The Significance of the Union Vote in 2012 and Beyond
Introduction
Only slightly more than 100 years ago, on March 25, 1911, the modern labor movement was birthed out of tragedy. That winter, in Manhattan’s lower east side; the Triangle Shirtwaist Factory experienced a devastating fire. 146 women and children died that day, either burning to death or jumping to avoid the flames. For those that tried to escape, the plummet from a ten-story window was the only way out, because the owners of the factory decided to lock the doors – a measure intended to prevent theft. This tragedy was the catalyst for a nationally supported labor union movement, intent on enhancing worker’s rights and protections. Shortly thereafter, the Secretary of Labor was made a cabinet level position and President Roosevelt’s “New Deal” brought about revolutionary labor laws, intended to enhance worker’s rights and protect them from many of the inhumane factory conditions that plagued America’s industrial age. It was Roosevelt’s policies that caused labor unions to align with democratic politics (Allen).
Although the United States never developed a “Labor Party,” protecting workers became a national and political issue and, a century later, unions are a consistent and politically active voting block with a great deal of influence. While union membership has declined in the last 20 years, it would be hard to tell simply by reading the paper in recent months. Both Ohio and Wisconsin, two key battleground states for the 2012 presidential election, have been faced with protests and unrest regarding collective bargaining rights. Republican governors Kasich (OH) and Walker (WI) have engaged in self-proclaimed ‘union-busting’ and the voters of Ohio and Wisconsin have taken a stand. This past November, Ohio voters supported the repeal of a union ballot measure, called “SB 5,” by 57-32 - almost a 2 to 1 margin. If SB 5 had remained law, it would have limited the collective bargaining rights of teachers, police officers and fire fighters by denying the ability to strike and to negotiate pensions and health benefits. Similarly in Wisconsin, Governor Walker faces a concerted recall election next year in response to his “union busting” behaviors (Hananel). Given the significance of labor unions to America’s historical and contemporary political environment, this paper will examine the national role of labor unions in presidential elections from 1960 – 2008, as well as take an in-depth look at key battleground states from the 2008 elections in order to frame a discussion on what implications the union vote could hold in the upcoming 2012 presidential race.
Theory
In Robert Brownstein’s article in the National Journal, he cites an AFL-CIO survey of the voting sentiments behind overturning governor Kasich’s union ballot measure in Ohio. This survey reveals that the desire to repeal was fairly universal among union members and non-union members alike. An explanatory 86% of union voters wanted to repeal SB 5, as well as 52% of non-union voters, 92% of liberals and 70% of moderates. This issue crossed color boundaries as well, as a majority in both white and minority communities voted to repeal the measure. The only group that did not support the repeal was conservatives, who voted against it by almost 2 to 1 (Brownstein). At a time where partisanship is at its highest in decades, this unifying issue - at least amongst the majority of Ohioans – reveals many interesting possibilities, particularly in an election year.
Cable news networks seem to be enforcing the idea that this recent union victory in Ohio symbolizes a comeback for union issues, and thus democratic issues, in national politics. However, many other factors have to be considered in making this determination. For instance, union membership has been declining in the past several decades and reduced numbers don’t bode well for political viability. Furthermore, as Anne Kim says in her article for the Progressive Policy Institute, unions “bridge the Democratic ‘base versus moderate’ divide (Kim 1)." While it is true that unions have historically aligned themselves with Democrats, a whopping 74% of union members self-identify as moderate or even conservative – often taking more socially conservative positions. However, according to Kim, “moderates in union households were much more likely to vote democratic than moderate voters in none union households (Kim 2-3)." While it is good for democrats to receive high numbers of moderate votes during a presidential election, union membership is declining, which means that democrats will have to become more creative if they wish to maintain their union base and the moderate support it holds (Kim 7). Something that could prove beneficial to President Obama in this endeavor is the fact that unions are gradually becoming more diverse, even compared to the general population. In 2008, African Americans composed 15% of voters in union households, about 2 percentage points higher than the national voting average. Furthermore, Hispanics composed 12% of the union vote, as opposed to their 8% national average (Kim 5). This increased diversity and decline in the ‘white-working-class’ block should allow President Obama to play to his strengths. After all, he was overwhelmingly victorious amongst minority voters in the 2008 election (Francia 156).
