STATE |
% OF VOTE |
DELEGATES |
TOTAL DELEGATES |
AS |
1% |
0 |
6 |
AL |
5% |
0 |
52 |
AR |
9% |
0 |
31 |
TN |
9% |
1 |
64 |
NC |
11% |
2 |
110 |
VA |
11% |
1 |
99 |
VT |
12% |
0 |
16 |
OK |
13% |
0 |
37 |
TX |
13% |
0 |
228 |
MN |
16% |
10 |
75 |
UT |
15% |
4 |
29 |
ME |
16% |
5 |
24 |
CO |
17% |
13 |
67 |
MA |
20% |
26 |
91 |
TOTAL |
|
64 |
929 |
64 delegates out of 929 is less than 7%. Looks pretty bad, right? So lets imagine that she’d dropped out. Suppose that 70% of her voters (including me) had switched to Sanders, the other 30% to Biden; so that Sanders would improve his delegate margin versus Biden by delegates equivalent to 70-30=40% of her vote. So:
STATE |
% Warren |
TOTAL DELEGATES |
(Biden +) |
(Sanders +) |
(Net Sanders,
hypothetical) |
Warren,
Actual |
AS |
1% |
6 |
0 |
0 |
0 |
0 |
AL |
5% |
52 |
1 |
2 |
1 |
0 |
AR |
9% |
31 |
1 |
2 |
1 |
0 |
TN |
9% |
64 |
2 |
4 |
2 |
1 |
NC |
11% |
110 |
4 |
8 |
4 |
2 |
VA |
11% |
99 |
3 |
8 |
5 |
1 |
VT |
12% |
16 |
1 |
1 |
0 |
0 |
OK |
13% |
37 |
1 |
3 |
2 |
0 |
TX |
13% |
228 |
9 |
21 |
12 |
0 |
MN |
16% |
75 |
4 |
8 |
4 |
10 |
UT |
15% |
29 |
1 |
3 |
2 |
4 |
ME |
16% |
24 |
1 |
3 |
2 |
5 |
CO |
17% |
67 |
3 |
8 |
5 |
13 |
MA |
20% |
91 |
5 |
13 |
8 |
26 |
TOTAL |
|
929 |
36 |
84 |
48 |
64 |
In other words, under these assumptions, if Warren had dropped out, Sanders’s margin against Biden would have improved by only 48/929=5%.
In other words, she got more net progressive delegates by staying in the race than she would have by dropping out.
Now, this calculation is quick and dirty. I haven’t accounted for how the delegate math would change if there were fewer votes wasted on sub-threshold candidates. But I also have assumed a 70/30 split of Warren voters, which I think is generous to Sanders; if I had to bet, I think 60/40 would be closer to the truth, which would mean less than 3% using the calculations above.
You can explain this with voting theory.
Since the late 90s, long before I’d even heard of either Elizabeth Warren or Bernie Sanders, I’ve been a student of election methods. I’ve organized an academic conference on the subject, written a peer-reviewed paper, gotten a doctorate in a related field, and helped implement a novel voting method in practice for a moderately high-stakes yearly election involving thousands of voters. So in what follows, I’m not just making shit up.
In studying voting theory, there are a few pathological scenarios that can occur across various voting methods — not just in our stupid U.S. choose-one voting method, but in improved single-winner voting methods such as STAR voting, Ranked Choice Voting, Condorcet, etc. Two of the most important pathological scenarios are called Chicken Dilemma and Center Squeeze. These are similar, but with a key difference.
In Chicken Dilemma, there are two candidates on one side — let’s call them A1 and A2 — and one candidate on the other — call them Z. The voter preferences are something like the following:
33: A1>A2>Z
22: A2>A1>Z
45: Z>A1=A2
In other words, 33% of the voters prefer A1 over A2 over Z; 22% prefer A2 over A1 over Z; and 45% prefer Z but are indifferent between A1 and A2. Thus, A1 and A2 combined can beat Z, but Z could win if they can divide and conquer the A voters. The “correct” strategy is for one of the A candidates — probably A2, since they have fewer strong supporters — to drop out so that the other one gets their votes and beats Z. But it’s like a game of chicken: both A1 and A2 would prefer to hold straight until the last minute and hope the other one drops out, but by doing so they risk a crash where Z wins.
In Center Squeeze, the situation is slightly different. Instead of A1, A2, Z, you have A, B, C; in other words, the first two candidates are not quasi-symmetric clones, because the second candidate is in some sense “in between” the first and the third. Thus, voter preferences are something like the following:
33: A>B>C
11: B>A>C
11: B>C>A
45: C>B>A
In other words, if B dropped out, their 22% support wouldn’t all go to A, but instead would split between A and C. (Here, I have their vote splitting exactly in half for simplicity; but the argument that follows would work similarly even if that split favored A.) In this case, the “correct” strategy is for A to drop out so that B can win. Even though A has more first-choice support than B, B is the safer candidate to beat C. If B drops out, C will just win.
I think it’s outstandingly clear that the current Democratic primary is more of a Center Squeeze scenario than a Chicken Dilemma one; Warren is candidate B, not candidate A2. If Warren dropped out and endorsed Sanders, at *most* 70% of her voters (including me!) would follow that endorsement; at least 30% would go to Biden. Thus, only at most 70-30=40% of Warren’s total would add to Sanders’s margin against Biden; not enough for Sanders to win.
(Note: the fact that this scenario is called “center squeeze” does not mean in this case that Warren is more-centrist/less-progressive than Sanders in practice; only in terms of preferences and strategies. I actually believe that Warren is the more-progressive option where it matters the most. But that’s an entirely separate debate; the point here is about which strategies would work in practice, not which are desirable in ideological theory.)
So the question is:
Will Bernie Sanders heed the call, drop out, and endorse Elizabeth Warren?
I believe that this is the only remaining way that a progressive candidate would have any substantial chance of getting the presidential nomination. I also, of course, realize that this just is not gonna happen, no way, no how.
I still want our 2020 ticket to be as progressive as possible. Frankly, at this point, it’s very probably down to the VP slot. I still trust Elizabeth Warren to use her leverage to ensure the VP is a progressive. (Not Warren or Sanders themselves; they’re both too old, and I’d prefer a person of color.)
I’m still with Warren. She’s not only the best progressive option on the merits; so far, she’s been the best strategically. Her dropping out prior to Super Tuesday would have been a strategic mistake for progressive goals. That may change but I trust her to recognize if and when it does.