The United States has reached an unprecedented level of inequality in presidential elections. In 2012, only 10 states drew the major party presidential candidates for postconvention campaign events, and those same 10 states attracted 99.6% of all general election television advertising spending by the campaigns and their allies. The remaining 41 spectator states (counting the District of Columbia) included all 38 states that had been similarly overlooked in 2008. This article details these inequalities and their roots in state statutes allocating electoral votes on a winner-take-all basis. It argues that states should end this inequality by enacting the National Popular Vote interstate compact, which would ensure that it is the popular vote in all 50 states and the District of Columbia that determines who becomes the president.
The Electoral College has been the subject of more proposed amendments than any other provision of the U.S. Constitution (National Archives and Records Administration 2012). For decades, a majority of Americans has backed moving to a national popular vote in presidential elections, and polls have consistently shown that some two-thirds of Americans, across a full range of states and political ideologies, support this change (National Popular Vote 2013; Saad 2011). The most recent example is a January 2013 Gallup poll indicating that at least 60% each of Democrats, Republicans, and independents would support a national popular vote for president (Saad 2013). In the wake of an election in which the presidential campaigns focused on fewer states than ever before in the modern era, the case for change has never been stronger—and a realistic roadmap for reform has never been closer than with the National Popular Vote plan, which has garnered support in state legislatures across the country.
The problems of the current Electoral College system are grounded in state “winner-take-all” laws that award all of a state's electoral votes to the candidate who wins the most votes in that state. Many of our nation's founders, including the Constitution's chief architect James Madison, strongly opposed such winner-take-all laws (McCarthy 2012), but by the 1830s, they had been enacted in nearly every state in order to ensure state partisans could provide maximum support for their party's nominee. Collectively, the winner-take-all rule violates fundamental principles of representative democracy.
The most obvious problem with states using the winner-take-all rule is that a candidate can lose the presidency despite earning the most votes nationwide. This reversal of the popular vote has occurred in four of our nation's 52 presidential elections, or one-in-13 times, and it has occurred one-in-seven times in close elections won by less than 10%. We experienced a “wrong winner” in 2000 and only narrowly avoided it in 2004, when John Kerry would have won the White House with a shift of fewer than 60,000 votes in Ohio despite George W. Bush's popular vote lead of more than three million votes. If one assumes a uniform swing in the popular vote across all states, Barack Obama would still have been elected in both 2008 and 2012, even if he had lost nationally by more than a million votes—for example, Obama's 2012 national margin was 3.7%, but his margin in the tipping point state of Colorado, which Mitt Romney would have needed to win in order to reach the 270 electoral vote threshold, was a larger 5.4% (National Archives and Record Administration 2013).1
Grounded in his analysis of these recent elections, New York Times analyst Nate Silver argues that in the last three elections, Democrats have had a distinct Electoral College advantage (2012). While this advantage is not immutable, the fact that voting rules could skew elections toward one party for any sustained length of time is a serious flaw.
There is potential for even more unrepresentative outcomes if the selection of the president were to be thrown to the House of Representatives, as it would be in the event that no candidate won an Electoral College majority. This could occur either in the event of an Electoral College tie or in the event of a three-way division of electoral votes in which a third-party candidate won a substantial number of electoral votes. At that point, the U.S. House would pick the president by the indefensible formula of one-state, one-vote, and the House vote would almost certainly be based along party lines, no matter who had won the national popular vote (Richie 2012a).
A second problem with the winner-take-all rule affects every recent presidential election: the geographic gerrymander that leads presidential campaigns to focus all their resources on a handful of potentially competitive states while ignoring most Americans in election after election. Due to the winner-take-all rule, modern presidential campaigns ignore any state where one candidate is comfortably ahead.
In this era of highly partisan voting behavior, the identity of these spectator states has become increasingly predetermined by recent electoral outcomes. Of the nine states that received campaign attention in 2012 beyond the national norm, every single one had also been among the 13 battleground states in the 2008 election. Of the 37 states (including the District of Columbia) that were treated as nonbattlegrounds in both 2008 and 2012, none are projected to be battleground states in 2016.
This increasingly rigid, shrinking presidential election battleground consistently marginalizes two-thirds of Americans. As the identity of swing states becomes more predictable, we can expect additional problems. Parties, candidates, and incumbent presidents will be all the more tempted to target their messages, policy preferences, and discretionary spending toward swing states, while partisans in swing states will be tempted to skew voting rules to their advantage.
Certainly a state's status as a potential presidential battleground affects where an incumbent president will spend time. As an example, Barack Obama earned 45% of the vote in South Carolina in 2008 but has not been back to the state since his primary victory there over Hillary Clinton. His 2008 election vote share in North Carolina was just four percentage points higher, but that small difference was enough to give him all of North Carolina's electoral votes. In sharp contrast to his treatment of South Carolina, the president was in North Carolina for 15 public events between his inauguration and the 2012 Democratic convention in Charlotte (Richie and Levien 2012).
In a paper on federal grants for the Center for the Study of Democratic Institutions, John Hudak found that “swing states receive about 9% more grants and almost 7% more grant dollars than non-swing states when an election is proximate” (Hudak 2011)—a pattern certain to increase as swing state status becomes more predictable. Providing insight into the differential treatment accorded to swing states, in 2011 the late U.S. Senator Arlen Specter candidly explained his opposition to Pennsylvania changing its winner-take-all rule: “I think it'd be very bad for Pennsylvania because we wouldn't attract attention from Washington on important funding projects for the state … In 2004, when I ran with [President George W.] Bush, he … came to Pennsylvania 44 times, and he was looking for items the state needed to help him win the state. … It's undesirable to change the system so presidents won't be asking us always for what we need, what they can do for us” (Kelly 2011).
