With President Obama’s victory in the 2012 Presidential election, the hopes of all those anticipating a precipitate electoral shift in Gary Johnson’s favor have been dashed. Governor Johnson, who stood for the Libertarian Party, nevertheless garnered a respectable 1.1 million votes—a number which seems somewhat less respectable when you consider it falls just shy of 1% of total votes cast.
Why did Governor Johnson so thoroughly fail to charm America’s electorate? His economic positions fall distinctly outside the field’s mainstream, but in that respect they hold substantial allure for the half of American voters who have embraced a Republican economic policy formulated largely by non-economists. Johnson’s thoughts on social issues are decidedly liberal: from his advocacy for drug decriminalization to his pro-choice stance on abortion, the Governor represents the values of America’s liberal voters more closely than do most Democratic candidates. He is also a triathlete. From a purely demographic standpoint, Gary Johnson should be President.
Johnson couldn’t have won, of course. The United States is home to a deeply entrenched two-party system, and has been for a century and a half. In the 2012 Senate race, all but two seats were taken by Democrats or Republicans. There are no independents in the House of Representatives.
What is the genesis of the two-party system? And why, in the United States, is it so exclusive? There are other countries dominated by two parties, but in few nations is two-party control so seemingly indelible as in the U.S. The last American President who was neither a Democrat nor a Republican was Millard Fillmore, who left office in 1853. The multi-party Congress gasped its last breaths in the elections between 1866 and 1872. The 1874 House featured only eight independents out of 293 Representatives, all of whom were either Democrats or Republicans. The Senate that year was dominated almost completely by those two parties, with the exception of one Independent Republican.
The prevailing explanation for two-party supremacy in political science circles has long been Duverger’s Law, which states that the voting system (“first-past-the-post”) used by the U.S. and several other democracies is mathematically ordained to eliminate third parties. First-past-the-post (FPTP) voting is simple: citizens vote for one candidate per seat, and whoever gets the most votes wins. The problem, according to the French political scientist Maurice Duverger, is that this system discourages third parties from standing for election and citizens from voting for them. The mechanism by which Duverger claimed this occurs is uncomplicated; if a party comes in third or fourth in a “winner-take-all” election in which the party with the most votes claims all the spoils, there is little incentive to stand again. Likewise, voters feel (frequently correctly) that they are “wasting their votes” if they choose a party that has been demonstrably uncompetitive.
This is a convincing enough narrative. But some political scientists have poked plausible holes in it as the source of the two-party system with one simple point: plenty of countries that use FPTP have multi-party systems. India (arguably the most vigorous multi-party system in the world) uses it, as do Japan and South Korea. India features six major political parties; two are dominant in Japan, but third parties hold 97 of of 480 seats in its House of Representatives and 63 of 242 in its House of Councillors. Likewise, South Korea features four major parties– although the two minority parties hold only 13 of 300 seats in its National Assembly.
There are a few things to consider about this alleged rebuttal of Duverger’s Law. First, here’s a list of countries with FPTP or some close variant thereof:
Antigua and Barbuda
Papua New Guinea
St. Kitts and Nevis
Saint Vincent and the Grenadines
Trinidad and Tobago
Turks and Caicos
All right. It’s worth noting, although it’s not particularly relevant to this conversation, that of these 66 countries, only seven (U.S., U.K., South Korea, Canada, Barbados, Bahrain and Japan) are “developed” (that is, lying in the top quartile of the U.N. Human Development Index). Common to nearly all of the rest of the countries on this list are a host of political, economic and social difficulties that necessarily disrupt any analysis of their electoral systems. Some are too small to exhibit determinate characteristics of a partisan system, while others are so corrupt as to make electoral analysis meaningless. Still others are in a state of transition. For the purposes of this assessment, we’ll use only the seven developed countries and India.
Of the eight countries under consideration, three (Korea, Bahrain and Japan) use a variant of first-past-the-post incorporating additional features. Japan and South Korea use parallel voting, in which FPTP is employed alongside a different electoral system. In South Korea, FPTP is used for 246 seats in the National Assembly, while 54 are filled with proportional representation (PR is relatively straightforward: seats are allocated in direct proportion to the percentage of the vote received). In Japan, 180 of 480 seats in the House of Representatives are allocated using PR, as are 96 of 242 in the House of Councillors. The National Assembly of Bahrain, which in many respects is only nominally a democracy (since the King appoints the powerful upper house of the legislature), uses first-past-the-post block voting in its lower house. In this system, voters select as many candidates as there are open seats. If there are X open seats, the X candidates with the greatest number of votes occupy those seats.
