OTTAWA -- "Despite Stephen Harper's tough-on-crime rhetoric, the number of police officers in Canada declined every year from 2010 to 2014. Stephen Harper's plan isn't working."

NDP Leader Tom Mulcair made that assertion Wednesday as he promised a New Democrat government would work with provinces, territories, municipalities and First Nations to provide stable funding to "put 2,500 new officers on the streets and keep them there."

"It's boots on the ground that fight crime, not empty Conservative promises."

How accurate is his assertion?

Spoiler Alert: The Canadian Press Baloney Meter is a dispassionate examination of political statements culminating in a ranking of accuracy on a scale of "no baloney" to "full of baloney."

This one earns a rating of "some baloney." It is partly accurate but important details are missing.

Here's why

It is not strictly accurate to say the number of police officers declined from 2010 to 2014.

According to Statistics Canada, the raw number of police across the country has actually increased every year since Harper became prime minister, except for the last two years.

In 2006, there were 62,461 police officers. That number crept up slowly year by year to 69,539 in 2012. It dropped to 69,250 in 2013 and 68,896 in 2014.

However, Mulcair is correct if increases in the general population are taken into account.

According to Statistics Canada, there were 191.3 officers for every 100,000 Canadians in 2006. That grew steadily to 203.1 officers per 100,000 in 2010.

Since then, the number has dropped each year, hitting a seven-year low of 193.9 officers per 100,000 in 2014.

The experts

Criminologists don't take particular issue with Mulcair's assertion about fewer police officers over the past four years, but they do dispute the assumption behind it: that more police means less crime.

"The baloney is thinking that the number of boots on the ground has anything to do with the crime rate," says Irvin Waller, criminologist at the University of Ottawa.

Since 2010, Waller says the number of police officers "hasn't changed that much" -- less than one per cent. Yet public expenditures on policing have "gone through the roof" -- more than doubling to $13.5 billion since 2000, primarily as a result of salary increases.

At the same time, the rate of police-reported crime has dropped dramatically.

Why the crime rate has dropped is the result of complex factors, including fewer people reporting crimes to the police. But it does not, in Waller's view, reflect the number of cops on the ground or the money spent on them.

Neil Boyd, director of the University of British Columbia's school of criminology, agrees.

"Irrespective of whether we are looking at absolute numbers of police or the rate of police strength over time, (Mulcair's) comment assumes, implicitly, that there is a clear and linear relationship between police strength and crime rates," says Boyd.

"The relationship is actually much more complex, with cities like Toronto, for example, having modest declines in police strength and dramatic drops in most forms of crime."

The verdict

It is not strictly correct to assert that the raw number of police officers in Canada has declined over each of the last four years. It would have been more accurate to say the number of officers per 100,000 Canadians has declined.

Regardless, criminologists say there's no direct correlation between the number of police and the crime rate.

For these reasons, Mulcair's assertion earns a rating of "some baloney."

Methodology

The Baloney Meter is a project of The Canadian Press that examines the level of accuracy in statements made by politicians. Each claim is researched and assigned a rating based on the following scale:

No baloney - the statement is completely accurate

A little baloney - the statement is mostly accurate but more information is required

Some baloney - the statement is partly accurate but important details are missing

A lot of baloney - the statement is mostly inaccurate but contains elements of truth

Full of baloney - the statement is completely inaccurate