Government of Canada | Gouvernement du Canada Government of Canada
    FrançaisContact UsHelpSearchHRDC Site
  EDD'S Home PageWhat's NewHRDC FormsHRDC RegionsQuick Links

·
·
·
·
 
·
·
·
·
·
·
·
 


An Evaluation Overview of Seasonal Employment

EI Evaluation
Evaluation and Data Development
Strategic Policy, HRDC

November 2001

 [Table of Contents] 
Executive Summary


Seasonal workers make up a significant and unique portion of Canada's labour market. According to the Canadian Out-of-Employment Panel (COEP) Survey database used in this paper, roughly fifteen percent of those who experienced a job separation were in jobs they described as seasonal. This estimate is substantially larger than the roughly 3 percent of total employment considered to be seasonal when looking at the regular seasonal fluctuations in total employment using the Labour Force Survey. We consider the COEP estimate of seasonal work to be more realistic for the purposes of policy as the use of fluctuations in total employment tends to hide the seasonal rises and falls of seasonal industries. This is because a seasonal fall in one industry may cancel out the seasonal rise in another. Our COEP estimates of seasonal work will vary, depending on province or industry being examined. It also varies among demographic groups. In general, seasonal work is more prevalent among less-educated males in primary industries in Eastern Canada. As well, an important distinguishing characteristic is that seasonal workers are less likely to be married or have children.

Evaluation evidence is available on the ability of the individual's ability to cope with spells of unemployment. In general, the seasonally unemployed were able to cope better than non-seasonal. They were more likely to find subsequent employment and less likely to experience a drop in their consumption after the period of unemployment. In spite of this, workers in seasonal industries were in general found to pay less into the EI system than they took out.

Past evaluations showed two aspects of seasonal claim behaviour that became the focus of EI reform. First, under UI, there was no evidence of an entrance requirement effect on employment durations for seasonal jobs. An evaluation did find, however, a strong statistical and economic significant effect (about 1 in 20) for seasonal job spells ending at the maximum year point, at which workers qualify for enough weeks of benefits to fill the remainder of a 52 week period. This suggested a significant tailoring of seasonal jobs to the UI system. As well, past evaluations found that UI was becoming more of a permanent income support for high repeat users since the early 1970s. Further evaluations found a high correlation between seasonal work claims and repeat claims. High frequency users represented only 7% claimants but 25% of all claims. Over the 20 years of claims, there was an upward secular trend of such income support dependency. By contrast, UI was found to protect low frequency users against the risk of unemployment. Low frequency users represented 62% of claimants but 25% of all claims. There has been a downward trend in their pattern of use, which increased during recessions.

Some evaluation evidence is now available on the impact of EI reforms that are relevant to seasonal workers. The move to the hours rule has led to seasonal workers working more hours per week, while forming a claim with slightly fewer weeks of work. However, preliminary monitoring analysis has also documented the differences in EI experiences within the seasonal worker group. Those with income of $12,000 or less per year fared considerably worse in meeting eligibility requirements and weeks of entitlement relative to other seasonal workers. These individuals generally worked fewer hours or were students. For seasonal workers overall, internal HRDC research has shown that after EI reform entitlements increased by 1 week and the percentage of experiencing periods where they are receiving neither benefits or employment income dropped substantially.

The EI community focus groups have also provided some social context for understanding the adjustments of seasonal workers to the EI reforms. Specifically, the EI system was considered generally to be fairer in its eligibility and entitlement rules, except where no alternative employment was available. In this context, seasonal workers perceived the intensity rule to be punitive and unfair in punishing them for living in areas where alternative, stable employment was not available.

 [Table of Contents] 
Introduction


It is in the nature of the economy that the duration of some forms of employment will be determined by seasonal factors. The frequency of this is a function of the industrial structure and the incentives that have been put in place. A key incentive is the EI system. By reducing the cost of the periods of unemployment that may follow seasonal jobs, the incidence of seasonal work may be higher than otherwise. The purpose of this paper is to provide an overview of the impact of EI on seasonal employment from a strategic evaluation perspective.

The paper will first start with an assessment of the extent of seasonal work. The next section will provide a description of the seasonal workers. As well as the basic demographics, the evaluation literature will also be surveyed to develop a sense of the nature of seasonal work from a larger perspective.

