that 70 percent of Americans believe both the Democrat and Republican parties should be broad-based and pursue compromise rather than polarization.  That same poll found that a solid majority believes the two parties are too focused on their respective base voters and that, as a result, compromise has become impossible in Washington, DC. Looking at recent partisan battles over Social Security, presidential appointments, Terry Schiavo, the culture war, the Iraq war and more, one can understand why so many Americans hold these views.  There appears to be no end in sight to this bare knuckles partisanship that substitutes for politics today. Amidst all the ink reporting on the partisan wars, however, what has been missing is recognition of how extreme positions on policy issues are incited by certain fundamental realities of our nation and our political system. Consequently it will be difficult to change without dealing with these root causes. The first reality is simply demographics — where people live, and how partisan demographics have been lining up in recent years along regional lines.  The infamous Red versus Blue America map only begins to encapsulate the consequences of this recent trend.


The second fundamental reality is how those regional partisan demographics are funneled as votes through our winner-take-all electoral system. The electoral process of electing our representatives one winner-take-all district at a time is carving our nation into red and blue one-party fiefdoms. Given the regional partisan demographics, in 90 percent of legislative districts at federal and state levels it’s only possible for one side to win because there are simply too many of one type of voters packed into that district (this effect is occurring outside whatever shenanigans take place as a result of partisan gerrymanders during the redistricting process). In combination, these two are greatly contributing to the toxic brew that we know today as partisan politics. This article will explore the subtleties and ramifications of these phenomena in greater detail and make predictions for the future of American politics.

Winner-take-all is making most of us losers

Under our winner-take-all rules, we elect one district representative at a time.  Whichever party’s candidate has the most votes, wins, and all other parties and their candidates lose.  Hence, the term “winner-take-all,” because one side wins all the representation while everyone else wins nothing. Of course, this is how our system has operated for a long time, but over the last 15 years something new and different has begun to occur. In particular, three factors are changing the course of politics.


First, in this era of “Red” versus “Blue” America, the nation has begun to balkanize along regional lines with heavy partisan overtones.  Like other large winner-take-all democracies, such as India and Canada, entire regions of the U.S. have become one-party fiefdoms.  The Democrats control the cities, most of the coasts and sizable chunks of the Midwest, while the GOP dominates rural areas and most of the flyover zones between the coasts. On the other’s turf, it is difficult for the minority party to win any elections, leading to not only a loss of political competition but even a minimal debate of important issues.


Note that this is not the result of partisan gerrymandering during the redistricting process, which some observers cite as an explanation for a lack of political competition.  No, this is something new and different in which partisan residential patterns — where people live — is outstripping the ability of the mapmakers to greatly affect the outcome of elections. Each legislative district, whether at federal or state levels, has been branded either red Republican or blue Democrat before a single vote is even cast, or before the partisan line-drawers sit down at their computers, purely because of where people live.


Also note that this phenomenon is not happening uniformly in every state.  Certainly in states like Texas, Florida, Ohio, Pennsylvania and Michigan, partisan gerrymanders have impacted election results since 2001.  But in most other states, even states like Arizona, Iowa and Washington which employ independent redistricting commissions of one sort or another, regional partisan demographics have bedeviled any attempts to create more competitive elections.  In most states, it turns out, demography has become destiny.


If the country is now politically Balkanized along regional lines, this raises a question of whether or not American voters themselves are also polarized into partisan camps.  The answer, perplexingly, is “yes and no,” which leads us to our second significant change over the last 15 years.


The University of Michigan’s National Election Studies, a series of public opinion polls taken over the last three decades, gives an interesting snapshot of political attitudes since the early 1970s.  For a whole host of questions, including on abortion, government intervention in the economy, the size of government, the amount of services government ought to provide, affirmative action, the desirability of a national health plan and more, the general public has barely budged in its attitudes — except in two of these areas, where Americans indeed have become more split since the mid-1980s. The first is over government spending, where more Americans today are opposed to big government and higher taxes; and the second is over government aid to African-Americans and minorities, i.e. affirmative action, with more Americans today opposed than previously. And in the public mind these two are closely fused, half of the NES respondents now agreeing that “the government should not make any special effort to help blacks because they should help themselves,” up from a third of the population in 1984.[1]


A brief outline of American electoral history for the past 35 years reveals how these two interrelated domains strongly reflect the degree to which national politics and partisan competition have become centered around appeals to culturally and racially-conservative white voters, who still comprise the bulk of the American electorate. This attitudinal shift once formed the basis of Richard Nixon’s successful “Southern strategy” in the early 1970s, which used coded words and symbolic gestures and actions directed at white people in the South, particularly white men, to paint the Democratic Party as the party of racial minorities, rioting cities and civil rights agendas. Ronald Reagan continued the Nixon strategy, peeling off white Democrats, particularly in the South, with his attacks on race-based policies, government spending, and stereotyping of cities. President George H.W. Bush used his infamous Willie Horton ads and coded language to finger Dukakis and the Democrats as the party of liberals, big spending and, again, as being “soft” on crime, i.e. racial minorities.  New Democrat Clinton hoisted his finger in the air and, taking note of the political winds, tried to expropriate parts of the GOP strategy by taking public stances that distanced the Democratic Party from the racial tag, with high-profile dissings of rapper Sister Souljah and Reverend Jesse Jackson.  Even today, much of current U.S. politics still can be explained by the dynamics unleashed during this not-so-distant era by Nixon’s tactics.


