As the ideological sorting of American voters continues, liberals find themselves outnumbered in four out of five states.
The election of Donald Trump and the Republican victories in the House and Senate were a real eye-opener for many. Today’s political map remains dominated by shades of red. Indeed, conservatives outnumber liberals in 44 out of 50 states, according to new polling results from the Gallup Organization.
In five states, conservatives outnumber liberals by a whopping 30 percentage points; in 13 more, they hold a 20-to-30 point advantage, and in another 15 right-leaners hold 10 point leads, among them Ohio, Pennsylvania, Michigan, Wisconsin, and Florida. Liberals hold a numeric advantage over conservatives in just four states: Massachusetts, New York, Vermont, and Connecticut.
The map below by Taylor Blake of the Martin Prosperity Institute charts the ideological identification of states compared to the national average—where conservatives make up 36 percent of the country and liberals make up 25 percent of the country.
A broad conservative belt cuts across much of the middle of the country, being especially pronounced in the Deep South and Plains. The only remaining solidly liberal pockets are in the Northeast and New England, along the West Coast, and
Before the surviving American liberals book their flights to Sweden, an important caveat: The difference is not as sharp when we combine moderates and liberals. Thirty-four percent of Americans identify themselves as moderates, according to Gallup, compared to 36 percent who identify as conservative and 25 percent who identify as liberal. In fact, the conservative share of the electorate has tracked down somewhat from its peak in the mid-2000s and the liberal share has tracked up a bit. As the report notes, “Despite the dominance of conservative states, the conservative advantage has decreased within most states, with 42 states becoming at least marginally less conservative since 2009.” Together, the share of moderates and liberals outnumber conservatives in 46 states. Only in Wyoming do conservatives make up a greater share of the polity than moderates and liberals combined.
This process of self-segregation along class and ideological lines has been happening in the U.S. for a decade or more: It’s the “Big Sort” that Bill Bishop described in 2004. The deep and overlapping cleavages of class and geography—knowledge and density—were the crucial fault lines of the 2016 election. But what are the key economic, demographic, and cultural factors that account for the persistent conservatism of American states?
To get at this, my colleague Charlotta Mellander ran a basic correlation analysis on the Gallup poll numbers, comparing political identification (conservative, liberal, and moderate) to a variety of key economic, demographic, and cultural factors by state. (This updates previous analyses we conducted in 2011 and 2012.)
As usual, I point out that correlations simply reflect associations between variables and do not imply causation. Still, taken with our previous findings, they highlight the increasingly conservative tide across American states and the deep cleavages of class, income, education, urbanity, and religion on which they turn.
First, these divisions of political identification line up with the actual votes for Trump and Clinton across the fifty states. Across those states, conservative identification is positively associated with Trump votes (0.88) and negatively associated with Clinton votes (-0.90), while liberal identification is positively associated with Clinton votes (0.84) and negatively associated with Trump votes (-0.82).
Moderate identification is positively but more modestly associated with Clinton votes (0.31), and negatively but more modestly associated with Trump votes (-0.35). But, when we combine moderates and liberals, the pattern is similar to that of liberals, with the share of moderates and liberals being positively associated with Clinton votes (0.87) and negatively associated with Trump votes (-0.86).
Class—defined as income, education and occupation—continues to play critical role in America’s political divide.
Conservative states are significantly less educated than liberal ones. There is a strong negative correlation between conservative identification across states and the percent of adults who are college graduates (-0.74). This is about the same as in 2012. The opposite is for liberals and moderates. Education levels across states are positively correlated with liberals (0.69), moderates (.40), and moderates and liberals combined (0.75)
Conservative states are also much more strongly working class. Conservative identification across states is positively correlated with the percentage of a state's workforce in blue-collar occupations (0.74). And it is highly negatively correlated with the proportion of the workforce engaged in knowledge-based professional and creative work (-0.64). The creative class share of the workforce is positively associated with liberals (0.62), and moderates and liberals combined (0.65), but not significantly associated with moderates by themselves.
States with higher levels of conservative identification are considerably less affluent as well. Conservative political affiliation is negatively correlated with state income levels (-0.64) and economic output (-0.49). The opposite pattern holds for liberals and moderates. Income levels across states are positively correlated with liberals (0.57), moderates (0.32), and moderates and liberals combined (0.62).
Urbanity is a second key line of division. Conservative identification across states is negatively associated with both the share of a state’s population that lives in urban areas (-0.37) and the overall urbanization of a state, measured by the urban share of state land area (-0.45). Urbanization is positively associated with liberals (0.39), and moderates and liberals combined (0.41), but not significantly associated with moderates by themselves.
Conservative states are also considerably more religious. Conservative political affiliation across states is highly correlated with religiosity, measured as the share of state population for which religion is a very important component of daily life (0.76). Once again, the opposite pattern holds for liberals and moderates. Religiosity across states is negatively correlated with liberals (-0.75), moderates (-0.29), and moderates and liberals combined (-0.78).
Conservative states are less diverse. Conservative political affiliation is highly negatively correlated with the percent of the population that are immigrants (-0.60), or gay and lesbian (-0.68). Both immigrant and gay and lesbian share of the population are positively associated with liberals (0.50 and 0.73 respectively) and moderates and liberals combined (0.54 and 0.64), but not significantly associated with moderates by themselves.
In short, the trend I noted back in 2011 remains: Conservatism is most pronounced in America's least well-off, least educated, most economically hard-hit states. In many ways, it is the ideology of the economically left behind.
As I wrote then:
But the much bigger, long-term danger is economic rather than political. This ideological state of affairs advantages the policy preferences of poorer, less innovative states over wealthier, more innovative, and productive ones. American politics is increasingly disconnected from its economic engine. And this deepening political divide has become perhaps the biggest bottleneck on the road to long-run prosperity.
This is even more the case today. The ideological right now holds all the cards, but deep political shifts reflect an ongoing sorting along geographic lines. Despite the slow overall decline of conservatism, liberals have consolidated in fewer blue states and blue metros while conservatives have gained an edge in larger swaths of the country, especially critical swing states. The result: Narrow victories in less-populous states bring about dramatic change. Even if the country continues towards liberalism in the aggregate, the question is whether the states will reflect this change or remain red.