The Impact of Intent in the Characterization …...The Impact of Intent in the Characterization Analysis of a Worker Nabeel Peermohamed Master of Laws Faculty of Law University of
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The Impact of Intent in the Characterization Analysis of a Worker
by
Nabeel Peermohamed
A thesis submitted in conformity with the requirements for the degree of Master of Laws
The Impact of Intent in the Characterization Analysis of a Worker
Nabeel Peermohamed
Master of Laws
Faculty of Law University of Toronto
2013
Abstract
The Tax Court of Canada and the Federal Court of Appeal are regularly asked by taxpayers and
the Canada Revenue Agency to determine whether a worker is an employee or an independent
contractor. The distinction has significant tax consequences. The analysis and various legal tests
used by the courts to determine the characterization of a worker have been through significant
transformations over the last 15 years. The analysis remained objective for several years.
However, in 2002, courts began to consider the common intent held between the taxpayer and
the worker when characterizing that worker’s status as either an employee or an independent
contractor. Since its introduction, the courts have placed various levels of importance on this
common intent in the characterization process. This paper seeks to quantify the varied emphasis
placed by the courts on the intent held between a taxpayer and its worker.
iii
Table of Contents
List of Tables .......................................................................................................................................... v
List of Appendices ................................................................................................................................ vi
Part 1 Introduction .............................................................................................................................. 1
Part 2 The Characterization Process .............................................................................................. 3
1 The Objective Factors ................................................................................................................... 3
2 The Fall of Integration .................................................................................................................. 6
3 An Introduction to the Evolution of Intent ............................................................................ 8
4 A Detailed Account of the Evolution of Intent ....................................................................... 9
5 The Inconsistency caused by Intent ...................................................................................... 20
Part 3 The Study ................................................................................................................................. 22
1 Inspiration from the Wintermute Study .............................................................................. 22
2 Extrapolating Wintermute to the Present .......................................................................... 22
3 The “More Strongly than Royal Winnipeg” Standard ....................................................... 24
4 Methodology for Study .............................................................................................................. 27
4.1 The TCC or FCA’s Characterization (employee/independent contractor) ........................ 28 4.2 The Characterization of the MNR or TCC Below ......................................................................... 28 4.3 Finding of Common Intent (employee/independent contractor/none) ........................... 28 4.4 Finding Based on Factors (employee/independent contractor) .......................................... 29 4.5 The Courts’ Finding After an Application of each of the Four Wiebe Door factors (Control/Financial Risk/Ownership of Tools) ................................................................................... 29 4.6 Judge ......................................................................................................................................................... 29 4.7 Gender of the Judge .............................................................................................................................. 29 4.8 Province or Territory .......................................................................................................................... 30 4.9 Personal Service Business (PSB) ..................................................................................................... 30 4.10 Gender of the worker (male/female) .......................................................................................... 30 4.11 Nature of Organization (profit/non-‐profit) .............................................................................. 30 4.12 Industry of Organization ................................................................................................................. 30
the other factors were neutral. The court then found the intent of the parties was not consistent
with this finding. The court said:
The actual dealings between the parties, however, were not consistent with an independent
contractor relationship. If parties wish to have an independent contractor relationship respected for
purposes of the Employment Insurance Act and the Canada Pension Plan, their actions need to be
consistent with their intent. Unfortunately for the appellant in this case, the evidence as a whole
suggests that the parties did not act in a manner consistent with an independent contractor
relationship.50
Therefore, the court found the intent between the taxpayer and the worker did not reflect the
reality of their relationship, and thus characterized the worker on the factors alone. A review of
these recent cases demonstrates that the introduction of intent into the characterization process
has given rise to several varied approaches to characterizing a worker, which has resulted in
several inconsistent characterization approaches prescribed by the FCA.
5 The Inconsistency caused by Intent The problem with the introduction of intent into the characterization process is that over the last
decade, the courts’ treatment of intention has varied significantly creating substantial uncertainty
in tax litigation for both taxpayers and the tax authorities. The court’s frustration with the varied
approaches prescribed in previous FCA decisions was evident in Lang.51 After a review of the
relevant jurisprudence, Chief Justice Bowman of the TCC (as he then was) expressed his
frustration with the FCA’s inconsistent approach to the characterization process.
The court summarized the four approaches observed over time. The first approach observed was
that intention is irrelevant to the characterization process, as done in Wiebe Door. The second
approach observed was that intention served as a tiebreaker, as done in Wolf. The third approach
observed was that intention was determinative, as done in Royal Winnipeg Ballet. The final
50 Ibid, at para 25. 51 Supra, note 21.
21
approach observed was that the weight to be placed on intention depended on the circumstances.