President Obama seems to be aware of the importance of maintaining the union vote for his re-election campaign. This was made evident during the Ohio ballot vote, when he aided labor unions in their campaigns to motivate their workers to the polls. However, according to the New York Times, this aid was both “largely unnoticed” and “proof that Mr. Obama’s supporters can be revved up and turned out (Shear)." These two points seem to be at odds with each other. If Obama’s push was, in fact, “unnoticed” then why would this present proof that it was Obama’s “supporters” that came out in force? As union voters are typically considered to be more “blue-dog” in their political affiliation, it is highly possible that these Ohioans were devoted to unions but not necessarily President Obama. While the Ohio vote certainly demonstrated that unions still have a great deal of political capital, even with their declining numbers, Ohio’s significance as a battleground state makes understanding this election imperative to winning the 2012 race. As Sam Hananel posits in his piece for the Associated Press, “The question for many is whether to interpret Tuesday's Ohio referendum as simply a rejection of Republican overreach in a heavily unionized state or, more broadly, as a barometer of a battleground state that could resonate with voters nationwide (Hananel)." Certainly party leadership, such as Democratic National Committee Chairwoman Debbi Wasserman Schultz, has come out in full force to claim this, not just as a union victory, but as a democratic victory and a victory for the president as well. In one quote, she even associates the repealed measure with “the likes of Mitt Romney and the millions of dollars in tea party and special interest money that poured into Ohio” (Hananel).
While President Obama did issue a statement congratulating “the people of Ohio for standing up for workers,” he also has a challenging election coming up. On one hand, he must maintain his base, which traditionally includes unions, while also appealing to the broader public. In some regards, the fact that unions are typically comprised of “moderate” voters, should allow him to appeal to his base and the broader public. However, white working-class voters are the majority demographic in unions and they don’t happen to be Obama’s strong suit. While Obama was able to win the 2008 election with somewhat of a landslide, at least some of his victory was a result of President Bush’s devastatingly low approval ratings. If the 2012 election is as close as the 2000 or 2004 election, Obama may not carry this demographic, particularly if “social issues” are allowed to take center stage. As Francia concludes in his article entitled, “What’s the Matter with the White Working Class?” it seems that “the future success of the democratic party in presidential elections will be tied, at least somewhat, to the strength of the labor movement and its continued ability to focus white working-class voters on economic issues (Francia 156)."
While Francia’s assessment is probably correct, with “Occupy Wall Street” and the “99% movement” at the forefront of national dialogue (particularly in regards to the republican candidates for the 2012 nominee) this “focus” does not seem like it will be very hard to attain. Furthermore, although Obama may have a harder time with white working class voters, unions are considered a democratic base because they tend to vote the party line, regardless of the candidate. Thus, I hypothesize that while the union vote is declining, and may very well be rendered insignificant within the century, it is still statistically significant in swaying elections toward the democratic candidate. Furthermore, I hypothesize that through analyzing union data for specific battleground states, it is possible to predict what percentage of the union vote President Obama will need to mobilize in order to be re-elected in 2012.
Research Design
In order to test this hypothesis, I have collected two sets of data. One is composed of statistics that represent national, as opposed to state wide, union election statistics dating back to the Kennedy election in 1960. This first set of data is from a variety of sources, including ANES, the Bureau of Labor Statistics and the Roper Center. I believe that having this level of historic data is beneficial in framing the union vote, both in its decline throughout the years and its potential sway in national elections. The second set of data was compiled using CNN exit polls for the 2008 elections that were conducted in traditional battleground states. The five battleground states I chose were Ohio, Pennsylvania, Michigan, Wisconsin and Minnesota. Each of these states is a potential victory for either President Obama or the Republican nominee and a victory in any of them, especially Ohio, would go a long way to determining the outcome of the election. Each of these states has significant Electoral College points, high union density and Obama happened to win all five of them in 2008.
While, unfortunately, CNN has only made the 2008 exit polls by state available and not 2004, I have included additional data from Ruy Teixeira’s November, 2011 article for the Center for American Progress. This data is not necessarily useful for running regressions, however it does provide helpful background information for each battleground state and has the new distribution of Electoral College votes. Pennsylvania has 20 votes, and the democrats have consistently won them since 1992. However, Gore only won by 4 percent in 2000 and Kerry only won by 3 percent in 2004 (Teixeiria 17). Ohio lost two votes, bringing it down to 18. Democrats lost Ohio by 4 points in 2000 and 2 points in 2004 (Teixiera 20). Michigan has 16 votes, which have gone to the democrats in the last 5 elections. However, in 2000, Bush came within 5 points and in 2004 Bush came within 3 points (Teixiera 24). Wisconsin has 10 Electoral College votes and has been won by the democrats since 1988. However, this margin has been slim, with the democrats only winning by .2% in 2000 and .4% in 2004 (Teixiera 31). Finally, Minnesota has 10 votes and has gone blue for the last 9 presidential elections, however, only by 2% in 2000 and 3% in 2004 (Teixeira 25). This additional data helps to frame the 2008 exit polls, as it demonstrates how high Obama’s margin of victory truly was.
