Japan from the inside out

Margin of error, or erroneous margins?

Posted by ampontan on Tuesday, June 1, 2010

THOSE WHO TAKE politics for a game and confuse the national news section with the sports pages must have been thrilled by the volley of media polls released on Monday showing that support for the Hatoyama Cabinet had sunk below 20%, with the interval between air bubbles lengthening. Every Cabinet whose ratings plummet to that level has collapsed within a year, and usually much sooner.

The most widely cited was the Kyodo poll, which pegged the support/non-support figures at 19.1% / 73.2%. The other surveys from the Big Media outlets had the support ranging from 17% (Asahi) to one outlet at 20%.

Some reports claimed this was the first time the polls supporting Mr. Hatoyama came in at less than 20%. That’s incorrect, and an explanation of that error is important, but we’ll get to that shortly. First, let’s take a closer look at some of the other numbers.

DPJ back to #2

The Kyodo poll had another result that must have been a bowel-clutcher for the ruling Democratic Party of Japan. The breakdown of support rate by party shows that the percentage of the electorate backing the DPJ is now below the support for the opposition Liberal Democratic Party. The numbers were 21.9% for the LDP, 20.5% for the DPJ, and 10.5 % for the surging Your Party, led by Watanabe Yoshimi and Eda Kenji.

It would be a mistake to say the public was falling in love with the LDP all over again. The mudboaters haven’t changed a bit. What’s happened is that the DPJ lost its footing and tumbled down the hill past them.

When asked which party they would vote for in the proportional representation phase of the July upper house election, the Kyoto respondents answered 20.9% for the LDP, 19.9% for the DPJ, and 11.4% for Your Party.

In contrast, the Fuji TV / Sankei Shimbun poll had the DPJ maintaining an edge on the LDP in party support at 22.0% to 19.5%, but showed a higher ranking for Your Party at 14.0%

The response for the question of whether Prime Minister Hatoyama was right to fire Fukushima Mizuho from the Cabinet was intriguing. A majority of 51.4% agreed, but a surprisingly high (to me) 40.8% thought otherwise. Of course she should have been canned—had she a sense of honor, she would have resigned before it came to that—but I suspect this is another indication that the Japanese electorate respects politicians who take strong stands and stick to them, regardless of their positions or philosophies.

Keep that in mind the next time someone mindlessly repeats the old chestnut about how the Japanese hammer in the protruding nail. The Japanese love protruding nails.

The Sankei Shimbun also reported on a confidential poll commissioned by the DPJ leadership. A total of 120 seats will be there for the taking in the upper house election, and the DPJ wanted to win 60 of them. That would give them an outright majority and preclude the necessity of putting up with any more nonsense from their cranky coalition partners. One party rule, baby.

The Sankei claims the secret poll shows the best they can hope for today is about 45 seats, with a worst-case scenario as low as 29. The results were particularly bad for the regions west of Tokyo. The article also contained the reminder that poor showings in upper house elections can spell doom for governments–the Abe administration lasted little more than a month after the LDP took only 37 seats in 2006, and the Hashimoto administration was finished after winning only 44 seats in 1998.

The Jiji poll

It’s incorrect to say these are the first polls to have the Hatoyama Cabinet support rate under 20%, however. On 14 May, the Jiji news agency released their poll showing the Cabinet’s support rate at 19.1%. Though some might dismiss it as an outlier, blogger Kibashiri Masamizu on 16 May explained the reasons he thinks the Jiji poll is the most accurate of the media polls in Japan. His blog post is well worth reading, but it’s long, contains several charts, and is in Japanese, so I’ll summarize it here.

The Jiji polls are conducted by Central Research Services (English website), Jiji’s research division. They conduct marketing and advertising research in addition to public opinion surveys. They began monthly polling of political preferences in 1960.

The element differentiating the Jiji poll from the others is the method. The other surveys are conducted using RDD (random digit dialing). In contrast, CRS selects potential respondents from voter and residential registrations in 157 districts nationwide and conducts in-person interviews.

Many polls in the United States also use the RDD method, but it’s been frequently criticized for inaccurate results caused by such factors as the poor weighting of age groups and its inability to reach those with only cell phones. Those who use RDD in the U.S. include the old mainline national TV networks and Gallup. It’s not coincidental their polls were among the least accurate in forecasting the results of the 2008 American presidential election. (They regularly gave Barack Obama a double-digit lead over John McCain. His actual margin of victory was seven points.)

