It’s too soon for definitive labels, but the economic downturn reported thus far in 2015 looks similar to Canada’s recession in 1980.

There is plenty of interest in whether or not Canada was, or is, in recession.[1] With a federal election underway, the political stakes of a recession call are elevated.

On September 1st, Statistics Canada will release GDP data for the second quarter of 2015. If they report that growth in the first two quarters was negative, then the phrase “technical recession” will be used endlessly. But even if they don’t, the great recession debate will continue.[2]

The main reason that Canada may have avoided a recession is because of its resilient labour market. Despite declining output, total employment (in the Labour Force Survey), held up relatively well over this period, growing — albeit slowly.

But increased employment doesn’t rule out a recession. A precedent for this was set in 1980, for which the CD Howe Institute Business Cycle Council agreed that Canada experienced a mild recession (“category 1” on their 1-to-5 scale of severity), when GDP fell as employment grew.

That recession provides a convenient benchmark to compare with the recent situation. As we’ll see below, these two episodes look similar based on three common metrics used to classify recessions: depth, duration and dispersion.

Depth: Was the fall in economic activity big enough to be a recession? Taking the recent peak of economic activity as December 2014 (from the current vintage of monthly GDP data), the decline in output was similar in both episodes. The main difference is that the 2015 episode shows a more consistent march downward, whereas the 1980 episode is choppier.

OutputAggregate employment grew slowly in both cases.

EmploymentWith falling output but rising employment, output per worker — a crude measure of productivity — declines by more than output. Stephen Gordon’s characterization of this recent episode as a negative productivity or negative supply shock for Canada is apt.

Output_per_workerDuration: Was the downturn long enough to be a recession? GDP was down over a period of five months in both cases. This isn’t long, but was enough to make the January to June 1980 episode a recession in Canada. The NBER recession dates for the US include a similar period of January to July 1980 — just a six-month recession, which was the shortest in their recent history. Thus, it looks like Canada’s recent episode was long enough to be a recession.

Diffusion: Was it only the energy sector that was negatively affected? A second important argument against a recession call is that the 2015 downturn was concentrated in the resource sector and didn’t include a meaningful slow-down in other sectors of the Canadian economy. One way to measure this idea of ‘diffusion’ is to calculate the percentage of industries that grew and contracted.[3]

In the recent period, using 20 industries (at the NAICS two-digit level), shows that 11 grew (55%), while 9 contracted (45%). Unfortunately, I can’t compare this directly with the 1980 episode, but it looks like about half of all industries grew in that period (when 51% and 45% of industries grew in the first and second quarter).[4] Again the results look quite similar in each episode.

An easier way to see if this downturn has been wide-spread is to examine Canada’s GDP excluding the energy sector. This measure is also down relative to December 2014 (though by only 0.4%, whereas GDP including energy is down 0.8%). So, while this episode has largely been a shock to the energy sector, other sectors have also contracted on net. Durable manufacturing industries were particularly hard hit, which is typical in recessions.

GDP_ex_energyThe CD Howe council doesn’t seem to explicitly include income-side measures in its business cycle dating approach, focusing more on production measures. However, an important feature of this recent episode is that the large — and thus far sustained — drop in energy prices, particularly oil, is a big negative income shock, which operates through our terms of trade.

Canada’s real gross domestic income (GDI is a key indicator of Canadians’ international purchasing power) fell substantial in the first quarter of 2015 (-4.9% at annual rates, after falling by 1% in 2014Q4). Moreover, nominal GDP was also down quite significantly (-2.9% at annual rates in 2015Q1).[5] It will be important to see if these measures rebounded in the second quarter.

Preliminary assessment: Canada’s mild recession at the start of 1980 is a useful reference point. The recent episode looks similar in several respects, so if current data patterns hold,[6] then this cursory analysis suggests that Canada likely experienced at least a category 1 recession in the first five months of 2015. Importantly, the negative hit to Canada’s income started earlier and was larger than the hit to production.

Of course, we won’t know what “really happened” for sometime, but I’m curious to see a counter-argument that Canada didn’t have a recession at the start of 2015.

******

Now, I’ll digress by answering some frequently asked questions on this recent episode: whether we are, or were, in recession, and whether any of this matters:

Q1) Are we in a recession?
Strictly speaking, we can’t ever really answer this question, so the best answers are probably that we can’t or don’t know.

We don’t know simply because aggregate GDP data isn’t available on the current state of the economy. We’re always looking through the rear-view mirror as we don’t first know what happened until two months after the fact.[7]

So when economists say “we’re in recession” (present tense), what they often mean is that they are forecasting continued negative growth in the current quarter.

