Friday, November 16, 2018

British Columbia's Electoral Reform Referendum

The first thing that needs to be said about this is BC may well vote to change its electoral system. Much of the polling to date has been favourable to electoral reform although the margin has been tightening and the most recent survey has those favouring the existing first-past-the-post system ahead by one point 50.5% to 49.5% - effectively a tie.

The question on the ballot is in two parts. The second part asks the electorate to choose among three options. When I looked at the choices it struck me that of three there is only one realistic choice and that is Option Two.  Here is my analysis of the three options.

Option One is called dual member proportion, explained by Elections BC here. One key aspect of the proposal that is also its greatest weakness is that parties would normally be expected to nominate a primary candidate and a secondary candidate in a series of two member districts. Without going into the details the result is proportional (indeed I find the design rather ingenious), but we have a political culture that would not at all understand the idea of primary and secondary candidates. The system is new and is not in use anywhere, so it is highly likely to have some unforeseen consequences.

Option Two, which I would support if I had a vote, is Mixed Member Proportional (the Elections BC page on it is here). In this system you vote for a local candidate in a local constituency and that constitutes about half the legislature. Remaining members are chosen from lists in regional districts on the basis of overall province-wide proportionality. This system is in place in many countries around the world, notably Germany and New Zealand, but also in sub-national parliaments such as Scotland. There is deep experience with it and it is probably the system that best reflects generalizations made about PR by political scientists such as Arend Lijphart:
"The conventional wisdom concerning the choice between majoritarian electoral systems and proportional representation (PR) – and, more broadly, between majoritarian and consensus forms of democracy – is that there is a trade-off: PR and consensus democracy provide more accurate representation and better minority representation, but majoritarianism provides more effective government. A comparative analysis of 18 older and well-established democracies, most of which are European democracies, shows that PR and consensus democracy indeed give superior political representation, but that majoritarian systems do not perform better in maintaining public order and managing the economy, and hence that the over-all performance of consensus democracy is superior."
A recent poll by Mainstreet Research reports that this is the preferred option in BC.

Option Three, which Elections BC labels as Rural-Urban Proportional actually would deploy two systems, the Single Transferable Vote in multi-member urban districts, while Mixed Member Proportional as described above would be used in rural areas. This makes little sense to me and it baffles me to see it as an option. The Single Transferable Vote requires voters to rank their choices, 1, 2, 3 etc. It then requires a complex, indeed an opaque system for counting the votes. It has been used in the past in a number of Canadian jurisdictions but no longer. In my view it is not strictly proportional (political scientist Rein Taagepera has demonstrated how it can actually produce a non-proportional outcome) but more often than not, and certainly more than FPTP, it generally does produce proportional outcomes. Apart from the complexity a new system will be seen as more legitimate if it is the same everywhere, a virtue that characterizes the first two options. However, it was recommended by a Citizen's Assembly in BC in 2005 and supported by a majority in a subsequent referendum but failed as the government of the day required a super-majority of  60% and the YES vote obtained was 57.7%, so it fell just short. This would explain its presence on the ballot here, although I think it would make more sense as a province-wide choice.

If one views this period of Canadian history as a whole one can see numerous attempts at fundamental electoral system reform.  Even if this one fails it appears increasingly likely we will see successful electoral reform in the near future. In Quebec the newly elected Coalition Avenir Québec government has promised a new electoral system by the time of the next election without a referendum.



Tuesday, November 13, 2018

Proportional representation - Arguments in favour based on the results of the 1993 election.

This is an essay I wrote in 1997 titled Ten Arguments for Proportional Representation. I used the 1993 federal election as a starting point. Some of the discussion is clearly dated but the extent of vote distortion in 1993 makes it an ideal candidate for analysis of this kind.




The tables above illustrate the difference between the results of the 1993 federal election and what the House of Commons might have looked like had it been elected by proportional representation.

[The PR table distributes seats based on provincial vote shares from the 1993 Election using the St. Lague method of allocating seats.  It uses a 5% threshold as the minimum requirement for parties to gain seats in the House of Commons.  The Yukon and the N.W.T. results are combined for the purposes of this exercise as PR ordinarily requires that a larger number of members be elected from a single district to achieve proportionality.  Provinces are used for the calculation here with the exception just noted, but most PR systems use multi-member local constituencies.]