Finally, there will then be a focus on the impact of EI on seasonal work. A few evaluations deal directly with seasonal workers. Many evaluations contain results that are at least to some extent relevant to the issue of seasonal work. The goal of this will be to identify firm conclusions that can be drawn on the subject of seasonal work.

 [Table of Contents] 
The Extent of Seasonal Work


This paper uses a definition of seasonality based on the COEP database1. The COEP survey is based on individuals who have experienced a job separation. Each individual is asked to provide the characteristics that best describe their job. The response of individuals to this question is used to define seasonality in this paper. Based on the responses of roughly 32,000 individuals, over the 1995 to 1997 period, seasonal workers make up 15.5 per cent of job terminations.

These numbers simply will not be the same as those based on other databases or methodologies. In some cases, seasonal workers have been defined as those who work in industries that are considered seasonal. Another popular definition measures seasonality by examining the seasonal fluctuations in total employment.

Based on aggregate Labour Force Survey data, Marshall (1999) found that the total level of employment fluctuated, on a monthly basis, an average of 2.8 per cent from its expected mean due to seasonal fluctuations, in 1997. From this it can be claimed that 2.8 per cent of total employment is seasonal. It is important to note that this measure tends to ignore fluctuations at the more detailed levels. For example, seasonal employment at ski resorts may balance out seasonal job losses in the tourist industry that supports cottage owners.

This number is twice as high as the U.S. (Rydzewski et al. 1993). If rather than looking at the average value, the difference between the highest and lowest is examined, then the 2.8 per cent for Canada goes up to 6 per cent. This number is still considerably lower than the 15.5 per cent that comes from COEP.

In both the U.S. and Canada, the total economy-wide estimates of seasonality indicate that it is falling. The fall in Canada was from 3.4 per cent in 1976 to 2.8 per cent in 1997, or 18 per cent. However, this may be due to increases in the extent to which fluctuations in individual industries cancel each other out.

 [Table of Contents] 
Characteristics of Seasonal Workers


The 15.5 per cent of the COEP sample who are considered seasonal workers can be characterised in many ways. Table 1 shows the basic demographic distribution. Then the province, industry and communities are examined. The next table gives the description of the seasonal jobs, followed by a description of the adjustment process when the jobs are over. The final characteristic to be focussed on is the planning framework faced by seasonal workers.

Seasonal Work is Unevenly Distributed

Seasonal workers are distinctive in many ways. Even an examination of the basic demographics is revealing. In Table 1, it is shown that 64.92 percent of seasonal workers are male, whereas only 50.5 are male in the non-seasonal jobs. Although, the average age of the two groups is the same, there are significant differences in the underlying age distribution with youth being far more prominent among seasonal workers. Seasonal workers also appeared to be less tied down by family obligations as they are less likely to be married or divorced by about 8 percentage points. In addition, they are 5.6 percentage points less likely to have children under 16. Finally, seasonal workers are more likely to be less educated, as 33.9 per cent of them have less than high-school education compared to 19.9 per cent for non-seasonal.

Table 1
Percent in Demographic Categories
Seasonal Yes No
Male 64.9 50.5
Age (years) 35.5 36.0
15-24 24.3 17.7
25-54 66.9 73.6
55+ 8.8 8.6
Disabled 7.4 8.4
Married now or in past 61.8 69.8
Have Children under 16 29.4 35.0
Less than Highschool 33.9 19.9
Highschool 28.4 26.4
More than Highschool 37.7 53.6
COEP:  Source COEP, cohorts 1-8
All data is in per cent unles noted

Table 2 shows the extent to which seasonal work tends to be concentrated in certain industries and regions. For example, in Newfoundland 31 per cent were seasonally employed. This is twice as high as the national average, indicating a distinctly different industrial structure than the rest of the country. The range over the whole country is immense, from 42.4 per cent in Prince Edward Island to 11.4 per cent in Ontario. In general, seasonality is highest in the east.

Table 2
Seasonal Employment by Province and Industry
 % Seasonal within Province  % Seasonal within Industry
Newfoundland  31.0 Agriculture 59.0
Prince Edward Island  42.4 Primary 45.2
Nova Scotia  25.7 Manufacturing 11.6
New Brunswick  31.1 Construction 29.1
Quebec  17.0 Transportation 21.6
Ontario  11.4 Communications 4.8
Manitoba  15.8 Utilities 18.3
Saskatchewan  17.1 Trade 7.4
Alberta  13.0 Finance Insurance and Real Estate 3.7
British Columbia 12.2 Education 9.5
  Business Services 13.5
Government Services 14.8
Public Administration 16.6
COEP 96: Cohorts 1-8

According to this table, the variations in seasonality are more pronounced by industry. The industries range from 59 per cent in Agriculture to 3.7 per cent in the Finance, Insurance and Real Estate industry. In general, the high levels of seasonality are in industries that have not been growing as quickly, thus contributing to an overall fall in the concentration of seasonal work in the economy.