Our greater ability to predict which handful of states could decide the presidency also creates new incentives for partisans in swing states to game laws governing access to the ballot and allocation of electoral votes in what has become a political form of trench warfare, fighting over each of the relatively few electoral votes in play. There have been disputes in swing states over electoral rules and election administration that appear to be efforts to reduce or increase voter participation of different demographic groups (O'Toole 2012). In addition, efforts by lawmakers in several swing states to allocate electoral votes proportionally or by congressional district are clearly designed to produce a partisan advantage in the 2016 election (Richie 2012b).
This article focuses in particular on the problem of the vanishing swing state. It shows how pronounced the divide has become between swing states and the rest of the nation, and reviews its perverse impact on elections and governance. It then explains why the best path to reform is for states to join the National Popular Vote interstate compact. As of February 2013, this reform plan has been passed into law in states representing 132 of the 270 electoral votes necessary for it to govern the 2016 election. If successful, it will guarantee the White House to the candidate who wins the most votes in all 50 states and the District of Columbia.
Measuring Inequality on the Campaign Trail
The 2012 presidential election was both the most expensive election in American history and the least equitable in its treatment of voters since states moved to holding popular elections. Although more than 129 million people voted for president in 2012, fewer than a third of them lived in battleground states where the campaigns treated their votes as meaningful.2 This reality is in direct conflict with fundamental American values of “one person, one vote,” “consent of the governed,” and “we the people.” As the president is the only American leader elected nationally, every voter should have equal power to hold an incumbent accountable and to elect a preferred candidate—a far cry from today's reality.
The campaigns in 2012 were candid about how few states they planned to target. During the May 2012 Boca Raton fundraiser at which Mitt Romney made his oft-cited “47 percent” comment, the presumptive Republican nominee explained, “You'll see the ads here in Florida … all the money will get spent in 10 states. And this is one of them” (NBC News 2012). In September 2012, President Obama's campaign manager Jim Messina said, “What I care way more about is Ohio, Colorado, Virginia, Wisconsin, etc. In those states, I feel our pathways to victory are there. There are two different campaigns, one in the battlegrounds and one everywhere else. That's why the national polls aren't relevant to this campaign” (Miller 2012).
The campaigns' actions confirmed the accuracy of these admissions. Obama, Romney, and their two running mates held public campaign events in only 12 states after the party conventions, and only nine states met our battleground definition of earning more attention than warranted by their population size (as opposed to the 13 states meeting this criterion in 2008 and the 14 states doing so in 2004). The campaigns and their allied interest groups3 targeted 99.6%4 of their advertising money at voters in the 12 states where candidates campaigned in person. Additionally, 87% of Obama's campaign field offices and 92% of Romney's offices were in those dozen states—with the remaining field offices primarily focused on generating votes in battleground states.
The degree of inequality between states in 2012 was breathtaking. While the Obama and Romney campaigns and their allied groups spent an average of more than $30 per eligible voter in New Hampshire and Nevada after April 10 (when Romney became the presumptive Republican nominee), they spent less than one cent per eligible voter in 34 states that together represented two-thirds of all eligible voters (Figure 4).5 Although Florida, Ohio, and Virginia represent only one-eighth (12.8%) of the American eligible voter population ( Figure 1), they were the target of more than half (53.5%) of advertising money (Figure 2) and 58.9% of public campaign events6 (Figure 3) during the peak campaign season.7 In contrast, voters in four of the six most populous states—California, Texas, New York, and Illinois, with 28.5% of the nation's eligible voter population—were targeted with a total of $58,760, just 0.007% of the total advertising money spent nationwide, and had no public campaign events with any of the four major party nominees (McDonald 2013).8
To compare how much relative attention states received from presidential campaigns, FairVote's “attention index” measures campaign attention relative to the average attention a state would have received if campaigns had spent money and allocated numbers of candidate events based only on population.9 Each state would have an attention index of 1.0 if states were treated exactly equal relative to population, meaning that an index above 1.0 indicates relatively more attention than a state's population would warrant, and an index below 1.0 indicates less attention. Only nine of the 51 states (including Washington, DC, as a state in this analysis) earned our definition of “battleground state” by having an attention index above 1.0. The remaining 42 states received less attention than warranted by their population, including 35 states that received less than 1% of the attention warranted by their population.
The Table 1 chart summarizes subclasses based on electoral vote totals per state.10 This comparison of state population and campaign attention highlights how all spectator states were ignored regardless of their numbers of voters and electoral votes. Contradicting claims that the current system helps small states, only three of the 27 states with fewer than nine electoral votes had an attention index above 1.0, and the median attention index for the 21 states with six or fewer electoral votes was only 0.0002. Wyoming, the state said to benefit the most from its number of electoral votes per capita, did not have a single non–fund-raising campaign event with a major party presidential or vice presidential candidate during either the presidential primaries or the general election.
Table 1. 2012 Attention Index—Averages by State Size
States in list
States w/ index > 1
6 largest states (20 or more electoral votes [EV])
Analyzing the location of campaign field offices provides another valuable way to measure inequality in campaign treatment,11 as campaign staff on the ground can track likely voters and organize volunteers to make personalized appeals to potential voters. Of Barack Obama's 790 field offices, more than half were in just five states (Ohio, Florida, Iowa, Wisconsin, and Colorado), including 131 in Ohio alone.12 Mitt Romney's campaign opened field offices in only 16 states, with 41% of those offices in Ohio, Florida, and Virginia ( Figure 5).
As we can see, inequality based on competitiveness penetrates multiple levels of presidential campaigns. From polling potential swing voters for their opinions on policies to making direct appeals to them, candidates treat voters in swing states as if they were the only people who mattered in presidential elections. Furthermore, all evidence suggests that the percentage of targeted voters will not grow any time soon, unless we reform the system through which we elect the president.
