Do election outcomes matter?

Lane Kenworthy, The Good Society
April 2017

Most Americans identify as either a Democrat or a Republican. As figure 1 shows, the Democrats currently enjoy an advantage of about ten percentage points. Vote totals in elections for the presidency, the Senate, and the House of Representatives are often closer than this, and in recent elections American voters have been fairly evenly split between the two parties.

Figure 1. Party identification
Share of US adults. Question: “Generally speaking, do you usually think of yourself as a …?” ANES has seven response options: strong Democrat, weak Democrat, independent leaning Democrat, independent, independent leaning Republican, weak Republican, strong Republican. GSS has eight response options: strong Democrat, not strong Democrat, independent leaning Democrat, independent, independent leaning Republican, not strong Republican, strong Republican, other party. “Democrat” here is the three Democrat groups. “Republican” is the three Republican groups. Those choosing “other party” in the GSS, usually just 1% or 2%, are excluded. Data sources: American National Election Studies, electionstudies.org, series party identification; General Social Survey, sda.berkeley.edu, series partyid.

The political left and right tend to differ along three main axes. One is economic, with the left preferring more government support for security and fairness and the right prioritizing freedom for individuals and firms. A second is social-cultural, with the left here emphasizing individual liberty and the right privileging order, tradition, and community. A third is foreign policy. Here the left has tended to be more isolationist, the right more favorably disposed to intervention abroad. In the United States, the Democrats and the Republicans have differed on the economic axis since the early 1930s and on the social-cultural and foreign policy axes since the late 1960s.

Given these differing aims and priorities, election results should produce differences in economic and social outcomes. Do they?

PARTISAN EFFECTS: CROSS-COUNTRY COMPARISON

In the rich longstanding-democratic nations, governments headed by a social democratic or Christian centrist party have been far more likely to expand public insurance programs and public services, resulting in greater economic security. Social democrats have tended to structure government transfers in an especially egalitarian fashion, producing less income inequality and poverty. This is a big part of the reason why countries such as Denmark and Sweden have more expansive and generous welfare states and less economic insecurity and inequality than the United States.1

PARTISAN EFFECTS: THE UNITED STATES OVER TIME

In the US, the left party has been in power a good bit since the early 1930s, but with limited public insurance to show for it. Is the United States an exception? One reason to suspect so is that our “left” party has been more centrist than most of its European counterparts. That isn’t surprising: in a two-party system there is an electoral incentive for each party to move toward the center in order to maximize its vote share. Indeed, a recurring theme in American politics is the notion that the two parties aren’t meaningfully different in their commitments. “There’s not a dime’s worth of difference between the two,” George Wallace famously opined in 1968. “You’re dealing with tweedledum and tweedledee,” said Bernie Sanders in 1986. “Our two parties are basically one corporate party wearing two heads and different makeup” is how Ralph Nader put it in 2000. There are additional reasons why the US might not follow the cross-national pattern: Americans don’t like the idea of big government, the US labor movement is relatively weak, and our government has multiple veto points that allow blockage of proposed policy changes.2

On the other hand, over the past century a lot of the progressive legislation has come when the Democrats held the presidency, a majority in the House of Representatives, and a large (filibuster-proof) majority in the Senate — particularly 1933-38, 1964-65, and 2009. Some, but less, has passed when the Democrats weren’t firmly in charge.

What does a systematic look at over-time patterns in the US tell us about the impact of election outcomes?

I’ll consider 30 outcomes for which we might anticipate a partisan difference. Where possible, I’ll begin in 1933, when the two parties separate clearly on the economic axis, and go up through 2016. A partisan difference should have two features: an observed robust difference in the outcome between Democratic and Republican periods and a plausible story about how the difference is genuinely causal.

In defining “Democratic” and “Republican” periods, it makes sense to focus on presidents. One reason is that, as figure 2 shows, the presidency, unlike the congress (especially the House), has shifted back and forth between parties, so party effects are less likely to be conflated with period-specific factors. Another reason is that the ideological differences between Democratic and Republican presidents have been large and consistent since the 1930s. Figure 3 shows the ideological position of each president, on a left-right scale, along with averages for each party in the Senate and House. It’s only in recent decades that the differences between Democrats and Republicans in the two congressional chambers have matched those of presidents.

