Main menu:

History of RPE Thought

Posts by Tag

RSS New from the CCPA

  • A critical look at BC’s new tax breaks and subsidies for LNG May 7, 2019
    The BC government has offered much more to the LNG industry than the previous government. Read the report by senior economist Marc Lee.  
    Canadian Centre for Policy Alternatives
  • The 2019 living wage for Metro Vancouver April 30, 2019
    The 2019 living wage for Metro Vancouver is $19.50/hour. This is the amount needed for a family of four with each of two parents working full-time at this hourly rate to pay for necessities, support the healthy development of their children, escape severe financial stress and participate in the social, civic and cultural lives of […]
    Canadian Centre for Policy Alternatives
  • Time to regulate gas prices in BC and stop industry gouging April 29, 2019
    Drivers in Metro Vancouver are reeling from record high gas prices, and many commentators are blaming taxes. But it’s not taxes causing pain at the pump — it’s industry gouging. Our latest research shows that gas prices have gone up by 55 cents per litre since 2016 — and the vast majority of that increase […]
    Canadian Centre for Policy Alternatives
  • CCPA welcomes Randy Robinson as new Ontario Director March 27, 2019
    The Canadian Centre for Policy Alternatives is pleased to announce the appointment of Randy Robinson as the new Director of our Ontario Office.  Randy’s areas of expertise include public sector finance, the gendered rise of precarious work, neoliberalism, and labour rights. He has extensive experience in communications and research, and has been engaged in Ontario’s […]
    Canadian Centre for Policy Alternatives
  • 2019 Federal Budget Analysis February 27, 2019
    Watch this space for response and analysis of the federal budget from CCPA staff and our Alternative Federal Budget partners. More information will be added as it is available. Commentary and Analysis  Aim high, spend low: Federal budget 2019 by David MacDonald (CCPA) Budget 2019 fiddles while climate crisis looms by Hadrian Mertins-Kirkwood (CCPA) Budget hints at priorities for upcoming […]
    Canadian Centre for Policy Alternatives
Progressive Bloggers

Meta

Recent Blog Posts

Posts by Author

Recent Blog Comments

The Progressive Economics Forum

Ontario Falls Off a Cliff

The Ontario economy fell off a cliff last month as the US meltdown intensfied the already virulent manufacturing and forest jobs crisis.

An almost unprecedented 42,000 manufacturing jobs were lost in November in Ontario alone – that’s one in twenty of the total, and more than the total of manufacturing employment in either Oshawa or Windsor. And 20,000 jobs were lost in other sectors in Ontario. Unemployment rates soared in hard hit industrial communities, particularly Kitchener, Hamilton, St Catharines- Niagara and Oshawa. The overall Ontario unemployment rate leapt from 6.5% to 7.1%.

Nationally, 75,000 employees lost their jobs – led by the loss of manufacturing jobs and 26,000 related transportation and warehousing jobs – and the unemployment rate rose from 6.2% to 6.3%. Unemployment would have risen much more if the labour force had not shrunk as many workers, particularly older workers dropped out. The participation rate fell from 67.9% to 67.6%.

These numbers are at least as bad as those we saw in the brutal 1989 to 1992 manufacturing led recession. With almost 400,000 manufacturing jobs now lost since the peak in 2002 this downturn is likely to be far worse… unless governments respond to this crisis very quickly and in a very determined way.

The numbers also leave little room for doubt that the Canadian economy as a whole has now entered a recession.

Enjoy and share:

Comments

Comment from travis fast
Time: December 5, 2008, 10:57 am

Affirmative on your conclusion. The decline in labor force participation matched by an increase in the unemployment rate really tells the story. The regional break downs out west also show a deterioration in the participation rate with only Manitoba bucking the trend. So we can probably call a peak to the employment rate out West.

My conjecture is that as lotus land land transforms itself into locust land we will perhaps start to see a change in regional attitudes towards stimulus and thus perhaps federal politics.

Comment from Salty Crackers
Time: December 5, 2008, 12:48 pm

I do believe we all have been hoodwinked by Statcan again. For some reason we have three nice months that nothing happens, and then suddenly it all falls off the mountain. Seems like there was some funky number tom foolery going on.

I mean really, numbers do not jump around in such a fashion, you had to be there I guess to realize the process.

It is a travesty that the fundamental mechanisms within our country were massaged in such a manner and those within statcan should be found and held accountable for such behaviour. It is pathetically banana republic like.

Given the behaviour of Mr. Harper lately one can only imagine what kind of deception and subtrefuge was going on behind the scenes to manipulate many things during the last electoin.

I mean he is willing to incite national disunity to save his job, so ordering up some lame assed numbers that we witnessed before the election was well within his tool chest. Spare me with the Statcan god like, untouchable sermons, as that image has been forever tarnished after the last election with the biased GDP and employment numbers that came out of that place.

