This is my third post in Dividing Howard week on my blog, as I discuss some topics related to my new book on the history of council council redistricting in Howard County, Maryland, and the broader events of Howard County politics from 1960 on. Previous posts discussed the role of Columbia in spurring creation of a county council, and the struggles of Howard County Republicans under the council district system. In today’s post I take a step back and look at the overall impact of having a council district scheme with drawing of district lines primarily controlled by one party.
There are really two questions here: First, has there been gerrymandering going on with respect to council district lines? I think the answer to this is yes, as evidenced by the past behavior and statements of the people engaged in drawing district lines; see the later chapters of Dividing Howard for many examples. (Although arguably the gerrymandering in question has been less egregious than in other jurisdictions.) That’s not to say that it’s simply a matter of evil Democrats and victimized Republicans; there’s no question that Howard County Republicans would return the favor if they were ever in a position to do so. (And in fact Republicans in other jurisdictions have happily engaged in blatant gerrymandering against Democrats when given the chance.)
Second, why exactly is gerrymandering bad? There are many answers that people have given to this question: It reduces competition and prevents having a healthy two-party system, it advantages incumbents and discourages “new blood” in politics, it violates principles of fairness, and so on.
However I think a better answer is that gerrymandering works against the nominal purpose of government, to provide for the greater good of all. A good explanation for why this is the case comes from “selectorate theory”, an idea in political science recently discussed in the popular work The Dictator’s Handbook: Why Bad Behavior is Almost Always Good Politics. Selectorate theory provides a simple general model of how those who exercise power are motivated to behave, no matter the type of political system in which they operate.1
In our context the important characteristics of the model are as follows:
The primary goal of all leaders is to remain in power. (They could certainly also be motivated by more noble motives, like serving the public, but if they do not remain in power then they will be unable to act on those motives.)
Given point 1, leaders will act first in the interests of those whose support is needed to stay in power, and only secondarily in their own interest. The interests of everyone else will always come last.
When the number of needed supporters is very small relative to the size of population (or, in general, the size of the group over which the leader exercises power) then leaders will attract supporters primarily by providing them private goods not made available to others. As the number of needed supporters increases to a significant fraction of the total population then providing private goods to supporters becomes less and less feasible and leaders will provide support increasingly in the form of public goods that benefit supporters but are generally available to others as well.
In this model the difference between a dictatorship and a representative democracy is not that leaders of democracies are more moral and public-spirited, rather it’s that unlike dictators they must rely on coalitions of supporters that are much larger: A typical dictator might require the support of only a few dozen or few hundred people (key members of the military, intelligence service, personal guard, and various cronies), while the typical leader of a populous representative democracy might need the support of at least a few million people to gain and hold power.
In this light the problem with gerrymandering is this: It lowers the size of the “winning coalition” needed to put a leader into power, and therefore increases the chance that the leader will focus on the needs of the supporters in that coalition to the detriment of the needs of everyone else. This is most clearly seen in cities like Baltimore that are heavily dominated by Democrats or states like Utah that are heavily dominated by Republicans: In such jurisdictions the general election is essentially irrelevant, the outcome having already been decided in the party primaries, in which the number of people voting is relatively small. As long as politicians in these jurisdictions can keep their primary voter base happy it doesn’t matter whether anybody else is satisfied.
Closer to home, let’s look at the 2010 general election results for the Howard County Council. In Council Districts 2, 3, and 4 the Democratic candidates won the general election with approximately 67% of the vote, and in Council District 5 the Republican candidate won the general election with 67% of the vote. Only District 1 was relatively competitive, the Democratic candidate winning with 53% of the vote.2
In 2010 there were almost 180,000 registered voters in Howard County. However most of these were irrelevant to the final results. In particular, since Democratic dominance of Districts 2, 3, and 4 was so complete those races were arguably decided at the time of the party primaries. For example, in District 4 there were approximately 37,000 registered voters (the so-called “nominal selectorate”), but the race was essentially decided by the roughly 6,200 voters in the Democratic primary (the “real selectorate”), so that Mary Kay Sigaty’s winning coalition could be as small as about 3,100 voters—less than 10% of the total voter population in the district.
Similar calculations could be done for the other districts. As it happens both Calvin Ball in District 2 and Jen Terrasa in District 3 had no primary opposition, but if they had it’s likely that their winning coalitions could have been roughly the same size as Mary Kay Sigaty’s. The net effect is that three out of the five council members, and thus a council majority, could likely be selected based on the votes of as few as 9,000-10,000 people, or about 5% of the total number of registered voters in Howard County.
This is not to say that Calvin Ball, Jen Terrasa, or Mary Kay Sigaty don’t care about the other 95% of Howard County voters; I think they, like Courtney Watson and Greg Fox, in general are sincerely working for the good of Howard County as a whole. However if there’s something to selectorate theory, and I think there is, then I think it makes sense to arrange things so that politicians need as large a winning coalition as possible in order to get elected. This minimizes any incentives to favor a limited set of supporters at the expense of others, and maximizes the chances that their actions will be to the good of all.
Minimizing gerrymandering also can help prevent situations where a relatively small minority of voters can thwart the will of the majority and in essence demand special favors for themselves. For example, consider the current structure of the U.S. Senate: Because of the Senate’s rules on filibusters a minority of 40 senators can prevent legislation from passing unless it is modified to meet their demands (which, per selectorate theory, are really the demands of their winning coalitions). That minority of senators could then be elected from as few as twenty states, and those could be states with relatively small populations. If those states are also dominated by one party then in effect the direction of the country as a whole could hinge on the votes of at most a few million people.3
In a local context the Taxpayer Protection Initiative promoted by Howard County Republicans could have produced a similar effect had it been passed, allowing two council members out of five to block tax measures. Those members in turn could be elected with as few as 13,000-14,000 voters—about 2,500 voters in the Republican primary in District 5 and about 11,000 voters to elect a Republican in the general election in District 1—and would have the opportunity to hold spending measures hostage in order to extract special favors for their own supporters.4
So what would I suggest we do with respect to the current system of drawing council district lines? That will be the topic of my next two posts.
In the meantime I encourage you to check out Dividing Howard if you haven’t already; it’s only $2.99 from Amazon or Barnes and Noble, and all royalties go to the local charity Voices for Children, which recruits and trains volunteer advocates to represent the best interests of abused and neglected children in Howard County courts.
Selectorate theory isn’t restricted to political systems; it can also be applied in the context of business, for example to explain why CEOs act the way they do. Also note that for purposes of this post I’ve simplified an already simple model even further.
To take an extreme example, Wyoming is the least populous state in the U.S., with less than 600,000 people. It is also reliably Republican; in the 2008 race for the Wyoming Senate seat Republican Mike Enzi won election with over 75% of the vote. The number of Republican voters in the primary that year was about 70,000, so as few as 35,000 voters could determine Wyoming’s two U.S. senators. (As it happens Enzi was unopposed in the 2008 primary, but the general point stands.)
Had I thought of it at the time I would have added this to the list of the reasons why the Taxpayer Protection Initiative was and is a bad idea.