how to handicap city council races

To me, the city council races are at least as interesting as the mayoral races.  Oh sure, they are nowhere near as consequential and most people don’t pay any attention, but for me, the complexity and uncertainty of a multi-seat SNTV race beats the glamour of a single seat executive race any day.  But you could probably tell that I’m a little strange by my long, long tables.

Today I’m going to try to spell out the mental model I have for these local races.  This is not a precise statement.  Most of these ideas have been floating around in my head in a fuzzy haze for several years, but I’ve never tried to lay them all out on paper.

City council races are hard to think about because they live in an intermediate level.  They aren’t as partisan and issue-oriented as legislative races.  But they aren’t as exclusively local as some of the more rural county assembly races (or as most people think they are).  You need ten to twenty thousand votes to win a race, which means you can personally know more of your voters than a county assembly member but fewer than a legislator.  And these politicians are not always ready for the national spotlight, but not all of them are simply locally oriented scions of influential families or small-time local power brokers.  I also think that the common Taipei stereotype of a county assembly member – a betel nut chewing guy with a bit of money, a construction company, ties to organized crime, and no political ideals at all – is very often quite misleading.  Politics is a difficult game to master, and these guys deserve a bit more respect, or at least a bit more careful examination.

So my question for the day is, how do I look at a field of candidates and determine which ones are going to do well and which ones are not?  Alternatively, how do you win a city council election?  The best way would be to immerse myself in the race, talking to people and learning the intricate stories of each candidate in the race.  However, I don’t do this because (a) I’m not a very good soaker-and-poker and (b) there are too many soap operas for any single observer to follow.  Instead, I look at a lot of less telling variables.  This inevitably means that I will miss a lot of important factors, but hopefully I understand a little bit about a lot of races.

I weight the following factors most heavily:

1. Party and partisan factors
2. The candidate’s past electoral record
3. Family connections
4. The geographic structure of a candidate’s vote coalition
5. How the candidate matches up with the rest of the field

There is a common notion that parties don’t matter in local elections.  Partly this is because candidates almost never talk about party positions or advertise their party affiliations in county assembly races.  Partly this is because the DPP doesn’t get anywhere near the number of votes in local elections as in national-level elections (or county/city executive races).  However, city council elections in urban areas are not quite like rural county assembly elections.  In Taipei City, the results in city council elections are not that dissimilar to legislative elections.  In Kaohsiung, there is a bit less similarity, and as you go out to the small towns in Tainan, Kaohsiung, and Taichung, partisan factors will matter less and less.  However, in these elections, partisan factors never stop mattering.  There are always a significant number of voters who will vote for a DPP or a KMT candidate.  All the other factors matter a lot more here than in national elections, but the partisan structure of a district is still a good place to start.

So how do you judge the partisan balance of a district?  I use three different things.  I start with the partisan results from the last city council elections.  I compare that to the partisan balance of the district in national elections.  The spread between these two gives me something of an indication of how much partisanship matters and how much everything else matters.  In the long run, I expect the gap to narrow significantly.  Usually this means that I expect independents to lose ground, the KMT to stay about the same or lose (depending on whether independents are strong in that district), and the DPP to make gains.  The third thing I look at is the larger partisan trend in that particular year.  For example, DPP city council candidates in Tainan should have a few extra votes to work with this year because this looks like a good year for the DPP in general and particularly in the Tainan mayoral race.  In concrete numbers, how much should this be worth?  Well, that’s hard to say, so let’s not try.

Second, I look at what the candidate has done in the past.  Elections are hard, and the best way to convince me that a particular candidate can win votes is to have previously done it.  In fact, one of the first things that every candidate in an SNTV election has to do is to convince voters that he has a chance to win.  If voters get a notion that you might be a turkey, you are in trouble.  Incumbents have a leg up in that voters usually aren’t scared that they will be throwing away votes on incumbents.  If you are a brand new candidate in your first election, that’s a different story.  Unless I know something else about you, I’ll probably assume you are a turkey.

Incumbency is good, but it isn’t the best thing to have on your resume.  My rank-ordering is as follows:

1. Legislator
2. Town mayor
3. Incumbent city councilor
4. Town councilor
5. Li (village, ward, neighborhood) head
6. Non-elected position

The legislature is a much better job than the city council, and most legislators would sooner eat glass than stoop to such a lowly position.  However, not everyone feels this way.  There are four former legislators running in this years election (Qin Huizhu, Wang Shijian, Wang Shuhui, and Tang Bi’e).  They will all be heavily favored to win.  One more former legislator, Wang Shixun, is already out.  Wang shocked everyone by losing in the DPP primary.  I guess not every legislator is a sure bet.

