Japan 2016 - July 10

[jaichind]

広島(Hiroshima) (elected 2)

LDP            49.76% elected
DP              23.15% elected
ORA            13.82%
JCP               7.75%
PJK               2.47%
HRP              1.60%
DP rebel        1.46%

PR section was

LDP+        55.12%
DP+          22.77%
JCP             8.52%
ORA+        11.85%
NPB            1.05%
HRP            0.70%

If we do a Right/Left breakdown we have for district 67.65/32.36 and for PR 67.67/31.29

Using my algorithm on assigning flows of the PR vote between different blocs (also assign rebel vote share to be from the PR vote of said party) we get the net flow in or out of each bloc from PR to district vote to be.

LDP+        -4.98%
DP+           1.32%
JCP           -0.77%
ORA+         4.44%

which in turn we can derive the following inferred net PR vote bloc to district vote flow

JCP -> DP           0.77%
LDP+ -> ORA+   3.68%
DP+ -> ORA+     0.75%
LDP+ -> DP        1.30% (mostly from KP)

Hiroshima is dominated by the LDP so the fight here is for the second seat.  DP has a clear advantage over ORA and JCP.  ORA ran a very strong candidate which provoked some JCP tactical voting for DP as well was LDP+ tactical voting for ORA.  There were also some LDP+ tactical voting for DP which is mostly from KP since KP is not running a candidate here and some KP votes moved to DP due to the constitutional issue.  In the end DP beats out ORA conformably despite a good ideal of LDP+ tactical voting for ORA. 

[jaichind]

福岡(Fukuoka) (elected 3)

DP                30.67% elected
LDP              29.30% elected
KP                21.40% elected
JCP                8.95%
ORA               4.29%
SDP               2.52%
PJK                1.41%
LDP rebel       0.73%
HRP               0.72%

PR section was

LDP+        53.67%
DP+          24.60%
JCP            10.78%
ORA+         9.13%
NPB            1.13%
HRP            0.69%

If we do a Right/Left breakdown we have for district 57.85/42.14 and for PR 63.49/35.38

Using my algorithm on assigning flows of the PR vote between different blocs (also assign rebel vote share to be from the PR vote of said party) we get the net flow in or out of each bloc from PR to district vote to be.

LDP+        -2.77%
DP+           8.02%
JCP           -1.83%
ORA+        -3.43%

which in turn we can derive the following inferred net PR vote bloc to district vote flow

JCP -> DP+        1.83%
ORA+ -> LDP     0.51%
ORA+ -> DP+    2.92%
LDP+ -> DP+    3.28%

LDP KP and DP easily won the the 3 seats as ORA and JCP are relatively weak here.   DP ran a very popular media personality which in turn drove up turnout and led to very large cross voting by all blocs in favor of DP.  As a result DP surged to first place even though it made no difference in the result in terms of seat distribution.

[jaichind]

For Tokyo governor race Kyodo poll has LDP rebel Koike running ahead of united opposition candidate ShuntaroTorigoe with LDP-KP-PJK backed Masuda running third which seems to match CW view of the race.

------------------------------------------------
Koike, Torigoe leading Tokyo gubernatorial election race: Kyodo poll
TOKYO, July 17 Kyodo
Former defense minister Yuriko Koike and veteran journalist Shuntaro Torigoe are leading the race to become Tokyo governor, with former internal affairs minister Hiroya Masuda in third place, a Kyodo News survey indicated Sunday.
But the two-day telephone survey conducted from Saturday, just two days after official campaigning started for the July 31 election, showed around 40 percent of respondents had not decided who to vote for.
Supporters of the ruling Liberal Democratic Party appear split as LDP lower house member Koike, 64, is running without the backing of the party, which is supporting Masuda, 64, also a former governor of Iwate Prefecture.
Koike was backed by over 30 percent of LDP supporters and around 30 percent of unaffiliated voters, while proving very popular among women voters, the survey showed.
Despite the backing of the LDP, only about 30 percent of the party's supporters said they would vote for Masuda, who is also backed by the LDP's junior coalition partner Komeito and the minor Party for the Japanese Kokoro.
Torigoe, who is running as the joint candidate of four opposition parties, was backed by around 60 percent of the supporters of the main opposition Democratic Party and the Japanese Communist Party. The former TV anchorman also had solid support among unaffiliated voters, the survey indicated.
In answering to a multiple response question regarding which areas the new governor should prioritize, 41.9 percent of respondents selected education and child-rearing support, followed by 39.9 percent who cited medical and nursing care. Administrative reform was considered a top priority by 28.8 percent and preparations for the 2020 Tokyo Olympic and Paralympic Games by 9.7 percent.
The telephone survey covering 1,521 randomly selected households with eligible voters in Tokyo received valid responses from 1,033 people.