Q2) Were we in a recession?
Asked in the past tense, this becomes an answerable question. Unfortunately, “we don’t know yet” is probably the right answer. We need more time to pass, and more data releases and revisions, before we can be reasonably confident about any assessment.

In its recent meeting, the CD Howe council wisely determined that, as of July 22, it didn’t have sufficient information to put a date on the peak of economic activity, i.e., the council couldn’t declare whether or not a recession had started. (Note this isn’t saying that a recession had not started, just that they couldn’t definitively determine if one had).

And, if they ultimately determine that a recession had started, they’ll need even more time to place a date on the trough, i.e., to say when the recession ended.

In the US, for instance, the NBER business cycle dating committee is cautious in its proclamations and waits for a while to call the start and end of recessions. “The committee waits long enough so that the existence of a peak or trough is not in doubt, and until it can assign an accurate… date.” The last time they called the US recession start-date was 11 months after the fact and the recession end-date was 15 months afterwards.

So we will likely have to wait until after the election for the CD Howe council to weigh in again on whether or not a recession had begun.

Interestingly, the co-chair of the council recently provided his views (not those of the council) which suggested that, “Canada may yet escape an oil-induced recession.” And today, another council member, Stephen Gordon, wrote that he “subscribe(s) to the view that we’re not in a recession, at least based on available data”.

My guess is that the council will eventually determine that Canada experienced a recession in 2015. If things improve in the second half of the year, this episode will likely be classified as a mild category 1 recession, largely relegated to the first half of the year (and perhaps eventually including part of the fourth quarter of 2014 to acknowledge the income shock).

But we’ll just have to wait and see. The good news is that we should have a better idea on September 1st with the Q2 GDP data. I’m not only looking to see if the quarter is negative for real GDP, but also whether there appears to have been a rebound in real GDI and/or nominal GDP.

Q3: Does it really matter if we call it a recession?

Yes and no.

If there was a recession, then the opposition parties will obviously try to score some political points against the government.

I liked Nick Rowe’s take on how shallow such discussions can be. He wrote that, “We gain absolutely nothing by adopting some arbitrary cut-off to convert a continuous variable into a binary variable {recession; not in recession}.”

That’s true, which is why the CD Howe council’s categorization from 1-to-5 that scores the severity of recessions is a more useful description.

As economists we should emphasize that “we’ve had a recession” does not equal “everything is terrible” and “we didn’t have a recession” does not mean that “everything is awesome”. Prolonged periods of slow, marginally positive, growth should probably be just as concerning. Which may be why Armine Yalnizyan is concerned about “slowth”. Trend growth is just as important as the cycle, maybe more so.

However, there is at least one way in which the recession label could have real implications. The new federal balance budget law uses the two consecutive negative quarter definition for a recession. If we get that, and if the government complies with its law, then the federal government would not “need” to balance its budget this fiscal year.

In there was a deficit, then starting in April 2016, there would be “no increase in the operating budget of any (federal) government entity to fund annual wage increases”. So wage bills would essentially be frozen across the federal government. In addition, there would be a pay freeze for the PM, all Ministers (of State), and all Deputy Ministers. However, this is actually good news for these folks, because if a deficit is incurred outside of a recession or other extraordinary situation, then they would otherwise face a 5% pay cut.

Incidentally, my former PBO colleague Scott Cameron suggests that we could then call it a “statutory recession” and avoid the over-use of the “technical recession” label.

 


[1] Among many others see this article, or this one, this blog, this tweet, etc.
[2] Disclosure: A few months ago, in my wisdom, after GDP was negative in the first quarter of 2015 and the recession chatter started, I bet Jim Stanford a beer that we wouldn’t get a second negative quarter in Q2, so I now have a small pecuniary stake in the outcome.
[3] This exercise is known to be problematic, because in some sense it includes too many goods-intensive industries, which makes it too cyclical.
[4] Statistics Canada doesn’t currently have an active time series that goes back as far as the 1980 episode, so I don’t have historical data by industry over the first period. However, the CD Howe paper (Table 11) reports diffusion results that can provide a rough comparison.
[5] Once again, I don’t have comparable data for the 1980 recession, but I’d be surprised if the 1980s episode was worse for these measures.
[6] Of course, caution is warranted because these data will be revised, perhaps quite heavily when additional information becomes available.
[7] Even then data revisions could materially change our understanding of history. LFS employment data are released quicker — within a few weeks of collection — but as demonstrated above, on their own, they are insufficient to rule out a recession.