Vanishing Swing States and Growing Stability
The most reliable guide for where campaigns will target resources is a state's past presidential results relative to the national average. With great predictability, candidates target states based on whether they are among a decreasing pool of swing states that in the previous election demonstrated a relatively balanced partisanship, that is, states where percentages of the vote for the major party candidates were close to their national percentages.13 In 2008, only nine states had results within a six-percentage point partisanship range of 47% to 53% (meaning the candidates' percentages in those states were on average within three percentage points of their national percentages), resulting in the fewest number of states in the modern era that met our definition of being a likely swing state in the next election.
Seven of those nine swing states in 2012 ended up being among the nine battleground states in 2012 with an attention index over 1.0. The only projected swing states to be relatively ignored in 2012 were Pennsylvania and Minnesota, although both states ended up being close enough to stay on the swing state list for the 2016 election. The two additional states to be 2012 battlegrounds had been battlegrounds in 2008: Wisconsin, a Democratic-leaning state that Mitt Romney's campaign sought to bring into play with the selection of Congressman Paul Ryan as a running mate, and North Carolina, where Obama had earned an upset win in 2008.
Increasing partisan rigidity between elections further reduces the number of Americans with any chance of drawing presidential campaign attention in the next election. From 1960 to 1984, an average of more than 19 states shifted their partisanship by 5% or more from one presidential election to the next. In the four presidential elections from 1984 to 2000, that average declined to just over eight. In the three presidential elections from 2004 to 2012, an average of only three states shifted their partisanship by more than 5% (Table 2).
Table 2. Number of States Shifting Partisanship by 5% or More Between Presidential Elections
Number of States
In 2012, just three states shifted their partisanship by more than 3.9%, all of which were small and ignored by the campaigns: Alaska, which became more Democratic without its governor on the Republican ticket, and Republican-trending Utah and West Virginia. Only five states in 2008 had outcomes that deviated at least 3% from their 2004 partisanship: two states moved sharply towards Democrats (Hawaii, where Barack Obama grew up, and Indiana, where Obama benefited from his 2004 Senate campaign in neighboring Illinois and from building a campaign operation in a fiercely contested presidential primary), and three Southern states that moved sharply toward Republicans (Louisiana, Tennessee, and Arkansas).
As a result, while campaigns may monitor potential new swing states with occasional polling—keeping an eye out for a wildcard like Indiana in 2008—even that broader review of states is severely limited. At the 2004 Republican convention, George Bush's campaign analyst Matthew Dowd admitted that the campaign had not polled a single person outside of 18 potential battleground states since 2002 (Frontline 2005).
In addition to being more rigidly defined, today's presidential election swing states are also far fewer in number and less populous than a generation ago. In 1960, for example, the major party candidates' vote percentages were within 3% of the national average (swing state status) in 23 states, with a total 319 electoral votes. In 1976, 24 states controlling a total of 345 electoral votes met this same swing state definition. As recently as 1988, there were still 21 swing states that together represented more than half the population and a total of 272 electoral votes.
In our most recent elections, however, the number of swing states and their total number of electoral votes has decreased dramatically—down to just 12 states after 2004 election, nine after the 2008 election, and 11 after the 2012 election, in each case with no more than 160 total electoral votes. Recent election results suggest the list of swing states is unlikely to grow by more than one or two states any time soon, thereby continuing the marginalization of a large majority of Americans (see Table 3).
Table 3. Shifts in Number of Swing States and Uncompetitive States(Swing states = 47-53% partisanship. Uncompetitive states > 58% partisanship.)
11 (w/140 electoral votes)
25 (w/247 electoral votes)
9 (w/116 electoral votes)
26 (w/275 electoral votes)
13 (w/159 electoral votes)
20 (w/163 electoral votes)
16 (w/167 electoral votes)
20 (w/166 electoral votes)
13 (w/206 electoral votes)
13 (w/90 electoral votes)
22 (w/207 electoral votes)
5 (w/20 electoral votes)
21 (w/272 electoral votes)
8 (w/40 electoral votes)
21 (w/260 electoral votes)
9 (w/44 electoral votes)
15 (w/221 electoral votes)
13 (w/58 electoral votes)
24 (w/345 electoral votes)
9 (w/46 electoral votes)
22 (w/235 electoral votes)
9 (w/46 electoral votes)
19 (w/273 electoral votes)
11 (w/57 electoral votes)
17 (w/204 electoral votes)
13 (w/100 electoral votes)
23 (w/319 electoral votes)
9 (w/64 electoral votes)
Absent a National Popular Vote, More of the Same in 2016
Leading up to the 2012 election, President Obama's middling popularity ratings and a stagnating economy suggested that the presidential race would be closer than in 2008, when Obama won by 9.5 million votes (7.3%). Although Obama ultimately won the 2012 election by nearly five million votes, most polls during the year indicated a closer outcome. The campaigns acted on that calculation, focusing primarily on the swing states as had been defined in 2008.
Looking ahead, the partisan landscape has barely changed: every swing state as defined by the 2008 election maintained its swing status after 2012, and only two new states (the 2008 and 2012 battlegrounds of North Carolina and Wisconsin) earned 2016 swing state status. Eight of these 11 swing states (all but Colorado, North Carolina, and Virginia) were also among the relatively few swing states coming out of the 2000 elections, and all but North Carolina was a swing state after the 2004 elections. It is unclear whether five of these 11 prospective swing states (Minnesota, Nevada, North Carolina, Pennsylvania, and Wisconsin) would be true battlegrounds in the event of a close presidential election in 2016, meaning that real presidential campaign attention might be limited to as few as six states (Table 4).
Table 4. Projected 2016 Swing States and Partisan Trends (Republican Partisanship)
* Signifies that a state has been a swing state in the last four elections consecutively.