Fortunately, all four approaches yielded the same result in Lang so the Chief Justice did not have
to decide among the four. However, his expressed frustration with the emergence of four
approaches from the FCA struggling to place common intention alongside the four objective
factors in the characterization process is reflective of the litigation uncertainty caused by the
introduction of intent. The magnitude of the impact of intent on the characterization process can
be determined through an empirical study.
22
Part 3 The Study
1 Inspiration from the Wintermute Study In his study, Wintermute52 conducted a trend analysis between 2000 and 2007 regarding the role
of intent in the characterization process. He took a random sample of 40 cases in each of the four
periods he studied: 2000-2002, 2003-2004, 2005-2006, and 2006-2007.53 He characterized each
case based on whether the decision made reference to a common intention held by the
contracting parties and on the ultimate decision reached.54 He found that after Royal Winnipeg
Ballet was decided, the courts’ reliance on contractual intent increased significantly.
2 Extrapolating Wintermute to the Present If Wintermute’s study were extrapolated to analyze the present case law from the TCC and FCA,
we find a marked emphasis on intent after Royal Winnipeg Ballet and a significant de-emphasis
on intent placed afterwards.
Following Wintermute, I selected a random sample of 40 cases in each of eight periods
demarked by the landmark cases from the FCA described above. I analyzed each case based on
the previous ruling, the ultimate decision, and reference to a common intention between the
contracting parties. However, I also analyzed the cases for incidents where the factors and
common intent pointed in opposite directions. In such cases, I recorded whether the factors or
common intent prevailed.
52 Kurt G Wintermute, “A Worker’s Status as Employee or Independent Contractor: Recent Case Law, Trends, and Planning,” Report of Proceedings of Fifty-Ninth Tax Conference, 2007 Tax Conference (2008) 34:1. 53 Ibid, page 16. 54 Ibid.
23
Table 1: Findings based on Common Intent Era Common Intent of
Independent Contractor Found and Independent Contractor Found
Common Intent of Independent Contractor Found and Employee Found
January 1, 2000 to March 15, 2002 (Wolf) 50% 50% March 16, 2002 to May 3, 2004 (Tremblay) 73% 27% May 4, 2004 to March 2, 2006 (Royal Winnipeg Ballet) 71% 29% March 3, 2006 to October 24, 2007 (B-Pro Grooming) 86% 14% October 25, 2007 to October 29, 2008 (Kilbride) 57% 43% October 30, 2008 to December 14, 2009 (D.W. Thomas) 53% 47% December 15, 2009 to September 22, 2011 (TBT) 65% 35% September 23, 2011 to December 31, 2012 (Present) 50% 50%
As seen from Table 1, prior to Wolf and the introduction of intention as a consideration in the
characterization process, whether a common intent was found or not made no difference to the
court – the court was 50% likely to characterize a worker as an independent contractor and 50%
likely to characterize a worker as an employee.
Over the years however, that trend grew significantly one-sided up to March 2, 2006. This makes
sense because Tremblay said common intention could be implemented into the four-factor test
and specifically deployed common intention into the integration test. As seen from the review of
the case law above, taxpayers generally argue their intention to treat their workers as
independent contractors to avoid EI and CPP payments. A melding of intention and the four
factors explains the increased likelihood for a court to find in favor of that intention. The next
period shows an even greater likelihood of the court to find in favor of a common intent. After
Royal Winnipeg Ballet, which said a common intention was determinative of the characterization
of a worker, the trend shows the courts were 86% likely to find in favor of the intention held
between contracting parties. Subsequent courts then started to apply Royal Winnipeg Ballet as a
standard to assess the degree of factors indicating an employment relationship.
24
3 The “More Strongly than Royal Winnipeg” Standard During my review of the case law, I observed an interesting emergence of a standard used by the
courts in the characterization process – the “more strongly than Royal Winnipeg Ballet” standard.