It’s too soon for definitive labels, but the economic downturn reported thus far in 2015 looks similar to Canada’s recession in 1980.

There is plenty of interest in whether or not Canada was, or is, in recession.[1] With a federal election underway, the political stakes of a recession call are elevated.

On September 1st, Statistics Canada will release GDP data for the second quarter of 2015. If they report that growth in the first two quarters was negative, then the phrase “technical recession” will be used endlessly. But even if they don’t, the great recession debate will continue.[2]

The main reason that Canada may have avoided a recession is because of its resilient labour market. Despite declining output, total employment (in the Labour Force Survey), held up relatively well over this period, growing — albeit slowly.

But increased employment doesn’t rule out a recession. A precedent for this was set in 1980, for which the CD Howe Institute Business Cycle Council agreed that Canada experienced a mild recession (“category 1” on their 1-to-5 scale of severity), when GDP fell as employment grew.

That recession provides a convenient benchmark to compare with the recent situation. As we’ll see below, these two episodes look similar based on three common metrics used to classify recessions: depth, duration and dispersion.

Depth: Was the fall in economic activity big enough to be a recession? Taking the recent peak of economic activity as December 2014 (from the current vintage of monthly GDP data), the decline in output was similar in both episodes. The main difference is that the 2015 episode shows a more consistent march downward, whereas the 1980 episode is choppier.

OutputAggregate employment grew slowly in both cases.

EmploymentWith falling output but rising employment, output per worker — a crude measure of productivity — declines by more than output. Stephen Gordon’s characterization of this recent episode as a negative productivity or negative supply shock for Canada is apt.

Output_per_workerDuration: Was the downturn long enough to be a recession? GDP was down over a period of five months in both cases. This isn’t long, but was enough to make the January to June 1980 episode a recession in Canada. The NBER recession dates for the US include a similar period of January to July 1980 — just a six-month recession, which was the shortest in their recent history. Thus, it looks like Canada’s recent episode was long enough to be a recession.

Diffusion: Was it only the energy sector that was negatively affected? A second important argument against a recession call is that the 2015 downturn was concentrated in the resource sector and didn’t include a meaningful slow-down in other sectors of the Canadian economy. One way to measure this idea of ‘diffusion’ is to calculate the percentage of industries that grew and contracted.[3]

In the recent period, using 20 industries (at the NAICS two-digit level), shows that 11 grew (55%), while 9 contracted (45%). Unfortunately, I can’t compare this directly with the 1980 episode, but it looks like about half of all industries grew in that period (when 51% and 45% of industries grew in the first and second quarter).[4] Again the results look quite similar in each episode.

An easier way to see if this downturn has been wide-spread is to examine Canada’s GDP excluding the energy sector. This measure is also down relative to December 2014 (though by only 0.4%, whereas GDP including energy is down 0.8%). So, while this episode has largely been a shock to the energy sector, other sectors have also contracted on net. Durable manufacturing industries were particularly hard hit, which is typical in recessions.

GDP_ex_energyThe CD Howe council doesn’t seem to explicitly include income-side measures in its business cycle dating approach, focusing more on production measures. However, an important feature of this recent episode is that the large — and thus far sustained — drop in energy prices, particularly oil, is a big negative income shock, which operates through our terms of trade.

Canada’s real gross domestic income (GDI is a key indicator of Canadians’ international purchasing power) fell substantial in the first quarter of 2015 (-4.9% at annual rates, after falling by 1% in 2014Q4). Moreover, nominal GDP was also down quite significantly (-2.9% at annual rates in 2015Q1).[5] It will be important to see if these measures rebounded in the second quarter.

Preliminary assessment: Canada’s mild recession at the start of 1980 is a useful reference point. The recent episode looks similar in several respects, so if current data patterns hold,[6] then this cursory analysis suggests that Canada likely experienced at least a category 1 recession in the first five months of 2015. Importantly, the negative hit to Canada’s income started earlier and was larger than the hit to production.

Of course, we won’t know what “really happened” for sometime, but I’m curious to see a counter-argument that Canada didn’t have a recession at the start of 2015.

******

Now, I’ll digress by answering some frequently asked questions on this recent episode: whether we are, or were, in recession, and whether any of this matters:

Q1) Are we in a recession?
Strictly speaking, we can’t ever really answer this question, so the best answers are probably that we can’t or don’t know.