Legislators are kind of a curiosity; town mayors are everywhere.  Mayors are almost always strong county assembly candidates.  Since executives have a budget to spread around, almost all politicians would rather be a mayor than an assembly member.  In fact, many, perhaps most, mayoral candidates are incumbent county assembly members looking to move up the ladder.  In the past, mayors generally only changed to assembly races after they had served their two mayoral terms.  This year, since towns will no longer be independent administrative units and there will be no more mayors, a lot of single term mayors are also moving into the city council elections.

Incumbent city council members are next in my hierarchy.

Fourth are town council members.  Town councils are usually ignored, even by people like me.  They don’t have much power at all, and it usually only takes a few hundred votes to win a seat.  I look at these people as minor leaguers trying to move up to the major leagues.  What they have done is good practice, and they have shown rudimentary abilities, but winning a city council seat is much, much harder than winning a town council seat.

Li heads are the lowest on my ranking of elected jobs.  (It’s hard to translate li.  A li is a neighborhood with anywhere from 500 to 10000 voters.)  Li heads are geographically limited.  While town council candidates have to develop ties across a dozen or two li, li heads are focused in one small area.  It is very hard to expand on this base.

All elected positions are better than all non-elected positions.  When I see that someone is head of the local fireman’s association, parents association, or social welfare NGO, I give this almost zero weight.  These positions are something like Treasurer of the Spanish Club on an American student’s college application.  Everyone has an impressive list, but they are usually just filler.  They don’t tell you much about the politician’s ability to persuade voters to support him.  Even positions in the local farmers association, which is an important organization in electoral politics, don’t tend to tell you much about how well a candidate will do.

The resume has a very short shelf life (or if you prefer, a high discount rate).  If you were elected mayor eight years ago but lost a county assembly race four years ago, I’m not very impressed.  I look a bit more favorably on incumbents who have been re-elected three or four times than on those who have only won once, but the difference is not overwhelming.  Lots of longtime incumbents lose in every election cycle.  In other words, just because you proved repeatedly you could do it in the past doesn’t mean that you will be able to do it again.  (If you’ve never done it before, the odds are really stacked against you.)

Third, I look at family connections.  This is the same logic.  If a family member has been able to collect lots of votes, you probably will be able to as well.  Family members are usually able to transfer votes from husband to wife or from father to son/daughter.  (I’m not sure why, but non-family support (within factions or to a non-family protégé) doesn’t seem to transfer.  I don’t understand this very well, but blood seems to matter a lot.)  So if your father is a mayor or incumbent county assembly member, you have a pretty good shot.  If your father was mayor a decade ago, that may not be so helpful.  If your father has passed away, you might be in real trouble.  Children almost never have the wide set of connections and influence that their parents, who are successful politicians after all, have.

Fourth, I care about how a candidate’s support is geographically distributed.  This variable matters mostly for more rural areas.  It is almost entirely irrelevant in Taipei City, but it can be decisive in other areas.  The geographical distribution reflects the type of coalition a candidate is building.  Candidates who rely on networks of friends, relatives, and neighbors in a traditional organization-dominated campaign will usually have their votes concentrated in a subset of the entire district.  Often these bailiwicks are organized along township lines.

Bailiwicks are not necessarily a good thing or a bad thing – sometimes it doesn’t matter if you get your 10000 votes in a small area or spread across the entire district.  But it can be important.  In a very complex district with lots of candidates, a well-defined bailiwick can be an effective defense against strategic voting.  An ability to get lots of votes in a small area is also one way to break through partisan structures.  However, they can also be a prison; many candidates find it extremely hard to find new sources of votes outside their traditional bailiwick.  If there are no other candidates in the race from your bailiwick, you might have an easy election.  If there are lots of candidates fighting for pieces of the same small pie, you might be in real trouble.

This brings us to the fifth factor, the rest of the field.  Depending on how many candidates are incumbents, from your party, or from your home area, your chances of winning might be significantly higher or lower.

I don’t think of this as a deterministic model.  I’m not going to tell you that so-and-so is a sure thing to win.  This is a probabilistic model.  So-and-so has a pretty good chance of winning.  There are always a lot of things that I don’t see that are critical.  These factors are crude proxies for some of the things that really matter.

I should probably illustrate this mental model with an actual race, but I have no energy to do so tonight.  I’ll probably get around to that sometime in the future.