So the situation we are left with is this: along two key axes, that of big government/higher taxes, as well as government-sponsored race-based affirmative action programs, the public indeed has become more polarized. And more conservative, prompting the GOP to aggressively exploit this shift. Add to this picture the recent GOP targeted attacks on cultural issues, such as gay rights, abortion, prayer in the classroom, Terry Schiavo and more, and you have the major axes upon which American politics turn today (though the cultural issues are not necessarily reflective of a shift in American attitudes, Americans have always been queasy about gay marriage, and overall Americans show little shift in their attitudes about abortion.  What it reflects is a newfound willingness on the part of the GOP in this era of the 49-49 nation/Red vs. Blue America to exploit these issues in an effort to mobilize their base).


The third significant change over the last 15 years, greatly related to the previous two, is the decrease in ticket splitting among voters.  It was once common for voters to vote, say, for a Republican for president and a Democrat for Congress.  For decades, Southern white voters would pick a Republican like Nixon for president but a Southern Democrat like Sam Nunn for Senate or Sam Ervin for the House; Montanans would vote for Ronald Reagan as well as Democrat Majority Leader Mike Mansfield for U.S. Senate.  But that doesn’t happen as much anymore, voters have become extremely predictable in their voting patterns — by region, by office, and by party.  Republican voters are not likely to vote for a Democratic candidate these days, no matter how dissatisfied they may be with the Republican candidate, or vice versa. The two sides are more intractably apart than ever, with a smaller proportion of voters straddling the middle between the two parties.

Polarize, Not Compromise

This tidal shift regarding the three factors of 1) regional partisan Balkanization, 2) voters’ more conservative attitudes about big government and racial programs, and 3) an increase in voter consistency along partisan lines (i.e. a decrease in ticket splitting) — and all of this expressed as votes through the winner-take-all electoral system — has given rise in recent years to an audacious GOP strategy.  It can be summed up as “polarize, not compromise.” Rather than engaging in Bill Clinton-like triangulation in which candidates step to the middle to woo swing voters — supposedly one of the moderating influences of our winner-take-all system — the Republicans have figured something out:  in a winner-take-all system, they don’t need to win everybody’s vote, they just need one more vote than their opponent. And sometimes they can increase their vote totals more easily by mobilizing bigger chunks of their base voters than by bending over backwards to woo undecided swing voters (as long as the efforts to mobilize their base do not alienate too many swing voters).


The winner-take-all system is reminiscent of the following joke: two hikers are walking in the woods when suddenly they encounter a hungry bear.  The bear growls. One of the hikers calmly bends down and reties her shoe.  The other hiker, quite agitated, says: “What are you doing?  We gotta get out of here, we have to outrun that bear!”  To which the other hiker calmly replies, “No I don’t, I just have to outrun YOU!” and then jumps up and dashes off, leaving her companion in the dust.


So too with a winner-take-all system: you just have to be the last one standing. And with the 49-49 nation, with red and blue America so evenly divided, a political party can do that by either mobilizing more of their base voters or by convincing undecided swing voters. Or a bit of both, so long as the attempt to mobilize more base voters does not alienate your swing voters, or vice versa. Compared to even five years ago, GOP leaders have clearly embarked on a strategy that relies more on mobilizing their base voters, as opposed to Clintonian triangulation in which Democrats in the 1990s tried to capture undecided swing voters (“soccer moms” and “NASCAR dads”, for example) by co-opting certain Republican policies. The GOP attempted a bit of that swing voter serenade in the late 1990s with their re-labeling of George W. Bush as a “compassionate conservative.”  But increasingly the GOP strategy seems to be to mobilize its base. Hence, the incentives now are to polarize, to sharpen the differences between the parties, rather than compromise since that might blur those differences. In this new era, that’s the way to win elections, figure the GOP strategists like Karl Rove.  And their strategy seems to be paying off.


So the incentives to compromise, supposedly an advantage of our winner-take-all system, are largely missing in action at this point.  And it is the winner-take-all system itself that is exacerbating the problem, because given the regional partisan demographics it allows the electoral map to be easily carved up into solidly red or blue legislative districts, or in a presidential election into solidly red or blue states.  Party strategists then can target their resources to the handful of districts or states where there is likely to be a close race.  But there are fewer and fewer of these up-for-grabs swing districts and states in recent years.  In the 2004 elections there were perhaps 30 out of 435 House seats — less than 7 per cent — considered competitive, and perhaps a dozen swing states out of 50 in the presidential race (though many pundits, including myself, were saying six months before the election that John Kerry would win if he captured either Ohio or Florida, meaning an effective battleground of two states). Of over 7000 state legislative races, 40 percent of them had no candidate from one of the two major parties because the district was considered a waste of time and resources for the minority party. That’s two of five races for which there was not even nominal competition, or any campaigning or debate of ideas, to interest voters.  And most of the remaining races were won by huge landslide margins.


