« A Reminder About GOP Absentees | Main | Snub This! »

Delegate Allocation: The California Example.

29 Jan 2008 10:53 am

Over the next few days, I'm going to write quite a bit about the complexity facing Democratic campaign targeters over the next several days. It's akin to running 22 simultaneous presidential campaigns in 22 different counties where the winner is based not on the popular vote in the country but on the delegates selected by congressional districts.

So let's start by taking one such "country," like, say, California.

In some ways, it makes sense to concentrate resources in some areas and skip others.

Why?

Because some districts send an odd number of delegates to the national convention, campaigning there is more efficient than in districts allocating an even number of delegates.

Why?

Because even without campaigning or concentrated television ads, the split in most of the even districts is not likely to allocate more delegates to the winner than to the second place finisher, especially if the number of total delegates allocated is 4.

Each congressional district in California has between 3 and 7 delegates to give; a total of 241 pledged delegates. The popular vote statewide determines the allocation of an additional 81 delegates, and 48 more are PLEOs -- but forget about the PLEOs for now.

So it makes sense for each candidate to maximize turnout in the larger odd-delegate congressional districts, right?

Not necessarily.

In states like New York, where Hillary Clinton will almost certainly win, and Illinois, almost certainly an Obama state, it makes more sense for the candidates to target the smaller-delegate-allocating congressional districts because they can increase turnout to boost their statewide totals AND win extra delegates at the same time.

It's easier, in other words, to extract an additional delegate by winning a smaller, odd-delegate congressional district than by trying to winner a larger, odd-delegate congressional district.

A further layer of complication is demographic.

Even though some advisers concede that Hillary Clinton will probably win California, Barack Obama's campaign will heavily target a number of large-and-small, odd-and-even congressional districts in the Bay Area (think Oakland, Berkeley, Marin County) because Democrats there tend to be more educated and younger -- and black -- exactly the demographic profile Obama has used to success in earlier states. But wait -- if you're in charge of Obama's California spending, do you spent, say, $100,000 extra in the 6th Congressional District, which comprises Marin County and Somona County north of San Fransisco? It allocates an even number of delegates -- six. Unless there's a landslide, both Obama and Clinton will get 3, each.

Why not spend that money trying to beat Clinton in the 7th congressional district across the bay -- Solano County and parts of Contra Costa counties, where the congressman, George Miller, has already endorsed Obama? CD 7 allocated 5 delegates, an an extra effort there might give Obama one extra delegate.

Comments (17)

" . . . districts in the Bay Area (think Oakland, Berkeley, Marin County) because Democrats there tend to be more educated and younger -- and black -- exactly the demographic profile Obama has used to success in earlier states." Pointing out they are predominately black may get you labeled a racist in today's charged political climate. Better you than me or anyone else, I suppose.

I think Marc is a education elitist and a ageist because he pointed out that younger and more educated people vote for Obama. Just kidding of course.

The Democratic party rules are just crazy! That said, I prefer them to the winner-take-all Republican side which doesn't paint a perfect portrait of the party's feelings nationwide.

LawSchoolDem,

It is racially charged now because of the Clintons attempt to paint Obama as a Jesse Jackson candidate who appeals to only black people.

If we are playing identity politics, then Obama should naturally appeal to black and white men since he is biracial. He should also appeal to recent immigrants (a lot of Latinos), since his father was an immigrant...

So that should put him in a pretty strong position, no?

...and to continue, Obama shoud also appeal to people under age 50, as they are closer to his age than Clinton...

...and that leaves Clinton with middle age/older women

It gets pretty absurd when one takes the identity politics argument to its fullest - we are all pods blindly voting for people who look like us. It is demeaning to any group to be characterized that way...

It's problematic when despite the analysis the comments even veiled as a joke speak of race. But I guess so did mine.

Marc, we are waiting for more of your analysis. Some of us actually do care about the procedure :)

It doesn't have to be much of a landslide.

If a district awards six delegates, you get the sixth delegate if you have a larger remainder than the second candidate. So, if you get 59% of the two-person vote, you get 3.51 delegates, while your opponent gets 2.49 delegates, meaning you get four and your opponent gets two.

