The Audacity of Hope

I read Obama’s book about his political program during the election and inauguration.

I don’t usually read books by politicians, especially about current politics. However, being an Obama supporter and volunteer, I thought I should read his book, The Audacity of Hope: Thoughts on Reclaiming the American Dream.

I had already read Dreams From My Father and loved Obama’s writing style. The Audacity of Hope is also written well. Of course, I liked Dreams From My Father better, but that’s because of its subject of autobiography and identity.

The Audacity of Hope was written when Barack Obama was a US Senator and I started reading it a bit before the election. I finished it around inauguration time. As to why I didn’t read it earlier? I got it early in 2008 but then most of my spare time was consumed by the election campaign. I started the book only when I went to Pakistan for a couple of weeks in October.

In this book, Obama explains his views and his political program. Having followed his career since his keynote speech at the 2004 Democratic convention, I was familiar with a lot of his views. But two things still stood out. One is how Obama is actually fairly moderate. The other is Obama’s tendency to give an honest airing to conservative views and even agreeing partly before arguing for his liberal viewpoint on any issue. This quality, the so-called post-partisanship, was evident throughout the book.

Dreams from My Father

I usually do not read autobiographies. Or books by politicians. But Dreams from My Father: A Story of Race and Inheritance is an exception.

First of all, it’s written by Barack Obama and was written before he got into politics. So it’s refreshingly honest. Secondly, Obama writes well, very well.

Dreams from My Father is basically about the young Obama’s quest for identity. As a child of a Kenyan black man and a Kansan white woman who was raised by his mother and maternal grandparents, he was somewhat of an outsider to both the African American experience and the general White majority.

It is a work of self-reflection and Obama comes across as a thoughtful and honest person.

I read the book on the way to Pakistan (more about that in my next blog post) and during my stay there.

If you want to understand Obama the man and where he comes from, I think this book is essential reading.

Georgia Election Analysis III

Today I dig deeper into the Presidential election results in Fulton county and specifically the precincts in Alpharetta, Milton and South Atlanta where our team of Obama volunteered worked.

I did some analysis of the Presidential election in the state of Georgia in two installments. Today I’ll look at Fulton county where I live as well as the specific precincts in Alpharetta, Milton and South Atlanta where our team worked as volunteers for the Obama campaign.

Year D R Turnout %Turnout
2000 57.53% 39.68% 264,276 65.06%
2004 59.35% 39.99% 330,791 74.15%
2008 67.07% 32.09% 405,531 73.11%

There was a 15.3% swing towards Obama in Fulton county compared to Kerry in 2004. The turnout, based on the number of registered voters, stayed constant from 2004 to 2008. However, if we use the voting age population estimates (VAP) for Fulton county, we get the following turnout rates:

Year Turnout based on VAP
2000 42.30%
2004 48.00%
2008 52.72%

This shows a bigger turnout in 2008. One reason for the discrepancy between the turnout in the first table and this one is that the Obama campaign focused a lot on voter registration this year and thus got more people registered.

Also, please note that the VAP turnout estimate is probably lower than the real turnout which should be calculated as a proportion of the voting eligible population.

Out of the 3,924,440 votes cast for President in Georgia this year, 2,084,179 (or 53.11%) were cast during advance/early voting or by absentee mail-in. In Fulton county, 184,240 votes (45.42%) were cast early, absentee or provisional out of a total of 405,628 votes cast. This is very unusual for Georgia and Fulton as can be seen in the table below for Fulton county. (This shows the numbers for the 2004 general election for President, the 2006 election for Governor, the 2008 Presidential Primary on Super Tuesday and the 2008 general Presidential election.)

Year Absentee/Early/Provisional votes
2004 Prez 12.68%
2006 Gov 11.76%
2008 Prez Primary 7.65%
2008 Prez 45.42%

In the 2008 general election, the Obama campaign tried to get everyone to vote early. The effect of this can be seen in the early voting numbers in Fulton county where Obama had a 49.5% lead in early voting compared to a 22.9% lead on polling day itself.