In this kind of political landscape, the strategy of “polarize, not compromise” makes perfect sense.  It is following the logical trajectory of the incentives of a winner-take-all system, given the regional partisan demographics in red vs. blue America, voters’ more conservative attitudes toward government spending and taxes, and the increase in voter’s partisan consistency.

Will Democrats begin mimicking GOP strategy?

Over the years, Democrats and Republicans have been notorious in ripping off each other’s successful tactics and issues in their bid to win elections. Bill Clinton was a master at co-opting Republican issues as a way to win votes.  And Republican strategists soon caught on to the tactic of soft money ads brilliantly advanced by Clinton and his consultant Dick Morris for his 1996 re-election over Bob Dole.  So will the Democrats similarly follow the Republicans in adopting the tactic of “polarize, not compromise”?


Certainly there have been discussions of this sort within Democratic Party circles.  These sorts of trends have been floated as one reason why a polarizing candidate on the Democratic side like Hillary Rodham Clinton actually might beat a Republican candidate like John McCain or Rudy Giuliani. However, there are striking demographic differences between the GOP base and the Democratic Party base that make it unlikely that the Democrats can succeed by simply copying the GOP strategy.


For starters, the Democrats’ base is more complicated than the GOP base.  The Democratic base is largely composed of a diverse array of minorities, young people, and unions who want the government to lend a helping hand, whether for affirmative action, lower tuition, health care and pensions, greater regulatory enforcement, and a host of other ways that can only be realized by an activist government.  All of these mean government programs, which means government spending, which equates in many peoples’ minds to higher taxes.


But that desire conflicts with the Democrats’ need to not alienate and lose swing voters in their bid to mobilize their base.  Swing voters tend to be more fiscally conservative and not very welcoming of higher taxes or more government spending. They are part of the group of Reagan Democrats that Bill Clinton tried to woo back with his pronouncements in the early and mid-1990s along the lines of “the era of big government is over.” Thus, the Democrats attempt to motivate their base conflicts with their need to not alienate swing voters in the process.


Republicans, on the other hand, can motivate their base with an infinite array of cultural and religious issues, whether Terry Schiavo, gay marriage, abortion or school prayer.  Those sorts of issues cost relatively nothing in terms of government taxes or programs, and consequently they don’t alienate the fiscally conservative swing voters.  The swing voters may not completely agree with Republican leadership on these cultural matters, but that disagreement does not greatly motivate their voting choices the way that higher taxes do.


An integral part of this GOP strategy is a keen awareness of the intensity with which voters hold a particular belief, or agree or disagree with a particular issue.  As key GOP strategist and NRA board member Grover Norquist has said, in commenting on the fact that national polls show that a large majority of Americans want gun control, “Who cares?  The question is intensity versus preference. You can always get a certain percentage to say they are in favor of some gun controls. But are they going to vote on their ‘control’ position?”


“Intensity versus preference” is a key cornerstone concept that many pundits and political analysts on the Democratic Party side seem to ignore or not understand.  The fact is, fiscally conservative swing voters may roll their eyes at the extremes of the Republican Party and their ploys to throw red meat to their party base, but that does not decide their vote the way higher taxes and distaste for big government do. They will vote against a John Kerry, pegged as a liberal from Massachusetts, because they do not trust him in terms of the federal budget, spending priorities, taxes or minority programs, regardless how nutty the wings of the Republican Party act. While the NES study reveals that Americans’ attitudes have changed little on most issues and topics, the areas where they have changed and become more polarized — over big government/higher taxes and minority programs — tend to have produced intensely held viewpoints that are predictive of their voting preferences.


So in this gambit the Republicans have the advantage, at least for the foreseeable future, because when it comes to the all-important issue of big government/higher taxes and minority programs, the GOP base has more of a similar attitude with swing voters than the Democrats’ base has. The Republicans enjoy a bridge between their base and swing voters that the Democrats do not have access to. At some point the Republicans run the risk of going too far in motivating their base, risking the type of backlash that resulted in the early 1990s from the antics of cultural conservatives such as Pat Buchanan and Marilyn Quayle on national TV at the 1992 Republican National Convention. That backlash pushed enough swing voters over to the new kid from Arkansas saying, “It’s the economy, stupid.” But that was 13 years ago, when the trends I am identifying were not as far along.  Also, Republican strategists have learned a lot since then about public relations and messaging. The Pat Buchanan’s don’t make it to the microphone anymore, at least not during national conventions.


Also, say what you want about George W. Bush, he has proven to be extremely adept at appearing as a “compassionate conservative” for a national TV audience, but when the cameras aren’t paying close attention at private meetings and GOP rallies he can talk like a fundamentalist preacher or cozy friend of the rich.  The GOP has become extremely sophisticated at such targeting techniques, and this has allowed them to mobilize their base while not greatly alienating either a national audience or fiscally conservative swing voters.