As shown in Table 3, seasonal jobs have significantly different characteristics. For a start, the wages on the job lost by seasonal workers is somewhat lower. However, this understates the relative disadvantage of seasonal workers as they will earn less in a typical year. This is shown by comparing the total earnings given on the T1 in the year before the job loss. The seasonal workers reported an average of $12,303 on their T1 tax form, whereas the non-seasonal reported $20,114. In addition, seasonal workers were less likely to have medical benefits or a pension plan. These additional factors combine to portray seasonal workers in a less favourable light. This is particularly true when it is considered that seasonal workers have to work more hours to get less. As well, seasonal workers are less likely to be in unions. Still it is important to remember that the figures given on Table 3 are only averages and that seasonal workers are highly varied. Although the average seasonal worker does make less than their non-seasonal counterparts, there is a significant portion that are well-paid.

Table 3
Job Characteristics
Seasonal Yes No
Weekly wages of last job 486 571
Actual Earnings in
Previous Year 12, 303 20, 114
% union membership 24.7 32.3
% have medical benefits 19.6 45.0
%have pension 14.1 29.6
Hours 44.2 38.9
Tenure N/A N/A

Certain Communities Can Take on Strong Seasonal Characteristics

To fully appreciate the nature of seasonal work, it is necessary to understand that it can be highly concentrated in certain communities, most often as a result of a dependence on one particular industry in the community. In order to illustrate this, Table 4 provides data on four selected communities, which are available within the COEP survey.

Table 4
Seasonal Employment in Selected Communities
  Clarenville,NF Miramichi Calgary Kelowna
% Seasonally Employed 33.3 34.4 9.7 13.7
Most Prominent Seasonal Industry Food Processing Logging Construction Construction
% Prominent Seasonal Industry of All Seasonal 33.2 25.0 64.4 38.8
2nd Most Prominent Seasonal Industry Construction Construction Oil & Gas Wood Products
% 2nd Most Prominent of All Seasonal 31.7 20.9 8.5 18.5
Annualized Earnings 11,755 11,574 15,096 15,124
Expected to return to employer 81.1 85.0 48.8 74.1
Have Pension 12.9 2.6 6.4 9.9
Have Medical Plan 19.5 11.3 20.1 14.5
In Union 26.9 13.0 25.4 9.7
Source: COEP 96, Cohorts 1-8

What is first striking is the variation in seasonal work among the communities. The two more rural communities in Atlantic Canada have seasonal rates far higher than the other two. For example, in Clarenville Newfoundland, the share of seasonal workers in COEP is 33.3 per cent compared to the 9.7 per cent found in Calgary.

The distribution of industries was examined to obtain a sense of the nature of the community labour market. This analysis was restricted to primary, manufacturing and construction as the distribution of the service industries were similar among communities. In the communities where there is a higher concentration of seasonal work, a primary industry is found to dominate seasonal employment. However, in the lower concentration communities, construction tends to dominate the seasonal employment that exist. In addition, in the lower concentration communities, seasonal workers are less likely to expect to return to their employers after the job loss.

Table 5
Labour Market Outcomes After Job End
Seasonal Yes No
Length of Time Unemployed 23.6 24.2
% Unemployed > 52 weeks 15.5 26.6
% experiencing wage drop 55.5 53.4
% experiencing wage drop who were unemployed less than 12 weeks 14.3 21.5
% who received EI 52.9 53.5
% who went on SA 4.5 3.4
% who experienced a drop in consumption 9.0 13.7
Source:  COEP 96, cohorts 1-8

Seasonal Workers Can Cope Better with Job Separation

In general, the evaluation evidence argues that seasonal workers cope at least as well, if not better, with job loss than non-seasonal. Audenrode and Storer (1998, Table 5) reported that the likelihood is higher of finding another job for those who lost a seasonal job. This is confirmed on Table 5, where the seasonal workers are shown to have a 15.5 per cent chance of being unemployed for more than 52 weeks, as compared to a 26.6 per cent chance for non-seasonal. In addition to being more likely to get a new job, Audenrode and Storer found that they perform just as well as non-seasonal in terms of labour force participation and wage gains. A possible contributing factor may be that seasonal workers are more likely to have a recall date when their job terminates, as shown in Table 6.

