The Green Threat (Part 3): Consequences for this Election

In my last two posts, I looked at what a Green surge would mean for the NDP in two scenarios. One where the NDP does as well as they were in April 18th polls, and a second where they did as poorly as in the 2013 election. Both analyses suggested that a Green surge would have a limited impact on the outcomes of the election. The NDP would lose some seats on Vancouver Island, but unless the vast majority of the increase in the Green vote comes as a result of NDP defections, in neither scenario do the Liberals pick up a substantial number of seats. This is not to say that a Green surge cannot contribute to a Liberal victory, but rather, that it will only do so if the election is already very close. Potential strategic voters should be aware of this when deciding how to approach this election.

The first danger that the NDP faces is that the Green party will win enough seats to deny them a majority in spite of a win over the Liberals. The previous posts suggest that this is unlikely to happen. Neither suggests that the Green party is likely to take more the 10 seats (all on Vancouver Island from the NDP). Indeed, most election forecasters would likely be surprised if the Green party managed to take that many. Yet, only three times since 1945 (1952, 1979, and 1996) has the difference between first and second in a BC election been less than 10 seats. One of those elections, in 1952, used a different electoral system and another, in 1996, saw the party with the second most votes win the most seats. BC’s first past the post system tends to give whichever party wins the plurality of the vote a healthy advantage over the second place party. If the NDP defeats the Liberals by a substantial margin, it is unlikely that loosing a handful of seats on Vancouver Island will prevent them from forming a government.

It is also worth considering the danger that voters switching to the Green party will allow the Liberal party to win seats by splitting the anti-Liberal vote. The last two posts suggest that, while this certainly matters in some ridings, defections from the NDP to Greens are unlikely to affect the outcome of most ridings. Where the Greens are strongest, on Vancouver Island, they tend to be stronger than the Liberal party. In most ridings on the Island, defections from the NDP to the Greens are more likely to a Green victory than a Liberal one. Many seats off the Island are either safe Liberal or NDP seats. In safe seats, the gap between the two parties is too large for a surge in Green support to take enough votes from the NDP to give the Liberals the riding.

It is important to note that in highly competitive seats, particularly those that the NDP are trying to take from the Liberals, a Green surge may end up costing the NDP seats. There is thus some merit to anti-Liberal strategic voting in some parts of the province. Anti-Liberal voters in ridings that have historically been close, or where it looks like a New Democrat might be able to unseat a Liberal, should consider backing the NDP. It is unlikely, however, that most voters live in those kinds of close ridings. Anti-Liberal voters in safe ridings should not be concerned that a vote for the Green party will hand the Liberals an extra seat.

There are two important points from previous posts that should be highlighted. The first is that the election needs to be very close for the growth in Green support to have an impact on the result. It is very possible that it will be that close. The NDP have only won the popular vote in the province twice (in 1996 they won a third election despite losing the popular vote). Even though they are leading in the polls, the NDP has to deal with the fact that they are fighting an election in a province where historically most voters have supported other parties. This does not mean that one should discount the likelihood of NDP victory, but one should also not be surprised if the election ends up being closer than the polls currently predict.

The second important point is that strategic voting is a lot less likely to have an impact on the election than is often assumed. In certain, very competitive ridings, strategic voting can affect which party wins. In a large number of ridings, however, the margins between the two strongest parties are simply to big for strategic voting to change anything. In these ridings, anti-Liberal voters ought to back the party they prefer regardless of which one is stronger. They should pay careful attention to the competitiveness of their ridings when deciding whether or not to vote strategically.


The Green Threat (Part 2): A Green Surge Would Have Changed Little in the 2013 Election

In my last post, I looked at how the rise of the BC Greens might affect the NDP if they manage to hold on to their current levels of support. Polls in British Columbia are not always as reliable as one would hope, and with the main leadership debate still to come, there is plenty of time for the polls to change. In this post, I will look at what would have happened if a Green surge similar to the one happening now occurred in 2013. Where in my last post I looked at what would have happened had the parties had their current poll numbers, in this post I look at what would have happened had the Liberals and NDP won vote shares similar to their 2013 numbers. Together, these posts provide a sense of the possible outcomes of two different scenarios. This one provides a look at a case where the Liberals are ahead by a substantial margin and the last post a scenario where the NDP has a slight lead. A Green surge in the last election would have changed very little.