The large numbers of early voters have complicated my precinct level analysis. The precinct level data does not include early or absentee votes which are listed separately as one per county. Still let’s see what we can conclude for our precincts.

For the Get Out the Vote (GOTV) campaign in October and then the first four days of November, we were based in the South Atlanta precincts, 12E1, 12J and 12T, which are just north of Atlanta’s Hartsfield-Jackson Airport. These precincts are heavily Democratic (more than 90%) but there are a lot of sporadic voters there. So the main task was to get the vote out.

Election 12E1 12J 12T
Voted %Turnout Voted %Turnout Voted %Turnout
2004 518 49.5% 1024 59.7% 566 62.3%
2006 224 22.1% 473 29.2% 581 35.8%
2008 Primary 280 27.5% 564 34.6% 609 39.1%
2008 747 55.1% 1,360 67.3% 1,232 68.0%

Thus, turnout as a percentage of registered voters increased from 57.4% to 64.3% while actual number of people who voted in these three precincts increased from 2,108 to 3,339, a 58% increase in the number of voters. Compare this 58% increase to a 22.6% increase over the whole of Fulton county and a 19.6% increase over all of Georgia. I am actually pleased at these numbers and I am sure all of our Obama team would be too.

Let’s now look at the precincts in Alpharetta and Milton where we made the most calls, starting in August, for identifying voters, persuading them and then getting them out to vote. The table below shows the Republican advantage over Democrats. For the 2008 Presidential Primary, I added the votes for all the Republican Presidential candidates together and did the same for the Democrats. Let’s take the example of ML01A in the 2008 general election, McCain got 73.5% and Obama got 25.3%, which means a Republican advantage of 73.5-25.3=48.2%.

Precinct 2006 2008 Primary 2008
AP07B +66.0% +33.2% +40.9%
ML01A +64.0% +34.4% +48.2%
ML02A +67.4% +40.8% +62.6%
ML02B +68.0% +45.2% +57.5%
ML03 +66.9% +45.7% +59.5%
ML04 +65.2% +38.5% +56.5%
ML05 +48.9% +14.9% +37.7%
ML06 +29.9% -11.4% +8.2%
ML07 +65.1% +41.8% +52.0%

I couldn’t find the data for the 2004 Presidential election for the Milton (MLxx) precincts but AP07B had a 42.9% Republican advantage in 2004.

Before I opine on those results, let’s look at the turnout (as a percentage of registered voters) in these precincts:

Precinct 2008 Turnout
AP07B 44.6%
ML01A 44.6%
ML02A 55.3%
ML02B 55.3%
ML03 52.5%
ML04 44.7%
ML05 44.2%
ML06 41.9%
ML07 54.2%

We already know that about half of Georgia and Fulton county voters voted early. The early voters are not listed in the last two tables for the Alpharetta/Milton precincts because they are not reported that way by Fulton county. Looking at the election day turnout for Milton, it seems that fewer than half of the voters voted early.

Because we only have the vote breakdown by party for those who voted in person at their polling location on November 4 and a significant number of voters had voted early in 2008, we cannot really say what the margin between McCain and Obama was in Alpharetta and Milton. There is no reason to believe that the early voters had the same partisan distribution as those on election day. In fact, there is reason to believe that the early voters were more likely to be Obama supporters. The Obama campaign had been working hard asking people to vote early. Also, in the whole of Fulton county, Obama had a 49.5% advantage in early voting compared to a 22.9% advantage on polling day. Thus, the very large Republican margin in the Milton precincts in 2008 is most likely not correct. I can say with certainty that McCain won all those precincts except ML06 but probably with a somewhat smaller margin. Even then, Republicans had a huge advantage here in North Fulton. Unfortunately, we don’t have any way of finding out whether our team was able to reduce that Republican advantage or not.