This standard demonstrates the effect the FCA’s Royal Winnipeg Ballet decision had on the
characterizations of a worker in the subsequent jurisprudence. On September 13, 2007, the TCC
decided Maxwell C. Bishop o/a Ultra-Max Construction v. MNR.55 In Maxwell, the taxpayer
carried out a general construction contracting business and hired five workers to assist. The
Minister argued the workers were employees and that the taxpayer owed EI and CPP payments
for the relevant taxation years. The taxpayer argued the workers were independent contractors. In
its decision the court employed a “more strongly than Royal Winnipeg Ballet” standard and said:
The application of the facts to the Wiebe Door factors in Royal Winnipeg Ballet was not sufficient
to alter the arrangement from that which was intended by the parties. Therefore, unless the
application of the facts in this case to the Wiebe Door factors would more strongly indicate an
employer-employee relationship than in the case of the Royal Winnipeg Ballet, it seems to me
that the individuals, for whom the intention was to create an independent contractor relationship,
would be independent contractors.56 (emphasis added)
The court found the contracting parties mutually intended the workers were independent
contractors. The court then proceeded to determine if the application of the Wiebe Door factors
could alter this intention. The court found that control, ownership of tools, or financial risk, did
not suggest an employer-employee relationship “more strongly than did the facts in Royal
Winnipeg Ballet.”57 Thus, the court found the workers were independent contractors and allowed
the taxpayer’s appeal.
Interestingly, this “more strongly than Royal Winnipeg Ballet” standard was adopted by the
courts for exactly four years. On September 13, 2011, the TCC decided SIP Distribution Inc. v
55 2007 TCC 541, 2007 CarswellNat 2996. 56 Ibid, at para 9. 57 Ibid, at para 18.
25
MNR,58 and employed the “more strongly” standard in its characterization of a worker as an
employee or an independent contractor. In SIP Distributions, the taxpayer imported and sold
building materials and retained a worker to handle the marketing and administrative tasks. The
Minister characterized the worker as an employee and argued the taxpayer owed EI and CPP
payments. The taxpayer argued the worker was an independent contractor. The court found the
contracting parties mutually intended the worker was an independent contractor and deployed the
“more strongly than Royal Winnipeg Ballet” standard. The court said:
In the Royal Winnipeg Ballet case, the facts related to the dancers and the circumstances of their
work were not sufficient to alter the arrangement from that which was intended by the parties.
Therefore it seems to me that “in keeping with the approach set out in Royal Winnipeg Ballet”, the
relevant facts in this case, as determined by the factors as set out in Wiebe Door and Sagaz,
would have to more strongly indicate an employer-employee relationship than did the facts
in the case of the Royal Winnipeg Ballet in order for Erin Hrushowy to be considered to be an
employee. In both the Royal Winnipeg Ballet case and in this case, there was an intention to create
an independent contractor relationship and not an employer-employee relationship.59 (emphasis
added)
The court found the control and financial risk factors did not indicate an employment
relationship more strongly than that in Royal Winnipeg Ballet. The court found ownership of
tools was not useful in the analysis because very little equipment was required. Thus, the court
characterized the worker as an independent contractor and allowed the appeal.
After SIP Distributions, the “more strongly than Royal Winnipeg Ballet” standard was not
applied by the courts. However, referring to Table 1, during the standard’s four years in the
limelight, a reduced emphasis on intention can be observed. It is logical to find a reduced
likelihood of the courts to find in favor of intention after cases such as B-Pro Grooming and
Kilbride, which demoted the role of intention. Finally, a return to the 50% likelihood is observed
after TBT Personnel Services Inc., which characterized the relevant workers according to the
58 2011 TCC 423, 2011 CarswellNat 3519. 59 Ibid, at para 11.
26
objective factors alone. Although this trend reflects the history of the jurisprudence, I was
interested to learn whether the objective factors or subjective intent prevailed in cases where the
conclusions from each led in opposite directions.
Table 2: Court Findings based on Common Intent and Objective Factors Era Independent
Contractor Intent Overrides Factors
Independent Contractor Factors Override Intent
Employee Intent Overrides Factors
Employee Factors Override Intent
January 1, 2000 to March 15, 2002 (Wolf) 0% 55% 0% 95% March 16, 2002 to May 3, 2004 (Tremblay) 0% 39% 0% 95% May 4, 2004 to March 2, 2006 (Royal Winnipeg Ballet) 0% 62% 0% 100% March 3, 2006 to October 24, 2007 (B-Pro Grooming) 17% 25% 6% 81% October 25, 2007 to October 29, 2008 (Kilbride) 15% 60% 0% 95% October 30, 2008 to December 14, 2009 (D.W. Thomas) 5% 45% 0% 89% December 15, 2009 to September 22, 2011 (TBT) 8% 38% 0% 88% September 23, 2011 to December 31, 2012 (Present) 8% 25% 0% 100%
Based on Table 2, the objective factors override intent exclusively from January 1, 2000 to
March 2, 2006. However, when after Royal Winnipeg Ballet, we see the start of intent overriding
the objective factors. This makes sense because that case said intention was determinative.