[jaichind]

Sankei poll also has similar results and has partisan breakdown



It has LDP rebel Koike (Green) splitting the LDP vote with LDP-KP-PJK backed Masuda (Red). 
United opposition candidate ShuntaroTorigoe (Blue) would capture the majority of DP votes which is expected but Koike getting a solid share of KP vote as well.  Masuda wins a clear majority of KP votes who are loyal to party direction.  For the JCP vote ShuntaroTorigoe would take a majority but Koike doing surprisingly well with JCP voters.  For the ORA vote Koike has a clear lead. 

It seems Koike has capture the anti-establishment mood of the electorate and is able to have solid cross party appeal giving her the lead in this race.

[jaichind]

2016 Upper House election exit poll which projected results in terms of seats if only a certain age group could vote



Real result was LDP-KP 70 Opposition Parties 51

　     　LDP-KP    Opposition 
10s        72             49
20s        82             39　   -> strongest age group with ruling parties
30s        75             46
40s        71             50
50s        71             50
60s        61             60     -> strongest age group with opposition parties
70s        70             51

[World politics is up Schmitt creek]

20s        82             39　   -> strongest age group with ruling parties

[Inks].

[jaichind]

Raw votes on the PR section when compared to 2013 (in units of 10,000)



KP (dark pink) pretty much had exactly the same number of votes but with higher turnout has a lower vote share.  This pretty much confirms the view that KP does well in low turnout elections.

DP (Blue) surged since 2013 when compared to 2013.  It added to it a bloc of the 2013 JRP vote  which ORA (Red) lost since 2013 and also got part of the 2013 YP vote (Yellow)

LDP (light pink) went up a bit since 2013 and mostly because it got some of the 2013 YP vote.

JCP (lighter blue) also went up since 2013 but it was way less than expected.

[jaichind]

20s        82             39　   -> strongest age group with ruling parties

[Inks].

Nothing for the opposition to worry about.   Just like the GOP should not really be that worried over its deficit in the Youth vote relative to Dems.  Partisan view change over time and age, especially for people in their 20s.

[Famous Mortimer]

Nice charts but Jesus, can't anyone just give the current party standings?

[jaichind]

Nice charts but Jesus, can't anyone just give the current party standings?

I never got around to doing that mostly because the ground realities keeps on changing.  First, the day after the election the Ind(LDP) winner was retroactively nominated by LDP so he has to count as LDP.  Then a day after that a DP/PLP rebel MP from the class of 2013 who had caucused with NPR also joined LDP giving LDP a simple majority for the first time since 1989.  Then there is the issue of when PJK NPR or AEJ will wind up.  Also it is not clear which of the opposition Grand Alliance independents will join the party they are actually from.  Anyway as of today the standing of the Upper House post-election is

Ruling Bloc
 LDP                           122   -> 1 seat above majority
 KP                               25
---------------------------------------------------
                                 147

Pro-Constitutional change opposition
 ORA                           12
 PJK                              3
 ex-YP Ind.                    3
 ex-PFG Ind.                  1
 Pro Const Chg AEJ        1
--------------------------------------------------
                                  20

Anti-Constitutional change opposition
 DPJ                            50
 SDP                             2
 PLP                              2
 JCP                            14
 OMSP                          1  
 Ind (OMSP)                  1
 Ind (PLP)                     2
 Ind (DP)                      1
 ex-YP                          1
 ANTI Const Chg AEJ      1
---------------------------------------------
                                 75

Total                        242

[jaichind]

Manichi poll

Abe Cabinet Approval 44(+2)/35(-4)
Support talks on Constitutional changes 51/32

[jaichind]

Looking over the results one can identify which prefecture most mimic all Japan results in terms of votes for different blocs.

For PR 静岡(Shizuoka) most matches that of all Japan.
For district votes 福岡   (Fukuoka) most matches that of all Japan
For a hybrid of both 千葉(Chiba) and 神奈川(Kanagawa) most matches that of all Japan

This makes sense.  Chiba and Kanagawa are the suburb and increasing urbanizing area of a growing urban Greater Tokyo while it contains some rural parts which are similar to the rest of rural Japan.