Overall, the seasonal workers are better able to maintain their well-being. Browning (Browning 1998, p. 25, Table 3.7) makes this point, as less seasonal workers experience drops in consumption a year after the job loss. They are able to do this without making significantly different use of EI or Social Assistance than non-seasonal, as is shown in Table 5.

A lot of the reason for their relative success can be explained by the examination of the circumstances surrounding their job loss. Table 6 reveals that seasonal workers are far more likely to be laid-off than other workers, 72 per cent as versus 39.2 per cent. This by itself should have made it more difficult to cope with job terminations. However, 73.0 per cent of seasonal workers expected to return to the same employer compared to only 47.1 per cent for other workers. As well, other indicators suggest advantages for seasonal workers as they were more likely to receive notice and have return dates. On the down side, they were less likely to receive some form of financial compensation, such as severance, following the job loss.

Table 6
Nature of Job End
(percent)
Seasonal Yes No
Reason for Job Loss
Layoff 72.0 39.2
Return to School 7.7 4.2
End of Contract 8.5 11.9
New Job 3.3 9.8
Maternity 0.1 5.7
Other 8.4 29.3
Received Notice 34.5 26.8
Expected to Return to Employer 73.0 47.1
Had Return Date 32.9 23.0
Had Firm Return Date 24.1 18.6
Received a Package 23.1 32.7
Source: COEP 96, cohorts 1-8

The higher expected return rates is in line with the finding that seasonal workers did not put as much time into their job search (Crémieux et al. 1995 p. 20). This is likely due to the nature of the labour market work, as some seasonal workers will just return to jobs that they are familiar with each year, which would not involve an extensive job search.

Seasonal Workers have a Different Planning Framework

In Green and Sargent (1995), the claim behaviour of seasonal workers are examined under the UI system. The initial view was one in which seasonal workers have complete freedom to select the number of weeks that they work over the 52-week period for which they plan. It was found that if the seasonal worker tried to maximise both his annual income and leisure3, then it was possible to work just enough weeks to earn benefits that would last until the end of the 52-week period. The evaluation estimated that 1 in 20 seasonal jobs ended at a point when sufficient benefits are obtained to fill the income requirements for the remainder of the year. If the seasonal worker was employed for fewer weeks, then there would be some weeks in the year for which he would not be receiving any income.

From this evaluation, it could be argued that many seasonal workers would be less likely to work only the minimum number of weeks to qualify for EI, than the non-seasonal. This is because if they worked only the minimum number of weeks, there would be weeks in the year for which they would not receive any income. Non-seasonal workers would not face this constraint and would be more likely to work just the minimum number of weeks.

A number reasons were suggested, by Green and Sargent (1995), why this analysis would not be directly applicable to all seasonal workers. For one, it assumes that the seasonal worker has complete control over the number of weeks of work. This is of course not applicable in all cases, as the demands of employers will play a significant role in determining the number of weeks worked. However, Green and Sargent (1995) do show that it may be in the employers interest to come to some work arrangement with the seasonal employees so as to maximise EI benefits received.

Another complicating factor is that some seasonal workers may find it desirable to minimise on the number of weeks that they are unemployed. Apart from the natural desire to work, they may want to avoid the deterioration of work-skills caused by prolonged periods of unemployment. There may also be a concern that the periods of unemployment would cause them to appear less valuable to future employers.

EI Reform has a Significant Impact

COEP evidence suggests that EI reform has had a significant impact on the EI eligibility. This primarily comes as a result of the move to the hours legislation. Table 7 below summarises these impacts. Basically seasonal workers are less likely to be eligible for EI than non-seasonal but this gap narrows as a result of EI-Reform. The first row gives the percentage point difference between seasonal work compared to non-seasonal work. For the sample, selected seasonal workers who lost their job were 5.6 percentage points less likely to have enough hours to be eligible for EI compared to other workers. They were also entitled to 4.594 less weeks of entitlement, as given in the fifth row.

