what happened in Hsinchu?

One of the major motivations for writing this blog is that sometimes I have to write through things to figure out just what I think.  This post is going to be one of those searching essays.  I’m not sure what the conclusion will be, or even if there will eventually be a conclusion.  If you don’t want to be frustrated by sharing my thought process, I suggest skipping to something else.

The puzzle I’m struggling with is what happened in the Hsinchu County legislative by-election last week.  Here are the recent election results (leaving out minor candidates and parties):

Feb 2010 LY by-election
Valid votes	Zheng 鄭永堂 (KMT)	Peng 彭紹瑾 (DPP)		Zheng 鄭永堂 (KMT)	Peng 彭紹瑾 (DPP)
127967	56342	71625		44.0	56.0
Dec 2009 County executive
Valid Votes	Qiu 邱鏡淳 (KMT)	Peng 彭紹瑾 (DPP)	Zhang 張碧琴 (IND)	Qiu 邱鏡淳 (KMT)	Peng 彭紹瑾 (DPP)	Zhang 張碧琴 (IND)
252424	97151	76254	76254	38.5	30.6	30.2
Jan 2008 LY district
Valid votes	Qiu 邱鏡淳 (KMT)	Xu 徐欣瑩 (IND)		Qiu 邱鏡淳 (KMT)	Xu 徐欣瑩 (IND)
192253	127892	60209		66.5	31.3
Jan 2008 LY party list
Valid votes	Pan-blue	Pan-green		Pan-blue	Pan-green
195144	136854	45009		70.1	23.1

How did Peng Shaojin win Hsinchu County when there hasn’t been a hint of DPP strength in this district for over a decade?

Recall the background.  Two KMT heavyweights, Qiu Jingchun and Zheng Yongjin have been slugging it out for a decade and don’t particularly like each other.  In 2009, Zheng stepped down as county executive after two terms, and the KMT nominated Qiu to succeed him.  Instead of supporting Qiu, Zheng threw his support to the speaker of  the county assembly, Zhang Biqin, who ran as an independent.  Qiu won the ensuing bitterly contested three-way race.  For the by-election, the KMT attempted to make all sides happy by nominating Zheng’s younger brother, Zheng Yongtang.

Idea #1: mobilization

In another post, I have already argued against a pure mobilization hypothesis.  The pure mobilization hypothesis says that the DPP turned out the same 70000 voters as in Dec 2009, but the KMT suffered a massive failure of mobilization, and only about half of its voters showed up.  I won’t recount this argument.  The conclusion was simply that there must have been some conversion.  Some people who have voted for blue candidates in the past must have voted for the DPP candidate in this by-election.

Idea #2: the KMT reconciliation was fake

On the surface, the KMT got all sides to sit down and pretend that they had forgiven each other for the past nastiness.  For example, both sides withdrew their libel lawsuits, and Qiu accompanied Zheng on campaigning trips to local markets.  However, there is a suspicion that this was all fake, that Qiu actually worked hard against Zheng.

The main problem I have with this idea is that it doesn’t feel right.  The tone of the newspaper stories wasn’t what I am accustomed to seeing in these situations.  Again and again, instead of reporting on either side’s protestations that everyone was now working together, the papers reported lamentations that the voters didn’t believe there was actual peace.  Instead of insisting, they were whining.

But let’s assume that Qiu did in fact tell his network to throw their support to Peng.  It’s a little hard, but not impossible, to figure out how we should expect things to change from 2009.  Zheng and Zhang were allies, and presumably most of Zhang’s votes came from Zheng’s support.  Thus, we would expect Zheng to retain most of this support in 2010.  Of course, Zhang would have had a little bit of her own support, which would have been concentrated in her former county assembly district.  That district included only Zhubei Township.  However, since Zhang and Zheng were close allies, we would expect her to throw all of her support to Zheng as well, though she might not have mobilized quite as extensively on behalf of someone else as she did for herself.  So we would expect Zheng to get all of Zhang’s votes, with perhaps a small discount in Zhubei.  Qiu’s votes have to be divided into two parts.  Some people supported him because he was the KMT nominee.  These people should be expected to support Zheng as well, since Zheng was also the KMT nominee.  Other people supported Qiu for personal reasons.  These are the people that Qiu’s network should have been able to mobilize his network to throw over to Peng’s side.

That’s not what the pattern of conversion looked like.  From TVBS surveys, we have the following (which you may remember from my election results post):

			LY race
County Executive		total	Peng (DPP)	Zheng  (KMT)
	total	N=827	44	34
	Qiu (KMT)	38	32	51
	Peng (DPP)	15	89	4
	Zhang (IND)	17	37	44

Defections to Peng come fairly evenly from both Qiu and Zhang supporters; they are not concentrated among Qiu supporters as we would expect from the factionalism story.  The factionalism story has no way to explain the 37% of Zhang supporters who shifted over to Peng in 2010.