There is no doubt that we are finally in a recession, officially. and there is no doubt that Mr. Harper has corrupted many delicate national entities to secure his undemocratic hold on power.

ST

Comment from Mound of Sound
Time: December 5, 2008, 12:51 pm

“With almost 400,000 manufacturing jobs now lost since the peak in 2002 this downturn is likely to be far worse… unless governments respond to this crisis very quickly and in a very determined way.”

I guess you didn’t get the memo. Government has been cancelled for two months.

Comment from Stuart Murray
Time: December 5, 2008, 1:41 pm

Odd that manufacturing is collapsing even amidst the plunging dollar. Maybe the plunging dollar was masking a would-be collapse for three months, then when we hit purchasing-power-parity a backlog of shutdowns hit us like a shock.

Is it too cruel to point out that we are still above the Bank of Canada’s NAIRU of 8.0%? As such, things will probably get far worse before everyone agrees that it’s as bad as ’89-92. I agree that this one will be comparable, it’s just that it will take a few months before this becomes a consensus.

The upside is that everyone is scrambling to buy safe bonds, which means that this time around, lenders won’t be able to extort public policy concessions before buying government bonds. Which means that if Bob Rae becomes PM in January and he sticks it to the poor and workers, we will know that he is really just doing this for fun.

Comment from Stuart Murray
Time: December 5, 2008, 1:41 pm

Below NAIRU, rather

Comment from Stephen Gordon
Time: December 5, 2008, 3:04 pm

I do believe we all have been hoodwinked by Statcan again. For some reason we have three nice months that nothing happens, and then suddenly it all falls off the mountain. Seems like there was some funky number tom foolery going on.

I mean really, numbers do not jump around in such a fashion, you had to be there I guess to realize the process.

It is a travesty that the fundamental mechanisms within our country were massaged in such a manner and those within statcan should be found and held accountable for such behaviour. It is pathetically banana republic like.

This advances your cause not at all.

Unless your cause is advancing stupid conspiracy theories.

Comment from Stuart Murray
Time: December 5, 2008, 3:45 pm

I assumed it was intended as irony.

Comment from Andrew Jackson
Time: December 5, 2008, 4:48 pm

Any thoughts of statscan conspiracies should take into account that there is variability in month to month numbers simply due to fact that the series is based on a sample survey. Likely some of these stunning manufacturing job losses in Ontario reflect under reporting in the previous couple of months, and could even be slightly overstated. The basic picture is still prettry grim.

Comment from Travis Fast
Time: December 5, 2008, 6:25 pm

“Unless your cause is advancing stupid conspiracy theories.”

The inverse would be of course that there are smart conspiracy theories.

Comment from Nick Rowe
Time: December 5, 2008, 9:47 pm

Just to put numbers on Andrew Jackson’s comment above, StatsCan reports a standard error (from sampling variance) of 12,500 around its estimate of Ontario manufacturing employment for any month. So I think there’s about a 90% probability that employment (estimated to fall by 42,200) did in fact fall. (Somebody check this; I was never much good at stats.)

Comment from Salty Crackers
Time: December 5, 2008, 9:57 pm

There is a variability undoubtedly, and given the sampling plan and the standard errors, we could see some fairly large swings from one month to the next. However, I find that the way the stats from a couple of key sources have went over the past 4-5 months to me suggests something is a foot. But hey we have already been down that path with nothing but dumb ass comments like the one coming from Mr. Gordon.

It is not a conspiracy theory, a conspiracy theory ties together many levels and layers of actions together in some loosely linked outlandish thread. All I am suggesting is that a small handful of people that produce the estimates in question who are fairly high ranking in the government have made some err in judgment and have twisted around some numbers. You cannot tell me that one is not suspicious of the GDP and the LFS numbers leading up to and and now after the election. Leading up to the election we have some quite rosy numbers, then they fall in a hockey stick type manner. To me that is suspect. It would only take a few twists and turns by a handful of people sympathetic or have something to gain. With the amount of BS that is going on within Mr. Harper’s camp I wouldn’ t put it past them. I mean the “PM” is now instigating disunity to seek his power, a few calls to statcan and the right people could easily bring in some nice numbers just before the election.

If you think that is not possible, then you are being quite naive about how the political system works in this back woodsy political space we have in this country. Anything goes.

I am still quite suspect and the numbers offer some evidence of it. I find the estimate and the errors just a little too conveniently in favour of Mr. Harper before the election. We had some devilish looking LFS numbers happening in Q1 and a couple months prior and then flatlined for two and then back to the downward trend.

Same goes for the GDP. All too convenient for me.

However, standard error ranges do have a probability and the estimates are supposedly random about the mean.

Anyway the point is now the truth finally comes out and the unemployment deluge is “officially” upon us.