To look at the impact a rise in the Green vote I use April 18, 2017 polling data from CBC’s poll tracker to estimate party’s vote shares in each riding. Rather than estimating the support for the Liberals and NDP though, I take their 2013 results as is. I consider two scenarios when reducing Liberal and NDP support to account for Green growth. First, I consider a scenario where the Greens get 60% of their increase from the NDP and 40% of their increase from the Liberals. Second, I consider a scenario where the Greens get 80% of their increase in support from the NDP and 20% from the Liberals. This takes into account two different cases, one where the Greens draw only a slightly higher share of their increase from traditional NDP supporters and a second where the overwhelming majority of Green growth comes at the expense of the NDP. As in the previous post, it is important to note that these are crude estimates. They are designed to demonstrate the potential impact of a rise in Green support on the other parties, not to give a definitive account of which parties would have won which exact seats.

In the first scenario, the effect of the rise in Green support is small and largely confined to Vancouver Island. When only 60% of the growth in Green support comes at the expense of the NDP, only 5 seats change hands. Only one of those, in Burnaby/Coquitlam, is not on the Island. In this scenario, there are also 8 seats whose competitiveness changes. These seats either go from being safe seats (the winner no longer has a 5 percentage point lead over the second place finisher) to competitive ones, or from competitive to safe seats. All of these seats are on the Island (the riding in Burnaby Coquitlam was competitive both before and after the rise in Green support was estimated).

Seat Changes Due to Growth in Green Support (2013 Result)

The impact of the Greens is more significant in the scenario where they draw 80% of their increase in support from the NDP. Under these circumstances, 9 seats (or 11% of the BC legislature at the time) would change parties. Again, these changes are largely confined to Vancouver Island. Only two seats off the Island (both in Burnaby/Coquitlam) would have elected a different MLA. A substantial number of seats, 14, would see a change in competitiveness, though most of those would also be on Vancouver Island.

The likelihood that a Green surge would lead to the NDP losing a large number of seats to a Liberal party taking advantage of vote splitting is also limited. Of the 5 seats that the NDP would lose if the Green party got 60% of their increase from the NDP, 4 would be won by the Green party. The Liberals would take only the seat in Burnaby/Coquitlam. If the 80% of the Green rise would have come from the NDP, the Liberals would take an additional two seats. There would also be two tied races, one between the Greens and the NDP, and one between the Greens, Liberals, and NDP. In both scenarios, the Greens and not the Liberals are the main beneficiaries of a rise in Green support.

Gains by Parties Due to Green Growth

Had the Green surge that is happening now occurred in 2013, not much would have changed. Vancouver Island would have elected more Green MLAs, and would have been more competitive, but for the most part, the rest of the province would have elected the same candidates. This is not to say that a rise in Green support cannot work to the advantage of the Liberals, but the Liberals and NDP have to very close in support in order for that to happen. In 2013, the gap between the two major parties was large enough that a rise in Green support would have had a relatively small impact on the election result.


The Green Threat (Part 1): 2017 Polling Numbers suggest the Greens are Unlikely to Prevent an NDP Victory

Polling in the BC election is showing a remarkable increase in support for the provincial Green party.   The party that won less than 9% of the vote in 2013 and 2009 is now polling at around 20% across the province and over 30% on Vancouver Island. The potential for the Greens to significantly increase their vote raises important questions about what effect a Green surge would have on the NDP. If the Greens draw large numbers of left wing voters to them, vote splitting between them and the NDP could allow the Liberal party to pick up significant numbers of seats. It is thus important to examine where the Green party is having success, and test the its potential impact on NDP electoral prospects. In this post, I look at what would have happened in 2013 had polling numbers been where they were April 18, 2017.  If the NDP can establish a solid lead over the Liberals, they should be protected from a Green surge. In a close election, however, a strong Green party is likely to be a problem for the NDP.