Slowly, however, we see the incidence of intention overriding the objective factors reducing to
present day. However, it is interesting to find that even after TBT Personnel Services Inc. was
decided, there are still cases where intention has been followed over the objective factors. Cases
such as Zoltan can explain this trend, where the TCC has struggled to place a consistent weight
on intent in the characterization process.
Although this survey is not that sophisticated, it serves to preliminarily demonstrate a general
trend. However, for a more precise and reliable trend analysis, more sophisticated tools and
analysis must be implemented. I therefore expand my study to cover a greater period of time, a
greater number of cases, and a greater number of variables for a more complete and reliable
4 Methodology for Study I have conducted and analyzed a survey of all the decisions from the TCC and FCA
characterizing a worker as an employee or an independent contractor. I have expanded the
timeframe of my survey beyond my previous analysis to include the five years prior to Wolf in
order to gain a perspective of the relevant jurisprudence leading up to the introduction of intent
in the characterization process (i.e. from October 14, 1997 to March 14, 2002)60. There are 540
decisions over the 15-year period I cover, which have been broken down into the following eras:
• October 14, 1997 to March 15, 2002 (pre-Wolf)
• March 16, 2002 to May 3, 2004 (Wolf to Tremblay)
• May 4, 2004 to March 2, 2006 (Tremblay to Royal Winnipeg Ballet)
• March 3, 2006 to October 24, 2007 (Royal Winnipeg Ballet to B-Pro Grooming)
• October 25, 2007 to October 29, 2008 (B-Pro Grooming to Kilbride)
• October 30, 2008 to December 14, 2009 (Kilbride to D.W. Thomas)
• December 15, 2009 to September 22, 2011 (D.W. Thomas to TBT Personnel Services)
• September 23, 2011 to December 31, 2012 (Post-TBT Personnel Services)
Using STATA 12.0’s probative regression function, I hope to find statistically significant results
that reveal any trends in the jurisprudence. Specifically, I set out to determine what statistically
significant factors contribute to the probability of the court characterization of a worker as either
an employee or an independent contractor. I have set out to establish how the relative emphasis
on intent has fluctuated over the past 15 years. I have also set out to determine if any other
factors contribute to the courts characterization of a worker. Therefore, I consider the following
variables and offer an explanation of their importance.
60 The first case I found dealing with the characterization process was W.A. Pacific Rim Co. v. MNR, 1997 TCC 942355, 1997 CanLII 96, decided on October 14, 1997.
28
4.1 The TCC or FCA’s Characterization (employee/independent contractor)
This is the responding or dependent variable in the regression. The court’s final characterization
of the worker will be extremely relevant to the entire analysis, as it will show how the courts
finding is affected by the controlled or independent variables. A dummy variable of 0 for
employee and 1 for independent contractor has been used. Dummy variables are useful because
trends can be detected when each variable is recorded in the same fashion. Thus, if I use 0s and
1s for all the variables, I am essentially comparing apples to apples rather than apples to oranges.
Several cases reported at the TCC level have been appealed to the FCA. I record both the TCC
and FCA decisions in my survey for a more complete analysis of the jurisprudence. I assume the
court’s ultimate characterization is dependent upon the following independent variables.
4.2 The Characterization of the MNR or TCC Below This variable will show which finding is appealed more often in the jurisprudence. A dummy
variable of 0 for employee and 1 for independent contractor has been used. Maintaining a
consistent dummy variable for the dependent variable and the independent variables will reveal
any trends after the regressions are conducted. Examining this variable may show a trend
regarding which decisions (either decisions that a worker is an employee or independent
contractor) have been appealed more so over time. As stated earlier, I have included decisions
from the TCC (appeals from decisions by the MNR) and decisions from the FCA (appeals from
decisions by the TCC) for a more complete picture of the characterization process in the
jurisprudence.
4.3 Finding of Common Intent (employee/independent contractor/none)
This factor is still highly relevant to the analysis. I record the courts’ finding on intent to
determine how heavily the courts rely on intent when characterizing a worker. The dummy
variable of 0 for employee, 1 for independent contractor, and 0.5 if common intent is not found
in the decision shall be used. Many times, the courts are presented with evidence of common
29
intent. I record this variable based on the courts’ finding regarding intent. If a common intention
is found, I have recorded the result. If a common intention is not found or not relied upon in the
decision, I record the result as neutral. The regressions will show the trend between this variable
and the courts’ ultimate characterization.