Potentially they really need to think about raising up the sample size of the LFS so we can have a much more accurate estimate on a monthly basis. THe current range on the error is quite unacceptable, especially during times such as which we are experiencing now.

Comment from Paul Tulloch
Time: December 5, 2008, 10:43 pm

In Ontario employment dropped from 6,732.1 to 6,666.1 (in thousands) with a standard error of 18.8. That is a drop of 66,000 net jobs. i.e. some could have been added and some lost but more lost than added.

Two ways of looking at the measurement. First is the change significant. Handy rule of thumb, if the change is twice as big as the Standard error than you definitely have significant change (95% confident).

Method two: take the change and add and subtract double the amount of the standard error and you will have the confidence interval at 95%. In this case we can estimate that the real change in the employment 19 times out of 20 (95%), is between -28,400 and -107,600 employed. The estimate is -66,000 in net lost employment.

This of course has a couple of caveats, the first and most important is that the previous month had a similar trend. If not then the estimate is a lot less accurate. This also assumes that the entire poluation was included in the sample, and we know there is bias as not everybody has a telephone and the bias, and studies have shown that the unemployed have a tendency to be more mobile and harder to contact with a telephone survey. We also have the sample rotation effects.

The LFS is made up of 6 annual rotations, and this can effectively help smooth out some estimates of the survey, however I believe it has some inbuilt bias as well that produces some difficulties. Then we ahve the seasonal adjustment process and again the numbers get a treatment.

Not very accurate when you think about what we are talking here. And that is Ontario with a 15K sample size. As suggested above it would be helpful if we bumped up the sample sizes.

Comment from Nick Rowe
Time: December 6, 2008, 12:04 am

Paul:
“Two ways of looking at the measurement. First is the change significant. Handy rule of thumb, if the change is twice as big as the Standard error than you definitely have significant change (95% confident).”

I’m thinking about one-tailed and two-tailed tests. We are 95% confident that it lies between plus or minus two standard errors. Aren’t we 97.5% confident that employment falls if it falls by two standard errors? (2.5% confident that it increased, 95% confident that it id between plus or minus two standard errors, and 2.5% confident that it fell by more than two standard errors plus the estimate.)?

I’ve been looking at employment data for years, and confess this is the first time I’ve tried to get a real sense of sampling variation.

Comment from Nick Rowe
Time: December 6, 2008, 12:17 am

And I’ve just read that footnote in StatCan LFS. The standard error is the standard error of the change in employment, not in the level of employment, as I first thought. So a fall in Ontario manufacturing of 42,200 with a standard error of the change equal to 12,500, is very statistically significant, whether one- or two-tailed.
Oh well. I’ll leave it to you stats guys.

Comment from Paul Tulloch
Time: December 6, 2008, 11:35 am

you have to look at two tailed so you must include 4 standard deviations or errors to get a 95% range. So basically you need to double it to get a fix on a 95% significance in the change. You are close Nick. You are right though, it is a bit subjective as to what objective measure we should use, 95%, 90% or 66%. I guess it depends on the importance of what you are making a decision on.

The bottom line is, we have had some very bad news and we need to get some real action towards saving jobs within the core of our economy. Manufacturing and forestry are two of the pillars we have built our economy on and we must focus on them right now. Not two months from now. As far as I am concerned by the government locking the doors on parliament, Harper and the gang have run away from their responsibility.

Sure they can still make some one sided decisions, but to have a balanced response from the government, one that is evaluated and critiqued by the opposition coalition we will have to wait until February. We could have had action in a democratic fashion if our democratic process was not interferred with in such an undemocratic fashion.

It is quite an unhealthy situation having our PM spreading such foul ideas around the country such as separatism and the tory notion of what democracy is supposed to be. I think the people of our country need to understand what the difference is betwen a majority and a minority government is. And I am sure Mr. Harper should sit in on that class as well.

How have we let ourselves become victims of such undemocratic actions. I am beside myself in trying to understand how such idiocy is allowed to prevail across the land.

Comment from Stephen Gordon
Time: December 6, 2008, 5:10 pm

All I am suggesting is that a small handful of people that produce the estimates in question who are fairly high ranking in the government have made some err in judgment and have twisted around some numbers.

Hee. Got anything other than paranoia to back up that suggestion?

Comment from Travis Fast
Time: December 7, 2008, 2:23 pm

Well I am not going to say one way or other on Stats Can and the LFS. My sense is that they are pretty honest bunch and understand themselves as above the political fray.

That said, as Foucault demonstrated long ago enumeration is part of the broader technology of power and control so any claim to be “apolitical” on their part would be naive in the broader sense.

However, I did take the plane out west last summer after the cons announced budget cuts to Stats Can and the person I was sitting next to on the plane just happened to work at stats can. A real sense that the cons did not much care about Stats Can or the services it provides. It would not be hard to get from there to a climate of fear and a certain cow-towing.