Few additional states have a chance to become new battlegrounds in 2016. While the identity of the core swing states has been largely consistent, states on both ends of the political spectrum are trending away from potential competitiveness—and with that shift are even less likely to receive attention from the presidential candidates.
On the Republican side, all but one of Mitt Romney's state wins were by margins of more than 7.7%, meaning that states with a total of 193 electoral votes have a partisanship of at least 55.8% Republican. North Carolina, the one state Romney won more narrowly, could again be competitive in 2016, but all remaining Romney-won states are unlikely to be swing states. For example, Missouri had the single closest vote margin in 2008, but its 53.7% partisanship in that election led to it being treated as a spectator state in 2012 for the first time in more than four decades (Table 5). In 2008, the candidates held 20 events and spent almost $10 million in ad money in Missouri during the fall campaign, but in 2012, neither campaign held even one peak season campaign event in the state, and they and their allied groups spent a total of only $127,560 (representing a decline of more than 98%) on advertisements targeting Missouri voters between April and November of 2012. Given Romney's comfortable win, Missouri is even less likely to be a 2016 battleground.
Table 5. Formerly Competitive/Targeted States (Republican Partisanship)
Indiana was another Republican-leaning state to fall from battleground grace in 2012. After experiencing a brief shift in its partisanship to near swing state status and all of the attention that came with it, Indiana returned to its normal position of being completely ignored in the general election by the candidates. Its 56.8% Republican partisanship ensures continued irrelevancy in 2016.
For all the talk of the potential emerging swing state status of long-time Republican bastions like Arizona and Georgia, it is hard to make a case for them as 2016 battlegrounds. Arizona had a Republican partisanship of 56.5% in 2012 and trended slightly more Republican despite the fact that its home state senator John McCain was no longer on the ticket. Georgia Democrats are now virtually obsolete in state elections, not even fielding candidates in half of its 2012 state legislative races, and its Republican partisanship of 55.7% makes it an unlikely 2016 presidential campaign target, particularly without the high African American turnout that was driven by Obama's presence on the ballot.
More Democratic-leaning states are closer to partisan balance, but none seem likely to become new swing states in 2016. Although New Mexico and Michigan had enjoyed swing state status since the 1980s, Obama's big wins there in 2008 deviated too far from the national norm to make them real targets in 2012. With the 2012 elections confirming their Democratic tilt, they are likely to remain on the sidelines in 2016. Other Democratic states like Washington and Oregon are even less likely to become swing states.
The Landslide States: A Stark and Growing Divide
One of the single most striking phenomena of the last seven presidential elections has been the sharp increase in states with large partisanship leans that make it nearly impossible for them to become swing states for several decades (Table 4). The number of such definitively noncompetitive states has increased more than threefold in the past 24 years. In 1988, only eight states, with a total of only 40 electoral votes, had a partisan lean of at least 58% toward one party. In 2012, however, a whopping 25 states, controlling a total of 247 electoral votes, met that definition (Figure 6).
The partisan divide between the 10 most Democratic and the 10 most Republican states has steadily increased (Figures 7 and 8) during this time. The disparity between the average partisanship of these groups of states was 16% in 1988, but by 2012 had nearly doubled to 30%. A Democrat is now projected to win the 10 most Democratic states by an average margin of 28% in a nationally competitive election, while a Republican in the same election would be projected to win the 10 most Republican states by an average margin of 32%.
Regionally, states are becoming increasingly set in their ways, with Democrats as the favored party of coastal states and interior states with large metropolitan areas, and Republicans dominating the South and interior states with more rural voters. Since 1992, for six straight elections, the same party has won in 32 states. Nineteen of these 32 states are Democratic strongholds, with a total of 242 electoral votes. Republicans have won 22 states with a total of 179 electoral votes in all four elections in 2000 to 2012–including 13 states won in all nine elections since 1980. Idaho, Nebraska,15 Utah, and Wyoming have not been swing states for at least 60 years. The last election in which Utah was won by a competitive margin of less than six percentage points was 1912.
Effect on Voter Turnout
Unsurprisingly, the differing ways that Americans in swing states and safe states are treated affects voter participation (Figure 9). In this election cycle, 129,072,347 of the nation's more than 221,925,820 eligible voters recorded ballots for president, for a national turnout rate of 58.2%, 3.4 percentage points down from 2008.16 While presidential election turnout varied greatly from state to state, from a low of 44.2% in Hawaii to a high of 75.75% in Minnesota, swing state voters on average had a higher rate of turnout than voters in spectator states.17
Eight of the nine battleground states had an increase in their number of voters from 2008. In the remaining 42 states (including Washington, DC), only 11 states had an increase in their number of voters. Instructively, two of the largest turnout declines (more than four percentage points each) were in Michigan and Pennsylvania, 2008 battlegrounds that in 2012 were relatively ignored. The 42 spectator states' collective turnout was 56.9%, nine percentage points lower than the 65.7% turnout in the nine battleground states. The turnout gap of nine percentage points between the 12 closest states and the rest of the country is larger than the six percentage point disparity in 2008 and eight percentage point disparity in 2004.18
Turnout was especially low among nonswing Republican states in 2012 (Figure 10). While nonswing Democratic states in 2012 had a turnout rate of 58.8%, turnout in nonswing Republican states was only 54.8%, a difference of four percentage points that marks an ongoing trend. Since 1988, the safest Republican states have had lower turnout than the safest Democratic states in every election. Presidential campaigns have been ignoring voters in these entrenched spectator states for so long that many of their eligible voters may have come to believe that their votes do not matter.