4.4 Finding Based on Factors (employee/independent contractor)
This variable will also be relevant to determine the emphasis placed on the factors over intent. In
their decisions, the courts find on balance which direction the Wiebe Door factors point to. The
dummy variable will be used (0 for employee, 1 for independent contractor, and 0.5 for neutral).
4.5 The Courts’ Finding After an Application of each of the Four Wiebe Door factors (Control/Financial Risk/Ownership of Tools)
This variable will be useful to determine which factor the courts most heavily weigh. The
dummy variable for each factor shall be used (0 for employee, 1 for independent contractor, and
0.5 for neutral or no consideration shall be used). I record the courts’ finding on each individual
factor to determine which factor is most heavily relied upon. I also hope to establish an ordering
of the factors for each time period and determine if that priority of factors changes over time.
4.6 Judge This variable will tell me which judges from the TCC or FCA are more likely to characterize a
worker as an employee or an independent contractor. This will indicate how widespread the
positions of bench are. The names of the judges will be recorded for the purposes of this
variable.
4.7 Gender of the Judge The gender of the judge may reveal a trend where either male judges or female judges from the
TCC or the FCA are more likely to characterize a worker one way or the other. A dummy
variable will be used – 0 for a male judge and 1 for a female judge.
30
4.8 Province or Territory As the TCC and FCA sit in each province and territory, analyzing this variable may show a trend
that indicates workers from certain provinces are more likely to be characterized one way over
the other. The name of the province will be recorded for the purposes of this variable.
4.9 Personal Service Business (PSB) PSBs (or incorporated employees) are used by taxpayers for the ability to deduct more expenses
from income than normally allowed. However, there are anti-avoidance rules in place that
depend on the characterization process between employee and independent contractor to prevent
PSBs. Therefore, this variable may show the courts’ treatment of PSBs in an effort to prevent tax
avoidance. A dummy variable will be used – 1 for found, 0 for not found, and 0.5 if PSBs were
not relevant to the decision.
4.10 Gender of the worker (male/female) The gender of the worker may indicate a trend, or it may not. Thus, I will record the gender just
in case. A dummy variable of 0 for male and 1 for female will be used.
4.11 Nature of Organization (profit/non-profit) The courts are cognizant of the financial restraints on non-profit organizations. The courts’
finding may be affected in this regard. Therefore, to examine this trend I will record the nature of
the organization in this regard using a dummy variable of 0 for profit and 1 for non-profit.
4.12 Industry of Organization A trend may emerge regarding the industry of the organization. Therefore, I will record the
industry of the organization the worker provided services to. The cases I have surveyed can be
placed in the following industries:
31
Accounting Information technology/telecommunications Arts and entertainment Insurance Auto repair Legal Broadcasting/marketing Manufacturing Church/charity Native affairs Cleaning Natural resources (lumber, oil and gas, etc.) Construction/landscaping Property management Consulting Restaurant/bar Education Retail/sales Employment placement service Safety/security Engineering Salon/spa Finance Software development Fitness/recreation Taxi service Food processing/farming Technical service and repair Health services/residential care Trucking/delivery
After conducting regressions on all of these variables, I not only hope to determine how the
reliance on intent and the Wiebe Door factors have shifted over time, but also if there are any
other factors that affect the courts’ finding in any statistically significant way.
5 Assumptions In my approach, I have surveyed the 540 cases over the last 15 years for the purposes of
identifying any discernable trends, specifically with regards to the emphasis or de-emphasis on
intention as an analytical factor used by the courts. Grouping the last 15 years of jurisprudence
on the characterization of a worker as either an independent contractor or employee in this
manner requires certain assumptions to be made.
I assume that both taxpayers and the CRA litigate the same issue in the same way over time even
though the law gradually shifts. I believe this is a reasonable assumption to make for the
purposes of this survey because I separate my analysis of the 15 years of case law by eight
separate eras. In each of these eras the legal analysis used by the courts changes. Different
trends, such as the “more strongly than Royal Winnipeg Ballet” standard, emerge in the case law
during these eras. The purpose of my survey is to quantify the effects of these trends on the
outcome ultimately reached by the courts.