Next: A look at the Senate runoff between Jim Martin and Saxby Chambliss.

Georgia Election Analysis II

I look at the detailed county level results for the 2008 Presidential election and also try to quantify the effects of race.

Continuing with my election analysis in Georgia, my data source is the Secretary of State’s website:

Let’s start with a graph showing Obama’s share of the vote in each Georgia county versus Kerry’s share four years ago. Please note that the size of the circles is proportional to the total number of registered voters in that county.

Obama vote vs Kerry vote

Since there was an 11.4% swing towards Obama compared to Kerry, a majority of the counties show a larger percentage for Obama as compared to Kerry. This is especially true of large (in population) counties. Since Georgia has 159 counties, let’s redo the same graph separately for the large, middle and small counties.

In the largest 53 counties (min number of registered voters 17,354 and max 554,682), only the most Republican counties got more Republican compared to 2004.
Largest counties swing

In the middle 53 counties (registered voters ranging from 8,182 to 17,018), the situation was a lot more mixed, but Obama did lose a lot of ground in some.
Swing in middle counties

In the 53 counties with the lowest populations (registered voters ranging from 1,371 to 7,988), Obama did better in some and worse in others. However, these rural counties don’t have many people and probably didn’t matter much at all for the Presidential election.
Swing in least populated counties

Since one major reason for Obama’s better performance was the higher turnout and vote from African Americans, let’s look at the same data, i.e. Obama’s total vote share compared to Kerry’s, but by categorizing counties by how white they are. I am using the percentage of registered white voters in a county as a proxy for the percent population that’s non-Hispanic White in a county.

The whitest third of the counties are at least 77% white. Do note that the scales of this graph are different from the earlier graphs. Obama did worse in the smaller, whiter counties than Kerry and better in the bigger white counties. But he didn’t get more than about 35% of the vote in any of them.
Swing in the whitest counties

The middle third counties by white percentage of the population have about 64% to 77% whites. Obama generally did better than Kerry here. Obama’s share of the vote varied from 27-45% in these counties except for Clarke county (65%) which contains Athens and the University of Georgia.
Swing in the mid-white counties

In the bottom third counties, the white percentage is 18-64% and Obama did very well here.
Swing in least white counties

Let’s look at Obama’s share of the vote in each county plotted against the percentage of whites and blacks in that county.

White percentage vs Obama vote share

Black percentage vs Obama vote share

The red line in both graphs is the least squares fit. As is obvious, Obama’s share of the vote decreases as the percentage of whites increases in a county.

In the previous two graphs, we looked at Obama’s share of the total vote. So we cannot completely tease out the effects of a large African American turnout or the how the white voters voted in each county. Since we can’t have the numbers for Obama’s share of the white vote in each county, we’ll make some simplifying assumptions to guesstimate it:

  1. The African American turnout in each county was the same as the total voter turnout in that county. This is approximately true at the state level, but could be wrong for some counties.
  2. 98% of the African American voters voted for Obama. This is the percentage at the state level.
  3. The number of voters who are not African American or White is small enough (5%) that we can approximate the white vote share by simply excluding the African American vote.

We know this procedure gives us only a guesstimate of the nonblack vote since in one small county (Clay) this results in Obama getting -1.85% of the nonblack vote. In the other 158 counties, Obama’s share of the nonblack vote varies from 4.4% (Randolph county) to 54.6% (Dekalb county).

White population percentage vs Obama share of nonblack vote

Black population percentage vs Obama share of nonblack vote

The red lines in both graphs are LOESS fits.

The first thing apparent in these two graphs is that Obama did particularly well among whites in most of the populous counties as they show up outliers at the top. Among the other counties, there is not a significant pattern, but it does look like Obama’s nonblack vote share is lowest in counties where African American population is in the 20-30% range.

Razib did a similar analysis of racial polarization in Mississippi.

Charles Franklin analyzed the white vote for Obama at the state level. Nate Silver saw that data and noticed that most of the states where Obama performed worse than Kerry among whites were those where Obama did not campaign.