The adverse effect of being a spectator state on voter turnout is likely to endure for decades because of its particular impact on the youth vote. Young Americans (those under 30) in spectator states are treated differently from their swing state counterparts solely based on where they live. A comparison of youth (ages 18 to 24) turnout in 1972 and 2004 found that while there was an overall decrease of five percentage points, every single one of the 10 states with the large decrease was a solid spectator state in presidential elections, including five states that have been firmly Republican and five firmly Democratic (Pearson 2006).19
In 2000, turnout among eligible voters under 30 in the 10 most competitive states was 51%, compared to only 38% elsewhere, a difference of 13 percentage points. In 2004, this disparity rose to 17%, meaning that young adults in the battleground states were a third more likely to vote than their spectator state counterparts (Pearson 2006). Barack Obama's relative appeal to young voters reduced this disparity in 2008, but the lower turnout in spectator states is likely to have a lasting impact, as studies show that voting behavior early in life is a decisive indicator of a person's future participation (Franklin 2004).
Impact on Racial Minority Voting Rights
While the question of whether the current Electoral College rules consistently harm or help racial minorities is complex, it is indisputable that racial minorities make up a smaller share of the swing state electorate than of the national electorate. Using 2010 census numbers,20 racial minorities disproportionately live in spectator states—particularly Latinos and Asian Americans (Figure 11). More than two-thirds (68.4%) of the nation's citizen voting age population (CVAP) lives in the 39 spectator states where the candidates did not hold a public campaign event after the conventions, which is comparable to the 65.6% share of whites and 69.6% share of blacks in those states. But the percentage of Latinos and Asian Americans who live in those spectator states is far higher: 83.7% of the Asian CVAP and 80.5% of the Latino CVAP. In other words, one out of three white voters lived in states that received campaign attention, but fewer than one in five Asian or Latino voters did.
Although the presidential campaigns targeted Latino voters in Florida, Nevada, and Colorado, the Latino CVAP is underrepresented in every other 2012 election battleground. Looking forward, many of the states with the largest populations of people of color, such as California, Illinois, Maryland, New York, and most southern states, show no sign of becoming battlegrounds.
While African Americans are not as underrepresented in swing states as Latinos and Asian Americans, their power as a swing voting bloc has diminished dramatically in recent decades. In 1976, 73% of African Americans were in a classic swing voter position: they lived in highly competitive states (with partisanships between 47% and 53%) in which African Americans made up at least 5% of the population. By 2000, that percentage of potential swing voters declined by two-thirds to 24%. In 2004, it was just 17% and in 2008, only 19%.21 Latinos and Asians have had a similarly sharp downward trend in their population percentages in presidential campaign battlegrounds; for example, in 1976, 73% of the nation's Latino population lived in an electorally competitive state with a population that was at least 5% Latino, but today only 13% of Latinos are in such a position.
New Temptations to Manipulate Voting and Allocation Rules
One important consequence of the stagnating identity of swing states is the temptation for partisans to seek advantages through manipulation of election rules, given that a shift of relatively few votes in a single state can tip control of the White House. In both 2000 and 2004, for example, the national outcome hinged on the outcome in a single state that both major party candidates anticipated might be in that position—Florida in 2000 and Ohio in 2004. In those elections, Democrats accused those states' Republican secretaries of state of partisan manipulation of voting procedures.
Battles over election laws raged again in several swing states in 2012. Expressing concern about voter fraud, Republicans in control of state governments in Ohio, Pennsylvania, and Florida passed laws that opponents claimed would curtail suffrage, such as decreasing the number of early voting dates, requiring photo identification at the polls, and making it harder to register new voters. Sen. Bill Nelson (D-FL) testified to Congress that, “Florida's 2011 election law changes were politically motivated and clearly designed to disenfranchise likely Democratic voters.” (O'Toole 2012). The Department of Justice, civil rights groups, and Democrats relied on section 5 of the Voting Rights Act22 and other legal challenges to keep many of these laws from being implemented, but most such laws were still set to go into effect before the 2016 election.
Both major parties have a sad record of willingness to manipulate voting laws for partisan gain, but for the moment, Republicans are in a better position to act in states where it might affect the presidential election. Of the 12 states hosting major party campaign events after the party conventions, Republicans control the state legislature and the governorship in seven (Florida, Michigan, North Carolina, Ohio, Pennsylvania, Virginia, and Wisconsin). Three states have divided control, and the Democrats have narrow control over only two (Colorado and Minnesota).
It is likely that both old and new laws that restrict suffrage will be implemented before 2016. Congress is unlikely to override them, as the leadership of the House of Representatives will almost certainly be more focused on voter fraud than voter access. State legal challenges will run their course, and federal courts may stand aside, given that the Supreme Court chose to uphold voter identification requirements in Indiana despite a lack of evidence of a history of voter impersonation fraud in the state. The Supreme Court also may strike down or weaken section 5 of the Voting Rights Act, thereby making it easier for section 5 states, such as Florida, Virginia, and North Carolina, to pass new laws and implement procedures affecting suffrage.
Laws that reduce turnout could shift which party wins all of a swing state's electoral votes, but an alternative approach beckons that would lessen the impact of a Republican losing a state. Urged on by Republican National Committee chair, Reince Priebus (Marley 2013), Republican lawmakers in several closely contested states that voted for Obama in 2008 and 2012 have debated plans to replace winner-take-all rules with methods that would guarantee electoral votes to both major party nominees. The most partisan proposal is to allocate electoral votes based on the presidential vote in congressional districts and then award the two statewide electoral votes to the candidate winning the most districts. If this plan had been in place in 2012, Romney would have won more electoral votes than Obama in all of these seven Republican-controlled swing states and would have therefore easily won the Electoral College despite losing the national popular vote by nearly five million votes. If this plan were passed in all seven of these Republican-controlled swing states, the 2016 Democratic nominee would be unlikely to win the White House without a national popular vote margin of more than seven percentage points, representing some 10 million popular votes (Richie 2012b). Although Republican enthusiasm for such changes has waned as of February 2013, it could return as we near the 2016 election.