The prefecture which is the least like Japan is of course 大阪(Osaka) given its ultra-strong performance of ORA, the weaker performance of LDP-KP, and ultra-weak performance of DP there.

[jaichind]

Looking through the results of the PR section, especially at the votes for each individual candidate by prefecture allowed me to pretty much reverse engineer the KP election strategy.    

As I mentioned before, the Upper House PR vote allows someone to vote for a name or just the party.  The PR seat allocation by party is based on the total accrued votes for a party (be it for party or a name) but seat distribution within a party list is based on the ranked order by name.   KP's strategy was to concentrate its hardcare votes around 6 names and let the non-hardcare votes just vote for the party.  The idea is that 6 was the minimum target PR seat for KP (they actually won 7).  The purpose of doing this was to demonstrate to the LDP the power of the KP political machine and the power it has over its hardcore KP voters.

Looking at the votes for KP (be it for a name on the list or for the party itself) is distributed by prefecture it is clear how KP went about doing this.  For each prefecture, around 55%-70% of the KP vote was for one particular KP candidate, around 1% for some other candidate and the remaining for the party itself.  So what the KP high command did was to assign a prefecture to one of its 6 candidates and told the KP machine in that prefecture to mobilize its hardcore voters to for the candidate assigned to said prefecture.  The hardcare voters would then lobby their friends and family to vote KP which then would most likely vote for KP the party.  Using this logic of analysis it is interesting that the KP hardcore vote is smaller as a percentage of the total KP vote in urban areas where it should be KP's base and larger in rural areas where KP is traditionally weaker.  It is more likely that in rural areas there are a lot less marginal KP voters.  

Now there are 4 very significant exception to this vote distribution.  In 4 prefectures 埼玉(Saitama), 神奈川(Kanagawa), 愛知(Aichi), and 兵庫(Hyōgo) the percentage of KP votes for a named candidate is in the single digit or teens percentages.  In other words there was no effort in these 4 prefectures to have the hardcore KP votes mobilized to vote for a particular KP PR candidate.  Why ?  The answer is obvious.  These are exactly the 4 districts where the KP candidate running in the district seat was in danger of losing.  So what took place in KP high command is now in retrospect obvious.  Their strategy was for each prefecture KP machine to push up the vote for the KP PR candidate assigned to it and get marginal voters to vote KP the party on the party list.  For these 4 districts, forget to PR section and focus all efforts to get friends and family to vote for the KP district candidate.  In Saitama, Kanagawa, and Aichi the main danger was the perceived JCP surge (which never materialized but KP high command did not know that) and in Hyōgo the danger was JCP tactical voting for DP.  In the end KP won in all 4 prefectures which validated their strategy.  

To show the power of the KP mobilization in these 4 targeted prefectures it is interesting to look at all 7 prefectures where KP ran district candidate and look at the KP district candidate vote as a percentage of the KP PR vote.

4 targeted prefecture for KP mobilization
埼玉(Saitama)              130.9%
神奈川(Kanagawa)        126.5%
愛知(Aichi)                   139.4%
兵庫(Hyōgo)                 146.1%

3 non-targeted prefectures
東京(Tokyo)                 108.5%
大阪(Osaka)                 111.7%
福岡(Fukuoka)             126.6%

In the non-targeted prefectures the extra KP district votes are most likely tactical voting by LDP voters to help ensure their ally wins.   In the targeted seats for KP mobilization it is LDP votes plus significant number of DP and ORA tactical votes to help bloc JCP based on the work of the KP election machine to generate these tactical voters.  In  Hyōgo it seems the threat of DP plus the JCP tactical vote for DP was the greatest and the KP machine went into overdrive and got 46%+ more votes for the KP district candidate than the KP PR vote.  I projected Hyōgo for DP over KP and while I was wrong I am glad to see that KP saw it the same way I did: that if KP was going to lose a district seat it was going to be in Hyōgo.  In fact if KP just got 110% of its KP vote in Hyōgo just like in Tokyo then it would have lost to DP for the third seat.  Only a frantic effort by KP saved the seat validating that my minority opinion (very much so) projection had legs.