I will use this illustration to advance the following ten arguments in favour of  proportional representation, some of them specific to the situation which emerged following the 1993 election:

1. The most compelling argument is that Proportional Representation (PR) more faithfully adheres to the democratic ideal. Unlike our current first past the post majoritarian system, PR would elect a House of Commons which accurately reflects the real preferences of voters.  Each party’s share of House of Commons seats would roughly approximate their share of votes. As the size of a party’s contingent of MP’s depends on its total vote, each voter can take comfort in the knowledge that their ballot contributed in some measure to that result. Currently many voters see their votes “wasted” on individual candidates who simply lose. In a PR election, in a very real sense, every vote counts. The sole exception is votes cast for parties who fail to achieve a minimum threshold. Such minimum levels of support (5%, for example) are common in countries with PR. As a consequence of a greater sense of efficacy on the part of individual voters, countries with PR have on average higher voter turnouts - about 10 percentage points more.

2. PR always produces a more balanced outcome.  What astonished many about the 1993 election were the extraordinarily lopsided results - 98 of 99 seats in Ontario going to the Liberals being the example that springs first to mind.  In fact, the outcome was lopsided in most provinces and Liberals weren’t the only beneficiaries. Reform won 75% of the seats in B.C. with 36% of the vote while the Bloc won 72% of the ridings in Quebec with 49% of the vote.  Even the NDP managed to win 36% of the seats in Saskatchewan with just 27% of the votes.  Only the Tories lost out everywhere (having benefited from the system in 1988).  The 1993 election was probably somewhat extreme in the lack of symmetry between the distribution of votes and constituencies won but such outcomes are characteristic of our current electoral system, and by definition impossible under PR.

3. The Bloc would have been denied its status as official opposition. The tradition of the parliamentary system is that the party with the second largest number of seats becomes the official opposition.  That it should fall to the BQ after 1993 seemed both inappropriate and bizarre. The Bloc was a purely regional party whose raison d’être was to exit the institution to which it had just been elected following what it expected to be a YES vote in a Quebec referendum. If PR had been in place not only would Reform have assumed its rightful place as official opposition but there would have been another equally important outcome in Quebec itself. Under PR the Bloc would have been entitled to 38 MP’s, only half of Quebec’s ridings - a more realistic reflection of Quebec opinion. The Liberal party did elect some francophone federalist MP’s but many were elected in ridings with large English speaking or multi-ethnic populations. Under PR francophone federalists would occupy a number of seats commensurate with their real numbers in Quebec society - around 40 per cent. Note that some would likely have been among the 11 Tory MP’s that would have been elected in Quebec.

4.  The Progressive Conservatives and the NDP would have retained party status and held on to a significant number of seats in the House of Commons. The PC’s and the NDP between them commanded the support of over 3 million Canadians in 1993 but won a total of just 10 seats. Immediately following the 1993 election the media all but ignored both parties because reporters naturally equated House of Commons representation with the real preferences of Canadians. Under PR they would have continued to enjoy the status in Canada’s political scene to which their millions of voters were entitled.

5. The real political diversity of Canada that exists within the various regions would be reflected in political representation in Ottawa. Regional tensions are exacerbated when the first past the post system systematically elects full slates of one party.  For example, under PR Reform would have had more members from Ontario than Alberta or B.C., and even two MP’s from Atlantic Canada. The Conservatives, all but shut out in 1993, would have had representation from every province.  The House of Commons would lose its appearance of being divided between a Liberal party with a largely Ontario face, a western Reform party and a Bloc Quebecois apparently representing most of francophone Quebec.

6. PR would encourage the formation of new parties and groupings that represent real social trends in Canada and help them get into Parliament.  If European experience is any indicator, it is likely that the environmentalist Green party would win seats at least in B.C. if not elsewhere. Many Canadians experience a sense of alienation from the political system that is caused in part because the electoral system rewards older well-established political parties and punishes upstarts. The only exceptions historically have been regionally rooted parties such as the Bloc and Reform.

7. PR allows for a clearer expression of real ideological differences among Canadians. The Winds of Change Conference held in Calgary in 1996 represented a major effort to bridge the differences between the Progressive Conservatives and the Reform Party. It collapsed partly because both sides have an interest, in the context of the current system, in driving the other out of business. As well, despite their similarities, there are genuine differences between the two parties. For example, Reform and the PC’s have radically different approaches to national unity. Under PR they would have the opportunity to coalesce around issues, and perhaps someday to form a government, while retaining the integrity of their philosophical outlooks. It was dissatisfaction on the part of some small ‘c’ conservatives with the Mulroney government which led to the creation of the Reform party. Under PR, this diversity among conservatives could be reflected openly in political discourse. Under the current system, the alternative to perpetual opposition appears to be that the two parties will be forced to amalgamate. [Note: as it turned out coalition of the two parties occurred via what amounted to a takeover of the PCs by Reform with the new party adopting the name Conservative.] But this is likely to lead to an ongoing war between right and left for the surviving party’s soul. Under PR both could continue to reflect the divergent views that led to the current situation in the first place.

8. PR would diminish the opportunities for arrogant or capricious governance. Our majoritarian government vests too much power in the leader of the party which wins a majority in an election. The Prime Minister has the power to hire and fire and cabinet ministers (and even to appoint candidates in local constituencies).  The greatest difference that PR would have made in 1993 is that the Liberals would not have formed a majority government.  In fact, majority governments would become rare under PR.  This inevitably forces inter-party bargaining and compromise and power becomes more diffuse. No single individual can dominate.

9. PR systems better represent minorities and other under-represented groups such as women. Political Scientist Arend Lijphart found that PR parliaments have about four times as many women as non-PR legislatures. PR could provide better representation of the different strands of an increasingly diverse Canadian society and so help ease social tensions.

10. Last on this list of arguments but of enormous importance in a country with a history of constitutional paralysis: the political system can be reformed and improved by Parliament acting alone. After trying out proportional representation and finding it not to their liking, Canadians can, if they so desire, restore our old way of doing things. Unlike constitutional change, legislated institutional change can be used to experiment with a better way of doing things.

In my next post I will discuss the options facing citizens of BC in their current proportional representation referendum.

Monday, November 05, 2018

U.S. Midterms

If entirely conventional benchmarks were what mattered in the U.S. elections on November 6 the current strength of the economy - essentially the pinnacle of the growth cycle that started with Barack Obama's stimulus bill in 2009 - then the Republicans ought to be doing extremely well. In 1998 in the midst of another period of strong economic growth - the roaring nineties as Joseph Stiglitz dubbed them - the Democrats made gains in the House of Representatives despite the impeachment proceedings against Bill Clinton. This happened although off year elections following a presidential re-election are generally likely to damage the incumbent party. Credit may be due Obama for today's economic circumstances, but it is the current incumbent who gets credit or blame for the economy in U.S. politics.

However, Trump has managed to alienate large swathes of the U.S. as he essentially doubles down on the extreme white racism and misogyny that has been hallmark of his administration. It is a source of strength in rural and small town America but a source of weakness in the parts of the U.S. that are urban and diverse.

My overall impression is that this racial/ misogynistic extremism will lead to large Republican losses in urban and suburban areas, while Republicans will be either as strong as before or stronger in more rural communities. We saw this a year ago in the statewide election in Virginia where the Democrats won the Governorship by an unexpectedly large margin - the polls were wrong then,  underestimating the Democrat margin. The Democrats made huge gains in the state legislature almost gaining control. However, when I looked closely at the results I saw that the Republicans had held their own in rural areas. This year the Republicans may well gain a few rural seats in the House of Representatives at the Democrats expense, but that will be more than offset by gains in suburban areas.

In the Senate just one third of the institution is being elected, and strong rural support for Trump in places like North Dakota may keep the Senate under Republican control. We won't know what will happen until after the polls close.

The most interesting and significant Senate race is the one in Texas where the highly successful run by Democrat Beto O'Rourke may fall short in its effort to oust the execrable Republican Ted Cruz. However, it matters in the long term. That is because Texas is already similar to California in having a population composed of a majority of minorities. According to 2016 population estimates 44% of the Texas population is white. Republican control of its politics rests on lower turnout among the 37% of the population that is Hispanic. Population trends in California, now a reliably blue state, preceded Texas. In California the white population is 39%, while black, Asian and Hispanic populations total 59% of the population. Texas is headed this way and a key political indicator was the shift in support for the Democratic presidential candidate Hillary Clinton - she shrunk the size of the Democratic loss from 16% for Obama in 2012 to 9% in 2016. O'Rourke has raised a lot of money and is spending on turning out Latinos. There are similar population trends in Arizona and Nevada, two states that may switch from Republican to Democratic Senators this year.

Trump's racism and misogyny is the very opposite of what would serve Republicans in the long-term. According to Pew research there is a large gender gap in party identification with younger women supporting the Democrats by a huge margin. The racial and ethnic shifts in the composition of the American population along with large shifts in party identification and ideology, particularly among younger women, will register in Tuesday's results but the greatest impact of these trends will be felt in the years to come.