Republican strategists have become very professional at using these divisive winner-take-all tactics to win elections, and they are quite comfortable with winning by only “one more vote than the opposition.”  The Democrats have been caught flat-footed, with little capacity to respond, in large part because the dynamics of winner-take-all elections combined with the regional partisan demographics and voters’ more conservative attitudes regarding big government/taxes and minority programs put them at a decided disadvantage.


Given that the tactic of “polarize, not compromise” is working so well for Republicans, we can expect more partisan division and little compromise in the years ahead. We can expect more polarization, more Balkanization, which is kind of odd because that’s not what the political science textbooks taught us that the winner-take-all electoral system was supposed to produce.  Winner-take-all is supposed to produce moderation, centrism and compromise.  Yet anyone looking at American politics in recent years can see this is not the case.  What happened?  Could it be that the textbooks were wrong?

Faulty Textbooks:  The Strip-mining of Anthony Downs’ Economic Theory of Democracy

Anthony Downs’ heralded Economic Theory of Democracy is a seminal work often cited since its 1957 publication by the captains of American democracy as one of the theoretical underpinnings for the political centrism and moderation allegedly resulting from the American system. Downs, an economist, used game theory and spatial equilibrium models to say some interesting things about competition in a winner-take-all, two-party system.


According to the Downs theory, a two-choice/two-party system — considered a relative proxy for a winner-take-all electoral system which tends to foster two major parties — will cause the two parties to move toward the two ideological points where the most numbers of voters are amassed. Of particular interest to Downs were two specific voter distributions: 1) when most voters are amassed at the ideological center, i.e. the median-middle (illustrated by a one-humped curve), and 2) when voters are polarized into two ideological camps to the right and left of the center (a two-humped curve).


In the first instance, Downs posited that, given a voter distribution where most voters are amassed at the middle of any particular political spectrum, the two political parties, because they are vote maximizers, will tend to converge in the center where most of the electorate is concentrated.  Both parties will tend to adopt the views of what is called the “median” voter — the voter closest to the middle of the political spectrum. That’s because the strategic incentives for the two parties, and the rational choices for voters picking the best of what’s available in a limited two-choice field, act together to provide victories for the party that is closest to the median.[2]  A party that fails to converge to the median can always be defeated by a party that does (with one important addendum, that fear of losing their own extremist voters would keep the two parties from drifting completely to the median-center, “becoming identical”).[3]


Thus, according to this Downsian model, what has been called the “tweedle dee” or “tweedle dum” choice of near-identical Democrats or Republican candidates actually is rational behavior by politicians trying to win elections in a two-choice field.  Given a “one-humped” distribution of voters amassed in the middle of the political spectrum, if either of the two political parties stray too far toward the margins, they will yield more of the political turf to their opponent. So both parties stride toward the middle. This reality has given rise to a classic political strategy: in the primaries run to your base, but in the general election run to the middle — different elections with different median voters, with primaries determined by median voters considerably more partisan than in a general election.  Certainly Bill Clinton and the “compassionate conservative” version of George W. Bush in 2000 masterfully utilized such a strategy, and the Downsian spatial equilibrium model purports to tell us why.


But Downs proposed an important exception:  what if, he asked, the distribution of voters is not humped in the middle of the ideological spatial market? What if instead the electorate is polarized?  In that case, Downs showed, the two parties will “diverge toward the extremes rather than converge on the center.  Each gains more votes by moving toward a radical position than it loses in the center.”[4] When the electorate is polarized, Downs’ model portrayed the distribution of voters as two-humps, equal in size, like a Bactrian camel’s back. In such a situation, Downs goes on to predict that “regardless of which party is in office, half the electorate always feels that the other half is imposing policies upon it that are strongly repugnant to it” (that prediction should sound familiar to many diehard supporters of both the Democrats and Republicans today, who are convinced that the other side is tantamount to evil).  But in addition, Downs wrote “In this situation, if one party keeps getting reelected, the disgruntled supporters of the other party will probably revolt; whereas if the two parties alternate in office, social chaos occurs, because government policy keeps changing from one extreme to the other.”[5]  Thus, Downs concluded, the two party system “does not lead to effective, stable government when the electorate is polarized.”


Downs’ work was enormously influential over the ensuing decades. His rational-choice formulation of politics permeated throughout the academy, even the culture at-large. The pandemonium known as politics had been modeled and reduced to “natural law,” just like economics or physics, in a way that seemed to make sense out of the chaos. It is hardly possible to overstate the influence of Downsian theory, cited and re-cited by political scientists, judges, journalists, pundits, politicians and editors. Testifying to its enduring clout, a search of Downs’ Economic Theory of Democracy in the Social Science Citation Index turns up over 1000 entries in the 1990s alone.[6]


But in the ensuing four decades following Downs’ seminal work, various neo-Downsian editors, pundits and politicos, and even researchers in the academy, selectively strip-mined Downs for their own purposes.  Their interpretations of the Downs model ignored what apparently did not fit in with their preferred vision of America as a nicely-organized, single-humped electorate.  Thus, Downs’ formulation of what happens to a democracy when the electorate is polarized into two voter distribution humps somehow was left on the cutting room floor.