Figure 2. Partisan control of government
Blue: Democratic control. Red: Republican control.

Figure 3. Policy makers’ political orientation
Average DW-nominate liberalism-conservatism dimension scores. For the House of Representatives and the Senate, scores are based on votes cast. For the president, scores are based on their announced support or opposition on congressional bills. The scale runs from liberal (negative scores) to conservative (positive scores). Data sources: “The Polarization of the Congressional Parties,” voteviewblog.com, 2016; “The Presidential Square Wave Through the 113th Congress,” voteviewblog.com, 2016.

Of course, presidents aren’t omnipotent. They can’t control large economic and social developments, and much of what they propose can be blocked by congress. But they can introduce legislation, issue executive orders, guide regulatory agencies, nominate the Federal Reserve chair and Supreme Court justices, and shape the public narrative on issues, and they have a good bit of control over foreign policy.

Let’s begin with an outcome on which we’d expect to see a clear partisan difference: income inequality. Several recent studies conclude that this is indeed what we observe, with income inequality rising more under Republican presidents than under Democratic ones.3

Prior to the late 1970s, presidents had substantial influence on both fiscal and monetary policy. Democratic presidents used this to prioritize rapid economic growth and low unemployment, whereas Republicans tended to aim for low inflation. Economic growth was faster under Democrats (see below). While a recent analysis finds that Democratics’ economic growth advantage owed mainly to luck rather than to policy choices, the low unemployment under Democratic presidents ensured that growth led to rising wages, and hence rising incomes, for Americans on the middle and lower rungs of the economic ladder. So up to 1979 income inequality trends did differ significantly between Democratic and Republican years, as Larry Bartels has documented.4

Since the late 1970s, the Federal Reserve has made monetary policy decisions independent of the president’s wishes, so presidents have had limited ability to influence unemployment. As a result, during this period presidents appear to have had less impact on income growth for households in the middle and below.5

The rise in income inequality in recent decades has occurred primarily between the top 1% and everyone else. Income growth has been very rapid for those at the top, less rapid for the upper-middle, and relatively slow for most others. Surprisingly, though, this surge in top-end inequality has been just as large under Democratic presidents as under Republican ones, as figure 4 shows.

Figure 4. Income inequality: top 1%’s income share
Share of income going to the top 1% of tax units. Pretax income, including capital gains. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). The second numbers are for 1980ff. Data source: World Wealth and Income Database.

Why don’t we observe a partisan presidential difference in the surge in top-end income inequality since the late 1970s? Democratic and Republican presidents have differed on tax rates for the top 1%, as we see in figure 5. But that has had a limited impact on the trend in inequality.6 Democratic and Republican presidents have both contributed to deregulation of finance. One noteworthy attempt by a Democratic president to slow the rise in income inequality ended up backfiring. In 1993, the Clinton administration changed the tax rules so that firms could only deduct high CEO pay from revenues if such pay were pegged to CEO performance. This led to a surge in firms’ use of stock options, which became a key source of the massive rise in CEO compensation.7 Some important contributors to inequality’s rise haven’t been strongly influenced by policy choices and so haven’t differed much between Democratic and Republican administrations: technology, globalization (figure 6), corporate governance, and union weakening (figure 7). As figure 8 shows, another, stock values, has increased much more under Democratic presidents than under Republican ones, though it’s unlikely that this owes to any difference in policy.

Figure 5. Effective tax rate on the top 1%
Federal tax payments as a share of pretax income for the top 1% of taxpayers. Federal taxes include personal income, corporate income, payroll, and excise. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change. Data source: Congressional Budget Office, “The Distribution of Household Income and Federal Taxes, 2013,” Report 51361, 2016, supplemental data, worksheet 1.

Figure 6. Imports
Share of GDP. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: St. Louis Fed, FRED database, series B021RE1A156NBEA.

