These counterfactuals on negotiating strategy are hard to prove -- or to disprove -- and one's opinion of them usually correlates strongly with one's overall position on the health care reform bill. Those, like me, who consider the even the "compromise" bill to be a reasonably important and impressive accomplishment will usually conclude that liberal constituencies did about all that they could reasonably do. But those who are dissatisfied with the outcome are likewise dissatisfied with the tactics.
It certainly is worth pointing out that negotiating "harder" is not a universal good. There are potential costs to taking an inflexible negotiating position in a complex, multi-party negotiation. Most obviously, people may conclude that you're wasting their time and you'll find yourself looped out of the negotiation. Although I would guess that people, as a rule, being social beings who want to please others, have somewhat of a bias toward being too accomodating in negotiations, they can certainly find themselves on the other side of the ledger with ample frequency.
Recently, I've had several conversations with NYU political scientist Bruce Bueno de Mesquita in preparation for my book project. Bruce, in addition to being a Silver family friend, is a really brilliant guy who is trying to lend some much-needed rigor to the political science community. He is best known for a model he designed (as documented in his book, The Predictioneer's Game) to predict the outcomes of complex negotiations from relatively simple inputs. There is a scaled-down version of his model available at his website; I decided to run the numbers for health care and see what it came up with.
Let me caution: although Bruce's algorithm has a very good track record, this is mostly for fun. His model is not a "black box". It requires the specification of several variables for each party to the negotiation and a significant amount of expertise may be required in selecting those variables. It provides for an unbiased way to evaluate this information, but the choice of the inputs may introduce bias.
In particular, for each party to the negotiation, the model requires the selection of their policy position (on a 0 to 100 scale), their relative degree of influence, their flexibility (how eager they are to compromise), and their engagement (how focused they are on facilitating their desired outcome). A party may also be deemed to have veto power over the negotiation, meaning that it may unilaterally reject any settlement.
I defined the possible outcomes on healthcare as follows:
100 - Single payer
70 - House bill with strong public option
60 - House bill with weak public option
50 - Senate Finance bill (no public option)
20 - Incremental reform
10 - No reform (status quo)
0 - Complete deregulation
In turn, I defined 12 parties to the negotiation with the following characteristics; the score in parenthesis represents each group's desired policy outcome.
- White House (70). Very highly influential but distracted (lower engagement). Extremely flexible; wants deal done. Veto point.
- Nancy Pelosi (75). Influential, highly engaged, quite flexible. Veto point.
- Harry Reid (67.5). Modestly influential (less than Pelosi), highly engaged, very flexible. Veto point.
- Progressive Democrats (85). Modestly influential, engaged, modestly flexible.
- Moderate Democrats (67.5). Modestly influential, reasonably engaged, very ambivalent/flexible.
- Blue Dogs (52.5). Influential, engaged, fairly inflexible. Veto point.
- Committee Chairs (60). Modestly influential, highly engaged, reasonably flexible. Veto point.
- Olympia Snowe (42.5). Low influence (although high relative to other individual members of Congress), but engaged. Modestly flexible.
- Republicans (20). Low influence, modest engagement. Completely inflexible.
- Insurance Lobbiysts (0). Comparatively low influence, but very engaged. Modestly flexible -- not totally averse to compromise.
- Liberal Activists/Unions (100). Low influence, engaged, somewhat inflexible relative to policy-makers.
- Voters/public sentiment (45). Highly influential, but very distracted/disengaged. Position somewhat amorphous/flexible.
Note that this is a very complex negotiation. At least five different parties effectively have veto power over the process, including the White House, the Blue Dogs (who cast the decisive votes in both chambers of Congress), and both the Floor and Committee Leadership.
When I ran the negotiation the first time through, it came up with a score of ... 52. That is, it predicted an outcome just slightly to the left of the Senate Finance Committee's bill, but which would probably lack even a weak public option. That actually looks like a very good prediction, given what is likely to come out of the reconciliation process (if the Democrats get a bill at all).
The real value in this exercise, though, may be in the scenario testing: we can tweak various parameters and see what influence they have on the model.
-- What happens, for instance, if we reduce the flexibility variable for the Progressive Democrats in Congress? That is, we make them more intransigent and demanding, as many bloggers suggested that they should have been? It turns out that nothing happens; the outcome of the negotiation is still a 52. Nor is there much change if we make the progressive position more flexible; the model comes up with a score of 53. (We can increase the score slightly if we make the progressives more influential -- to a score of about 56. But influence is not easily obtained and should probably be regarded as exogenous to the model.)
-- The White House also had somewhat marginal influence, although more than the Congressional Progressives. If we make the White House more engaged, we increase the predicted outcome of the negotiation to 55. If we make them both more engaged and less flexible, we increase the score further to 59. The model suggests, then, that if the White House were both less eager to broker a compromise and very actively involved/engaged in the negotiation from the very beginning of the progress, it might have been able to broker a public option, but probably only a weak one.
-- I also tested the strategy of the group that I call unions/activists. The model seems to think that the groups that define the endpoints of a negotiation can play a reasonably important role in determining the outcome (the same is true of lobbyists), even if they aren't terribly influential relative to other parties. What's interesting is that, if I give the activists a less flexible position, the score increases to 58, meaning that a weak public option might have been obtained. But if I give the activists a more flexible position than the one I assumed them to have in the status quo, the score also increases, to 56. Although this is highly speculative, what the model seems to be suggesting in that the activist community might conceivably have gotten the worst of both worlds. Had they been more willing to compromise on a good-but-not-great bill (better than the one they got), they might have expedited the process toward that outcome rather than letting other (less liberal) groups influence the outcome instead. But had they dug their heels in even more, they might have succeeded in applying enough pressure to move the outcome (somewhat) to the left. The frustration that a lot of activists feel, I think, stems from the fact that they put a lot of money into the pot but then had to fold. It sure sounds good to say: oh, we need both pragmatists and idealists in the liberal community, but it's also possible that activists would have done better had either of these groups unilaterally dictated the strategy.
-- I also tested what happened if certain groups were removed from the process. For example, the projected outcome increases from 52 to 58 if the committee chairs are bypassed, and from 52 to 57 if Olympia Snowe is ignored. It appears that dealing additional groups into a negotiation can have potentially unpredictable and deleterious consequences; even if those groups are not influential initially, they have a chance to accumulate influence once they're seated at the table.
This is fun stuff to speculate about. Still, perhaps the most important finding of the model is that the outcome was relatively robust. Although there are a number of things that Democrats could have done a bit better, essentially all of the scenarios that I tested produced a score between a 50 -- a bill something like Senate Finance Committee's -- and a 60 -- a weak public option. It would probably not have been possible to get a strong public option (much less anything resembling single payer) even if a number of variables were changed within reasonable boundaries.
This squares, in any event, with my intuition. No matter how clever progressives and activist groups might have been, they were enmeshed in a complex negotiation that:
(i) necessarily required the approval of a certain number of Blue Dogs;The influence of any one group in what is essentially a 10- or 12-way negotiation is liable to be fairly limited, no matter how wisely they select their strategy -- and to suggest otherwise probably reflects a certain amount of self-importance.
(ii) featured some parties -- Republicans and lobbyists -- who had limited but nonzero influence and who were actively trying to do undo any settlement;
(iii) was overseen by a series of party leaders (Pelosi, Reid, Obama) who have institutional incentives to broker a compromise, regardless of their (fairly liberal) personal preferences and,
(iv) was constrained by an ambivalent public.