We would also expect to see some clear patterns in the aggregate election results.  In short, we would expect to see higher growth in Peng’s support in places where Qiu was strong.  So let’s look at how Peng’s vote evolved, town by town.  (Ignore the last two townships.  Jianshi and Wufeng have large numbers of aborigines. Aborigines vote with everyone else in the county executive election, but they have their own district in the LY election.)

township	township	2009	2010	growth	Qiu	Zhang
新竹縣竹北市	Zhubei	19772	18,653	0.94	30.6	36.5
新竹縣竹東鎮	Zhudong	13233	11,871	0.90	49.6	21.1
新竹縣新埔鎮	Xinpu	7105	7,165	1.01	27.9	35.2
新竹縣關西鎮	Guanxi	6103	5,496	0.90	32.5	32.5
新竹縣湖口鄉	Hukou	10831	9,887	0.91	40.2	30.8
新竹縣新豐鄉	Xinfeng	7013	7,164	1.02	39.6	30.0
新竹縣芎林鄉	Qionglin	3781	3,598	0.95	36.0	32.0
新竹縣橫山鄉	Hengshan	2455	2,479	1.01	46.6	24.7
新竹縣北埔鄉	Beipu	2191	1,956	0.89	38.6	24.2
新竹縣寶山鄉	Baoshan	2494	1,984	0.80	42.7	23.6
新竹縣峨眉鄉	Emei	1286	1,214	0.94	38.7	26.5
新竹縣尖石鄉	Jianshi	538	98	0.18	53.0	33.8
新竹縣五峰鄉	Wufeng	324	60	0.19	60.3	28.1
total		77126	71625	0.93	38.5	30.2

The structure of Peng’s support didn’t change a whole lot.  With the exception of (tiny) Baoshan Township, his vote in 2010 was between 89% and 102% of his 2009 vote in every township.  Moreover, what we don’t see is a clear connection between growth and Qiu’s vote.  For example, Qiu’s two strongest townships were Zhudong and Hengshan.  In one, Peng’s vote growth was above average; in the other, below average.  Likewise with Qiu’s two worst townships, Zhubei and Xinpu – in one Peng’s growth was just about average and in the other it was well above average.  There just aren’t the clear patterns that we would expect to see if Peng were getting a lot of support from Qiu.

Idea #3: KMT reconciliation was real, but voters didn’t believe it

This idea seems partly right, but mostly still inadequate to me.  On the one hand, I rather tend to believe that the elites did actually more or less reconcile.  On the other hand, something didn’t work since the votes never materialized.  The tone of the newspapers indicated to me that this is the explanation that reporters believed, for what that’s worth.  So we hear stories about rumors that if Zheng won, Zhang would be blocked and would never have a chance to develop her career.  Supposedly Zhang’s supporters didn’t want this.  And I guess it’s possible that Qiu’s supporters took out their frustration on Zheng, even though Qiu himself told them not to.

On the other hand, we still have to deal with the numbers from above.  If there was some systematic trend from Qiu or Zhang supporters to Peng, shouldn’t we be able to see it in the data?  More than that, though, this doesn’t seem to accord with my mental model about how Taiwanese politics work.  Elites care about betrayal and loyalty, but voters don’t.  Voters are more likely to see things in simpler partisan terms.  After all, they aren’t personally affected by which KMT figure gets to sit in the office nearly as much as an elite is.  They don’t get to enjoy the power, and no one has ever looked them in the eye, made a promise, and then reneged on it.  For voters who don’t have much partisan attachment and vote for individual candidates, I guess I don’t believe that they care very much about the election last time.  Each election is considered anew.  That is, they might have liked Qiu last time, but he’s not running this time.  The choice is between Peng and Zheng, so Qiu’s anger at Zheng’s brother’s backstabbing isn’t likely to be a real important consideration.  In short, I simply can’t believe that the voters would hold a grudge when the elites had made peace with each other.

Idea #4: Peng Shaojin was a great candidate

The media reports that tried to explain the election from Peng’s angle focused on how hard he worked.  There were lots of stories gushing about how this guy (with roots in the judicial system, several years of experience in the legislature, and deep roots in the party) had humbled himself and worked very hard at developing grassroots support.  Peng’s career had formerly been in Taoyuan County, but he transformed himself from a parachute candidate with a national profile into a genuine local person by shaking lots of hands, talking with lots of people, and developing a true common touch.

Well, isn’t that heartwarming.  I don’t buy it.  He may have shaken lots of hands, but that does not change peoples’ minds.  He may have impressed people with his sincerity, but I am skeptical.  Besides, doesn’t it imply a very shallow model of voting to suggest that people will vote for anyone whose hand they have shaken?

For me, this argument really founders on the fact that there was an election a mere three months ago, and this new Peng Shaojin apparently did not yet exist then.  I simply don’t believe you can develop such thick grassroots support in only three months time.  Personal networks take years to develop, and they take extended contact and levels of trust that don’t come from simply shaking a lot of hands.