Figure 7. Unionization
Share of employees who are union members. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). The second number in parentheses for Democratic presidents excludes 1933-52. Data sources: 1933-82 are from Richard B. Freeman, “Spurts in Union Growth: Defining Moments and Social Processes,” in The Defining Moment: The Great Depression and the American Economy in the Twentieth Century, edited by Michael D. Bordo et al, University of Chicago Press, 1998, table 8A.2. 1982ff are from Bureau of Labor Statistics, data.bls.gov, series LUU0204899600, using Current Population Survey data.

Figure 8. Stock market
S&P 500. Adjusted for inflation. The lines are loess curves. The numbers in parentheses are average ten-year change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: St. Louis Fed, FRED database, series sp500.

The American economy has grown massively since the early 1930s, with GDP per capita rising at an average yearly rate of about 2%. As figure 9 shows, the rate of growth has tended to be much faster under Democratic presidents than Republican ones.

The standard account attributes this to macroeconomic policy: Democrats tend to prioritize economic growth and low unemployment, whereas Republicans favor low inflation, and presidents representing the two parties have used fiscal and monetary policy to pursue those differing aims.8 Since the late 1970s, things have changed: presidents have had little influence on the Federal Reserve’s monetary policy decisions, and fiscal policy under Republican presidents has been just as expansionary as under Democrats. During this latter period there has been no partisan difference in rates of economic growth.

A recent analysis by Alan Blinder and Marc Watson suggests that the macroeconomic policy explanation may not hold even for earlier decades. They find that the partisan gap in economic growth owes mainly to oil shocks, productivity shocks, military spending, wars, foreign economic growth, and consumer expectations, not to differences in macroeconomic policy.9

Figure 9. GDP per capita
Real gross domestic product (GDP) per capita. “k” = thousand. The lines are loess curves. The numbers in parentheses are average yearly growth rates, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). The second numbers are for 1980ff. Data source: Bureau of Economic Analysis, bea.gov, table 7.1, line 10.

Figure 10 shows the employment rate under Democratic and Republican presidents. Here we see a clear difference. The employment rate has increased in every Democratic period while holding steady or falling in all but one Republican period. This is in line with the notion that Democrats tend to prioritize jobs while Republicans worry more about keeping inflation in check.

Figure 10. Employment rate
Employed persons aged 25-64 as a share of the population aged 25-64. The lines are loess curves. The numbers in parentheses are average ten-year change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: Bureau of Labor Statistics, data.bls.gov, series LNU02000060, LNU02000095, LNU00000060, LNU00000095.

Poverty is another outcome for which we would expect to find a partisan difference. It has tended to be a core concern for Democrats, less so for Republicans. And Democrats have been more willing than Republicans to increase expenditures on transfer programs for Americans with low incomes. Figure 11 shows the official poverty rate since the measure’s inception in 1959. Poverty has fallen in all but one Democratic period, while under Republicans it’s either held steady or increased.

Figure 11. Poverty rate
Share of persons living in a household with income below the poverty line. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). The second number in parentheses for Democratic presidents excludes 1961-68. Data source: Census Bureau, “Historical Poverty Tables,” census.gov.

In a similar vein, Democrats have frequently pushed for an increase in the statutory wage floor, the “minimum wage,” whereas Republicans have tended to favor maintaining the existing level or reducing it. Figure 12 shows that the minimum wage has risen, in inflation-adjusted terms, in all but one Democratic period. During Republican presidencies an increase in the minimum has been the exception rather than the rule.

Figure 12. Minimum wage
Dollars per hour. Federal statutory minimum, adjusted for inflation. The lines are loess curves. The numbers in parentheses are average ten-year dollar change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: Bureau of Labor Statistics.

Do we see partisan differences in outcomes for the middle class? One indicator is median income. Figure 13 shows median family income. (Household is a better unit than family, as it includes adults who live alone, but household data aren’t available until 1967.) It has tended to rise much more during Democratic presidencies than Republican ones. This owes partly to faster economic growth, but employment and wage developments also have contributed.10

Figure 13. Median income
Real median family income. “k” = thousand. The lines are loess curves. The numbers in parentheses are average yearly dollar change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). The second numbers are for 1980ff. Data source: Census Bureau, “Historical Income Tables,” census.gov, table F-5.