We don’t know simply because aggregate GDP data isn’t available on the current state of the economy. We’re always looking through the rear-view mirror as we don’t first know what happened until two months after the fact.[7]

So when economists say “we’re in recession” (present tense), what they often mean is that they are forecasting continued negative growth in the current quarter.

Q2) Were we in a recession?
Asked in the past tense, this becomes an answerable question. Unfortunately, we don’t know yet is probably the right answer. We need more time to pass, more data releases and revisions, before we can be reasonably confident about any assessment.

In its recent meeting, the CD Howe council wisely determined that, as of July 22, it didn’t have sufficient information to put a date on the peak of economic activity, i.e., the council couldn’t declare whether or not a recession had started. (Note this isn’t saying that a recession had not started, just that they couldn’t definitively determine if one had).

And, if they ultimately determine that a recession had started, they’ll need even more time to place a date on the trough, i.e., to say when the recession ended.

In the US, for instance, the NBER business cycle dating committee is cautious in its proclamations and waits for a while to call the start and end of recessions. “The committee waits long enough so that the existence of a peak or trough is not in doubt, and until it can assign an accurate… date.” The last time they called the US recession start-date was 11 months after the fact and the recession end-date was 15 months afterwards.

So we will likely have to wait until after the election for the CD Howe council to weigh in again on whether or not a recession had begun.

Interestingly, the co-chair of the council recently provided his views (not those of the council) which suggested that, “Canada may yet escape an oil-induced recession.”

My guess is that the council will eventually determine that Canada experienced a recession in 2015. If things improve in the second half of the year, this episode will likely be classified as a mild category 1 recession, largely relegated to the first half of the year (and perhaps eventually including part of the fourth quarter of 2014 to acknowledge the income shock).

But we’ll just have to wait and see. The good news is that we should have a better idea on September 1st with the Q2 GDP data. I’m not only looking to see if the quarter is negative for real GDP, but also whether there appears to have been a rebound in real GDI and/or nominal GDP.

Q3: Does it really matter if we’ve call it a recession?

Yes and no.

If there was a recession, then the opposition parties will obviously try to score some political points against the government.

I liked Nick Rowe’s take on how shallow such discussions can be. He wrote that, “We gain absolutely nothing by adopting some arbitrary cut-off to convert a continuous variable into a binary variable {recession; not in recession}.”

That’s true, which is why the CD Howe council’s categorization from 1-to-5 that scores the severity of recessions is a more description.

As economists we should emphasize that “we’ve had a recession” does not equal “everything is terrible” and “we didn’t have a recession” does not mean that “everything is awesome”. Prolonged periods of slow, marginally positive, growth should probably be just as concerning. Which may be why Armine Yalnizyan is concerned about “slowth”. Trend growth is just as important as the cycle, maybe more so.

However, there is at least one way in which the recession label could have real implications. The new federal balance budget law uses the two consecutive negative quarter definition for a recession. If we get that, and if the government complies with its law, then the federal government would not “need” to balance its budget this fiscal year.

In there was a deficit, then starting in April 2016, there would be “no increase in the operating budget of any (federal) government entity to fund annual wage increases”. So wage bills would essentially be frozen across the federal government. In addition, there would be a pay freeze for the PM, all Ministers (of State), and all Deputy Ministers. However, this is actually good news for these folks, because if a deficit is incurred outside of a recession or other extraordinary situation, then they would otherwise face a 5% pay cut.

Incidentally, my former PBO colleague Scott Cameron suggests that we could then call it a “statutory recession” and avoid the over-use of the “technical recession” label.

 


[1] Among many others see this article, this blog, this tweet, etc.
[2] Disclosure: A few months ago, in my wisdom, after GDP was negative in the first quarter of 2015 and the recession chatter started, I bet Jim Stanford a beer that we wouldn’t get a second negative quarter in Q2, so I now have a small pecuniary stake in the outcome.
[3] This exercise is known to be problematic, because in some sense it includes too many goods-intensive industries, which makes it too cyclical.
[4] Statistics Canada doesn’t currently have an active time series that goes back as far as the 1980 episode, so I don’t have historical data by industry over the first period. However, the CD Howe paper (Table 11) reports diffusion results that can provide a rough comparison.
[5] Once again, I don’t have comparable data for the 1980 recession, but I’d be surprised if the 1980s episode was worse for these measures.
[6] Of course, caution is warranted because these data will be revised, perhaps quite heavily when additional information becomes available.
[7] Even then data revisions could materially change our understanding of history. LFS employment data are released quicker — within a few weeks of collection — but as demonstrated above, on their own, they are insufficient to rule out a recession.