What remained was a half-baked simulacrum of the Downs theory, and the two-party system transmogrified in many minds into a proxy for moderate, majoritarian, centrist government.  Down’s median voter acquired a reputation as a political moderate, and even more as an undecided, i.e. a swing voter. And the policy passed by that government was presumed to be preferred by the majority of voters. Centrist government had been more or less achieved, at least theoretically according to the neo-Downsians, and unsurprisingly it looked quite a lot like two-party government in the United   States.


But this was certainly a misreading. Downs’ himself had postulated an important exception: except when a polarized electorate leads to a more difficult brand of democracy. For Downs, the behavior of vote-maximizing parties was determined by the distribution of voters, not vice versa. One of the strengths of Down’s model was that it was flexible, and examined the reaction of political parties pursuing votes from different voter distributions. Yet an unsubstantiated assumption had crept into the neo-Downsians’ revisions of the model, from the perspective of policy determination: like a kind of political conveyor belt, the two-choice dynamic, in aggregate, was presumed to drag both parties inevitably to centrist positions of moderation, leading to a politically centrist, moderate, majoritarian government — regardless of the distribution of voters.


The “polarized electorate” exception inexplicably had been dropped from the neo-Downsian canon, but politics in the real world is not always very cooperative or kind to those who try to bend theory to their will. The outstanding question was: which distribution of voters was applicable to American politics?  The “one-humped” majority of voters amassed around the ideological median-middle?  Or a “two-humped” polarized distribution of voters?  Or perhaps a distribution somewhere in between the two extremes, or one that fluctuated between the two extremes over time?  There was a one-hump vs. two-hump uncertainty to figure out.  Only after answering that question could anyone begin to speculate whether the politics of policy-making would lead to moderation and centrism, or polarization and paralysis.


When USA Today’s red versus blue America map first was published following the 2000 presidential election, for many political observers — particularly those more honestly and soberly contemplating the Downsian model — that map of the national vote caused the hair on their necks to raise, precisely because of what it conveyed: a national two-hump distribution of polarized voters, cleaved along starkly regional lines. It was like some sort of two-humped Godzilla monster, rising up out of the sea.


But the same two-humped Godzilla appears when analyzing congressional elections as well. Political scientist Elizabeth DeSouza re-evaluated the Downs rational choice theory and some of its most basic assumptions for U.S. House races. In particular, she re-evaluated the neo-Downsian notion of the distribution of voters, challenging the depiction of the median voter as a neutral, undecided or swing voter in congressional elections. DeSouza’s model included the more real-world condition that most congressional districts actually were noncompetitive safe-seats, favoring one party or the other by landslide margins.  In examining House races from 1974 to 1996, DeSouza found a strong link between vote shares and partisan legislative behavior:  the safer the seat, the more partisan the legislator.  All other things being equal, a legislator in a safe district was more likely to vote positions corresponding to the party’s wings, not the center, compared to a legislator from a competitive district.  It is only as vote shares decreased i.e. as races became more competitive — fewer and fewer such contests every year — that representatives’ positions gravitated toward the political center.


In the modern era, where 70 to 80 percent of U.S. House seats are won by overwhelming landslide margins, and 90 percent are won by at least ten-point victory margins due to natural partisan demographics exacerbated by partisan gerrymandering that create huge numbers of safe seats for one party or the other,[7] “very few [House] members will have any incentive to moderate their positions,” and “the logic of party convergence collides with the unmistakable party wars,” DeSouza concluded. The overwhelming numbers of safe seats, as DeSouza pointed out, gave political parties and their candidates “the political equivalent of academic tenure without the tenure process,” and produced a lack of accountability that in turn produced “the rise of the ideologue.”[8]


Yet neo-Downsian rational choice models based exclusively on a single-hump voter distribution had predicted the opposite pattern. According to the neo-Downsians, the “drive to the center,” toward median voters — toward moderate voters — was a proxy for a two-choice, winner-take-all system that on the whole produced centrist political winners and majoritarian governments of moderation.


One conclusion was glaring:  for the past two and a half decades since the mid-70s at least, Downs’ two-humped model had been slowly becoming the more accurate descriptor for most U.S. House races, with one partisan hump larger than the other, usually drastically so, depending on the district (i.e. in landslide safe-seat districts, which typify three-quarters of U.S. House races). DeSouza concluded that “rather than having strong incentives for centrist politics, we have a system that favors partisan politics:  politicians are rewarded for staking out extreme positions because most owe their seats to ideologically skewed electorates.”


Thus, just as with the national vote revealed by the USA Today map of the presidential election, Downs’ two-humped model had become the operative voter distribution pattern for most congressional races as well.  Given the partisan shape of regional demographics in Red versus Blue America, combined with the dynamics of winner-take-all districts, where the highest vote-getter wins everything and every other vote-getter wins nothing, it should not be surprising that the U.S. House has split into two polarized camps, a more solidly liberal Democratic Party and a more solidly conservative Republican Party, like two tectonic plates drifting in opposite directions, with a dwindling number of moderates between them trying to bridge the gap.