Reforming the Electoral College
Given the myriad of problems associated with today's presidential election system, the question should not be whether to reform the system, but how best to do so. Although there has been some talk of constitutional amendments or of states acting on their own to divide electoral votes, only one approach is both politically feasible and effective: state-by-state adoption of the National Popular Vote interstate compact.
The idea of states acting on their own to allocate electors by congressional district or by proportional representation is particularly misguided. As described in the previous section, doing so is fraught with opportunities for partisan mischief, as nearly all politicians will support or disapprove of such approaches based on whether they will help or hurt their party nationally. They also fail the test of making every vote meaningful in every election. Using the district method, most districts and most states will continue to be uncompetitive, marginalizing their voters. As to proportional allocation plans, the fact that states are limited to allocating whole electoral votes would result in perverse dynamics that would advantage a handful of states over the rest. Campaign pollsters would determine whether their vote share was near a “tipping point” that could shift an electoral vote. Campaign activity in most states would be unlikely to affect even a single electoral vote.
Done nationally via a constitutional amendment, the congressional district proposal is indefensible through the lenses of both partisan balance and political equality (Richie et al. 2011). Most districts are fundamentally lopsided for one party, ensuring that most voters would continue to be spectators. We also would see an increase in candidates winning elections while losing the popular vote, as Republican nominees would win every election in which the Democratic candidate failed to win by more than about four percentage points in the popular vote. Because Democratic voters are more concentrated in cities and in majority–minority congressional districts—a pattern that has been true for years in plans drawn by both Republicans and Democrats, but that is even more pronounced under today's trends in partisan demographics—Republicans today have a partisan advantage in more than 240 of our nation's 435 House districts (Richie and McCarthy 2012). In 2012, for example, Mitt Romney won at least 17 more congressional districts than Obama despite his popular vote defeat. Given this partisan skew, it is inconceivable that a constitutional amendment to establish a district plan could pass.
A constitutional amendment to establish proportional allocation of electoral votes in all states based on decimal points (rather than keeping human electors casting one vote each) would be better than the status quo but could only be done after working out key details such as how to handle third party and independent candidates winning enough electoral votes to deny any candidate an electoral vote majority. It also would fall short of the most obvious proportion: one person, one vote.
That proportion of one person, one vote requires a national popular vote for president. Achieving a popular vote through a constitutional amendment is often proposed but would require two-thirds support in both chambers of Congress and the backing of legislatures in 38 states—both tremendous hurdles absent a catastrophic election we hope never occurs. Furthermore, national popular vote backers disagree on key provisions that would need to be part of a constitutional amendment. It is one thing for states to join together by statute to improve elections with a national popular vote for president; it is quite another to resolve all the necessary details that must be part of an amendment to the Constitution.
Anyone seeking to change our broken presidential system in time for upcoming elections should focus on the most plausible Electoral College reform proposal of the modern era: passage of the National Popular Vote plan in states.
What Is the National Popular Vote Plan?
The National Popular Vote (NPV) plan is an interstate agreement that would guarantee the presidency to the candidate who wins the national popular vote in all 50 states and the District of Columbia. The agreement is activated upon being passed into law by states representing a majority in the Electoral College. At that point, NPV will ensure that every vote will matter in every state in every presidential election. As a statutory reform, rather than a permanent change to the federal Constitution, states can enact the plan while preserving their power over how to allocate electors in future elections.
The NPV plan is grounded in two powers granted to states under the Constitution. First, states have the power to decide how to allocate their electoral votes, one characterized by the Supreme Court as “exclusive” and “plenary” in McPherson v. Blacker (1892). In the late 1700s and early 1800s, states used this power repeatedly. Indeed, during the nation's initial presidential elections, only a few states awarded all of their electoral votes to the statewide popular vote winner, and several states chose not to hold popular elections at all. It was not until Andrew Jackson's presidency that the winner-take-all unit rule based on a statewide popular vote became the norm, driven by state political leaders' partisan desires to give as many votes as possible to their preferred candidate. This winner-take-all rule is not in the Constitution, was not intended by the framers of our Constitution, and most certainly is not in the best interests of our nation.
Second, states have a constitutionally protected power to enter into formal, binding interstate compacts. There are hundreds of examples of these legally binding agreements, including the Colorado River Compact and the governance of the Port Authority of New York and New Jersey. Detailed in fewer than a thousand well-vetted words, the NPV compact establishes that the candidate who wins the national popular vote in all 50 states and the District of Columbia will receive all the electoral votes of the participating states. The agreement is activated if—and only if—the participating states collectively hold a majority of electoral votes (currently 270 of 538).
States enter the NPV compact one by one by passing a state law. If by July 2016, states adopting the compact collectively have a majority of electoral votes and if congressional consent of the compact has been achieved or found to be unnecessary, the agreement would be set in stone for the 2016 election. At that point, the candidates, the media, and the voters would know that the White House was guaranteed to the popular vote winner for the first time in American history.
Under NPV, electors would still elect the president, but everyone would focus on the popular vote, not state electoral vote margins. Gone will be the red and blue maps on election night and the early projections of a winner while many western states are still voting. Every vote will count the same, whether cast in Maine, Alaska, Texas, or Florida. Over time, the attention index will be close to 1.0 for every state, with new incentives for parties to spur volunteer activity across the nation rather than only in battleground states.
The NPV plan was launched in 2006 and passed its first legislative chamber that year. As of February 2013, it has passed into law in eight states and Washington, DC. Collectively, the participating states represent 132 electoral votes—nearly half of the 270 votes necessary to trigger its enactment in 2016. The bill has passed in both small states like Hawaii and Vermont and larger states like California and New Jersey. It has been introduced in every state, earned the votes or sponsorship of more than 2,100 state legislators, and been endorsed by entities such as the League of Women Voters; Brennan Center; Common Cause; FairVote; the National Association for the Advancement of Colored People, the New York Times; former members of Congress like Fred Thompson (R-TN), Jake Garn (R-UT), and Birch Bayh (D-IN); and independents like Rhode Island governor Lincoln Chaffee and 1980 presidential candidate John B. Anderson.
