One can often hear the term “EI Numbers” when the conversation is about assessing job health or looking for signs of an improving economy. With Canada and the rest of the world slowly recovering from the recent worldwide recession, everyone is constantly monitoring employment data from the various government agencies. The beginning of the month in many countries often sees the release of employment data for the previous month. In Canada, Statistics Canada is responsible for employment statistics whereas in US, the Bureau of Labour Statistics releases all employment and unemployment data.

Statistics Canada publishes two different sets of data for employment and the analysis of this data can very quickly get confusing when one encounters many discrepancies. The Labour Force Survey is released usually on the first or second Friday of the current month and it reveals the employment data for the previous month. These numbers often find their way into headlines and are widely reported through media. The other publication that is less followed is the Survey of Employment, Payrolls and Hours of employers which is released a month or two after the LFS.

The confusion arises when the data from these releases are followed and analyzed over a period of time. For example last year, the August release of the LFS reported a decrease in employment by 0.26% after employers reported a decrease of 79,000 in employees in July. However, the SEPH which was released at the end of September painted a different picture by indicating that the number actually increased by 74,000 and thus the employment rate was up 0.51%. Looking at this data for the first time might lead one to believe that this is an odd discrepancy that occurs rarely. But the figures for August do not help clear the confusion as LFS shows a gain of 28,000 jobs whereas SEPH reports a loss of 110,000 employees. Now these numbers result in a debate on which survey to believe. Is the LFS more accurate since it is released earlier? Should the SEPH be ignored? On the other hand, is the SEPH a better indicator than LFS? In order to fully understand the discrepancy, the manner in which the data is obtained must be considered.

The LFS is a household survey and has a large and varied sample size covering Canada whereas the SEPH is an establishment based survey that measures the payroll data from the administrative files provided by businesses. An important distinction between the two comes from the fact that the LFS numbers include self-employed individuals whereas the SEPH does not have a payroll for self-employment. StatsCan itself also clarifies that “SEPH shows payroll employment over the entire month, while the LFS reflects labour market conditions only for the middle of the month”, confirming that both surveys have their share of similarities and differences.

Looking back to 2009, the unemployment rate had an unusual way of explaining itself. It rose from 7.2% to 8.0% from January to April, which included a jump of 0.5% points just from January to February. March and April saw the same rate as there was no significant change in the employment numbers. However, May saw a large increase of 0.4% points to 8.4%. Unexpectedly, the rest of the year saw minimal upward and downward fluctuation and the unemployment rate was an overall 0.1% point up from May to 8.5% at the end of the year. It also decreased unusually a significant 0.3% points from August to September. The obvious question is; why was the unemployment rate going down when firms were firing rather than hiring? The answer lies in an aspect that is often overlooked; the affect of the number of “discouraged workers” on the unemployment rate. A discouraged worker is a person legally eligible for work, but not actively seeking employment and thus is not included in the labour force when determining the unemployment rate. The word “discouraged” lends itself to the fact that the individual has had no success in finding a job or given up looking for employment. The first half of 2009 saw major increases in the unemployment rate as more and more people were fired and their search for employment continued. However, towards the end of the year many people gave up their search and became discouraged workers. This is evident in the decrease to 8.4% in September and the lack of change in November and December. Intuitively, the lack of increase in the rate and even a decrease makes sense as the number of discouraged workers is more likely to increase the longer the economy is in a recession. The unemployment rate was down however in reality more people were unemployed and a significant number had given up their job search.

Coming back to the issue of discrepancy between the LFS and SEPH numbers, when they are analyzed over July and August, they seem to explain themselves better as the net result of LFS and SEPH over the two months is a loss of 41,000 and 36,000 jobs respectively. The bottom line is that in order to get more similarity between the surveys, an influx of resources is required. Budget cuts over the years have resulted in the quality of the data diminishing and thereby inconsistent data. If in doubt over which survey to use, StatsCan has some helpful and brief advice; “Since each of these surveys addresses different needs, the choice of data depends on the user’s goals”.