This paper (pdf) written by Donald Green and Alan Gerber of Yale is titled “Enough Already with Random Digit Dialing” and recommends procedures similar to those used by Jiji.

Mr. Kibashiri presents some examples of how RDD has exaggerated polling results in Japan by examining Jiji’s records.

* When the Hosokawa Cabinet was inaugurated, the other media outlets pegged his support rate from 71% to 75.7%. Jiji had it at 62.9%.

* The results were similar for the start of the first Koizumi Cabinet. The rest of Big Media’s Cabinet support numbers ranged from 78% to 87.1%, while Jiji’s figures were 72.8%.

* The gap was most pronounced last September when the Hatoyama Cabinet took office. Most polls had the Hatoyama support rate from 71% to 77%, but Jiji thought it was only 60.6%. Jiji also was the first to show the reversal of support/non-support figures at the administration’s fifth month in office in January.


According to the Jiji polls, those identifying themselves as non-aligned voters have been in the absolute majority every month for the past five years of polling (64 separate polls) with the exception of the months just before and after two of the three national elections in that period. The figures fell below 50% only in 2005, when voters flocked to Mr. Koizumi, and in 2009, when they voted for the DPJ. The majority of the electorate stayed non-aligned even during the election of 2007, when the DPJ became the party with the most seats in the upper house. Just as important, the percentage of independent voters invariably rises again within a few months and reverts to an absolute majority. (Incidentally, according to Jiji, the Hatoyama Cabinet approval rate is lower now than it ever was for the Abe Cabinet.)

These numbers for the average percentage of non-aligned voters over a full year have been stable since 2005. The figures are usually around 55%. Last year, the average slipped to 54%. So far in 2010, it is up to 55.9%.

In contrast, the latest Fuji / Sankei poll claims the independent voters constitute just 42.1% of the electorate.

Mr. Kibashiri draws several conclusions from his analysis of the numbers. The first is that the Japanese electorate is disillusioned with all political parties. The support for the LDP has been steadily falling for years, with the exception of the last year of the Koizumi administration. Their support is still trending downwards.

While there are several reasons for the flagging support of the Hatoyama Administration, including the lack of leadership and money politics, the Jiji polling numbers suggest another reason is a reversion to the mean.

According to Mr. Kibashiri, the outcome of national elections in Japan is determined when the non-aligned voters shift their support to one of the major parties a month or two before the election—i.e., right about now. The current crisis in the Hatoyama administration does not bode well for their electoral hopes this July.

They might not improve even if Mr. Hatoyama were to step down. The mid-December 2009 Jiji poll found that more than 70% of the respondents thought Ozawa Ichiro was the real power in government. Independent voters, particularly younger ones, are unlikely to flock to what they perceive as an Ozawa puppet regime. Rather, a consensus is developing that Ozawa Ichiro is a man whose time has come and gone. Some are referring to the upcoming election as the “Hatoyama-Ozawa Recall Election”.

Your Party

Kyodo’s RDD poll resulted in the following support numbers among independent voters at the time of last year’s election:

DPJ: 36.1%
LDP: 13.9%


DPJ: 11.4% (Even lower than the LDP last August)
LDP: 10.7%
Your Party: 10.9%

It will be worth keeping an eye on the support numbers for Your Party in this month’s Jiji poll. This has historically been the period when the shift of independent voters begins, and the recent RDD polls had the party’s support numbers all over the map:

Percentage of support for Your Party:

Asahi: 5%
Yomiuri: 5%
Mainichi: 9%
Nikkei: 9%
Kyodo: 10.5%

The FNN / Sankei poll also found that Your Party was not peeling off much support from either the traditional DPJ or LDP backers. Their strength came from the non-aligned voters—that group the Jiji polls suggest determines Japanese elections. Will the independents surge to Your Party this year? Leader Watanabe Yoshimi hopes the party wins enough seats to hold the balance of power in the upper house, preventing both the DPJ and the LDP from winning 40 seats each.

These numbers, incidentally, make it easy to understand why the leader of the LDP reformers, Nakagawa Hidenao, was imploring the LDP last year to pitch its appeal to the independents. (That was the same strategy used by Mr. Koizumi for his landslide win in 2005.) He was ignored.