Other recent studies and analyses have added further empirical evidence to DeSouza’s and others critiques of neo-Downsian dogma, not the least of which is the previously cited National Election Studies, which reveals a polarization in American attitudes about big government/higher taxes and minority programs, with many voters holding these views so intensely that it becomes predictive of their voting preferences. A great deal of empirical evidence and real-world behavior contradict any spatial voter distribution models that predict a two-party winner-take-all system will produce moderate, centrist or majoritarian policy.  As Harvard political scientist David King has written, “Political scientists may instinctively suspect that polarization is an irrational strategy for party elites running the parties. The party locating its policy positions closest to the preferences of the median voter is supposed to get the most votes, or so we have been taught. With that model in mind, it makes little sense to allow one’s own party to become extreme, but that is precisely what has been happening.”[9]


Downs’ original model, which included the possibility of a one-humped OR two-humped voter distribution, it turns out, was exactly right. But those who strip-mined the fullness of Downs’ theory had thrown their weight around for decades. Over time, their half-baked version of their mentor’s work acquired a vast degree of influence. Neo-Downsianism rose to the level of political dogma, indeed it became a kind of priestly orthodoxy. The dogma trickled down to popular culture; it was championed in the mainstream press, inserted into encyclopedias, taught in the universities and civics classes with a kind of nationalist pride, with the Wall Street Journal and New York Times acting as high priests of the canon.[10]  Typical is the claim of one political scientist, that our “exceedingly fluid two-party system” is built upon “tenuous compromises” between ideological factions and special interests “which are forced at every turn to moderate their claims.”[11] Even Supreme Court justices like Antonin Scalia and Sandra Day O’Connor, straining in some of their opinions to sound erudite, fell back on unsubstantiated neo-Downsian characterizations.[12]


In the end, Downs’ clever and provocative model had transmogrified into a stultifying stereotype:   everyone knew, it seemed, that the kind of government and winner-take-all political system used by the United States was the best in the world because it led to effective, centrist, stable, majoritarian government.  Those who still swear by winner-take-all’s alleged centrism would have us believe that today we can have a U.S. House of Representatives that is populated by more unabashed liberals and conservatives and fewer and fewer moderates, and yet somehow this two-humped Godzilla monster will find its way to enact centrist, moderate policy.  It would be a miracle if such were the case.


Given the partisan bent to regional demographics in Red versus Blue America, combined with more conservative voter attitudes about big government and minority programs, combined with a decrease in ticket-splitting, combined with the dynamics of winner-take-all elections where the highest vote-getter wins everything and every other vote-getter wins nothing, polarization makes sense as a winning GOP strategy, contrary to what the neo-Downsians would have us believe.  As our politics have fragmented and polarized, the fuller version of Downs’ model including the two-humped “polarized electorate” exception would have predicted such fragmentation and the distorted policy that results.


So the neo-Downsian revisionism is in trouble (presumably Downs himself has had to distance himself from these misinterpretations of his work; one can imagine a forlorn Anthony Downs proclaiming “I am not a Downsian”). Centrist, moderate, majoritarian government, it turned out, was not guaranteed at all by a two-party system.  In fact, under winner-take-all’s two-choice/two-party paradigm, it was somewhat of a crapshoot. And observations of national politics in recent years bear this out.

The Final Nail:  “Crafted Talk” and “Simulated Responsiveness”

For the two major parties, politics has become a tricky balancing act. Both major political parties are caught between the poles of their base voters and undecided, unreliable swing voters.  In the 49-49 nation, whichever political party can sufficiently stimulate both constituencies wins. An operation of slicing and dicing various constituencies of voters, that is simultaneously one of trial-and-error as well as one of extreme sophistication with polls, focus groups and data mining of voter behavior directed by highly paid political professionals, is now a mainstay of American politics. I’ve shown how the Republicans have settled on a new strategy emphasizing polarization rather than compromise as a tactic for mobilizing more of their base (without alienating too much of the middle), and why it will be difficult for the Democrats to mimic this strategy. But this does not mean that Republican nor Democratic strategists have abandoned their attempts to woo more centrist voters.


On the contrary, it is necessary to cast a gloss of centrism over their extremism, at the very least to prevent the undecided voters from swinging to the other party. These sorts of electoral behaviors not only further complicate the Downsian model but also can be easily manipulated in a way that further exacerbates tendencies toward polarization.


In their book Politicians Don’t Pander, political scientists Larry Jacobs and Robert Shapiro added a further challenge to the neo-Downsian notion that “competition for the median voter” would motivate parties and politicians to move to the center. Jacobs and Shapiro were trying to figure out how Americans could hold simultaneously two contradictory beliefs:  that, on the one hand, there is a public perception that the growing influence of opinion polls has increased pandering by politicians, and so in one sense increased politicians’ responsiveness to the public; but on the other hand there is a perception that the pernicious combination of money in elections and partisanship cause officeholders to ignore the wishes of the public in favor of pursuing the agendas of various special interests.  How can both be true, the researchers wanted to know?


