Paul's Internet Landfill/ lj/ MMP 102: Are people happier under MMP?

MMP 102: Are people happier under MMP?

Here's the entry I don't want to write. It asks a simple question: are citizens happier with MMP than they would be with FPTP?

As usual, I offer a long and complicated answer, focussing on three situations: the world at large (courtesy of Arend Lijphardt's well- known research), New Zealand (one of the few democracies that has made an explicit transition from FPTP to MMP), and Wales and Scotland (both of which use FPTP nationally, but adopted MMP when devolution came into effect in 1999).

General Comparisons

The Fair Vote freaks like to point to Arend Lijphardt's 1999 book Patterns of Democracy: Government Forms and Patterns in Thirty-Six Countries, which compared so-called "majoritarian" and "consensus" democracies. I still have not been able to get my grubby hands on a copy of Lijphardt's book, so like many others I have been depending on a summary prepared by Fair Vote Canada.

The summary highlights a number of findings, all of which just happen to favour proportional representation. Among the most interesting for me: a study which found that citizens in consensus (more proportional) systems rated satisfaction with democracy much higher than those in majoritarian democracies (p. 3). Lijphart also reports that those who voted for "losers" in consensus democracies tended to be more satisfied with democracy than their counterparts in majoritarian systems.

Other papers I have read support this idea that proportional systems result in governments that are more in tune to voter wishes. A (somewhat sketchy) paper by Bengtsson (Bengtsson 2005) looks only at proportional and semi-proportional systems, and finds that voters appear to rate inclusiveness (characterized by proportionality) over accountability (characterized by single-party governments) when ranking the efficacy of governments.

In contrast, a paper by Blais and Bodet (Blais Bodet 2003) finds that proportional systems are no better and no worse than majoritarian systems in electing governments that reflect voter views: in majoritarian systems like FPTP, all parties tend to be centrist, so they start out not that different from the median voter. In proportional systems the parties that form government tend to come from a wider distribution, but on average they reflect the views of the median voter well. (Again, I find this paper sketchy because it assumes that only the median voter matters, without taking into account distribution of voter views.)

Overall, the cross-country literature seems to indicate that governments in proportional systems tend to be more responsive to their voters, and that voters feel they are happier with democracy overall.

So far so good, but the picture gets a lot less rosy when we focus on New Zealand's public perceptions.

New Zealand

New Zealand adopted MMP in two stages. In 1992 the government held a referendum which asked voters two questions: whether they wanted to replace FPTP, and which of four alternatives they preferred the most: MMP, Single Transferable Vote (STV), a parallel system, or Alternative Vote (AV). 84.5% of voters recommended replacing FPTP, and 65% chose MMP as their favourite system. In 1993 the government held a second referendum which pitted FPTP against MMP directly. MMP again proved victorious, but this time only 54% of voters selected it over FPTP. (Nagel 1994)

It appears that MMP continued being more popular than FPTP until the first election under MMP, held in October 1996. This marked the first rocky and unpopular National-New Zealand First coalition. From late 1996 to 1999, MMP lost its favoured position: support for MMP hovered around 30%, while about 50% of those polled favoured FPTP instead. (Karp Bowler 2001, p. 24). Although I have not been able to access consistent polling data since then, a few snapshots indicate that support for MMP has remained low since then.

A 2001 government inquiry into MMP examined voter attitudes into MMP in some depth. It found that opinions towards MMP have been volatile: when politics is going smoothly support for the voting system goes up, and when governments get into trouble (as they did in 1997 and 1998 when the National-New Zealand First coalition was breaking apart) people tend to despise the voting system.

Two sets of polls conducted in 2000 and February 2001 show some of this volatility. In 2000 27% of respondents had a positive view of MMP and 44% had a negative view; by 2001 this had eased slightly to 32% positive and 43% negative. The committee also found that FPTP was ranked higher than MMP in all but one survey survey from late 1996 to 2001. The exception was in December 1999, when MMP was preferred to FPTP by a margin of 45% to 43%. In contrast, support for FPTP vs. MMP has been as high as 59% to 29% in November 1998. (MMP Review Committee 2001, p. 77).