Table 7 
Impact of EI Reform on Seasonal Worker’s Eligibility for EI
  Of Seasonal Work Of EI Reform on Seasonal Work Of hours less than 30 On Income Less than $12,000
% Impact on Probability of EI Eligibility 1 -5.62 (0.00)      
2 -5.329 (0.00) -0.693 (0.49)    
3 -5.327 (0.55) 0.550 (0.59) -21.447 (0.00)  
4 -5.322 (0.00) 4.908 (0.00)

-18.611

(0.00)

-12.571 (0.00)
Impact on Weeks of Entitlement 5 -4.599 (0.00)      
6 -5.186 (0.00) 1.660 (0.00)    
7 -5.186 (0.00) 1.770 (0.00) -2.635 (0.04)  
8 -5.186 (0.00) 2.651 (0.00) -1.818 (0.16) -2.376 (0.00)
Notes: Absolute P values in brackets
Income defined as T1 employment and EI Beneficiary Income in the Year before Job Loss.

The estimates do not change by much if seasonal workers are compared before and after EI reform for eligibility. However, the number of weeks of entitlements goes up by 1.6. These basic results, given in Rows 2 and 6, constitute the simple baseline impact of EI reform on seasonal workers

If only those seasonal workers who have less than 30 hours are examined, it is found that they lost significant amounts of eligibility, as they are 21 percentage points less likely to qualify for EI after reform and if they do qualify they receive 2.6 less weeks of entitlements. Rows 4 and 8 were added in response to a concern that was expressed that seasonal workers who made less than $12,000 a year were even more adversely affected by EI reform. The above table shows this appears to be true, although at this time it is not clear what impact EI reform could have on seasonal workers over and above the hours effect. Hence, this conclusion should be regarded as tentative at this time.

 [Table of Contents] 
Evaluation Evidence on EI and Seasonal Work


Three of the changes to EI, as part of EI-reform, are directly relevant to seasonal workers. First the move to count hours rather than weeks will have a greater influence on seasonal workers because they work 5 hours more per week than non-seasonal, as shown on Table 3. Secondly, the divisor rule will encourage those who establish claims with the minimum number of weeks to work at least an extra two. Lastly, the intensity rule will reduce the benefits for repeat users of EI, which will impact seasonal workers more than non-seasonal. This section will first review the earlier evaluation evidence, which help set the stage for these three changes. Then the most recent evaluation evidence will be examined to determine the impact of EI-reform on seasonal workers.

Earlier Evaluation Evidence

The early evaluations on subjects directly related to seasonal work did not touch on the subject of hours. However, a justification for the move to the hours system can be found in an evaluation of jobs excluded from the UI system (Lin 1995)5. In this study, it was found that 21 per cent of all employment in 1990 was not eligible for EI benefits. This share had been growing since 1986. There were many reasons for this such as self-employment, however, the exclusion of jobs with less than 15 hours was prominent among them. The counting of all hours of work, in EI, can be seen as move towards rectifying this situation.

Early evaluations, directly related to seasonal workers, were concerned with the behavior of claimants while forming claims. Substantial evaluation evidence was compiled which argued that a significant portion adjust the length of their employment while forming a claim to suit the requirements of the UI system. Christofides and McKenna (1995) found that there was a significant tendency for workers to terminate their employment once they had accumulated the minimum number of weeks to qualify for EI. When a distinction was made between seasonal and non-seasonal workers by Green and Sargent(1995 p. 45), it was found that seasonal workers are more likely to terminate their employment at the point which would provide them just enough benefits to cover the time that they would be unemployed during the off-season. Thus in most cases, they would work more than just the minimum to qualify. However in very high unemployment regions, low wage seasonal workers were found by Green and Riddell(1995) to be sensitive to changes in the minimum required weeks. As a means of discouraging these tendencies described in the evaluation and other studies, the divisor rule was implemented which reduces the benefits for those who only work the minimum number of weeks.

It has been observed that a substantial portion of UI use resulted from individuals who were repeat users of the system. Lemieux and MacLeod (1995) found that seven per cent of claimaints had more than eleven spells and accounted for 22 per cent of all UI claims. Overall, the incidence of repeat use was found by Wesa (1995 p. 27) to be trending upwards, as the share increased by 25.9 percent from 1982 to 1990. It was argued that the bulk of this change occurred in seasonal industries. This is particularly significant when it is considered that the overall trend towards seasonal work is downward, as was discussed above. It should be pointed out that the change that came from the seasonal industries was due to both an increase in the number of claimants in these industries as well as an increase in the incidence of repeat use by the seasonal industries.