Andrew Gellman, at his must-read blog Red State Blue State Rich State Poor State, looked at Obama’s vote share divided by race in the West, Northeast, South and Midwest. He also showed the estimated nonblack vote by county for all of the US.

Andrew Gellman also looked at the voting preferences of rich and poor voters in both red and blue states in the 2008 election, something which is the topic of his book Red State, Blue State, Rich State, Poor State: Why Americans Vote the Way They Do.

Finally, Michael McDonald estimates the voter turnout from the vote eligible population.

My revised national turnout rate for those eligible to vote is 61.6% or 131.3 million ballots cast for president. This number does not include an approximate 1.4 million under and over votes, for an estimated total 132.7 million total ballots cast. This represents an increase of 1.5 percentage points over the 60.1% turnout rate of 2004, but falls short of the 1968 turnout rate of 62.5%.

His estimate of turnout for Georgia is 61.1%, up from 56.2% in 2004.

Next: An analysis of the precincts in Alpharetta, Milton and Atlanta where our team worked.

Georgia Election Analysis I

Let’s look at the Georgia results and see how Obama turned a 16.6% defeat of Kerry in 2004 to a 5.2% loss this time.

This is the first in a series of posts looking at the Presidential election results in the state of Georgia.

Let’s start with the top line numbers from the last three Presidential elections:

Year Democrat Republican
Votes Percentage Votes Percentage
2000 1,116,230 43.2% 1,419,720 55.0%
2004 1,366,155 41.4% 1,914,256 58.0%
2008 1,844,137 46.9% 2,048,744 52.1%

In 2000, Al Gore lost Georgia by 11.8 points and Kerry lost by 16.6 points. Obama has reduced that to 5.2 points, which is a swing of 11.4 percent from 2004. This swing is greater than the national swing of 9.2%.

Let’s look at the Georgia exit polls from 2000, 2004 and 2008.

During the last 8 years, the composition of the voters has changed a lot. The table below shows the percentage of voters who belonged to a specific race or ethnic group.

Race 2000 2004 2008
White 81% 70% 65%
African American 19% 25% 30%
Hispanic 4% 3%

That’s quite a change in voter demographics, Compare that to the census estimate of population (in 2006) of 58.9% non-Hispanic whites, 29.9% African Americans and 7.5% Hispanics. This shows that African Americans, probably for the first time in Georgia history, voted in proportion of their population. The credit for that goes to the Obama campaign’s voter registration and Get-Out-the-Vote (GOTV) efforts.

Turnout this year was a little lower than 2004, mostly because of the high turnout among whites in 2004.

2000 2004 2008
Total 69% 77% 76%
White 71% 80% 78%
African American 63% 72% 75%

Since I am estimating White and African American turnout in 2008 from the exit polls, it might be prone to some error. We’ll know more when all the detailed data is released by the Georgia Secretary of State.

If African American turnout could be higher, Obama might have gotten closer by at most 1 point. However, do remember that in absolute numbers 41% more African Americans voted this year compared to 2004. The increase in the number of white voters was less than 9%.

The next table shows why Obama lost. It tells us what portion of whites and blacks voted for the Democratic Presidential candidate in the last three elections.

Race 2000 2004 2008
White 35% 23% 23%
African American 86% 88% 98%

Obama needed around 27% or more of the white vote. However, he got the same proportion as Kerry.

Looking at voter groups, 37% of all voters (i.e. 57% of white voters) were white evangelicals. Obama got only 10% among them. Since he got 23% among all whites, so Obama got 40% of the white non-evangelical vote. Compare to Kerry who got 16% of white evangelical vote and 30% of white non-evangelical vote. Also, compare nationally where Obama got 43% of the white vote and 24% of white evangelical voters, i.e. Obama got a majority (53%) of non-evangelical white vote nationally.

Young voters nationally flocked to Obama in this election, a lot more than in recent elections. This can be seen in the graphs here.

18-29 30-44
US GA US GA
All 66% 48% 52% 56%
White 54% 20% 41% 32%

2008 election US GA age race

The low number (20%) for 18-29 year old white voters in Georgia surprised me a lot. It shows how difficult it might be for Democrats to win Georgia even in the future.

Nationally, Obama did somewhat better with white voters than Kerry in 2004. However, there were regional variations. In Appalachia and parts of the South, Obama got a smaller percentage of the white vote than Kerry did.

Southern counties that voted more heavily Republican this year than in 2004 tended to be poorer, less educated and whiter, a statistical analysis by The New York Times shows. Mr. Obama won in only 44 counties in the Appalachian belt, a stretch of 410 counties that runs from New York to Mississippi. Many of those counties, rural and isolated, have been less exposed to the diversity, educational achievement and economic progress experienced by more prosperous areas.

The increased turnout in the South’s so-called Black Belt, or old plantation-country counties, was visible in the results, but it generally could not make up for the solid white support for Mr. McCain. Alabama, for example, experienced a heavy black turnout and voted slightly more Democratic than in 2004, but the state over all gave 60 percent of its vote to Mr. McCain. (Arkansas, however, doubled the margin of victory it gave to the Republican over 2004.)

Less than a third of Southern whites voted for Mr. Obama, compared with 43 percent of whites nationally.

Do look at this table and a map of the US showing county-level electoral shifts.

MSNBC First Read gives the details of Obama vote share among whites.

In 13 [states], Obama received less than 35% of the white vote. His three lowest performing states: Alabama (10%), Mississippi (11%), and Louisiana (14%). The other 10: GA (23%), SC (26%), TX (26%), OK (29%), AR (30%), UT (31%), AK (32%), WY (32%), ID (33%), and TN (34%). On the other hand, Obama won the white vote in 18 states and DC: CA, CO, CT, DE, DC, HI, IL, IA, ME, MA, MI, MN. NH. NY. OR, RI, WA, WI and VT. Obama’s lowest percentage of the white vote he received in a state that he won: NC (35%). The highest percentage of the white vote Obama received in a state he lost: MT (45%).

Razib has a few maps showing Obama’s performance among white voters in all the states compared to Kerry’s performance in 2004 as well as to the national average for white voters this year.

Pollster shows that in states with higher than about 20-25% of African American population, Obama performed poorly among white voters.

If you are interested in comparing Democratic or Republican performance over the years in Presidential elections in Georgia, Dave Leip’s Atlas of US Presidential elections has the historical results in a convenient table format.

All things considered, Barack Obama did very well in Georgia. I was always skeptical about the Democrats winning Georgia, except in a national landslide. How well did the Obama campaign bigwigs think they could do in Georgia can be gauged from this anecdote from August 27.

[Obama’s election manager David Plouffe] said that the campaign’s target in Georgia is about 47% of the vote, owing to Ex-Rep. Bob Barr’s ability to siphon votes away from John McCain.

In the end, Bob Barr didn’t do well at all, but Obama got his 47% of the vote. It looks to me like the Obama campaign knew what it was doing. Do you remember the Obama spreadsheet that leaked just after Super Tuesday and was remarkably accurate about the primary results?

There’s a lot more I want to analyze, including county and precinct level data, precincts in which our team worked. So tune in for more analysis and rambling soon.

Election Day Live Blogging

I am volunteering for the Obama campaign on election day and will be liveblogging all day (and night).

It’s November 4, Election Day. Why Tuesday? Because it made sense in the 18th and early 19th centuries. It does need to be changed now, however.

7:12am: I am at the Obama campaign Roswell office printing walk packets, volunteer lists and directions etc for our team. I’ll be on my way to South Atlanta as soon as the print job’s finished.

Voting here in Georgia started at 7am.

8:35am: I am at the staging location in South Atlanta. Already at one of the polling places here, the line is more than 2 hours long.