To examine the impact of the Green surge I compared aggregate polling data from CBC’s Poll Tracker to the results for the 2013 election. I looked at the numbers the tracker has estimated for three regions (Metro Vancouver, Vancouver Island, and the North/Interior) and compared them to the results the party received in 2013. I then adjusted the riding scores to account for the difference between 2013 and current polling numbers. For example, if a party is polling 5 percentage points above where they were in 2013 in Vancouver, I would add 5 percentage points to the parties’ vote percentage from 2013 in Vancouver ridings. This is a very crude measure of party support in different ridings. Unfortunately, however, without riding level polling, it is impossible to account for how popular opinion is spread across ridings in a particular part of the province. I classified ridings as safe if the gap between the first and second place party was more than 5% and competitive if the gap was less than 5%. For the 2017 numbers (though not for 2013) I also count any riding that looks would switch parties as unsafe, even if it looks like the margin between first and second place is higher than 5 percentage points. I do this because of the uncertainty surrounding polling, I am more confident in the 2013 analysis that uses actual election results. It is finally important to note that there has been redistricting between the 2013 and 2017 elections. This means that predictions for individual ridings for 2013 will not necessarily be accurate for 2017. This analysis can provide a general sense of where the election is competitive, but the exact numbers presented should be taken with a grain of salt.

A look at the 2013 results shows that the Liberals had a large number of safe seats. Of the 49 seats the Liberals won in 2013, 40 were by more than 5 percentage points. In contrast, the NDP won by more than 5 percentage points in only 26 seats, and the two parties were within 5 points of each other in 16 seats. The Greens were largely a non-factor. They won one seat in Victoria by a fairly margin and were competitive with both the Liberals and NDP in another.

2013 Safe Seats by Party

Adjusting to match 2017 polling numbers shows some remarkable changes. The Green party is somewhat more significant. It still has only one safe seat, but the 2017 numbers would have been made the Green party competitive with the NDP in 6 seats and competitive with both the Liberals and NDP in an additional one. The big change, however, is the number of seats in which the Liberals and NDP would have ended up competing against each other. There are 31 seats in which 2017 polling numbers would have put the Liberals and NDP within 5 percentage points of each other or in which the 2017 number would have changed the party that won. The NDP would have only needed to win 14 of these seats (in addition to holding on to their 29 safe seats) in order to win a majority government. In these 31 competitive seats, the NDP would lead the Liberals by an average of 42% to 38% and would have had the highest predicted vote share in 17. These number suggested that if the 2013 election had finished with the numbers the parties are polling at today, the surge in Green support would not have prevented the NDP from winning a majority.

Seats that Would Have Been Competitive with 2017 Numbers

An examination of how this would have broken down regionally demonstrates that the Green threat is generally limited to Vancouver Island. In the Capital Region (Victoria and the surrounding area), the NDP would have had no safe seats, and would have had to compete with the Green party to win each of the 4/5 seats it won in 2013. Off the Island though, the NDP has significant numbers of safe seats in the regions were the Greens are most likely to be competitive in Vancouver and Burnaby. The bigger gains to be made by the NDP are in areas where the Green party is weakest, Northern and Interior British Columbia, and the suburbs around Vancouver.

Seats that Would Have Changed Hands by Region

New Democrats should be concerned by the regional distribution of seats they would need to take from the Liberals in order to compensate for losses to the Greens. They would have to pick up competitive seats in regions such as the North Shore, Richmond/Delta, Langley/Abbotsford, and the Okanagan, where they have been historically weak. The party would also have to do well in Central British Columbia, a region where the Liberals have had an average lead of 8 percentage points over them in elections between 1991 and 2013. Even though aggregated numbers suggest that the NDP would have been able to survive a Green surge, doing so requires the party to have success in areas of the province where they have struggled in the past. As such, the rise of the Greens presents the party with a significant threat.

A Green surge would have a decisive impact on an election in two scenarios, both opf which are ver plausible. The first is a situation in which the NDP and Liberal seat shares end up being quite close. In this case, losses on Vancouver Island may keep the NDP from forming the government. The second, is a scenario in which the Liberals and NDP numbers end up converging, and even the small numbers of voters that the Greens might pull from the NDP in the Interior and suburbs of Vancouver could allow the Liberals to win those seats. In this vein, it is important to note polls have the NDP and Liberals within the margin or error of each other, and that a significant number of the competitive seats are in areas of the province that have been traditionally Liberal. In a close election, a Green surge may prevent an NDP victory, and this very well may be a close election. At the same time, if the NDP can win with a 4 or 5 percentage point lead over the Liberals, they should be able to protect themselves from losing as a result of vote splitting with the Greens.


Who Speaks For British Columbians? First Past the Post Leads to Poor Representation for Interior BC

I my last post I argued British Columbia’s regions are less different politically than they sometimes appear to be. The way that first past the post over-represents parties with a plurality of the vote in a particular region can overstate the extent to which the region backs a particular party. The Liberals, for example, appear stronger than they are in Interior BC, while the NDP appears to be stronger than it is on Vancouver Island. This creates problems with respect to representation in the legislature. The extent to which first past the post increases the regional seat share of a party with a plurality of the vote in a region leaves significant numbers of voters poorly represented. In this post, I look at which voters have been under-represented in BC elections. To do so I group ridings into the same regions as I did in my last post, and look at elections between 1991 and 2013.

A look at the 2013 election shows that parts of Interior BC, some of the suburbs around Vancouver, and Vancouver Island are poorly represented in the legislature. The graph below shows the difference between a party’s seat percentage and vote percentage across the different regions. A positive score means a party has a greater share of seats than its share of the vote and a negative score means the opposite is the case. The Liberal party is significantly over-represented in the Okanagan, Central BC, Langley/Abbotsford, Richmond/Delta, and the North Shore. In all of these regions, except for Richmond/Delta, the Liberals’ seat share is over 45 percentage points higher than their vote share. The NDP is over-represented in the Capital Region (in and around Victoria) and on the rest of Vancouver Island. In and around Victoria the NDP’s seat share was 40 percentage points higher than its vote share and on the rest of the Island it was 31 points higher.

2013 Disproportionality By Region

A look at average disproportionality across all elections between 1991-2013 shows a similar pattern. While the disproportionality is not as great as it was in 2013, the Liberals have significantly more seats than their vote share in the Okanagan, Langley/Abbotsford, Richmond/Delta, and the North Shore. The NDP, as in 2013, is over-represented in the Capital Region and on the rest of Vancouver Island. This suggests that the lack of representation that occurred in 2013 is persistent over time. It is not the case that in some years some voters lack legislative representation while in others different voters do. Rather, New Democrats in the Okanagan and some of the Vancouver suburbs are consistently under-represented. This is also the case for Liberals on Vancouver Island.

Average Disproportionality by Region

A look at 2013 NDP and Liberal vote and seat shares underscores how this regional over-representation can be problematic. In 2013, the NDP did not win a single seat in the Okanagan, Central BC, Langley/Abbotsford, Richmond/Delta, or the North Shore. This happened despite the fact that they won a quarter to a third of the vote in each of those regions. This meant that a quarter to a third of voters in each of those regions went un-represented in the last legislative session. The same is true for the same is true for the just over a quarter of voters in Capital Region that voted Liberal yet failed elect a single Liberal MLA.

2013 NDP Seat and Vote Percentage

2013 Liberal Seat and Vote Percentages

This has important implications for the way British Columbians are represented. A New Democrat in Vancouver or Victoria is not going to have the same political interests as a New Democrat in Interior British Columbia. Vancouverites and Victorians, for example, are likely to be more concerned about urban issues, such as transit policy, while Interior British Columbians are likely to more concerned with other issues, such as those surrounding resource extraction and logging. The same is true for Liberals. Liberals in Victoria are not going to be well represented by MLAs from the interior. At the same time, the fact that individuals are voting for a different party suggests that they will not be well represented by the MLAs elected in their constituencies. The result is a situation where significant numbers of British Columbians have no one to speak on their behalf in the legislature.

This is a fundamental problem with first past the post electoral systems. Because only one party can win a seat, there will always be groups of voters whose votes do not contribute to electing an MLA. In regions where a party is particular strong, there is always the danger that a party will be shut out and that voters backing that party will have no one in the legislature to represent their interests.

*Election data comes from BC Pundits Guide.


Just How Divided is BC: Differences Between Regions are not as Large as they Sometimes Appear

With an election in British Columbia provincial election taking place this May, it is worth looking at some of the historic voting patterns in the province. There is a tendency in BC politics to assume that Vancouver Island is largely New Democrat, the interior of the province is Liberal, and the Vancouver region is split between the two parties. The geographic distribution of the parties’ seats generally supports this. BC’s first past the post electoral system, however, tends to produce seat shares that exaggerate regional divides in the province. An examination of the vote shares parties win across the province suggests that different regions are more similar than it sometimes appears. It also demonstrates that the Liberals are strongest not in the Interior, but rather in the municipalities surrounding Vancouver such as Langley, Richmond, and North Vancouver.

To get at the vote support for the Liberals and NDP in different parts of the province, I divided it into 12 regions. In the Interior there is North BC, South East BC, the Okanagan, and Central BC. The Greater Vancouver area is split into Langley/Abbotsford, Surrey, Richmond/Delta, Burnaby/Coquitlam, Vancouver proper, and the North Shore. Finally, Vancouver Island is divided into the Capital Region, including Victoria and the surrounding municipalities, and the Mid/North Island. A list of which ridings are placed in which region is provided in the document below. In my analysis, I look at the results from elections between 1991-2013.

Ridings in Each Region

A look at the vote distribution of the 2013 election shows that, while there are regional discrepancies in parties’ vote shares, there are substantial numbers of Liberals and New Democrats in every part of the province. In no region did the Liberals take much more than half of the vote. The parties’ largest vote share was in the North Shore (including North Vancouver and West Vancouver), where it won just under 54%. The Liberals won over a quarter of the vote in every region of the province, and over a third in every region except for the capital region. Similarly, the NDP vote is reasonably evenly distributed across the province. The party won at least a quarter of the vote in every region and did not win more than half of the vote in any region.

2013 Vote by Region

If one looks at parties’ average vote shares in all elections between 1991 and 2013, a similar trend comes through. In only Langley Abbotsford, Richmond Delta, and the North Shore do the Liberals average more than half of the vote. In their weakest regions in the Capital Region and on the rest of Vancouver Island the party still averages just under 40% of the vote. The NDP has more weak regions than the Liberals, in part because its average vote share is lower across the province. Still, in its weakest regions, Langley Abbotsford and the North Shore, it still averages a quarter of the vote. In its strongest regions, (Burnaby, Vancouver, and the two Vancouver Island regions) the party averages around just 44%.

Average Vote Share by Region

A comparison of the Liberal’s and NDP’s vote share in regions to their overall provincial vote share can demonstrate where parties are particularly strong and weak. The two graphs below show these differences for the 2013 election. Positive scores show regions where a party is disproportionately strong while negative scores show regions where it is weak. The 2013 scores for the Liberals show a party that is strong in parts of the Interior, Langley/Abbotsford, and the North Shore. The Liberals’ success in the Interior, however, was less universal than is sometimes assumed. While the party was disproportionately successful in Northern BC and the Okanagan, it only outperformed its provincial vote share by slightly over one percentage point in South Eastern BC, and by only three percentage points in Central BC.

2013 Difference Between Liberal Regional and Popular Vote

In 2013, the NDP did not underperform in the Interior as much as one might think. Outside of the Okanagan, its Interior vote share was within 4 percentage points of its overall vote share. Where the party struggled was in some of the suburbs outside of Vancouver. The NDP significantly underperformed in Langley/Abbotsford, Richmond/Delta, and on the North Shore. As expected, the party exceeded their vote share in Vancouver (the fact that both the Liberals and NDP over-performed in Vancouver suggests that the Conservatives and possibly the Greens underperformed) and on Vancouver Island.

2013 Difference Between NDP Regional and Popular Vote

The two graphs below show the Liberals’ and NDP’s average performance relative to their total provincial vote share for the 1991-2013 period. The graphs show similar results to the ones that look just at the 2013 election. The exception to this is the Liberal’s scores for interior BC. Outside of the Okanagan, the Liberals have typically equalled or underperformed in the Interior relative to the rest of the province. Some of this underperformance is the result of the presence of Social Credit and Reform candidates in the Interior during the early and mid 1990s. These candidates would have taken support from the Liberals. Like in 2013, the Liberals’ strongest regions are just outside of Vancouver in Langley/Abbotsford, Richmond/Delta, and the North Shore. As in 2013, the NDP does particularly well in Vancouver and on Vancouver Island and particularly poorly in several of the suburban regions surrounding Vancouver.

Average Relative Liberal Success in Regions

Average Relative NDP Success in Regions

A look at the vote share of the Liberals and NDP demonstrates two important things. First, that there are substantial groups of Liberal and NDP supporters in almost every region in the province. Neither party is so dominant that it makes up more than slightly over half the voters in a particular region, and nor is either party so weak that it makes up less than a quarter of voters in any particular region. To a large degree, the way that the first past the post system turns votes into seats overstates the political differences across different regions. Second, it shows that the Liberals’ strongest region is not the Interior, but rather he municipalities just outside of Vancouver. It is in Langley/Abbotsford, Richmond/Delta, and on the North Shore that Liberal most outperforms its provincial vote share.

*Election data comes from BC Pundits Guide.