The growth of repeat use was most pronounced in the eastern provinces. Examples were found of identical industries which had higher incidence of repeaters in Atlantic Canada than they did in Ontario. For example, the logging industry in Atlantic Canada had a repeater rate of 87.8 per cent where for Ontario it was significantly lower but still high at 67.4 per cent.

Current Evaluation Evidence on EI Reform and Seasonal Work

A major innovation with EI Reform has been the implementation of a monitoring framework which has ensured that evaluative evidence would be available relatively soon after the reforms were put in place. The reforms were fully implemented in January of 1997, and as of the Fall of 1999, fourteen evaluations had been completed by external evaluators. In addition, a large volume of internal evaluative work had been completed by that time.

Some of the evaluation work focussed on the impact of the move to hours. Friesen and Maki (1999) found that the move to EI resulted in a reduction in the proportion of seasonal workers who had jobs of 30-40 hours in favour of those who had jobs of 40-48 hours a week. A possible reason for this was suggested by Green and Riddell (1999), as they found that the greater hours of work by seasonal workers were leading to more entitlements. This resulted from the flexibility of seasonal workers to concentrate hours in a week such that there was no change in the total numbers of hours worked per year but they were able to qualify for EI benefits with 1.5 less weeks on their jobs.

As indicated in Wesa (1995), the changes in behaviour that resulted in high rates of repeat behaviour occurred over a period of many years. Still Fortin and Audenrode (1999 p. 1 and 12) were able to find some indications in the first year that the repeaters were starting to respond to the change in EI incentives. This primarily occurred with individuals having a one percentage point greater probability of leaving unemployment in the 19th week, the last week before benefits drop takes effect, so as to avoid their benefit rate dropping the next time that the collect EI. However, it should be pointed out that these changes in the probability of leaving unemployment were very small. It is interesting to note that in focus groups that were conducted at the community level, the intensity rule was found to be punitive and unfair.

Internal HRDC evaluative work found that seasonal workers increased their entitlements by slightly more than one week after EI reform. There was also an 8 percentage point drop in the share of seasonal claimants who experience "gaps" or periods of time when they are neither collecting benefits or receiving employment income. These two changes were a result of a move to the hours rule which benefited seasonal workers as they worked an average of 5 more hours per week.

 [Table of Contents] 
Conclusions


Seasonal workers make up a distinctive portion of the labour market. They are more likely to be male, less-educated and have fewer family dependants. They appear to be more successful at avoiding the downside of the adjustment process. In addition, evaluations have shown that some of the seasonal workers have adjusted their weeks of employment with reference to the rules of the EI system. It is also interesting to note, that the industrial sectors whose employees receive more from EI than they pay into it tended to be seasonal (Corak 1995).

The move to the hours system, which was undertaken for other reasons, has encouraged seasonal workers to work more hours per week over shorter periods of time. This has led to more generous entitlements for EI being experienced by this group. Another benefit of the move to the hours system was a drop in the number of seasonal workers who experience periods when they are neither collecting EI nor receiving employment income.

 [Table of Contents] 
References


AUDENRODE M. V. and Storer P. (1998), "Long-term Employment Outcomes", Evaluation and Data Development, Human Resources Development Canada.

BETCHERMAN G. and Leckie N. (1995), "Employer Responses to UI Experience Rating: Evidence from Canadian and American Establishments", Evaluation and Data Development, Human Resources Development Canada.

BROWNING M. (1998), "Income and Living Standards During an Unemployment Spell", Evaluation and Data Development, Human Resources Development Canada.

CHRISTOFIDES Louis and McKenna Chris (1995), "Employment Patterns and Unemployment Insurance", Evaluation and Data Development, Human Resources Development Canada.

CORAK M. and Pyper W. (1995), "Firms, Industries and Cross-subsidies: Patterns in the Distribution of UI Benefits and Taxes", Evaluation and Data Development, Human Resources Development Canada.

CRÉMIEUX P., Fortin P., Storer P. and Audenrode M. (1995), "The Impact of Unemployment Insurance on Wages, Search Intensity and the Probability of Re-employment", Evaluation and Data Development, Human Resources Development Canada.

FORTIN P. and Audenrode M. (1999), "The Impact of Experience Rating of Unemployed Worker", Evaluation and Data Development, Human Resources Development Canada.

FRIESEN Jane and Maki Dennis (1999), "The Effect of Bill C-12 on Weekly Hours of Work", Evaluation and Data Development, Human Resources Development Canada.