The last time the nation had a similar focus on reforming the Electoral College was in the 1960s and 1970s, when a proposed amendment for direct election won the votes of 81% of U.S. House members, including future presidents Gerald Ford and George H. W. Bush, and had the backing of Jimmy Carter, Richard Nixon, the Chamber of Commerce, the American Bar Association, and the AFL-CIO. Today, the case for reform is even stronger. Far fewer states are now contested, and far more states are sure to be marginalized for years to come. The time to act is now.
Objections to the NPV plan can be grouped into six categories: (1) evidence-free arguments that current Electoral College rules have been essential for the success of the United States, (2) election administration concerns, (3) accusations that the NPV effort is partisan, (4) confusion about how the compact works, (5) belief that the “right way” to replace the current system is by amending the Constitution, and (6) support for alternative reform approaches that we have already dismissed as politically infeasible and problematic in practice.
Those who suggest the NPV plan is not the right way to achieve a national popular vote overlook the fact that states have driven nearly all major changes to American elections, from holding popular votes for U.S. senators to expanding suffrage rights, including establishing a vote for president. The fact that most states have long used the winner-take-all rule does not mean they have forfeited their power to replace it. Any suggestion that the NPV plan undermines state powers ignores the fact that it is grounded in states exercising those very powers reserved for them under the Constitution, and passage of NPV does nothing to upset the balance of power between states and the federal government.
There is also no basis to claims that NPV has a partisan impact or motivation. Although Democratic legislators have often led efforts to pass NPV, hundreds of Republican legislators have voted for it or sponsored NPV legislation. The 2000 election has created an emotional tie for some Republicans to the current system, but it is clear that Democrats in recent elections have had an advantage under winner-take-all rules and that over time both major parties have done equally well in the popular vote. The aggregate popular vote for the parties since 1928 has been almost exactly equal, and the parties have split the popular vote in the past 12 presidential elections (since 1968), 16 elections (since 1952), and 30 elections (since 1896)—with Republicans in their strong years tending to do better than Democrats in theirs. Within states, Republicans today certainly can be highly competitive in popular vote elections, as evidenced by the fact that 30 of the 50 states have Republican governors, including several states with big cities like Florida, Texas, Michigan, Ohio, Georgia, Tennessee, and Pennsylvania.
Defenders of the current winner-take-all rules also present a grab bag of arguments relating to the compact, election administration, and the alleged relationship of the current rules to the strength of the United States. When arguing that voters in heavily partisan states will not accept their state's electoral votes going to a candidate of another party, they fail to grasp that all media coverage and “horserace” talk will focus on the national popular vote rather than state results. They worry that the United States will not be able to administer close popular vote elections despite the many successful examples of well-administered popular elections in large states and other nations. Indeed, upon examination, every claim about the difficulties of administering elections under a national popular vote is one that is more likely to affect elections negatively under current state rules, from concern about recounts to the impact of balloting errors.
Opponents warn that third parties will have a more negative impact despite lack of evidence from the thousands of statewide popular vote elections for governor and Senate over the last century. They argue that candidates will only spend time in big population states and cities despite all evidence from our other popular vote elections, including presidential elections in swing states, in which focusing exclusively on urban areas would be a recipe for defeat. They fear greater power of campaign money, despite the fact that candidates already try to raise as much money as they can. The list goes on—and advocates of the NPV have detailed answers addressing all such concerns collected in Every Vote Equal.23
Roadmap to Reform
Looking forward, we expect more states will pass the NPV plan in the coming years. Crossing the halfway point toward national enactment will mark an important milestone. The closer the compact comes to having the 270 electoral votes necessary to activate it in time for the 2016 election, the more the proposal will capture the national imagination and trigger calls for action in more states from the two-thirds of Americans—including majorities of voters in both major parties—who support a popular vote for president.24
Growing support among Republican elected officials promises to make the difference. Given the bias of the current system against Republican candidates in recent elections and the relatively depressed turnout of Republican-leaning states in the 2012 election, we think more Republican leaders will realize that the 2000 election outcome is no reason to oppose popular vote elections. Already a Republican-run legislative chamber in New York has passed NPV by a wide margin, and hundreds of Republican state legislators have backed the plan. Once a Republican-run state adopts NPV, the path to 270 electoral votes will widen considerably.
By early 2015, we should know if NPV will be in place for the 2016 election. If not, however, the reform drive will continue. Adoption of the NPV plan is cumulative, without a time limit. Although a participating state could withdraw, no state has come remotely close to doing so, and a referendum effort in Washington State failed miserably in collecting signatures. Furthermore, if too few legislatures take action to enact the compact, backers may turn to ballot measures. We believe implementation of NPV is a matter of “when” more than “if.”
Enactment of NPV will bring a sea change in our presidential elections. At that point, any successful campaign will have to seek votes everywhere in a true 50-state effort. Every vote in every corner of every state will be equal, creating incentives for Americans to get involved in presidential campaigns in their own communities and removing incentives for partisan efforts to game the vote in swing states. All Americans will have equal voting power to hold their president accountable, and the candidate with the most popular votes will win every election. Just a decade ago, upholding such basic goals of representative democracy in presidential elections seemed like a distant dream, but the NPV plan brings the prospect of change by 2016 within our nation's grasp.
All 2012 election data in this article, except for data used in analyzing turnout, is based on the certificates of ascertainment filed by states before the December 2012 meeting of the Electoral College.