The latest snapshot I have found comes from the 2006 paper from Vowles, Banducci and Karp that I cite incessently. They say that MMP was favoured by a "small majority" after the 2002 elections, which suggests that a majority of citizens did not support MMP before or since. (Vowles Banducci Karp 2006, p. 282)

Case closed, right? Certainly, I was not very happy to learn these results, and I am sure my friends on the NoMMP side will be all over these numbers if they bother reading this blog post.

In fact, I think there may be more to this story. Even people who support FPTP over MMP don't advocate ditching MMP entirely. In the February 2001 set of polling, respondents were asked whether they should leave MMP as-is, whether they should keep the basic structure but make some changes to how MMP operates, or whether they should change to another electoral system altogether. Here are the results (MMP Review Committee 2001, p. 78):

Option | Overall | Support MMP | Support FPTP

Stay with MMP as-is | 17% | 36% | Not reported

Make some changes | 47% | 57% | 35%

Switch to another system | 31% | Not reported | 53%

Some caveats: the paper did not include some numbers, and the authors noted that some of the ways in which people would have "modified" MMP were infeasible (such as eliminating list MPs). However, this does demonstrate that support for FPTP (or hatred of MMP) is not as clear- cut as it may appear.

One striking example of this comes from the National party, which has not formed a successful government under MMP and for the most part despises it. The National party wants to hold another referendum on electoral reform, but they do not recommend a return to FPTP: rather, they propose using a parallel system instead. (Vowles Banducci Karp 2006, p. 282)

Even more striking is the following statement from Brian Nicholle, who was campaign manager for the pro-FPTP side in the 1993 referendum (Nicholle 2006):

Ten years on since the first MMP election on October 12, 1996, it's time that the people had the chance again to vote in a referendum to either retain or reject MMP.

FPP is a relic of the command and control economy but the SM (Supplementary system) which contains a measure of proportionality combined with FPP) would be the system to go head to head with MMP in a referendum.

Even the campaign manager who defended FPTP in 1993 is willing to call that system "a relic of the command and control economy", and to advocate a "supplementary system" (which I am pretty sure is a parallel system, aka MMM) in its place.

Why? I think the reason is simple: New Zealand voters like some aspects of MMP, even if they dislike the system as a whole. When polled for the 2001 review, lots of New Zealanders felt that MMP had been successful at getting more women into power (50% felt it had been successful as compared to 17% who felt it was unsuccessful), getting more Maori into power (49% successful, 18% unsuccessful), and creating a parliament that was more representative of all New Zealanders (45% successful, 27% unsuccessful). (MMP Review Committee 2001, p. 80)

Similarly, New Zealanders seemed to appreciate the increased choice of the new system. 62% agreed that MMP made it easier to vote for the best local candidate in a riding (vs. 19% disagreeing) and 54% agreed that it gave voters more options (vs. 21% disagreeing). (ibid, p. 83) Given this taste of increased representation and choice, it should come as no surprise that even MMP opponents like Brian Nicholle propose a system that keeps some elements of voter choice (albeit diluting it considerably).

There is another intriguing aspect to the New Zealand experience: even though support for the proportional effects of MMP remains low, overall citizen attitudes towards their government appear to be improving. Take a look at the following NZES data, reproduced from (Vowles Banducci Karp 2006, p. 278). The table shows some statements about democracy and government, and then the percentages of people who agreed with that statement each year:

Statement | 1993 | 1996 | 1998 | 1999 | 2001 | 2002

MPs out of touch | 61 | 53 | 76 | 52 | 56 | 49
People like me have no say | 63 | 57 | -- | 55 | -- | 46
Politicians do not care what people think | 66 | 57 | -- | 55 | -- | 50
Government run by a few big interests | 60 | 54 | -- | 50 | -- | 42
Satisfaction with democracy | -- | 73 | 45 | 57 | 60 | 67
Trust government to do what is right | 31 | 30 | 26 | 36 | 47 | 44
Trust in a political party (**) | 44 | 54 | -- | 59 | -- | 65
Trust in the Labour Party | 13 | 23 | -- | 36 | -- | 42

(**) Trust in a political party was gauged by asking the question "Would you describe $party as trustworthy or untrustworthy?" where $party took the value of established political parties in turn. The set of parties varied from survey to survey -- it looks like they asked the question for all parties that won seats.