KP did execute a basic principle of management.  Have one goal and focus the entire organization around that goal.  Once they realized that in 4 prefectures the KP candidate could lose they threw out this goal of showing its power to LDP in these 4 prefectures and told its machine in these 4 prefectures "Forget the PR section, go get votes for the district candidate !!"  Very effective management strategy.  The various business schools would approve and should use this as an example of effective organizational management.

[jaichind]

Looking through the PR vote section by prefecture also resolved another mysterious result which I could not explain before.  Telling the tale of my discovery also helps tell a bigger story of why LDP win.

The initial mystery was around the result in the 1- district and old LDP stronghold 富山(Toyama).  What took place was was turnout surged from 2013 by 5.4% when the rest of Japan turnout went up by around 2%.  Then on the PR section the DP PR vote surged to 24.22% from 10.3% in 2014 and 9.41% in 2013. As a result LDP+KP PR vote was at 54% which was a drop of 2.3% from 2014 when LDP+KP PR went up by 2.6% in the PR section since 2014 across Japan.  The in the district section for Toyama LDP completely overperformed its PR vote over the united opposition candidate.  Based on my regression model, which uses PR vote section and looks at incumbency effects, the LDP got over 8.4% vote share over what my model would imply.  Part of it can be explained by a popular LDP incumbant and weak opposition independent candidate but then would would there be a surge of DP PR vote with high turnout only for the same voters to then just vote LDP on the district vote.

This mystery got unraveled when I looked at the PR section by candidate and prefecture.  It turned out that an incumbant ex-YP MP 柴田 巧 (Shibata Takumi) who won in 2010 on the YP PR slate but then joined DP won a large amount of the PR vote in his home prefecture of Toyama.  Out of the 116.5K PR votes cast for DP in Toyama, 柴田 巧 (Shibata Takumi) himself won as a vote for him 50.3K.  So nearly half of the DP PR vote in Toyama was cast for Shibata.  Just for comparison, in other prefectures the DP PR vote is usually around 20% for candidate names (let alone just one candidate)  and 80% for the party itself.  Shibata lost reelection because even though he won 50.3K in Toyama he won only 73K over all of Japan which placed him 17th on the DP PR list and DP only won enough votes to win 11 seats.

柴田 巧 (Shibata Takumi)


Looking at  柴田 巧 (Shibata Takumi) himself.  He was with the LDP in the Toyama prefecture assembly but in 2009 ran as a LDP rebel in the Lower House elections and lost but winning 50K votes in 1 of the 3  Toyama Lower House districts.  Then he joined YP and ran on the PR list for YP in 2010.  In 2010  Shibata won 87.8K vote of which most of it must have come from  Toyama which place him first in the YP PR list of which 4 was elected based on a 8.9% vote share and over 4.7 million votes.  2010 YP PR list did not have that many well known candidates and 90% of the PR vote for the YP party itself so 87.8K was enough to place him first.  2016 DP PR list had a lot more well known candidates so a slight decrease in total number of votes for Shibata drove him to 17th place. So Shibata went from a big fish in a small pond to a small fish in a big pond and lost re-election.    It seems most of Shibata's votes in Toyama are his personal vote and would otherwise vote LDP.  Just to give an idea the impact Shibata has on the PR vote, in 2010 when Shibata ran on the YP PR slate, YP won 20.34% of the Toyama PR vote.  In 2012 and 2014 when YP did not have Shibata  as a personal candidate on the PR list (of course 2012 was a Lower House election where one cannot for a PR candidate anyway) YP won 6.67% and 4.96% in the PR section of Toyama.

So now the mystery is solved.  The DP surge in Toyama 2016 was based on the fact that  柴田 巧 (Shibata Takumi) by himself won over 10% of the PR vote which pushed to the DP PR vote share to above 24% when it recently had been in the single digits.  But most of this vote are the personal vote of  Shibata and when left to vote on the district section mostly voted LDP.  It seems given the size of Shibata's personal vote it is likely he will run in the next Lower House elections so he can add his personal vote with the center-left and possibly JCP and make his race potentially competitive.  

The story of 柴田 巧 (Shibata Takumi) is very instructive on why LDP wins.  For each 柴田 巧 (Shibata Takumi) who is a LDP rebel, many other local politicians with personal votes stay in LDP.  The dynamic is the following:

LDP is in charge of central government -> Local governments cannot issue debt -> Center Government distributes subsidies to local government via the MPs and MLAs of said prefecture -> local politicians join LDP so bring home the bacon -> local politician personal votes transferred to LDP for national elections -> LDP in charge of central government.  