Results
The first table presented below is a sample of the raw data that I collected for my first data set. I found it fascinating how large of a percentage of the national vote the unions have controlled throughout the years (see first column). However, what I found most interesting was the results for my “Union Sway Dem” variable, which represents what percentage of the total national vote the union voters were responsible for swaying. This vote always sways democrat and I was astonished to see that it has been as high as 13.18%. I coded dummy variables for “Dem Win w/ Union” and "Dem Win w/o Union" in an attempt to run a regression analysis on the significance of the union vote for democratic wins. However, no matter how many numbers I crunched, this always came up statistically insignificant. Even though there are clearly elections that the democrats never would have won without the union vote (1960, 1976, 1992, 1996).
| Year | % Union of Total Vote | Total Nation Sway Dem | Union Sway Dem* | Dem Win w/ Union | Dem Win w/o Union |
|---|---|---|---|---|---|
| 1960 | 17.36% | 0.20% | 10.34% | 1 | 0 |
| 1964 | 20.67% | 22.60% | 10.20% | 1 | 1 |
| 1968 | 20.78% | -0.40% | 13.18% | 0 | 0 |
| 1972 | 20.46% | -24.00% | 6.95% | 0 | 0 |
| 1976 | 20.92% | 2.00% | 10.67% | 1 | 0 |
| 1980 | 18.18% | -10.00% | 6.84% | 0 | 0 |
| 1984 | 14.78% | -18.00% | 6.72% | 0 | 0 |
| 1988 | 13.92% | -8.00% | 9.53% | 0 | 0 |
| 1992 | 13.34% | 5.40% | 8.52% | 1 | 0 |
| 1996 | 13.82% | 9.00% | 9.26% | 1 | 0 |
| 2000 | 11.86% | 0.10% | 6.47% | 1 | 0 |
| 2004 | 10.75% | -3.00% | 7.96% | 0 | 0 |
| 2008 | 9.83% | 7.00% | 4.18% | 1 | 1 |
I was, however, able to run one regression with this data set based on the significance of the union vote over time. See below for regression statistics and graph.
| Variables | Coefficient | Standard Error | P-Value | Intercept | 5.783 | .58 | 3.63E-068 |
|---|---|---|---|
| Total % of Union Vote | -.002 | .00 | 4.61E-06 |
This regression (above) is statistically significant because the P-Value is well below .05. This is not surprising, seeing as all of the research in my theory section suggests that the union vote is significantly declining over time. For every additional year since 1960, the union vote (as a percentage of the national vote) declined by .002 or .2%. This regression is represented again in the graph that I have included below.
The next table I have included is a sample of the raw data from my second data set. In the data below, I was again surprised to see how much of the total vote the unions control in these 2008 swing state elections, which is represented by my “% Union Sways D” variable. The union vote, once again, is shown to sway the entire state toward the left. This data reveals that without the union vote, Obama would have lost every state in 2008. This is evident because the percent the union sways the total vote democrat is greater than the election margin.
| State | % Obama total | % McCain Total | % Obama Union | % McCain Union | % Union Household | % Union Sways D* | % Election Margin |
|---|---|---|---|---|---|---|---|
| OH | 51.38% | 46.80% | 56% | 43% | 28% | 15.68% | 4.58% |
| PA | 54.47% | 44.15% | 62% | 37% | 27% | 15.74% | 10.32% |
| MI | 57.33% | 40.89% | 67% | 31% | 34% | 22.78% | 16.44% |
| WI | 56.22% | 42.31% | 61% | 39% | 26% | 15.86% | 13.91% |
| MN | 54.06% | 43.82% | 56% | 42% | 30% | 16.8% | 10.24% |
Even with how seemingly significant this data is, I was again having trouble coming up with a statistically significant regression. However, once I changed my confidence interval from 95% to 90%, I was able to determine that the percent of the union vote that voted for Obama had some effect on the % of total vote that Obama received in each state. I have included by summary statistics and regression table below.
| Variable | Mean | Variance | St. Dev | # of Obs |
|---|---|---|---|---|
| % Obama Union | 60% | .00 | .05 | 5 |
| % Obama Total | 54.69% | .00 | .02 | 5 |
| Variable | Coefficient | Std. Error | P-Value |
|---|---|---|---|
| Intercept | .30 | .09 | .05 |
| % of Union Dems for Obama | .41 | .15 | .07 |
Although not as strong of an indicator as a 95% confidence interval, if we accept a 90% confidence interval, it appears that this regression is statistically significant. This would be significant because the P-Value is below .1%. This regression suggests that for every 1% increase in the union vote Obama received in 2008, he gained .41% of the total vote.