These countries all feature multi-party systems, but their application of alterations to FPTP voting makes them poor tests of Duverger’s Law. Indeed, South Korea’s and Japan’s partial utilization of proportional representation means that third parties will not, in fact, be discouraged from standing for election and, in turn, that voters will not be discouraged from choosing them. Bahrain’s electoral system is marred by corruption, royal interference, and skewed and sectarian district lines: it is difficult to view the country as a laboratory for any type of democracy.
Of the remaining countries utilizing FPTP, Barbados and the U.S. both feature two-party systems that most analysts regard as fully entrenched. In point of fact, however, the fact of bipolarity is nearly as inescapable in other FPTP countries as in the U.S. and Barbados. Consider first the House of Commons, the British Parliament’s lower house: approximately 85% of seats are either Labour or Conservative. In the House of Lords, this number is much lower, at 57.5%. Do these numbers, particularly the latter one, disprove Duverger’s Law?
Not exactly. The mechanism of Duverger’s Law does not necessarily preclude the formation of third parties; rather, it predicts that, if formed, weaker parties will fuse with stronger ones in order to survive. For much of its existence, Britain’s parliament has been bipolar. Whig and Tory supremacy before World War I was broken only by the brief materialization of Irish nationalism in third-party form following the lifting of legal restrictions on Catholics in the 18th and 19th centuries. Parliament shifted comfortably into the Conservative/Labour dichotomy and held in that vein until the appearance in 1988 of the Liberal Democrats, a third party formed in a merger of the Liberal (formerly Whig) Party and the short-lived Social Democratic Party.
The formation of the Conservative – Liberal Democrat coalition (the “Cameron Ministry”) in 2010 seems to indicate that the Lib Dems are being absorbed into Labour as predicted by Duverger’s Law; Britain has not seen a coalition government for half a century, and the intervening time has been bipolar. With the Cameron Ministry counted as a unitary force, third parties hold only 6.3% of seats in the House of Commons. In addition, there is a sizable contingent of independent MPs–”crossbenchers”–in the House of Lords. Eliminating those who are not party-affiliated from the count, 90.2% of the remaining seats in the House of Lords are controlled either by the Cameron Ministry or Labour.
While Canada does enjoy a vigorous third-party atmosphere, it is hardly multipolar. 93% of seats in its Senate are either Conservative or Liberal, while 85% of those in its House of Commons belong to either the Conservatives or the New Democrats. The New Democrats are, as befits their name, newcomers to the Commons scene. Indeed, until the breakup of the coalition led by the Progressive Conservative Party in the 1980s, the PC and the Liberal Party were the two controlling groups. The dismantling of PC’s large coalition left an opening for third parties that lasted for a number of years, but the New Democrats have now closed this gap in achieving the status of Official Opposition.
Consider India, then, where there can be no assertion of real bipolarity: the two largest parties of the 15th Lok Sabha hold only 68.7% of seats. These two parties—the Indian National Congress (INC) and the Bharatiya Janata Party (BJP)–hold dominion over the country’s political landscape, and have done for years. Nevertheless, the profusion of political parties to be found in India is truly astonishing. Of course, India has been an independent democracy for less than seventy years, and for much of that time (until 1977), the INC was in control of Parliament.
Where does this leave Duverger’s Law? It’s certainly not proved, at least not without supplements or necessary corollaries. But hardly is it disproved. In the countries we’ve looked at, we have found features or phenomena that either confirm or make irrelevant the mechanisms of the Law. In the U.S., third parties hardly ever enter onto the national stage; in Britain, they swiftly form alliances that melt back into bipolarity. Which of these options actually occurs seems down to systemic idiosyncrasies and perhaps to something as ostensibly mundane as which electoral powers are devolved to the local level.
Obviously, there is something to the dictates of Duverger’s model. But all models are by definition incomplete and uneven, and this one is exceptionally so. Perhaps the most instructive information provided by Duverger’s Law is not how well it predicts party bipolarity, but how much data is so clearly left out. This in itself is invaluable. FPTP does ordain a two-party system at least in some sense, but not inevitably and only based on the states of other, hidden variables. The contribution of Duverger’s Law is that it exposes the existence and importance of these variables; the work to be done now is to discover what they are and how they affect party structure.