To really outperform his share of the electorate, Obama would need a weird result where he was always close to the cutoff ... always just above 76%, 59%, or 4l%, but never just below those numbers. And it just doesn't make any sense that you could exercise such fine control over the results.

I suppose the other way to do it would be if Obama could somehow win 33 districts 59-41, while losing 20 of them 20-80. But that just doesn't make any sense. The fact that he is targeting SF and not LA, when there are more people in LA, means that the deck is stacked against him to start.

The ability to game the CD count is really overblown. In South Carolina, the popular vote went 55-27-18. The delegates went ... 55-27-18. It's very unlikely that you will have a delegate vote that differs that much from the popular vote.

The bottom line is that Obama has to get from the current 10% gap to roughly a 6% or less gap.

What about John Edwards - any chance he will break 15% in some districts? From a Clinton / Obama standpoint, that would make an odd CD even or vice versa.

I'll be watching California's CD-04 (Doolittle's district) next week. It offers 5 delegates but that's not why it's interesting. It's interesting because it's just about the reddest district in California (not at all Clinton friendly), but has a very strong Democratic organization from Charlie Brown's 2006 (and now 2008) congressional campaign. If Obama's people are smart, and I'm thinking they are, they will replicate their Northern Nevada success in California's CD-04 and take the lion's share of the delegates. Either way, it should be bellwether of Obama's CA organization.

The "landslide" threshold is reduce further by nonviable votes. Say Edwards pulls 12%, Kucinich 2 and "scatter" 1 in the 6-seater Nic considers above.

Now 49.6% of the raw vote becomes 58.35% of the viable vote, winning 4 delegates - a 2-seat margin.

49.6-to-35.4% is still a landslide by any standard, and if you miss by a hair, your campaign has put out a 14 margin point effort for nothing (assuming you and your primary rival started out dead even in the district).

Likewise, SC was a landslide, and landslides will tend to produce proportional results. It's the close ones where strategic allocation matters.

Ow, my head hurts even thinking about this. Have fun Marc.

Another case of interest: a thin win, evenly distributed statewide, would extract roughly the same delegate margin from North Dakota's 21 delegates as from New Jersey's 127.

ND's districted delegates are allocated to 8 single-delegate districts. (The winner also takes 2 of 3 at-large delegates, plus the add-on delegate.) NJ's 70 districted delegates are allocated to 20 districts, of which 10 are 3-seaters.

>

I don't understand this at all. Aren't all congressional districts, y'know, the same size?

Congressional districts have roughly the same number of eligible voters, but they have varying numbers of democrats.

Not just different numbers of Democrats, but different numbers of delegates. The Republicans, in California have the same number of delegates--all allocated to the candidate with the most votes in the district--regardless of the number of Republicans in the district. So the voting weight of 27,000 Republicans in Becerra's district is the same as that of 200,000 Republicans in Campbell's.

Democrats will use variable "magnitude" (as we electoral-systems folks say), and (semi-) proportional allocation. Given a lot of 3 and 5-seat districts, targeting makes sense.

There are also statewide delegates, allocated proportionally to candidates with over 15% of the votes in the state. Several other states area using similar rules.

I move to have Mr Armbinder's entire post stricken as invalid since he failed to use the adjective "complex" or "complicated" before the phrase "system of proportional representation" as required by Title IX, Chapter 7, Section 3(c) of the Reportage of Elections Act 1947. The nearest he gets is "[a] further layer of complication is demographic" and that, sir, is not near enough.

@Tom Round, its the klaw because harrassed journalists would prefer to have a nice simple ultra-minority "front runner" to annoint.

And I believe that section 3(d) of the law you cite requires a customary reference to Republican "love of order" (or even "horror of disorder") in their predominantly winner-take-all primaries, by contrast to Democrats who clearly love anarchy.

James Taranto at the WSJ picked up on a Reuters report that explained how the GOP don't have PR because - unlike the Democrats - they "don't like to share".

I s'pose winner-take-all contests make the selection of delegates more like the selection of Presidential Electors in November: the Republicans are happier with the status quo of the Electoral College, whereas Democrats are more interested in changes such as proportional or district allocation of Electors, or even national popular vote.