There is no question that this data is frustrating in many ways. All but the 1998 and 2001 data were collected immediately following an election, when election results (and the associated campaigning) is fresh in people's minds.

Some of the missing data is infuriating (why wasn't satisfaction with democracy recorded for 1993?!), but maybe most importantly this snapshot does not include enough years worth of data. I personally crave more data from the pre-1993 era -- 1993 in particular represents a high point of dissatisfaction with FPTP, since the referendum pitting MMP against FPTO was held that year. In particular, I wish available data stretched back to the mid 1970s or earlier, since New Zealand experienced its first of two second-place majorities in 1978.

Accepting these severe limitations, I still believe we can draw a few conclusions from this data. Firstly, public attitudes in certain areas are volatile, and tend to change with current events. Overall satisfaction with democracy tends to go up and down, as do public attitudes towards their MPs.

The responses on other questions may indicate more stable trends. I find it interesting that three of the most clear trends have to do with questions related to government responsiveness: the "people like me have no say", "politicians do not care what people like me think" and "government is run by a few big interests" questions. I would like to assign credit for these improved attitudes to MMP, but there are many other possible explanations: the Labour government that won the election in 1999 remained fairly popular throughout the polling period (and won re-election in 2005); and the economy was pretty good throughout this period. When governments face fiscal crises and have to cut services and jobs, I am willing to bet that citizen impressions of government responsiveness declines dramatically.

If nothing else, we might say that citizen satisfaction with democracy is improving despite MMP. However, I would like to believe that MMP has contributed to these improved perceptions, even as citizens dislike MMP by name.

The 2001 review suggests that citizens do acknowledge that proportional representation leads to more responsive governments, and although limited the table above does suggest some upward trend.

Scotland and Wales

Determining how happy the Scots and Welsh have been with their regional voting systems has been tough. Scotland and Wales are not considered their own countries, so they are not included in international survey data. Furthermore, Scotland and Wales did not move from FPTP to MMP; they introduced MMP right off the bat with their regionalized government.

However, some survey data is available, and both the Scots and Welsh have familiarity with different kinds of voting -- both locations suffer from several levels of government elected in different ways. For example, Scotland uses STV for municipal elections, MMP for Scottish Parliament, FPTP for the UK parliament, and list PR for European Union elections.

This gives us a rough means of gauging how much the Scots and Welsh approve of MMP as compared to FPTP: given their experience with proportional representation, ask the people whether they would switch other institutions over. The Arbuthnott Commission reported on this issue using data from the 2003 "Scottish Social Attitudes" survey. The Scots were asked whether the UK should introduce proportional representation for the UK House of Commons, which uses FPTP now. Over 50% (it's hard to tell from the provided graph, but it looks like about 53%) agreed that the UK house should move to some form of PR. A little less than 30% felt neutral, and only 10% strongly disagreed. (Arbuthnott 2006, p. 7) None of this directly comments on Scottish views of MMP, but if Scots were deeply unhappy with their proportional voting system you might expect the percentage of people resisting change for the UK house to be higher.

The Richard Commission in Wales reviewed MMP for that region. In addition to asking whether the UK parliament should be elected via PR, it also asked the question for local government and for the National Assembly of Wales itself. Here are the results (Richard Commission 2004, p. 42):

Response | UK Parliament | National Assembly of Wales | Local (municipal) Goverment

Strongly Agree | 20.2 | 19.1 | 18.6

Agree | 33.6 | 39.4 | 38.4

Strongly Agree + Agree | 53.8 | 58.5 | 57.0

Neither Agree nor Disagree | 31.0 | 27.8 | 28.1

Strongly Disagree + Disagree | 15.2 | 13.7 | 14.8

Disagree | 10.7 | 9.7 | 10.8

Strongly Disagree | 4.5 | 4.0 | 4.0

The trend is pretty clear: the Welsh appear to have a good deal of soft support (and some strong support) for PR across the board. To the degree that the Welsh are aware of their MMP system (which, as I have noted before [URL], appears to be limited) it seems they are reasonably happy with proportional representation.

My thinking is that if the Welsh and Scots were as acutely aware of their voting system as New Zealanders appear to be, these numbers could be drastically different. But as it stands these two countries demonstrate that MMP needn't be a public relations disaster.