At least for now Shibata being part of DP has not lead to his personal base going over to DP as seen by the district vote but over time that could change.  
  

[Vega]

I'm rather shocked that Iwaki Mitsuhide, the Justice minister, lost his seat. He apparently wasn't on the other Proportional list either.

[jaichind]

I'm rather shocked that Iwaki Mitsuhide, the Justice minister, lost his seat. He apparently wasn't on the other Proportional list either.

This is because his district, 福島(Fukushima), went from a 2- member district to a 1- member district.  So back in 2010 when both he and his DP opponent won, this time around it is a 1 on 1 race with JCP backing DP.  As a result he lost.  BTW, in Upper House elections you cannot run in both the district and PR section at the same time, which you can in Lower House races.

[Vega]

I'm rather shocked that Iwaki Mitsuhide, the Justice minister, lost his seat. He apparently wasn't on the other Proportional list either.

This is because his district, 福島(Fukushima), went from a 2- member district to a 1- member district.  So back in 2010 when both he and his DP opponent won, this time around it is a 1 on 1 race with JCP backing DP.  As a result he lost.  BTW, in Upper House elections you cannot run in both the district and PR section at the same time, which you can in Lower House races.

Ah, that makes sense. I think he can still stay on as Justice minister though, as the PM can have some normal citizens in their cabinet.

[jaichind]

For the 4 opposition independents that were elected, the 2 Ind(PLP) winners from 岩手(Iwate) and 新潟(Niigata) will obviously join PLP.  The ex-DPJ MP  that won in 山形(Yamagata) will caucasus with DP and I am sure will join DP in due time.  The Ind(AO) winner from 沖縄(Okinawa) is still undecided and most likely will stay an independent since AO is not a party but just an alliance of anti-base Okinawa prefecture lawmakers. 

[jaichind]

The two biggest vote getters in this election are both women with a media background before joining politics with both getting over 1 million votes.  They are

DP's 蓮舫(Renhō) of 東京(Tokyo)

She might eventually run for a leadership role in DP







LDP's 三原　じゅん子 (Mihara Junko) of 神奈川(Kanagawa)

She is a former singer and actress.  It is hard to believe she is almost 52.  She actually have fairly strong revisionist and hawkish views and is very conservative on social issues.  She is pretty in your face about it which makes it harder to pursue a  large leadership role in LDP.  But given her personal pull she is a very large vote getter for LDP.

[jaichind]

Another set of data I was able to dig out of the PR vote distribution by candidate was able to explain the unexpectedly large PR vote share by NPR (1.04%) as well as a tale of how ORA threw away a PR seat.  

NPR is pretty much a defunct Right Libertarian LDP splinter party.  Without knowing too much at the ground I figured they were only good for 0.5%. but they actually managed 1.04% which was still not enough to win a PR seat but still higher than I expected.  I looked into why.  It turns out an ex-YP incumbent MP 山田太郎 (Yamada Taro) won single handedly over 290K votes which itself was 0.52% of total PR vote share making him the 13th largest vote getter of all the PR candidates running.  So NPR got around 0.5% of the PR vote but when added  by Yamada's vote put NPR past 1% vote share.  Of course since NPR failed to win a seat on PR Yamada was not elected.

山田太郎 (Yamada Taro)


山田太郎 (Yamada Taro) was elected as part of the class of 2010 YP  slate.  When YP became defunct at the end of 2014 he joined AEJ.  While there he took on the cause of freedom of speech when it came to laws that try to regulate Hentai comics books on the premise of stopping child pornargraphy. This appears to have won him the support of a Hentai reading male youth voting bloc the media call the Geek Vote.  In early 2016 Yamada quite AEJ when it was clear AEJ will not even try to contest in 2016 and joined ORA who nominated him to be their candidate for 埼玉(Saitama). Yamada quit ORA claiming that he joined the the premise he would be put in the ORA PR list a fact ORA disputed.  From Yamada's point of view running in Saitama would lead to sure defeat when his support is evenly spread across Japan which is a perfect match for a PR candidate.  ORA's concern is that having Yamada on its PR list would only cannibalize the ORA PR vote and offend other ORA PR list candidate who would see Yamada potentially large vote as squeezing them out of their sure seat.   So Yamada ran on NPR PR list winning a large number of votes from the "Geek voting bloc" but failing re-election.  ORA made a big mistake as Yamada's ability to win over 0.5% of the national PR vote shows that he as independent draw and had he run on the ORA PR would have easily added a seat, won by Yamada, to the ORA seat count at the expense of LDP.  