Jacobs and Shapiro discovered the answer in their conception of “crafted talk” and “simulated responsiveness.”  Politicians indeed rely heavily on opinion polls, focus groups and the like; but more importantly, the authors showed that politicians use those electoral tools, not so much to pander or to be actually responsive but instead to find words and phrases to manipulate and mislead voters, using “crafted talk” designed to woo voters by “simulating responsiveness.”  Using the modern campaign technologies to rhetorically trick their audience about their alleged centrism, politicians and political strategists attempt to prevail even when they diverge from the sentiments of the voting public. Consequently, the views of a partisan few — party leaders, party activists who hold the most extreme ideological views, and their donors — are the real purveyors of actual policy, and successful politicians and strategists are fully capable of using centrist rhetoric to cover up their non-centrist policy when it suits their purposes. They call themselves things like “compassionate conservative” and “New Democrat” when most of their targeted policies are anything but compassionate or new.


Not surprisingly, said Jacobs and Shapiro, “We have found a dramatic decline of political responsiveness to the wishes and preferences of the public on major policy decisions.”  In fact, Jacobs and Shapiro found that in four social policy areas (social security, health care, welfare and crime) Congress in the 1990s was on the same page as the public only 36 percent of the time, compared to 67 percent of the time in the late 1980s.  This finding critically undermines Downs’ model.  If there was one thing fundamental to the entire model, it was that the behavior of parties and the policies they gravitated toward were completely dependent on the distribution of voters — whether one-humped or two- or three-, etc. But with a Congress that is now on the same page with the public only about 36 percent of the time, and with the use of “crafted talk,” and “simulated responsiveness” developed with the use of modern campaign techniques to fool voters into thinking that the party is more centrist and more responsive than it actually is, the connection between Downs’ distribution of voters has become increasingly tenuous. The Downsian connection still exists between median voters and campaign rhetoric, but under the influence of the modern campaign technologies rhetoric has become divorced from policy. A schism has developed between Down’s median voter, between what politicians say they will do to attract that voter, and what they actually do once in office.


Here we arrive at the crux of the dilemma that is utterly undermining any neo-Downsian incentives toward centrism. It is as if the two political parties have slickly substituted centrist rhetoric for centrist policy when it suits them. The Democratic and Republican parties today, their leaders, candidates and their mad scientist consultants, use modern campaign techniques like polling and focus groups to figure out what to say to voters. The sound bites, the contrived images, the permanent campaign, all of these are designed to produce Jacobs and Shapiro’s “crafted talk” and “simulated responsiveness” to dupe voters. Hence, the distribution of voters today — the very foundation of Downs’ model — is attached to campaign rhetoric but not necessarily to policy. If there are any Downsian incentives toward centrist policy, increasingly they are “simulated” ones.

Yet in a two-choice system, and with fewer voters willing to split their tickets, voters have very options when it comes to “tossing the bums out.” Thus, contrary to popular and even political science stereotypes, when we examine our current politics from both empirical and theoretical angles we see that winner-take-all need not — indeed, has not — led to moderate, centrist, stable or majoritarian government or policy.  The theorized neo-Downsian centrism has been undermined by “crafted talk,” “simulated responsiveness” and modern campaigning techniques that are used to hoodwink the voters.  When combined with the partisan bent of regional demographics in Red versus Blue America, and with the polarization in voter attitudes about big government and minority programs revealed by the National Election Studies; and with a decrease in ticket-splitting so that partisans rather than neutral voters comprise the decisive voters in most elections; and all of that expressed via votes cast in a winner-take-all electoral system that has produced a fundamental lack of competition in most lopsided legislative districts and presidential states, we end up with a strange brew of national politics that has become a formula for polarization and balkanization.


Despite ample opportunity in recent years, there has been little discussion by the national punditry about how the political mechanics and calculations of the two-choice, winner-take-all system drives such pointlessly adversarial politics.  But the logic of politics dictates a singular ambition: that you beat the other side. It is instinctual to the winner-take-all mechanism, like a pit bull trained as an attack dog, salivating to go on the offensive.  One of the defining characteristics of the two-choice system is that it promotes pointlessly adversarial politics so that on a whole host of issues it is painfully obvious that the overriding agenda for both major parties is not policy, principle or ideology, but that each side stake out positions vis-à-vis the other.  Increasingly, Republican strategists and leaders have bet that compromising is a losing strategy, but polarizing will win elections.


Unfortunately, the soundness of national policy gets caught in the crossfire. A poisoned atmosphere of polarization will continue to undercut attempts to pursue sound national policy in areas such as Social Security, health care, national security and the global economy.  In all of these areas and more, deformed policies have emerged as a result of the trends and incentives of our two-choice system identified in this article, further confusing an already disengaged and disgusted voting public which has nowhere else to go but the sidelines.  And as more and more voters abandon the field, the game is left to the partisans to wage their battles using the politics of polarization, which then fuels a new round of voter alienation and withdrawal, which in turn cedes even more influence to the partisans. It is a race to the bottom, and the full version of Downsian theory predicted this when Anthony Downs wrote that a two party system “does not lead to effective, stable government when the electorate is polarized.” Truer words have not spoken in a long long time.