GREEN David and Riddell Craig (1995), "Qualifying for Unemployment Insurance: An Empirical Analysis of Canada", Evaluation and Data Development, Human Resources Development Canada.

GREEN David and Riddell Craig (1999), "The Effects of the Shift to an Hours Based Entrance Requirement", Evaluation and Data Development, Human Resources Development Canada.

GREEN David and Sargent Timothy (1995),"Unemployment Insurance and Employment Durations: Seasonal and Non-Seasonal Jobs", Evaluation and Data Development, Human Resources Development Canada.

HRDC (2001a), "EI Reform and Seasonal Workers", Evaluation and Data Development, Human Resources Development Canada.

HRDC (2001b),"The Impact of EI Reform on Gappers", Evaluation and Data Development, Human Resources Development Canada.

LEMIEUX Thomas and MacLeod Bentley (1995),"State Dependence and Unemployment Insurance", Evaluation and Data Development, Human Resources Development Canada.

LIN Zhengxi (1995), "Jobs Excluded from the Unemployment Insurance System in Canada: An Empirical Investigation", Evaluation and Data Development, Human Resources Development Canada.

MARSHALL, Katherine (1999), "Seasonality in employment", Perspectives (Spring 1999 pp. 16-22), Statistics Canada.

RYDZEWSKI L. G., Deming W. G., and Rones P. L. (1993), "Seasonal employment falls over past three decades", Monthly Labor Review (July 1993 pp. 3-14), U.S. BLS.

TATTRIE D. (1999), "A Financial Incentive to Encourage Employment among Repeat Users of EI: The Earnings Supplement Project", Applied Research Branch, Human Resources Development Canada.

WESA, Lesle. (1995). "Seasonal Employment and the Repeat Use of Unemployment Insurance", Evaluation and Data Development, Human Resources Development Canada.

 [Table of Contents] 
Appendix A: A Description of COEP


The Canadian Out of Employment Panel (COEP) is sponsored by Evaluation and Data Development primarily for the purpose of evaluating the impact of EI on individuals. This survey has been conducted three times. The first time, in 1993, the survey was conducted for EDD by Ekos to aid in the evaluation of Bill C-133. A sample of the unemployed was taken before and after the passage of the bill to allow estimation of the effects of the changes to the legislation on individuals. Likewise, Statistics Canada conducted a second version of COEP for EDD, in 1994, to study the impacts of Bill C-17. The most recent version of COEP was conducted by Statistics Canada to measure the impacts of EI-Reform.

This survey is based on a stratified random sample of Records of Employment, ROE. The use of the ROEs allows for the easy linking of records to administrative files, which is key to the evaluation process. The current version of COEP was based on individuals who had ROEs over the period third quarter 1995 to forth quarter 1997. Additional quarters of data are being collected to allow for the monitoring of EI. For the purposes of studying seasonality eight quarters of data were used, yielding 32,000 respondents. The full sample was not used so that the sample would contain the same number of respondents in each quarter.

It should be kept in mind when comparing COEP results to those of other surveys, such as the LFS, that the sampling frame is significantly different. The LFS would comprise a random sample of the whole population, whereas COEP would be a sample of those who have just left a job. As well, the COEP will not include all those who leave self-employment, although it would include Fishers who are eligible for EI.

Within COEP, there are two questions that can be used to identify seasonality. The first questions question asks for reasons for the job loss. One possible reason is "seasonal factors". A second question asks the respondent to chose a characteristic that best describes their job. One characteristic is seasonal. For the purposes of this study, the second question is used. However, both questions yield similar responses.

COEP also over-samples 13 specific communities. This allows COEP to support detailed analysis and evaluative work to be done on a very narrow community basis, as is shown in Table 4.

 [Table of Contents] 


Footnotes

1 See Appendix A for a brief description of COEP. [To Top]
2 The predominance of men in seasonal work was noted in Green and Riddell (1999). [To Top]
3 An individual would maximise their income by working the entire year, however for a seasonal worker this is assumed to not be an option. [To Top]
4 The estimate of 4.59 weeks difference in entitlements between seasonal and non-seasonal is highly sensitive to the definitions of the data employed. It is possible that this difference is as low as 1 week using a definition derived purely from the administrative data [To Top]
5 This is only one possible source of justification for the move to the hours system. [To Top]


 [Table of Contents]