When will they learn?

If there is such a thing as pathological myopia, it infects the Japanese political class, particularly the members of the LDP and DPJ. Here’s something I wrote on another website five years ago. I could have written it yesterday, changing only the names.

The long-suffering Japanese public has been subjected to politicians from the ruling party who don’t pretend to mean what they say, can’t be bothered to hide their disdain for the average voter, and save their remaining passion for their mistresses or money raising. When an eloquent politician appears with enthusiasm, energy, and ideas, and—most importantly—focuses his attention on the public’s concerns rather than trying to convince the public to focus on the politician’s concerns, the Japanese public repays that politician tenfold.

Desperate for a leader who acted like a real human being, still recovering from the disillusionment over the crushing of the first reform government during Hosokawa Morihiro’s term as prime minister in 1993-4, and believing that this was the last real chance to reform Japan’s political system, the public rewarded the off-beat, blunt Koizumi Jun’ichiro with soaring popularity ratings unprecedented in Japan.

He had a specific agenda, articulated that agenda to the public, appointed people to his Cabinet to implement that agenda, and worked to make his policies a reality.

…But none (of the pundits) will mention his most significant accomplishment. He has shown the public and a younger generation of lawmakers that…a strong-willed leader can confront the entrenched political and bureaucratic interests better suited for an undeveloped 19th-century country rather than the interests of the people in a 21st century democracy, and achieve some success.

He has been an icebreaker cutting a path through the frozen wastes of Japanese politics. Before Koizumi Jun’ichiro, no one would have believed it could happen.

The reactionaries

Some people like to say voters have a short memory, but it’s no shorter than that of the dinosaur media either in Japan or overseas, or many of the politicos themselves, for that matter. Not only have they forgotten what Mr. Koizumi did accomplish, it’s as if they’ve forgotten he ever existed. The Fukuda, Aso, and Hatoyama administrations have all been reactionary in the truest sense of the word. Indeed, a common complaint about the current government, primarily a combination of Ozawa fascismos, neo- and paleo-leftists of various shades of pink, and the unionistas of the public sector and the teaching profession, is, “They’re even worse than the LDP!”

Neither party has been able to overcome its origins as an amalgam of smaller parties and groups, becoming vessels for holding political power instead. That vitiates a commitment to a reform agenda, and if it’s anything the Japanese electorate has shown that it craves, it is a commitment to reform.

Had the political class enough mental tinder to start a fire, they would long ago have seen that the Koizumi model is a put-it-in-the bank winner. Instead, they seem to enjoy having brick walls fall on them–they keep finding new ones on the verge of collapse to stand next to.

The hallmark of the last three national elections in Japan has been the electorate’s clear preference for candidates who promise new politics and their punishment of those who support the old politics or fail to keep their promises.

There’s no reason to think their behavior will change in July.

Add to FacebookAdd to DiggAdd to Del.icio.usAdd to StumbleuponAdd to RedditAdd to BlinklistAdd to TwitterAdd to TechnoratiAdd to Yahoo BuzzAdd to Newsvine

2 Responses to “Margin of error, or erroneous margins?”

  1. bender said

    The hallmark of the last three national elections in Japan has been the electorate’s clear preference for candidates who promise new politics and their punishment of those who support the old politics or fail to keep their promises.

    Don’t make promises that you can’t keep- no matter how they appeal to the people. Or, make promises but fake that you’re following them.

    Use bureaucrats instead of confronting them, they are here to stay. Don’t be too affected by the media unless you like confusion.
    B: That’s easier said than done about the bureaucrats. Try 日本国の正体by 長谷川幸洋 as a primer on why it’s so difficult. (An earlier book of his, something with 700 days in the title, talks about being on a blue ribbon panel in the Abe administration and how that opened his eyes.) They are an active force asserting what they think are their interests, play just as dirty as anyone else, and know more than the press, which they also manipulate. Some people from the Koizumi administration wrote that it will take a decade of guerilla warfare to subdue. You have seek this information out about the bureaucracy to find out just how they operate.

    – A.

  2. bender said

    I was trying to say to “know thy enemy”, and not providing any antidote to kill the enemy. Actually, “enemy” might not be the right word, because I don’t agree with destroying them. No country can be reasonably run without them.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: