Stephen Harper's polling picture bleak, but history offers him hope

Over the past 35 years, only two incumbent federal governments have polled lower than Stephen Harper's Conservatives with just 12 months to go before the next general election.

One of those met disaster in 1993, when Kim Campbell's Progressive Conservatives were reduced to just a couple of seats in Canada's greatest electoral defeat in history.

The other? Brian Mulroney's 1988 majority win.

But history suggests the prime minister faces an uphill battle as the next election approaches.

For the purposes of this analysis, I have looked at polls carried out between 12 and 14 months before a federal election going back to 1979. These polls have been averaged where the results of multiple surveys have been found.

Eleven federal elections have been held since 1979. The Liberals were the incumbents in six of those elections, while the PCs or Conservatives were the incumbents in five.

The Conservatives are currently averaging just under 31 per cent support in polls conducted over the last two months, which compares poorly with recent history. Incumbent governments have averaged about 37 per cent support one year out from the end of their terms, and only twice has a governing party polled significantly lower than the Conservatives' current level of support.

Opposite lessons can be drawn from those two precedents.

The first was in 1987. Brian Mulroney's PCs were polling at just 27 per cent a year before the 1988 election, putting them about 10 percentage points behind John Turner's Liberals and in third place behind Ed Broadbent's New Democrats. A year later, however, Mulroney took 43 per cent of the vote and secured another majority government.

The second example is far more ominous for the current government. The PCs were polling at just 15 per cent in August to October 1992, putting them in a tie with the upstart Reform Party for third place. Again the NDP was polling in second, with the Liberals well ahead. The New Democrats dropped significantly by the time of the 1993 election, while the PCs did not recover and took just 16 per cent of the vote in their disastrous campaign. Jean Chrétien's Liberals cruised to an easy victory.

However, Conservative incumbents have generally polled lower than Liberal governments at this point in their mandates.

The last five Conservative incumbents were polling at around 30 per cent with a year to go, or 33 per cent if we exclude the exceptional case of 1992. In that regard, Harper's Conservatives are performing comparably well. But the Tories were at 35 per cent a year before their 2008 victory and at 33 per cent a year before their 2011 win, so by that measure the party is slipping.

This is particularly problematic when compared to the state of current Liberal support. Justin Trudeau's party has averaged a little under 39 per cent support over the last two months. Only twice before has an Opposition party polled more highly a year before the next election. The first was in 1979, which is complicated by the fact that the 1979 and 1980 elections were held less than a year apart. But the second was in 1983. At the time, Mulroney's Opposition PCs were one year away from their inaugural 1984 victory.

Chrétien's Liberals were polling at a similar level of support a year before the 1993 federal vote.

What a difference a year makes

Much can happen in the 12 months before a federal election. On average, support for the two main parties has shifted plus or minus 6.2 percentage points between the polls and the results in the year before a vote.

Incumbents have generally performed worse, dropping an average of 2.1 points in the last year of their mandate. The main opposition party has, by contrast, gained an average of 3.4 points.

But there has been a marked difference between incumbent Liberal and Conservative governments. On average, PC or Conservative governments have gained 4.5 points in the last year of their mandate while the Liberals in Opposition have lost just under a point. Liberal incumbent governments, on the other hand, dropped an average of 7.5 points, with Conservative Opposition parties gaining 4.4 points.

So we should probably expect a fair amount of change in voting intentions over the next 12 months.

Over the last 11 federal elections, polls one year out identified the eventual winner eight times. In three of those cases, there were some significant changes, just not enough to make a difference in the winner or whether or not it would be a majority government. In two cases, there were no significant shifts whatsoever.

Recent historical precedent, then, could be used to argue for or against a Conservative victory in 2015.

On election night, the Conservatives and PCs have tended to outperform their polling levels one year earlier by a sizable amount.

But overall, incumbent governments have suffered at the ballot box compared to where they stood 12 months before, Harper's Conservatives are polling lower than most incumbents at this stage of their mandate and polls have generally not shifted enough to flip the identity of the eventual winner this far out from an election.

The weight of evidence comes down against a Conservative re-election in 2015, but the case is far from conclusive.

Note: This article reviews trends in national public opinion surveys. Methodology, sample size and margin of error if one can be stated vary from survey to survey and have not been individually verified. Read more polling analysis at threehundredeight.com.