[jaichind]

Abe now has led the LDP to 4 straight election victories in a row of 2012 2013 2014 and now 2016.  2016 is not a landslide like 2012 2013 and 2014 but it is a solid victory.  The last time the ruling bloc has won at least 4 solid victories in a row was back in the 1960s.

[Vega]

I wonder if Iwaki will mount a comeback... perhaps running in the lower house elections? He was a pretty important member of the LDP.

[jaichind]

Final PM on my projection

1 seat districts LDP-opposition 21-11

                       PR           My estimated          District              Total
                                     PR vote share
LDP                16                 31.00%               36                     52
Pro-LDP Ind                                                     0                       0
KP                    7                 13.50%                6                     13
ORA                 5                   9.50%                 2                      7
PJK                  0                   1.00%                 0                      0
NRP                 0                   0.50%                 0                      0
NPB                 0                   1.00%                 0                      0
HRP                 0                    0.50%                0                      0
DP                 12                  24.50%               22                    34
SDP                 1                    2.50%                0                       1
PLP                  0                    1.50%                0                       0
VPA                 0                     1.00%               0                       0
JCP                  7                  13.50%               3                      10
Pro-DP Ind.                                                     4                       4    (1 DP, 2 PLP, 1 AO)

When the result result was

So the final result is

1 seat districts LDP-opposition 21-11

                       PR          PR vote share          District              Total                                      
LDP                19                 35.9%                36                     55
Pro-LDP Ind                                                     1                       1 (retroactively nominated by LDP)
KP                    7                 13.5%                  7                     14
ORA                 4                    9.2%                 3                       7
DP                 11                  21.0%                21                    32
SDP                 1                    2.7%                 0                       1
PLP                  1                    1.9%                 0                       1
JCP                  5                  10.7%                 1                       6
Pro-DP Ind.                                                     4                       4    (1 DP, 2 PLP, 1 AO)

Overall one mistake I made which seems to have been shared by most election projections I overestimated the JCP surge.  This was true for the PR section as well as the district vote.  I also completely underestimated the LDP PR vote and overestimated the DP PR vote. I actually got every district correct except for 3.  The 3 mistakes I made were

神奈川(Kanagawa)  -> I called for JCP to win the 4th seat instead of Ind(LDP)
大阪(Osaka) -> I called for JCP to win the 4th seat instead of ORA
兵庫(Hyōgo) -> I called the DP to win the 3rd seat instead of KP

All 3 can be attributed to poor DP-JCP tactical voting but for Kanagawa and Osaka my overestimation of the JCP vote was decisive.   In Hyōgo it seems that KP went into overdrive to save its seat and was successful against a DP that was unable to get its PR vote bloc consolidated behind its candidate.

My 3 mistakes pretty much rank among the lowest if not the lowest among all know district projections.  This might seem strange if I was wildly off on the LDP PR vote share.  This is because I actually got two separate trends incorrect and the two seems to have canceled each other out.  To show how I will first look at the LDP-KP district vote share (which includes LDP rebels) which came out to 48.77%.  Note my own projection of the LDP-KP district vote share (which includes LDP rebels) was 47.88%.  So in terms of district vote I was much closer to the mark despite be off by 5% on the PR vote for LDP (31.0% vs 35.9%).

What took place was my LDP PR vote share projection was based on discontent with the Abe regime.  I assumed the discontent would play out where voters would vote non-LDP on the PR vote but vote LDP in the district vote to block JCP backed opposition unity candidate.  It seems what took place was the anti-Abe but anti-JCP vote went to LDP on the PR but went to a combination of HRP ORA+ and DP in the district vote.  This dynamic is exactly the opposite of my model which produced results wildly off on the PR section but fairly accurate results in the district vote.  

One way to look at this is to break the districts into 3 zones.  Those that ORA did not contest (so the anti-Abe LDP PR voter did not have a natural outlet), those that ORA did contest in except Tokyo, and Tokyo.  I separated Tokyo out as there was clear signs of anti-Abe voting in the district vote even as ORA was running.  The diversity of the Tokyo district candidates gave it a different pattern LDP PR voter defection.  I also choose separation for analysis because it seems that ant-Abe defection by LDP PR voters were more intense in rural districts where ORA pretty much did not run and where clearly Abenomics is doing nothing to help.  Tokyo is the exception so I broke that out separately.  Then using my regression modeling I was able to look at the net gain and losses to the LDP+KP PR vote to get to the final LDP+KP district vote.

