I recently read Moneyball, the Michael Lewis book that is ostensibly about baseball, but really is about politics. OK, it's really about baseball, specifically about the Oakland Athletics during the 2002 season. Broadly speaking however, it's about a contest where money is important, and the different parties in this contest have drastically different amounts to work with, forcing the side with much less money to either lose badly, or find the inefficiencies everyone else is missing. Oakland did the latter. If you're uninterested in baseball, I still suggest reading the beginning, and in your head replace "baseball" with "politics", "Oakland A's" with "Democrats", and "New York Yankees" with "Republicans", and some questions should come up. In a matchup of unequal financial resources, where the other side can just throw money at things and we can't, are we using our resources efficiently? Democrats actually tend to match Republicans in spending by party units, by candidates, and even by independent groups who have to report their spending and donations. The difference is in dark money, though we can't know by how much, which is the point of dark money. We simply don't have the Republicans' ample supply of crank billionaires willing to spend unlimited funds on their favored candidates and ideological crusades. They throw money at campaigns. We can't afford to.
Lewis said he started out looking into how a team that was consistently near the bottom in payroll was consistently contending. He coined the term "moneyball" to describe Oakland's approach to competing by finding players they could afford who were still good enough, and doing that required finding what other teams were missing. After the 2001 season, they were pushed hard when three star players signed big contracts with other teams including, of course, the Yankees. The Athletics' management did this by being willing to question what they believed, differentiate between knowledge and assumptions, trust data over experience, and ask if they were measuring and valuing the right things. If you think I'm leading up to a suggestion that our political campaigns are measuring and valuing the wrong things, yes. Though at least on Team Blue, it seems we value and use research more than baseball did. Just my impression, which is ironic since I'm suggesting less reliance on impressions. Anyway, we can do better, and it's not like I, and probably most readers, are unable to cite instances of people in campaigns stubbornly refusing to reconsider conventional wisdom and change established habits.
So yes, I'm thinking of ways we campaign where I suspect we value and measure the wrong things, so sometimes we lack data and sometimes use less than optimal strategies.
I'm going to give a couple examples from the book, admittedly at the risk of indulging the liking I have for baseball and the enjoyment I got learning about this stuff, but I think non-baseball fans will get this too, and I ask us to apply this same thinking to politics.
Something the Athletics chose to value was how often a player gets on base. Baseball is blessed to be followed by people enthusiastic about both baseball and statistics, who tried to tell the people who run professional baseball that the frequency of getting on base is the most important aspect of judging a player. Scouts did not agree, and Oakland was the first team to take the statisticians' side on this. Scouts didn't think much could be judged by a young player's numbers, especially at a high school level (they have a point, so Oakland looked more at college players, which scouts tended to ignore), so they didn't measure how often a player got on base. They measured foot speed, throwing distance, and judged whether a player had an athlete's body. Oakland had conflicts between scouts and the staff who applied statistics, who came to different conclusions about which players had potential. The statistics won, and Oakland didn't care what a player looked like, as long as he could get on base. Thus they found players overlooked by other teams and thereby affordable to the A's, and the A's were able to contend while spending roughly a third of what the Yankees were spending.
Similar conflicts arose regarding strategy. Statistically inclined fans calculated the odds of scoring a run given different combinations of outs and runners on base, and determined that certain long-standing assumptions were wrong, leading to self-defeating strategies, such as the sacrifice bunt. That's where the batter bunts the ball and signals the bunt to the defense, making himself an easy out at first. In exchange for giving up the out, the team at bat gets to advance a runner on first base to second, or second base to third. However, the odds of scoring with a runner on first and one out turned out to be higher than with a runner on second and two out. In other words, outs were being undervalued. The sacrifice is a good deal for the defense. Some managers still haven't come as far as fans in figuring this out, thus why the smattering of boos when the home team sacrifices.
So how about we ask similar questions about our campaigns? Are there analogous situations where we're measuring the wrong things? Refusing to use data which conflict with experience and conventional wisdom? Those of you who have been at this politics stuff a while are probably formulating examples already. As I thought about it, I realized I was thinking on two different scales, macro and micro, by which I mean the grand scale like national strategy, and down on the ground where we're pounding pavement. Campaigns entail a lot of factors that make it hard to tell exactly why we won or lost, which strategies help, but maybe we can get beyond merely taking our best guess, and having a strategy of "just do everything because we don't know what works".
To answer the question of whether we value that right things, I asked what it is that we value, and what potentially we could value. Getting into the details on either the macro or micro levels would make this post awfully long, let alone covering both, so my thinking now is I'm going to save the details for separate posts on each level. But I do want to share the general direction now.