Discussion/Conclusion
As my data did not yield to any truly statistically significant regressions, it seems that my hypothesis was generally wrong. While I was correct in asserting that the union vote is declining over time, once I began my theory research this seemed to be fairly common knowledge. Additionally, I was unable to make my data reflect that, either nationally or in the 2008 Swing States, the union vote was statistically significant. However, while data doesn’t lie, this seems flawed to me. In each data set, it was clear - simply given the raw numbers - that democrats would have lost in more, if not every race without the union vote. For the 2008 swing states, it would have been very helpful to have more data and perhaps a larger data set would have yielded a statistically significant regression with a 95% confidence interval as opposed to a 90% confidence interval. Even the data I had acquired for this assignment was difficult to find and I was very pleasantly surprised when I found CNN’s exit polling broken down by state.
However, I do believe that given a combination of my research and my 90% confidence interval regression, which demonstrates a positive correlation between union turnout and total turnout for president Obama, it is safe to say Obama has a high likelihood of taking several (if not all) of the swing states in my analysis In 2012. Between the union bill in Ohio and Scott Walker’s disapproval rating in Wisconsin, Obama is likely to be seen by the voters as “moderate.” Additionally, with the recent democratic surge due to increased union activity as well as Obama’s past victory in each one of these states, he has a strong foothold for the 2012 election. Furthermore, with the dialogue around the 2012 election revolving around jobs and the economy, Obama is likely to see a similar, if not greater turnout of support for his re-election as we saw for the repeal of the union measure.
Bibliography
- Allen, Jodi. "A Century After Triangle, Unions Face Uncertain Future." Pew Research Center Publications. (Mar 2011): n. page. Web. 13 Dec. 2011. <http://pewresearch.org/pubs/1938/triangle-shirtwaist-factory-fire-public-opinion-about-labor-unions-trend>.
- Brownstein, Ronald. "A Model for Obama?." 2012 Decoded Presidential election insight and analysis powered by National Journal. National Journal, 11/09/2011. Web. 13 Dec 2011. <http://decoded.nationaljournal.com/2011/11/a-model-for-obama.php>.
- "Election Center 2008." CNN Politics. CNN, Web. 13 Dec 2011. <http://www.cnn.com/ELECTION/2008/results/polls/>.
- Francia, Peter, and Nathan Bigelow. "What's the Matter with the White Working Class? The Effects of Union Membership in the 2004 Presidential Election." Presidential Studies Quarterly. 40.1.
- Hananel, Sam. "Ohio Vote Shows Unions Still a Political Force." Associated Press [Cincinnati] 09 Nov 2011, n. pag. Web. 13 Dec. 2011. <http://www.wcpo.com/dpp/news/political/elections_local/ohio-vote-shows-unions-still-a-political-force>.
- Kim, Anne, and Stefan Hankin. "Union Voters and Democrats." Progressive Policy Institute Political Memo. May (2011): 1-9. Print.
- McDonald, Michael. "Voter Turnout Data." United States Election Project. George Mason University, 2010. Web. 13 Dec 2011. <http://elections.gmu.edu/voter_turnout.htm>.
- "Opposition to Ohio's SB-5 Grows." Quinnipiac University Poling Institute. Quinnipiac University, 25 Oct 2011. Web. 13 Dec 2011. <http://www.quinnipiac.edu/x1322.xml?ReleaseID=1665>.
- Shear, Michael. "In Tuesday’s Election Results, Democrats See Hope for 2012." New York Times [New York, NY] 09 Nov 2011, n. pag. Web. 13 Dec. 2011. <http://thecaucus.blogs.nytimes.com/2011/11/09/in-tuesdays-election-results-democrats-see-hope-for-2012/>.
- Teixeira, Ruy, and John Halprin. "The Path to 270: Demographics versus Economics in the 2012 Presidential Election ." Center for American Progress. Nov (2011): 1-68. Web. 13 Dec. 2011. <www.americanprogress.org>.
- "Union Membership: Overall (1948-2004)." Working Life. Union Sourcebook 1947-1983; U.S. Bureau of Labor Statistics, n.d. Web. 13 Dec 2011. <http://www.workinglife.org/wiki/Union%20Membership:%20Overall%20(1948-2004)>.
- "US Elections How Groups Voted in 2008." Roper Center Public Opinion Archives. The University of Connecticut, n.d. Web. 13 Dec 2011. <http://www.ropercenter.uconn.edu/elections/how_groups_voted/voted_08.html>.
- "Voter Turnout 1948-2008." The ANES Guide to Public Opinion and Election Behavior. The American National Election Studies, 05 Aug 2010. Web. 13 Dec 2011. <http://www.electionstudies.org/nesguide/2ndtable/t6a_2_2.htm>.