                      LDP/KP                   Net gain/loss by LDP-KP relative to                                  LDP/KP
                    PR vote      HRP       ORA+     DP+     LDP rebel     NBP     Other   ex-YP Ind.   District vote
Districts
where ORA
did not run   51.98%   -1.76%   4.55%  -3.10%    -0.52%     0.49%   0.00%  -0.14%        51.50%

Districts
where ORA
did run
(excluding
Tokyo)         47.17%   -0.43%   0.47%   0.48%     -0.23%    0.52%   0.00%   -0.22%       47.70%

Tokyo          45.81%    0.04%   -1.08% -8.14%      0.00%    0.57%  -0.24%     0.00%      36.96%

All               49.45%   -1.05%    2.36% -2.28%     -0.35%    0.46%  -0.03%   -0.16%       48.41%

Note that the vote share of 48.77% in the district voet by LDP-KP plus LDP rebels  when the LDP-KP PR vote is artificially high.   In the districts where ORA did not run the LDP-KP bloc got 4.55% of the vote from ORA+ PR voters which they would have most not gotten if ORA ran a candidate.    In these same districts one would have expected DP+ would have lost PR votes to the LDP since in most of them the opposition was running in an alliance with JCP would would have drove anti-JCP DP+ voters into voting LDP.  I am certain this did take place but the LDP PR defection from the LDP district candidate was much stronger.  A similar dynamic was taking place in Tokyo.  A LDP rebel is now doing well in the Governor race in Tokyo is part of the same dynamic.  There seems to be a strong undercurrent of discontent against Abe which manifest itself in voting anti-LDP in the district vote but still voting for LDP in the PR section.  The discounted LDP voter I guess is still voting for the non-Abe factions of LDP when they vote for the LDP in the PR but personified its dissatisfaction at Abe in the district LDP candidate.   The surge of HRP vote share in districts where ORA is not running is also a sign that there are a lot of discounted LDP voters that voted LDP on PR but wanted to protest against Abe somehow.

If fact if we back out most of the 4.55% LDP candidates got from ORA+ PR voters in districts where ORA was not running then the total LDP-KP district vote would be something like 46.23% which if you add the 0.35% for LDP rebels comes out to 46.58%.  This number is a lot closer to my projected 44.5% for LDP-KP PR vote.  This number will be lower once we back out the unquantifiable anti-JCP DP+ PR voters which most likely vote LDP in the district vote in the 1- districts where DP had an alliance with JCP. 

One extreme example of this dynamic is in 山形(Yamagata).  Here the TPP issue meant that it was clear LDP was in trouble.  One would expect similar trouble for LDP-KP on the PR section as well.  But instead we got the following results in 山形(Yamagata) in the district and PR vote sections

District
Ind(DP)         59.05%
LDP               38.34%
HRP                2.61%

PR
LDP              42.49%
KP                11.78%
HRP                0.79%
NPB                1.62%
ORA+             4.30%
DP+             29.56%
JCP                7.60%

The LDP PR vote, by itself, was greater than what the LDP district candidate got even without taking into account of the KP PR vote of which a lot must have went to the LDP candidate.   There was massive defections from LDP PR voters toward the opposition unity candidate.  It seems any concerns from the discontented LDP voter about JCP manifested itself by it voting for LDP on PR rather than voting for the LDP district candidate. 

[jaichind]

I wonder if Iwaki will mount a comeback... perhaps running in the lower house elections? He was a pretty important member of the LDP.

Possibly.  It would be hard as out of the 5 lower house seats in 福島(Fukushima) 3 are held by LDP incumbents and the other two will be hard to wrestle from DP.  I guess he can run in one of the two held by DP, lose narrowly and then get in via PR.  Not sure he wants to put himself through that.   Note that he still can be in the cabinet despite losing his seat even though it will be hard for Abe with so many LDP MP waiting for their turn to be in the cabinet. 
 

[jaichind]

The 2 沖縄(Okinawa) MPs one which is from OSMP elected in 2013 and the Ind(AO) MP just elected has formed a separate caucus call Okinawa Wind.  They will be dedicated to the USA base issue.