clan wars

In today’s useless post, I look at a question that might have mattered more two hundred years ago.  Which surnames are winning at politics?  If you are a certain party chair, you might think to yourself, “Why aren’t there more people named Su in politics?  There should be more Sus!”  If you are a certain southern mayor, you might think the same thing about Lais.  So who has the right to feel oppressed, and who should be smug about the wonderful past performance of their clan?

I am using my database of elections in Taiwan.  Keep in mind that the people you are familiar with count the same as the people even I have never heard of.  I haven’t started collecting village chiefs 村里長 data, but I have three election cycles of township council 鄉鎮市民代表 elections (plus a few earlier for a few counties).  These candidates make up about 47% of the database.  City and county councils are another 24%.  The famous people, including all the candidates for president, governor, municipal mayor, city mayor, and county magistrate (going back to 1957) are only about 1.6% of the total.

Of course, the most common surname in Taiwan politics is Chen 陳.  This is hardly a surprise, since Chen is the most common surname in Taiwan.  The interesting question is whether there are more Chens in politics than in society.  Do Chens win more than their fair share of seats?

In my short internet search, I only found population data for the top ten surnames.  So here is how the top ten did:

	Population%	Individuals%	Seat%	Seat/Pop
陳 Chen	11.13	11.36	11.47	1.030
林 Lin	8.30	8.72	9.18	1.106
黃 Huang	6.03	5.97	6.03	1.000
張 Chang	5.28	5.16	5.12	0.970
李 Lee	5.13	4.99	5.40	1.052
王 Wang	4.12	3.99	3.76	0.912
吳 Wu	4.04	3.82	3.71	0.919
蔡 Tsai	2.91	2.84	3.08	1.060
劉 Liu	3.16	2.87	3.02	0.956
楊 Yang	2.66	2.45	2.27	0.855

If your family name is Huang 黃, you should feel ok.  The Huangs didn’t do well, and they didn’t do badly.  Huang is represented at just about the appropriate level.  6.03% of the population is named Huang, and 6.03% of the seats were won by people named Huang.  Huangs are, well, average.

Among the top 10 families, Lins 林 are the big winners.  8.30% of people in Taiwan are named Lin.  If you look at individual politicians, 8.72% are named Lin.  (Individuals looks at individual people, not candidacies.  For example, LY Speaker Wang Jyn-ping 王金平 has run 12 times.  Here he counts as one person, not 12 candidacies.)  When Lins run in elections, they also do better than average, and they have won 9.18% of all the seats.  That gives a seat/population ratio of 1.106.  In other words, Lins have won 10.6% more seats than their fair share.

The biggest losers are the Yangs 楊, who only get 85.5% of the seats they might be expected to win.  In fact, in my dataset, the Yangs did so poorly that they aren’t even in 10^th place.  They were beaten by the Hsus 許, and by quite a large margin (741 to 633 seats).

 

Lin and Yang were the biggest winners and losers, respectively, in the top ten.  However, they don’t appear to be the biggest winners and losers overall.  This makes sense, as larger samples tend to have smaller variance.  Wang Jyn-ping and his twelve seats barely make a difference to the Wang statistics, as the other Wangs won 1034 seats.  However, if he were named Chin Wang-ping 金王平, his twelve wins would increase the Chin total from 31 to 43.  A single individual can make a big difference to a smaller name.

Since I don’t have population data, the way to evaluate the rest of the names is to see if the individuals who enter politics win lots of seats.  On average, each individual wins 0.988 seats.  There are two ways that individuals can win more seats.  First, they can run lots of times.  Second, they can have a good winning percentage.

My initial guess was that the Yu 余 family would be the most overrepresented.  I was thinking about Kaohsiung County and all the Yus who have run multiple times and almost always won.  Also, Yu is a fairly small surname, so it should be relatively sensitive to one outlier family.  I was wrong.  The Yu clan is just about average.  (So much for my expertise!)  Here are the top 50:

			Ind%	Runs/Ind	Win%	Seat%	seats/ind
all				1.91	0.515		0.988
1	陳	Chen	11.36	1.86	0.530	11.47	0.988
2	林	Lin	8.72	1.95	0.528	9.18	1.030
3	黃	Huang	5.97	1.92	0.514	6.03	0.989
4	張	Chang	5.16	1.92	0.505	5.12	0.971
5	李	Lee	4.99	1.95	0.544	5.40	1.058
6	王	Wang	3.99	1.85	0.496	3.76	0.920
7	吳	Wu	3.82	1.87	0.508	3.71	0.951
8	劉	Liu	2.87	1.92	0.535	3.02	1.029
9	蔡	Tsai	2.84	2.00	0.532	3.08	1.062
10	許	Hsu	2.54	1.94	0.528	2.66	1.023
11	楊	Yang	2.45	1.88	0.482	2.27	0.907
12	謝	Hsieh	1.90	1.84	0.509	1.82	0.935
13	鄭	Cheng	1.84	2.02	0.545	2.07	1.103
14	洪	Hung	1.60	1.94	0.543	1.71	1.051
15	邱	Chiu	1.57	1.85	0.523	1.56	0.969
16	曾	Tseng	1.56	1.88	0.487	1.46	0.914
17	郭	Kuo	1.41	2.01	0.538	1.55	1.083
18	徐	Hsu	1.32	1.85	0.524	1.31	0.968
19	賴	Lai	1.25	1.91	0.486	1.18	0.930
20	周	Chou	1.22	2.03	0.524	1.32	1.064
21	廖	Liao	1.17	1.88	0.522	1.17	0.982
22	葉	Yeh	1.09	2.00	0.519	1.16	1.039
23	高	Kao	1.04	1.85	0.447	0.88	0.828
24	蘇	Su	1.02	2.08	0.568	1.23	1.179
25	江	Chiang	1.00	2.08	0.517	1.10	1.077
26	呂	Lu	0.94	2.12	0.581	1.18	1.232
27	莊	Chuang	0.91	1.77	0.475	0.78	0.842
28	何	Ho	0.86	2.00	0.528	0.93	1.057
29	羅	Luo	0.86	1.92	0.491	0.83	0.943
30	簡	Chien	0.80	2.00	0.517	0.85	1.035
31	鍾	Chung	0.80	1.88	0.547	0.84	1.026
32	潘	Pan	0.75	1.85	0.519	0.74	0.958
33	蕭	Hsiao	0.71	1.80	0.514	0.68	0.926
34	朱	Chu	0.65	1.77	0.534	0.62	0.946
35	彭	Peng	0.64	1.78	0.448	0.52	0.798
36	游	You	0.63	2.09	0.557	0.75	1.168
37	詹	Chan	0.57	1.76	0.498	0.51	0.876
38	余	Yu	0.56	1.94	0.516	0.57	1.000
39	胡	Hu	0.56	1.74	0.462	0.46	0.804
40	柯	Ko	0.53	1.92	0.535	0.55	1.027
41	顏	Yen	0.51	2.06	0.542	0.59	1.116
42	盧	Lu	0.47	1.83	0.518	0.46	0.948
43	施	Shih	0.45	1.91	0.484	0.42	0.922
44	趙	Chao	0.43	1.87	0.425	0.35	0.795
45	沈	Shen	0.43	1.78	0.548	0.43	0.975
46	魏	Wei	0.36	1.97	0.547	0.39	1.078
47	梁	Liang	0.36	1.69	0.529	0.33	0.892
48	翁	Weng	0.33	1.96	0.522	0.35	1.021
49	戴	Dai	0.33	1.76	0.543	0.32	0.957
50	宋	Soong	0.32	1.97	0.453	0.29	0.890

 

The biggest loser on this table is the Chao 趙 clan, who only win .795 seats per individual.  However, Chao is a fairly small clan, so they will have larger swings.  For my money, the worst performance comes from the Kao 高 and Peng 彭 clans, which are far bigger than the Chaos and still only managed to win .828 and .798 seats per candidate.  To put that in perspective, if the Pengs had won the average of .988 seats per individual, they would have won 181 seats.  In fact, they only won 146 seats.  The Pengs and Kaos were slightly below average in running for office, but the main culprit is that the simply didn’t win enough times when they did run.  Their .447 and .448 winning percentages are the lowest on this table.  If we go off the table, #53 (Fang 方) is the first spot to do worse, at .440.  The worst winning percentage in the top 100 belongs to #92 (Chien 錢), at .354.  The most reluctant to run was #89 Kung 龔, at a mere 1.33 runs per candidate.  And the worst overall in the top 100 was #88 Tsou 鄒 at only .593 seats per individual.  If your family name appears in this paragraph, too bad for you!

What about the winners?  It’s time for the former vice president to gloat.  Lu 呂 is the champ.  Each Lu ran 2.12 times.  To beat that, you have to go all the way down to #61 (Lan 藍) at 2.19.  #96 Wu 伍 is the best in the top 100, at 2.33 campaigns per person.  Lu also had the best winning percentage in the top 50, at .581.  The first to top that was #56 Wen 溫 at .608.  #99 Chang 章 won 24 of 33 races, for the highest winning percentage in the top 100 (.727).  In all, members of the Lu clan won a stunning 1.232 seats per person.  With an average .988 seats per person, they would have had 264 positions.  In fact, they filled 329 offices!  To beat the Lu clan, you have to go all the way down to #97, where the (very small) Han 韓 clan won 1.611 seats per person.  Apparently, Lu is the ruling class!