10:39am: The morning rush at the polls has lessened now.

11:51am: Sent out volunteers to go knock on doors to get people out to vote. Also, sent phone lists to volunteers to call.

12:55pm: There are so many people here to volunteer, it’s difficult to even find work for them.

2:07pm: This liveblogging will probably only pick up later tonight when results start coming in. For now, do read my last post about how Georgia is competitive this election.

2:55pm: Intrade has McCain overall win at around 7.3 right now and a Dem win in Georgia at 27.5.

3:21pm: Turnout here in South Atlanta is okay today but add in the early vote numbers and it’s pretty high.

5:01pm: The last shift just went out. Polls close in 2 hours.

Looking at the midday numbers, 25-30% more people had already voted in our precincts than did four years ago.

5:49pm: I am making last minute Get-Out-the-Vote phone calls.

6:28pm: Not sure what to think of the data coming in from the precincts. Did we do good but not as well as needed in this area?

7:00pm: Polls close in Georgia. We are packing up and going home.

8:00pm: Home.

8:13pm: With Pennsylvania being called for Obama, the chances of McCain winning have dropped below 1%.

8:16pm: Godless Hagan defeats Liddy Dole. God has spoken!

8:38pm: NBC has called Georgia for McCain. Looking at the exit polls, African Americans made up 30% of the voters which is equal to their proportion among registered voters and 5% more than their proportion in 2004. However, Obama got only 25% of the white vote.

The Senate race between Chambliss and Martin is closer due to the Libertarian candidate Buckley and might head to a December run-off.

9:25pm: With Ohio projected as an Obama win by NBC and Fox News, I now project Barack Obama as the next President of the United States. Go Obama!

9:47pm: I didn’t blog it but told my fellow volunteers on the Obama campaign that Obama will net around 350 electoral votes and he looks on track for that.

9:59pm: I have been watching CNN since coming home. Now I am switching to Comedy Central’s Jon Stewart and Stephen Colbert special.

11:00pm: Polls close on the West Coast and it’s official. President-elect Barack Obama!

11:32pm: A very nice concession speech by McCain, despite some booing from the crowd.

Nov 5 11:50am: I went to sleep after Obama’s victory speech feeling dazed. I had full confidence from the start that Obama would win but the actual win did affect me a lot. I was a little bummed out last night about Georgia but feel better about it today. More thoughts on Georgia and the neighborhoods our team worked in during this campaign after we have official precinct results.

12:09pm: What’s wrong with Alaska? Did they just re-elect a convicted felon to the Senate?

Georgia: Battleground

Georgia has been reliably Republican for a while now. But this election is different. There’s a 1 in 3 (or more) chance that Obama might win tomorrow.

I have been really busy recently and haven’t had time to blog. Volunteering for the Obama campaign has been fun but also kept me busy. We had phonebanking events every day last week with one at our home on Friday. Then on Saturday and Sunday, we went door to door contacting voters in South Atlanta to go out and vote on Tuesday. That’s also what we are doing today.

So how did Georgia become competitive? In 2000, Bush won Georgia 55%-43% and in 2004 increased the margin to 58%-41%. But in 2008, you can see for yourself:

Georgia is now a battleground state. McCain is leading Obama by only 1.7% here according to Pollster which “averages” all polls. Also, the betting market Intrade has Obama winning Georgia at 27.5%.

In addition, in 2004 African Americans were 27.2% of registered voters and contributed 25.4% of the votes. According to the latest voter registration statistics, African Americans are now 29.9% of registered voters.

In early voting since September 22, we have had 2,020,839 voters already cast their votes, which is 61.5% of the total number of votes cast (3,285,140) in 2004. Assuming a turnout of 80% (higher than 77% in 2004), this would mean that 43.9% have already voted. Among these early voters, African Americans comprised 34.9% of the total. We are hoping that African American turnout remains high even on Election Day tomorrow and puts Barack Obama over the top in Georgia.