Data Sources

Finding good data about public attitudes towards democratic institutions has proven to be surprisingly difficult (and I would appreciate any pointers you have).

The good news is that political scientists do care about studying voter attitudes towards politics. The bad news is that data for the countries I care about tends to be missing and inaccessible.

Many countries participate in a standard survey, and those results are supposed to be compiled into a dataset called the Comparative Study of Electoral Systems (CSES). As it appears some raw data in this set can be had free of charge (albeit with registration), a responsible blogger would have downloaded the data, figured out how to read the SPSS format, and run some analysis on long-term trends of voter satisfaction in FPTP vs MMP systems. Unfortunately, in addition to being irresponsible I am also incompetent, and I doubt I will get around to this task by the end of the campaign. However, if somebody else wants to carry out some of this analysis I think it could be useful and interesting.

One rationalization I used for my laziness is that the Canadian component of the CSES data seems to be poor. The only survey that appears to be present is from 1997, which is pathetic given that it looks like the Canadian group has been collecting data after every federal election. Unfortunately, I have not been able to find downloadable datasets from the Canadian group online, although it might be available on request.

The New Zealand Electoral Studies (NZES) folks publish their data, but I did not look at it directly. Rather, I have relied on several academic papers that summarize the data nicely. I sincerely hope that similar data has been collected in Ontario; if we do adopt MMP I expect there will be a rich publishing niche in comparing pre-MMP attitudes to post-MMP ones.

For Scotland and Wales available data is even more piecemeal. There is a set of surveys known as the "Scottish Social Attitudes" surveys which apparently asks some questions about citizen attitudes towards democracy. For Wales the Richard report lists four possible survey sources: a 1997 Welsh Referendum Study, a 1999 Welsh Election Study, a 2001 Wales Life and Times Study, and a 2003 Wales Life and Times study. I have not tracked any of these sources down. It looks like they may be summarized in a paper by Scully and Wyn Jones (Scully Wyn Jones 2003).

Overall, I have found the UK Electoral Commission to be a good source of information.


(Arbuthnott 2006) Commission on Boundary Differences and Voting Systems (chair: Sir John Arbuthnott). Putting Citizens First: Boundaries, Voting and Representation in Scotland. Edinburgh: The Stationery Office. ISBN 0-10-888179-2. (Gah! It looks like they took the website down!)

(Bengtsson 2005) Asa Bengtsson. "Inclusiveness or Accountability: What brings about the most responsive system?" Prepared for the Nordic Political Science Association (NOPSA) conference, Aug 11-13 2005, Reykjavik, Iceland.

(Blais Bodet 2005) André Blais and Marc André Bodet. "Does Proportional Representation Foster Closer Congruence Between Citizens and Policymakers?", Comparative Political Studies, vol 39 no 10, p. 1243-1262.

(Karp Bowler 2001) Jeffery A. Karp and Shaun Bowler. "Coalition government and satisfaction with democracy: An analysis of New Zealand's reaction to proportional representation. European Journal of Political Research, no 40, p. 57-79, 2001.

(MMP Review Committee 2001) Rt. Hon Jonathan Hunt, chair. Inquiry into the Review of MMP: Report of the MMP Review Committee, New Zealand House of Representatives, August 2001.

(Nagel 1994) Jack H. Nagel. "What Political Scientists Can Learn from the 1993 Electoral Reform in New Zealand", PS: Political Science and Politics, vol 27, no 3, Sept 1994, p. 525-529.

(Nicholle 2006) Brian Nicholle. "Brian Nicholle: Put MMP to the Vote". New Zealand Herald, October 12 2006.

(Richard Commission 2004) Report of the Richard Commission: Commission on the Powers and Electoral Arrangements of the National Assembly for Wales. Spring 2004. Available from uk

(Scully Wyn Jones 2003) Scully, R and Wyn Jones, R. Public Opnions, the National Assembly and devolution -- briefing note on the latest evidence. Dept of International Politics, Aberystwyth 2003.

(Vowles Banducci Karp 2006) Jack Vowles, Susan A. Banducci, Jefferey A. Karp. "Forecasting and Evaluating the Consequences of Electoral Change in New Zealand", Acta Politica, vol 41, 2006, pp. 267-284. Available from NZES website:

Livejournal URL:

Mood: burned out