Median wealth data begin only in 1983 and are available only every third year. As figure 14 shows, there is no partisan effect. Median wealth has been effectively flat, soaring briefly during the housing bubble between the mid-1990s and the mid-2000s but then falling back to its previous level when the bubble burst beginning in 2006.11

Figure 14. Median wealth
Real average household wealth in the middle quintile. The lines are linear regression lines. The numbers in parentheses are average three-year dollar change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: Edward Wolff, using Survey of Consumer Finances data. The survey is every 3 years, beginning in 1983.

Homeownership is a key marker of middle-class status for Americans, and it’s also the chief source of wealth for the non-rich. Figure 15 suggests no partisan difference in the homeownership rate, which has been essentially flat since 1960.

Figure 15. Homeownership
Share of households that own their residence. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data sources: 1900-1960 are from Census Bureau, “Historical Census of Housing Tables,” using decennial census data; 1960ff are from Census Bureau, “Housing Vacancies and Homeownership: Historical Tables,” using Current Population Survey data.

Let’s turn now to two measures of inclusion. Figure 16 shows the ratio of median household income for African Americans to that for whites. The ratio has increased, on average, under Democratic presidents, while it’s been flat when Republicans have been in office. But the difference is entirely a function of the Clinton years in the 1990s, and there is no clear policy story to account for this. The most sensible conclusion is that there has been no genuine partisan difference.

Figure 16. Black-white income ratio
Ratio of black to white median household income. The lines are loess curves. The numbers in parentheses are average ten-year change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: Census Bureau, “Historical Income Tables,” table H-5.

The same is true of the female-to-male pay ratio, shown in figure 17. Here it’s Republicans that look better. But that owes entirely to the Reagan-Bush years in the 1980s, and the apparent progress during those years is misleading. Prior to 1980 median pay was rising for both women and men, at about the same pace, so the pay ratio held constant. In the 1980s pay for women continued to increase, but for men it flattened. That not-especially-desirable development accounts for the rising pay ratio.12

Figure 17. Female-male pay ratio
Ratio of female to male median annual earnings for persons employed full-time year-round. The lines are loess curves. The numbers in parentheses are average ten-year change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: Census Bureau, “Historical Income Tables,” table P-38, using Bureau of Labor Statistics data.

Have Democrats or Republicans done better at advancing educational attainment? One measure, shown in figure 18, is the share of the population who get a four-year college degree. It’s been rising at a steady pace since the late 1940s, with no partisan difference.

Figure 18. College graduates
4 years college or more. Age 25-34. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: Census Bureau, “Years of School Completed by People 25 Years and Over, by Age and Sex,” using Current Population Survey data.

The same is true for life expectancy, shown in figure 19.

Figure 19. Life expectancy
Years at birth. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: OECD.

Since the 1940s, Democrats have sought to use government to expand access to health insurance, whereas Republicans have preferred to leave this to individuals and firms. As figure 20 indicates, the two big jumps in the share of Americans with health insurance in the past half century occurred under Democratic presidents. Both owed to policy change: the creation of Medicare and Medicaid in 1965 and the Affordable Care Act in 2010.

Figure 20. Health insurance coverage
Share of the population with health insurance. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: Council of Economic Advisors, “Methodological Appendix: Methods Used to Construct a Consistent Historical Time Series of Health Insurance Coverage,” 2014, using data from the National Health Interview Survey and other surveys.

Republicans tend to emphasize crime control more than Democrats. Yet as figure 21 shows, there’s been no partisan difference in the homicide rate. The same is true for the incarceration rate, shown in figure 22.

Figure 21. Homicides
Share of the population killed by homicide. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: 1960ff are from FBI, “Uniform Crime Reports”; 1933-60 are from FBI, “Uniform Crime Reports,” as reported in M. Dwayne Smith and Margaret A. Zahn, Homicide: A Sourcebook of Social Research, Sage, 1998, figure 2.1.

Figure 22. Incarceration
Share of the population in federal prison, state prison, or local jail. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: Bureau of Justice Statistics, “Total Correctional Population.”

Republicans tend to favor few if any restrictions on Americans’ access to guns, whereas many Democrats prefer more restrictions. Figure 23 shows a meaningful partisan difference in the rate of gun ownership. Much of this owes to sizable drop in the 1990s, but that wasn’t a fluke. The Clinton administration’s 1994 crime law banned some types of assault weapons and put tighter limits on the purchase of handguns. The share of American homes with a gun fell by nine percentage points over the ensuing decade, but then the restrictions expired and gun ownership began to increase again.

Figure 23. Gun ownership
Share of homes with a gun. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: Gallup.

Republicans have expressed much more concern than Democrats about the decline of marriage and two-parent families. Figure 24 shows that the share of 35-to-64-year-olds who are married has declined more rapidly under Democratic presidents than under Republican ones. On the other hand, as figure 25 shows, the out-of-wedlock birth rate has increased more rapidly under Republican presidents than under Democrats. Both of these, I suspect, are correlations with no actual causal effect.

Figure 24. Marriage
Married persons age 35-64 as a share of all persons age 35-64. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: General Social Survey, sda.berkeley.edu, series marital.

Figure 25. Out-of-wedlock births
Share of children born to unmarried women. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: National Center for Health Statistics.

Republicans and Democrats also have disagreed, at least during the past decade, about whether homosexuals should be allowed to marry. It’s no surprise that, as figure 26 shows, the big expansion in access to legal same-sex marriage occurred under a Democratic president, though it was state legislatures and courts and then the Supreme Court that were responsible for the expansion.

Figure 26. Legal same-sex marriage
Number of states in which same-sex marriage is legal. The lines are loess curves. The numbers in parentheses are average ten-year change.

Abortion is another social issue on which Republicans and Democrats tend to be on opposite sides. Figure 27 suggests no partisan difference in the frequency of abortions. The rate increased under both Republican and Democratic presidents in the 1970s and then subsequently fell under both.

Figure 27. Abortions
Abortions per 100 women age 15-44. The lines are loess curves. The numbers in parentheses are average ten-year change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data sources: Rachel K. Jones and Kathryn Kooistra, “Abortion Incidence and Access to Services in the United States, 2008,” Perspectives on Sexual and Reproductive Health, 2011, table 1; Rachel K. Jones and Jenna Jerman, “Abortion Incidence and Service Availability in the United States, 2011,” Perspectives on Sexual and Reproductive Health, 2014, table 1.

Republicans have been much more likely than Democrats to favor laws benefitting organized religion. While the US remains a very religious country, religiosity has declined somewhat since the 1960s. As figure 28 shows, there is no apparent partisan contribution to this decline.

Figure 28. Religiosity
Share of the population with some religious preference. Question: “What is your religious preference — Protestant, Roman Catholic, Jewish, another religion, or no religion?” The numbers shown here are calculated as 100 minus the share who respond “no religion.” The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: Gallup, “Religion,” gallup.com/poll/1690/religion.aspx.

Immigration has become a partisan issue only recently. Prior to the second term of George W. Bush, divisions on immigration policy tended to be mainly within the two parties rather than between them. As figure 29 shows, there is no noteworthy partisan difference in immigration trends. The foreign-born share of the US population declined following the (bipartisan) restrictionist legislation passed in the 1920s and then began rising after the (bipartisan) 1965 reform.

Figure 29. Immigration
Foreign-born share of the population. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: Census Bureau.

One of the most consistent differences between Democrats and Republicans has concerned military spending, with Republicans generally favoring more and Democrats less. Figure 30 shows that we don’t observe this difference over the whole of the period from 1933 to 2016. If we focus on the years since 1980, we do.

Figure 30. Military spending
Share of GDP. The lines are loess curves. Years of extremely high spending due to World War II are omitted. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). The second numbers are for 1980ff. Data sources: 1914-48 are from Correlates of War Project, cow.dss.ucdavis.edu; 1949ff are from SIPRI Military Expenditure Database.

Each of the parties has gone through periods in which it prioritized a balanced government budget and others in which it downplayed this issue. Over the past generation Republicans have tended to express the most worry about the dangers of a large government debt, and yet as figure 31 indicates, it is Republican administrations that have contributed the most to that debt, largely due to combining tax cuts with hikes in military spending.

Figure 31. Government debt
Share of GDP. The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). The second numbers are for 1980ff. Data source: St. Louis Fed, FRED database, series GFDGDPA188S.

Trust is important in a liberal society. It underpins democracy, lubricates the economy, bolsters community, and makes life more pleasant.13 In the United States, trust has been declining since the early 1970s (perhaps since the mid-1960s). It isn’t clear why there would be a partisan difference, yet as figure 32 reveals, this decline has occurred almost entirely under Republican presidents.

Figure 32. Trust
Share of US adults saying “most people can be trusted.” Other response option: “you can’t be too careful in dealing with people.” The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: General Social Survey, sda.berkeley.edu, series trust.

Both parties surely want Americans to be happy. Has either done better than the other at contributing to happiness? Data from the General Social Survey, shown in figure 33, suggest that the share of Americans saying they are “very happy” has tended to increase during Democratic presidencies and decrease during Republican ones. Research using data from a large number of countries finds that economic security and individual autonomy tend to boost happiness.14 Democrats have been more likely than Republicans to pursue both, so it’s possible that this partisan difference is causal. But it’s difficult to be certain.

Figure 33. Happiness
Share of US adults saying they are “very happy.” The lines are loess curves. The numbers in parentheses are average ten-year percentage-point change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: General Social Survey, sda.berkeley.edu, series happy.

Last, though definitely not least in importance, let’s look at climate change. Over the past several decades the two parties have differed sharply in their orientation, with Democrats acknowledging human-caused climate change and Republicans tending to deny it. Democrats also have tended to favor aggressive policies to reduce greenhouse gas emissions. But they’ve had little success in getting such policies enacted. Even if they had, slowing climate change would inevitably take time. It isn’t too surprising, then, that in figure 34 we see no partisan difference in average change in the earth’s measured temperature.

Figure 34. Earth’s average temperature
Difference from the 1910-2000 average. Land and ocean. The lines are loess curves. The numbers in parentheses are average ten-year change, excluding changes from the last year in office of an outgoing party to the first year in office of an incoming party (e.g. change from 2000 to 2001). Data source: National Oceanic and Atmospheric Administration, ncdc.noaa.gov/cag/time-series/global.

So what have we discovered? While these patterns need to be investigated more thoroughly, the data suggest no clear difference between Democratic and Republican presidents on 20 of the 30 outcomes:

  • Income inequality: top 1%’s share
  • Economic growth
  • Median wealth
  • Homeownership
  • Stock market
  • Unionization
  • Black-white income ratio
  • Female-male pay ratio
  • College graduates
  • Life expectancy
  • Homicides
  • Incarceration
  • Marriage
  • Out-of-wedlock births
  • Abortions
  • Religiosity
  • Immigration
  • Imports
  • Trust
  • Earth’s average temperature

We do observe a partisan difference for 10 of the outcomes (the party achieving better performance is listed in parentheses):

  • Employment (D)
  • Poverty (D)
  • Minimum wage (D)
  • Median income (D)
  • Health insurance (D)
  • Gun ownership
  • Legal same-sex marriage (D)
  • Military spending
  • Government debt (D)
  • Happiness (D)

An important caveat: The focus here has been on partisan effects that show up within a fairly short period of time, but for some outcomes, such as economic growth, life expectancy, happiness, and climate change, the impact of policy changes may be visible only with a considerable time lag.

PARTISAN EFFECTS: CROSS-STATE COMPARISON

It would be helpful to also compare across the US states. But this would need to be done very carefully. A number of conservative southern states had Democratic governors and legislatures until relatively recently, due to the legacy of the Civil War. And some moderate or liberal states, particularly in the northeast and the west, have frequently had Republican governments. Devin Caughey and colleagues have created a measure of policy liberalism for state governments. Figure 35 shows the trend in liberalism for each of the 50 states since the mid-1930s. The red and blue lines running through the middle of the graph show the average for states that have, respectively, a Republican governor or a Democratic governor. According to this measure, it isn’t until the late 1990s that policies in states with a Democratic governor were more likely to be liberal than policies in states with a Republican governor.

Figure 35. State policy liberalism
State policy liberalism scores, based on nearly 150 policies (at least 43 in each year). Blue and red loess lines indicate the average policy liberalism of states with, respectively, Democratic and Republican governors. Source: Devin Caughey, Christopher Warshaw, and Yiqing Xu, “Incremental Democracy: The Policy Effects of Partisan Control of State Government,” 2016, figure 1.

SUMMARY

In the world’s rich democratic nations, party control of government has had a large effect on government policy, particularly the expansiveness and generosity of public insurance. This has contributed to cross-country differences in economic security, poverty, and income inequality.

In the United States, the “left” party is more moderate and presidents have less ability (due to separation of powers) to enact their preferred policies, so we would expect to observe limited effects of party control of the presidency. On quite a few outcomes, we do in fact see little or no impact. But we also observe real and sometimes strong effects of presidents on a significant number of outcomes.


  1. Walter Korpi, The Democratic Class Struggle, Routledge, 1983; Gøsta Esping-Andersen, Politics Against Markets, Princeton University Press, 1985; Esping-Andersen, The Three Worlds of Welfare Capitalism, Princeton University Press, 1990; Robert E. Goodin, Bruce Headey, Ruud Muffels, and Henk-Jan Dirven, The Real Worlds of Welfare Capitalism, Cambridge University Press, 1999; Alexander Hicks, Social Democracy and Welfare Capitalism, Cornell University Press, 1999; Evelyne Huber and John D. Stephens, Development and Crisis of the Welfare State, University of Chicago Press, 2001; Walter Korpi and Joakim Palme, “New Politics and Class Politics in the Context of Austerity and Globalization: Welfare State Regress in 18 Countries, 1975-1995,” American Political Science Review, 2003; David Brady, Rich Democracies, Poor People, Oxford University Press, 2009; Manfred G. Schmidt, “Parties,” in Oxford Handbook of the Welfare State, edited by Francis G. Castles et al, Oxford University Press, 2010. 
  2. Lane Kenworthy, “How Much Public Insurance Do Americans Want?,” The Good Society; Kenworthy, “Democracy,” The Good Society. 
  3. Larry Bartels, Unequal Democracy, 2nd edition, Princeton University Press, 2016, ch. 2; Thomas W. Volscho and Nathan J. Kelly, “The Rise of the Super-Rich: Power Resources, Taxes, Financial Markets, and the Dynamics of the Top 1 Percent, 1949 to 2008,” American Sociological Review, 2012; David Jacobs and Jonathan C. Dirlam, “Politics and Economic Stratification: Power Resources and Income Inequality in the United States,” American Journal of Sociology, 2016. 
  4. Bartels, Unequal Democracy, figure 2.1. 
  5. Lane Kenworthy, “How Much Do Presidents Influence Income Inequality?,” Challenge, 2010; Bartels, Unequal Democracy, figure 2.5. 
  6. Lane Kenworthy, “Income Inequality,” The Good Society. 
  7. Kevin J. Murphy, “Executive Compensation: Where We Are, and How We Got There,” in Handbook of the Economics of Finance, vol. 2A, edited by G.M. Constantinides et al, Elsevier, 2013. 
  8. Edward R. Tufte, Political Control of the Economy, Princeton University Press, 1978; Douglas A. Hibbs, The American Political Economy: Macroeconomics and Electoral Politics, Harvard University Press, 1987. 
  9. Alan S. Blinder and Mark W. Watson, “Presidents and the U.S. Economy: An Econometric Exploration,” American Economic Review, 2016. 
  10. Lane Kenworthy, “Shared Prosperity,” The Good Society. 
  11. Lane Kenworthy, “Wealth Inequality,” The Good Society. 
  12. Lane Kenworthy, “Inclusion: Women,” The Good Society. 
  13. Lane Kenworthy, “Trust,” The Good Society. 
  14. John F. Helliwell and Shun Wang, “World Happiness: Trends, Explanations, and Distribution,” in World Happiness Report, 2013.