On a macro level, we value getting votes --- yet getting the most votes isn't always getting our candidates into office. What we need isn't votes; what we need is seats. Votes are merely the main means of getting seats, and sometimes when we get the most votes, Republicans get the most seats. Probably many readers just remembered Florida 2000, when Al Gore won the election but somehow Bush Jr. got into the White House. I thought of that too, but I'm also thinking of 2012, when Democrats in aggregate got the most votes for Congress, but weren't even close to getting a majority of seats. The same happened in some state legislative chambers. Republicans got more seats with fewer votes. Apparently there are more ways to get seats than getting the most votes. What else might affect getting more seats? Might seem obvious, but let's think it through, because sometimes it's surprising what we find when we back up and think through what we think we already know. So, ways to get seats besides getting the most votes:
- Gerrymandering: this is the big one, at least in terms of winning back Congress, and the macro post will focus on gerrymandering, but to put it briefly, we try to counter it by winning control of state governments and trying to win gerrymandered seats. But there might be a more efficient way to counter gerrymandering. We don't need to control states to allow us to gerrymander, which many Democrats have ethical problems with anyway. We need to gain partial control, of big enough states, to block Republican gerrymandering. I'm thinking that's both cheaper and more achievable.
- Voter suppression: If you weren't reminded of Florida 2000 before, you probably just recollected the effect of a partisan state secretary of state purging tens of thousands of black voters from the registration rolls by falsely identifying them as felons. Democrats definitely have an ethical objection to suppressing Republican votes, but we have to block Republican suppression of likely Democratic voters. Keep in mind Republicans don't do this merely to be mean, but because it gets them more seats.
- Partisan election officials: hard to separate from voter suppression since it usually requires election officials to act in a partisan way to suppress the other party's votes, but it's also about having people who want the machinery of government to work, and who value more people voting. Even just wanting to block Republican voter suppression rather than wanting to suppress their voters, which I expect no Democrat would be ethically OK with, requires winning the elections for the officials that administer elections, or the officials that appoint those who administer elections. We don't have to stop a crooked secretary of state from wrongly purging voters off the registration rolls if we elect a secretary of state that won't do that.
- Partisan judges: Republican were many years ahead of us in their focus on controlling the judiciary, and having partisans on the right benches has gained seats for them, the presidency in 2000 obviously being the big one. I don't want judges who will be partisan to Democrats because partisan judges erode the legitimacy of the judicial branch, but we sure have to block partisan Republicans.
- Election rules: By this I mean how we hold elections, like instant runoff versus separate runoff versus plurality winner, or party primary versus jungle primary.
Again, I'm saving the dive into the weeds of details for a separate macro post, but do notice right now one thing: winning seats by these non-vote methods has proven cheap. The big one is Republicans gained the presidency in 2000 for the cost of winning a secretary of state election. Getting Republicans on the US Supreme Court gained Republicans a court that throws out campaign finance laws, gaining who knows how many elections so far or to come, which we have failed to counter by what we've tried so far.
On a micro level where, like probably most of you reading this, I spend my actual campaign time, I thought of the times I heard campaign staff or candidates boast of the numbers of doors their campaigns knocked on. However, we know that only the actual conversations really matter. So when we count just the doors knocked on, we're measuring the wrong thing. Since only the conversations matter, counting the doors where no one answered is useful only in terms of establishing a proportion of doors where we're getting the conversations. Yes, it's a potentially useful piece of data, but we're treating it like the goal. A supposed difference between baseball and politics is that baseball has loads of discrete bits of data while politics doesn't. That's true about baseball, where each pitch can be taken as a separate measurable unit. But it's true about politics too. Yes, it's hard to know if votes were won or lost from TV ads, posters, bumper stickers, or debates. But we do have a load of data available if we recognize that each voter contact is a measurable piece of data. Since knocking a door and getting no answer is no better than a lit drop, it seems we would want to analyze our doorknock data with the goal of maximizing the conversations and minimizing the unanswered knocks. Instead, by emphasizing the sheer number of doors, we're actually providing an incentive to do a poor job of doorknocking.
Doorknocking is where I spend most of my volunteer time, so that's the explanation of why my thoughts would go that way. There are a number of things about our doorknocks I question and as I'm pondering the post, even just in my head it's getting really long. Like a separate macro post, my intention is to have a separate micro post for a deep dive.
cross-posted on MN Progressive Project
Mon Apr 20, 2015 at 8:50 AM PT: This is the macro post.
This is the micro post