But like I said I am not taking a position either way.

I might add that the Rovian tactics of the Cons makes for a climate of paranoia. Salty might just be reflecting that.

Comment from Paul Tulloch
Time: December 8, 2008, 7:06 am

Dear Mr. Gordon,

The universe is expanding, we think, the Higgs particle may just be that which explains everything, but we don’t have the proof, and the LDC at CERN is broke again, this despite spending billions for that evidence.

Of course I do not have any proof, except to say that your naivety, and warm fuzzy beliefs in structure and order are your weakness.

The whole house of Statcan is but a large expenditure on a small attempt at measuring a whole lot of complexity. Even something as seemingly simplistic as measuring employment, can become routinely perceived simplicity. Yet hidden within the deep pockets of its truth are some rather stark complexity that can, in a mass production sense of things, be given a solution with a hammer. And when the output’s hammer marks are hidden by finely scored detailed work, the footsteps of the hammer holders are easily erased from the public spectacle.

You have to have a whole lot of faith when it comes to such mechanisms as measurements and the scientific process. Is it paranoia- hardly. It is but the raw rejection of reality that you breath in and structure into some nice neatly placed data byte , within a nicely ordered record, from a nicely formatted questionnaire, within a nicely edited and process data set, sitting on a neatly ordered desktop, on a well ordered series of cubicles, placated within a brave faced office tower, nestled within the skyline of the city scape within a nicely groomed community.

So much order stacked neatly on a big chunk of chaos, riveted and welded and hastily beaten together- by your faith.

Comment from travis fast
Time: December 8, 2008, 10:49 am

“So much order stacked neatly on a big chunk of chaos, riveted and welded and hastily beaten together- by your faith.”

Beautiful sentence…although I would take some umbrage with the word “hastily”. “Meticulously” would perhaps be more accurate.

Comment from Stuart Murray
Time: December 8, 2008, 11:29 am

I don’t think a conspiracy to mask a 0.1% increase in unemployment would have as much impact on staying in government as say, attempts to de-fund opposition parties, ban strikes, and suspend pay equity.

Remember, people don’t talk about the statistics themselves, they talk about their personal experiences and the experiences of their friends. Remember how the jobless recovery in the early nineties resulted in things being technically “good” but pretty lousy personally? In many jurisdictions there have been all sorts of bizarre games played with numbers, but election results tend to reflect how well the economy is doing in a general sense.

I predict there will be a lot of attribution theory flying around for the next few months. Who caused the recession, who attempted to fix the problem, who caused the government to fall, and as it relates to government was this all a coincidence or was there real causation? I’m pretty sure our only way out of this is with a stimulus package, but I’m also pretty sure that all government decision-making during the crisis is going to come out of focus groups and not from research about the economic fundamentals.

Comment from Paul Tulloch
Time: December 8, 2008, 11:48 am

your right, I am stuck in my thoughts on that. On the one hand culturally the societal do it without too much meticulousness for the individual. However for some reason, the individual feels its way into these crevasses so tightly and neatly that one has no choice but to think that the emotionally soothing societal fit, has been hand crafted with meticulous effort. I would have to rewrite that.

This in a modern mass produced culture is the one that has particular relevance for things such as a socially engineered numeracy based culture that statcan tries to support and promote.

Look I am not one to promote chaos, amidst a storm, but when a crisis hits us, we need to have a look around and see why the sails have ripped and why the boat is sinking so we can at least try and fix it.

I would suggest through all this quite lengthy thread, we need to have more refined instruments for measuring things such as employment. Given the importance of such a beast, we would think that policy makers would want something a bit more accurate than the standard errors currently in place. We need a much more precise estimate, especially when drilling down to understand such complicated issues as regional dimensions, transformations, and a whole host of quite imperative employment dynamics.

The proof is no further than the display of the employment estimates jumping around it has been in the last few months like a fish out of water. Potentially there are some short run underlying causation, creating these periodic mini trends. There is upheaval happening within the underlying population and in some cases survey vehicles are not design specifically to capture these changes with great accuracy, but instead focus on longer term trends. This would be the case of the LFS, as its rotational sampling, inhibits its ability to measure short run changes such as that which we are experiencing. It is designed for medium to longer term. This through the rotational basis of teh sampling plan limiting its coverage in the short run, but enabling a more accurate medium or longer term.

If we wanted more precise shorter term then we would have to expand the sample. Quite simply it is too small for such purposes. However most would argue that the use of the LFS is for longer term goals and therefore sample sizes are adequate. That is true to a degree, but I would argue that during times of change such as what we are experiencing now, having a lot better gauge of the ups and downs in the here and now of today, would be a lot more helpful, i.e. monthly would need to have a lot more weight (pardon the pun).

Paul

(I do not think I have said anything bad about statcan here.)

Write a comment





Related articles