We use the terms “close state,” “swing state,” and “battleground state” to describe three different measures of competitiveness. A “close state” is one where the vote margin was 6% or less in that election. A “battleground state” is one that, regardless of its final outcome, drew a higher share of campaign attention than warranted by its share of eligible voters. A “swing state” is one that, based on the previous election, had a partisanship of between 47.0% and 53.0%, thereby giving it a realistic chance of being the decisive state in a nationally competitive election.
For a list of interest groups, see Andrews et al. (2012).
Advertising money is spent in media markets rather than statewide, and, therefore, numerous ads were spent in areas that crossed state boundaries. However, 99.6% of ad money spent nationwide was spent in areas that reached swing voters. Therefore, although many people who lived in areas of spectator states that bordered swing states saw campaign advertisements on their televisions, we categorize these ads as “targeted at” swing state voters. For example, Vermont was one of the safest Democratic states in the 2012 election. Ad money spent in Burlington, Vermont, was targeted at potential voters in the swing state of New Hampshire.
The figure is calculated by adding the total amount of money spent targeting voters in a state divided by the state's eligible voter population. As the figure varies in different parts of a single state, we use a figure that shows the average amount spent per potential voter.
For the purposes of this article, a campaign visit will be a public event held between September 7, 2012 (the day after the Democratic national convention), and November 6, 2012 (Election Day), that was used to woo voters in the area in which the event was held. A fund-raiser or nationally televised event does not count as a campaign visit.
We classify peak campaign season as the end of the Democratic national convention (September 7), at which point both parties had officially nominated their candidates, and Election Day (November 6).
Eligible voter population is calculated by Michael McDonald and represents the number of residents in a state who are citizens, over 18 years old, and are not disenfranchised from a felony conviction (McDonald 2013).
The formula to determine the attention index is ((% of ad money spent in state + % of public campaign events held in state)/2)/% of state's share of the eligible voter population.
The categories are unevenly distributed because of the varying number of states that have many or fewer electors. Dividing states into 10 groups of five would have resulted in the electoral vote ranges of over 15, 15-10, 10-7, 6-4, and below 4, with some states with the same number of electoral votes falling into different categories.
These numbers were determined from the websites of the Obama and Romney campaigns. The Obama offices can be found on each state's page under the “find your office” feature, but the Romney page is no longer active.
Sixty-eight percent of field offices were in the eight states in which both Obama or Romney held public events during peak campaign season
A state's partisanship represents the degree to which a state's percentages of the vote for the two major party president nominees deviates from the national vote shares for those candidates. A state's Republican partisanship is the vote share a Republican presidential candidate would likely receive in a given state when the national popular vote is tied. This approach would need to be modified if a strong third-party or independent candidate clearly drew more votes from one major party and ran disproportionately well in some states, but such candidacies did not affect the predictive value of this measure in elections from 1960 to 2012.
We base 2012 partisanship in Michigan and Oklahoma only on Obama's relative vote percentages in the state compared to the nation. Libertarian Party candidate Gary Johnson, the former Republican governor of New Mexico who briefly sought the Republican nomination in 2012, won 1% of the national vote and was on every state ballot except Michigan and Oklahoma. Data suggests more Republican-leaning voters backed him than Democratic voters. As a result, Mitt Romney received a larger percentage of the vote in these states than if Johnson had been on their ballot.
Nebraska, like Maine, awards its electoral votes by congressional district, with two votes going to the candidate who wins the state. In 2008, its second congressional district voted for Barack Obama. This is the only election in which either Nebraska or Maine has split their electoral votes.
Turnout signifies the number of votes for president out of the eligible voter population for a state. Eligible voter population is a measure of citizens residing in a state of at least 18 years on Election Day, excluding those who have been disenfranchised by a felony conviction. Eligible voters who live oversees are included in the calculation of national turnout, but not in calculations of statewide turnout, as these voters do not reside in any state. We rely on data provided by McDonald (2013).
Because New York tallied a large number of ballots after the filing of the certificates of ascertainment and other states had minor changes, turnout analysis was based on the December 31, 2012, spreadsheet made available by David Wasserman of the Cook Political Report at http://www.twitter.com/redistrict (accessed January 22, 2013).
See also Center for Information and Research on Civic Learning and Engagement (CIRCLE), http://www.civicyouth.org (accessed February 14, 2013).
Census, Voting Age Population by Citizenship and Race, 2006-2010 American Community Survey five-year estimates.
This is based off of the five states with a swing partisanship in 2008 in which the African American citizen voting age was at least 5% of the state's total CVAP (Florida, Nevada, Ohio, Pennsylvania, and Virginia). Including the three other states with large black populations, which were targeted to some extent in 2012 but did not have a swing partisanship after 2008 (Michigan, North Carolina, and Wisconsin), the number is 29%.
Section 5 of the Voting Rights Act requires jurisdictions with histories of racial discrimination to have any changes to their election laws preapproved by the Justice Department, so as to ensure that the new voting procedures will not disadvantage racial minorities. The Supreme Court granted writ of certiorari to a challenge of this provision of the Voting Rights Act in late 2012 and will likely decide the provision's fate in 2013.
For detailed discussion of concerns about NPV, we draw attention to the latest edition of Every Vote Equal posted at http://www.everyvoteequal.com (accessed January 22, 2013).
Rob Richie has directed the nonprofit electoral reform and policy group FairVote since 1992. His writings have appeared in nine books, including as coauthor of Every Vote Equal and Reflecting All of Us. He serves on the Corporation of Haverford College, where he earned a B.A. in philosophy.
Andrea Torres Levien is a research fellow at FairVote. She graduated from Yale University in 2012 with a B.A. in political science. Her senior thesis analyzed past scholarship on the effects of the Electoral College on the political power of racial minorities.