Perhaps Peng Shaojin had already started developing his grassroots support before, but we didn’t see it because Qiu and Zhang had even deeper grassroots support?  With Qiu I could accept this.  He has been working the county for a couple of decades.  With Zhang, not so much.  The problem is that Qiu, even with a party nomination, barely got more votes than Zhang.  For his part, Peng doesn’t seem to have gone much beyond the traditional DPP vote.  If he already had such strong grassroots support, it should have shone through.  Besides, Peng’s competition in the by-election was Zheng Yongjin’s younger brother.  Zheng Yongjin has been working the county as long as Qiu and should have a roughly equivalent level of grassroots support.  So how would Peng’s grassroots support be muffled in the first election but not the second election.

I think this story is an Imelda Marcos story.  Why did the protesters storm the Presidential Palace in the 1986 People Power revolution?  Many observers, including serious academics, have claimed that they were enraged by her 3000 pairs of shoes.  The problem is that no one knew about all the shoes until after they stormed the palace.  This is a post-hoc imagined memory created to fit a story that you are comfortable with.  Just as we find it satisfying that the indignant people righteously overthrew the decadent dictator, it is also satisfying to believe that Peng’s victory was created by his humility and hard work.  Beware of things that you want to believe.

Idea #5: Zheng Yongtang was a terrible candidate

Now I’m getting into more speculative ideas.  The previous explanations have come from media reports.  These come from my little brain.  They’re probably wrong.

So this idea is simply that Zheng Yongtang lost the election, Peng did not win it.  There are several variants of this.

Variant 5a is that Zheng Yongtang is not equal to Zheng Yongjin.  While people liked and supported the older brother, they were not very impressed with the younger brother.  Perhaps he is dimwitted, uncharismatic, not likeable, lazy, corrupt, pinheaded, or has some other personality problem.  Perhaps they just didn’t like the implication that they were getting a second-best brother.  (I don’t have any evidence for or against this.)  For whatever reason, Zheng Yongjin’s popularity was not transferable to his brother.  This seems plausible to me.

Variant 5b is that Zheng Yongjin’s former popularity has evaporated.  Zheng has been around for a long time.  He was a member of the county assembly, then the speaker, then a legislator, and just finished eight years as county executive.  Maybe people are just tired of him.  Maybe he has cashed in all his political capital and exhausted his support.  Maybe he literally cashed in, let the county government drift, or made a lot of bad decisions during his second term, and this has sapped away at his popularity.  (I don’t have any evidence for or against any of this.)

One problem with variant 5b is that Zheng felt strong enough a few months ago to sponsor his own candidate against the very popular Qiu Jingchun.  She didn’t win, but Zhang Biqin got 76,000 votes.  If we assume that much of her strength came from Zheng’s patronage, then Zheng can’t have been popular enough to sway tens of thousand of votes in December but unable to produce anything in February.  There are only a couple of ways that I can think of to get around this.  One is that Zhang Biqin, not Zheng Yongjin, was responsible for almost all of those 76000 votes.  I’m not sure I believe that, but it’s plausible.  County Assembly speakers are usually strong enough to lose respectably in county executive elections.  (Indeed, Zheng himself lost the 1997 race when he ran as the incumbent speaker.)  The other explanation belongs in variant 5c.

Variant 5c says that maybe Zheng didn’t buy votes, either because he was following orders from the new KMT reform initiative or because he and his brother were literally bankrupt.  Zheng Yongjin has a long history of winning votes, and Hsinchu County has seen its share of vote buying over the years.  I would be very surprised if there were no intersection between these two facts.  Perhaps this time, Zheng simply couldn’t seal the deal because he did not hand the bit of cash to his voters.  (Again, this is pure speculation.  I have no knowledge of whether there was or was not vote-buying in the by-election.)

I don’t have a very high opinion of the KMT’s “reform” process.  They talk of nominating image candidates and practicing cleaner politics, but they nominated two local faction candidates in the recent by election (Zheng in Xinzhu, Wang in Hualian).  Call me when their county assembly speaker and vice-speaker candidates don’t buy votes and aren’t associated with organized crime.  For now, this is more of the same cynical cheap talk that I’ve been hearing for the past 20 years.  So I’m not real optimistic about this possibility.

The possibility of Zheng being out of money is also shaky, though slightly more plausible.  It is possible that Zheng leveraged himself so heavily on the county executive election that he didn’t have anything left for the by-election.  On the other hand, wouldn’t the KMT have known about that and declined to nominate his brother?

I’m rather out of explanations, and I’m not satisfied with anything.  I think maybe I believe in a combination of them.  Mobilization mattered, Peng’s hard work helped, and factional rifts couldn’t have helped Zheng.  If you force me to identify which explanation I believe the most, it’s probably that Zheng Yongtang was not perceived as being as good as his brother.  Zheng Yongtang was simply a lousy candidate.  Of course I don’t have any evidence for this.  Also, it’s a classic spin tactic to blame things you don’t understand on a fantastic or lousy candidate.  So I’m still not convinced I have any idea how Peng won the election.

(Here we run up against my methodological limits.  I soaker and poker would go to Hsinchu, talk to a lot of people, and find some answers.  I, not being much of a field worker, am stuck.)

Tags: 新竹縣