Steven Hill is director the Political Reform Program of the New America Foundation and author of “10 Steps to Repair American Democracy” (www.10steps.net).



[1] See David C. King, “The Polarization of American Political Parties and Mistrust of Government,” Harvard University Working Paper, www.ksg.harvard.edu/prg/king/polar.htm.  King’s paper was included as a chapter in Joseph S. Nye, Philip D. Zelikow and David C. King, editors, Why People Don’t Trust Government (Harvard University Press: Cambridge MA, 1997).


[2] Downs assumes that political parties are motivated by winning elections, and that “to avoid defeat, the government must support the majority on every issue.” Anthony Downs, An Economic Theory of Democracy (Harper & Row: New York, 1957), p. 55. Previously, Joseph Schumpeter had made similar claims in his classic work, Capitalism, Socialism and Democracy, (New York: Harper and Row, 1942).


[3] As Downs put it, “We can turn Harold Hotelling’s famous spatial market into a useful device for analyzing political ideologies…This model confirms Hotelling’s conclusion that the parties in a two-party system converge ideologically upon the center, and Smithies’ addendum that fear of losing extremist voters keeps them from becoming identical.”  Anthony Downs, An Economic Theory of Democracy (Harper & Row: New York, 1957), p. 140.


[4] Anthony Downs, “An Economic Theory of Political Action in a Democracy,” Journal of Political Economy, Volume 65, Issue 2, April 1957, p. 143.


[5] Anthony Downs, “An Economic Theory of Political Action in a Democracy,” Journal of Political Economy, Volume 65, Issue 2, April 1957, p. 143.


[6] One writer opining for the libertarianesque online zine Taipan claims that Anthony Downs’s An Economic Theory of Democracy is the most cited book in the discipline of political science. In the span of 40 years, a book written by an economist is cited more often than any other “political” book (or article for that matter) written by a political scientist, more than de Tocqueville’s Democracy in America, or Woodrow Wilson’s Congressional Government, Charles Beard’s An Economic Interpretation of the Constitution, even more than The Federalist Papers. See “Why people don’t vote,” http://taipanonline.com/archive/politics/politics-novote.html


[7] Lack of competition is even worse at the state legislative level, where 41 percent of over 7000 state legislative seats were uncontested by one of the two major parties in both 1998 and 2000.


[8] Elizabeth M. DeSouza, “When Tweedle-dee and Tweedle-dum Disagree: Position-Taking Strategies Under Varying Electoral Conditions,” unpublished paper prepared for delivery at the 2000 annual meeting of the Western Political Science Association, University of North Florida, p. 1-3, p. 7, and p. 28.


[9] David C. King, “The Polarization of American Political Parties and Mistrust of Government,” Harvard University Working Paper, www.ksg.harvard.edu/prg/king/polar.htm.  King’s paper was included as a chapter in Joseph S. Nye, Philip D. Zelikow and David C. King, editors, Why People Don’t Trust Government (Harvard University Press: Cambridge MA, 1997).


[10] To catch a glimpse of the type of dross that has trickled down into mainstream culture and education, here is a passage from Encyclopedia Britannica:  “Majority or plural methods of voting are most likely to be acceptable in relatively stable political cultures. In such cultures, fluctuations in electoral support, given to one party or another from one election to the next, reduce polarization and make for political centrism (my comment:  like in Angola, India and Algeria, which use winner-take-all, one wonders?). Thus the “winner take all” implications of the majority or plurality formulas are not experienced as unduly deprivational or restrictive.” Encyclopedia Britannica, 2001 Deluxe CD-Rom Edition, article on “Elections: systems of counting votes: proportional representation.”


[11] Paul Rahe, “The Electoral College: A Defense,” article published by the Oklahoma Council of Public Affairs, a public policy research and education organization, on the Web at www.ocpathink.org/Governance/TheElectoralCollegeDefense.html). Rahe is the Jay P. Walker Professor of American History at the University of Tulsa. A version of Professor Rahe’s article was published in the American Spectator on November 14, 2001.


[12] See Justice O’Connor’s concurring opinion, joined by Chief Justice Burger and Justice Rehnquist, in Davis v. Bandemer, 478 U. S. 109, 144-145 (1986), where she declared “There can be little doubt that the emergence of a strong and stable two-party system in this country has contributed enormously to sound and effective government.” Also see Scalia’s dissenting opinion, joined by Chief Justice Rehnquist, and Justices O’Connor and Kennedy, in Rutan v. Republican Party of Illinois (88-1872), 497 U.S. 62 (1990). In the latter, Scalia was opining on the patronage system, and in a loose tangent wrote about the differences between two political parties being moderated by Downsian-like incentives “as each has a relatively greater interest in appealing to a majority of the electorate and a relatively lesser interest in furthering philosophies or programs that are far from the mainstream. The stabilizing effects of such a [two party] system are obvious.”


Previous Article
Next Article