32
I also ignore whether the taxpayers or the CRA change their litigation strategy, such as the
evidence called or the arguments used, over time. I have narrowed my survey of the case law to
only the characterization process used by the courts. I believe this is also reasonable. Courts will
base their ultimate characterization of a worker on the findings of fact and application of the
prevailing jurisprudence to those facts. This approach remains constant. What does change is the
characterization analysis the court employs and thus, this is the target of my survey.
6 How to Interpret the Results I have used STATA 12.0 to run the regressions on the 540 cases I gathered. Characterizing each
case based on the variables previously described, I was able to run regressions on different
combinations of variables to determine if any had a statistically significant effect on the courts’
characterization of a worker. However, before I present and interpret the results, I think it is
helpful to first demonstrate how results from the output generated by STATA are presented in
the Appendix and how they should be interpreted.
Consider the following example. If I want to examine how the outcome of a case, represented by
X, was affected by the independent variables A, B, C, D, E, F, G, and H, I would record each
variable in a spreadsheet. I would take each case, record the X value, and then record the value
of A, B, C, D, E, F, G, and H gathered from each case’s decision. STATA is then able to run
regressions on that data set. From the output generated, any trends STATA detects are then
displayed in the output. As an example, consider Table 3. Table 3 provides an example output of
data from a regression.
33
Table 3: Example Variable
Coefficient z P-value 95% Confidence Interval
A 0 (omitted) B 0.010
(0.179) 0.06 0.954 -0.340 0.361
C 0.397 (0.174)
2.27 0.023 0.054 0.740
D 0.403 (0.198)
2.03 0.042 0.014 0.791
E 0.070 (0.228)
0.31 0.759 -0.377 0.517
F 0.132 (0.237)
0.56 0.576 -0.332 0.598
G 0.272 (0.267)
1.02 0.308 -0.251 0.796
H 0.055 (0.179)
0.31 0.758 -0.296 0.407
constant -0.132 (0.113)
-1.17 0.241 -0.355 0.089
As can be seen in Table 3, each of the independent variables I chose to consider have been
displayed. I have run the “probit” regression using all the variables to determine how the
probability of X changes based on changes in the value of all the independent variables. Each
variable has been listed, along with a constant value, for the regression equation. STATA detects
trends by using certain variables it determines are appropriate to use as a base. In Table 3,
STATA has determined that A is an appropriate base from which to detect trends that variables B
through H have on the probability of X’s outcome. The coefficient, standard error in parentheses,
z-score, P-value, and 95% confidence interval is generated by STATA. The standard error of
each coefficient is used to calculate the z-score, which in turn is used to calculate the P-value.
It is inappropriate to interpret results that are not statistically significant. From the outset, it was
determined that statistically significant results would have a P-value of less than 0.05. The
convention is to only consider results with a P-value of less than 0.05. I have thus set the level of
significance to 5%. Doing so eliminates results that are considered extraordinary. In other words,
results with P-values grater than 0.05 are observed by chance, and are not statistically significant.
Significant results (those with P-values less than 0.05) are in blue.
The coefficient of each variable indicates by how much the probability of X will change with
one-unit change of the observed variable. Thus, the probability of X of changing (if C is
34
increased by one unit) is 0.397, or 39.7%. However, it is inappropriate to interpret the results
with this much certainty, and therefore, one must also consider the 95% confidence interval.
The 95% confidence interval provides an upper and lower limit on the coefficient to 95%
confidence. Considering the variable C, using A as a base, I am 95% confident that with one unit
increase in C, the probability of X will increase between 0.054 and 0.740, or between 5.4% and
74%. These principles should be kept in mind when interpreting the following results. After
running the regressions, these are the results I encountered.
7 Results and Interpretation The following tables and those in the Appendix show the affects of the listed independent
variables on the responding/dependent variable, which is the courts’ characterization of a worker
as an employee or an independent contractor. Table 4 shows the probability of the courts’ finding
in each of the eight eras described above.
Table 4: The Probability of the Courts’ Characterization by Era Era
Coefficient z P-value 95% Confidence Interval
Pre-Wolf 0 (omitted) Wolf-Tremblay 0.010
(0.179) 0.06 0.954 -0.341 0.362
Tremblay-Royal Winnipeg Ballet 0.398 (0.175)
2.27 0.023 0.055 0.740
Royal Winnipeg Ballet-B-Pro Grooming
0.403 (0.198)
2.03 0.042 0.015 0.792
B-Pro Grooming-Kilbride 0.070 (0.228)
0.31 0.759 -0.378 0.518
Kilbride-D.W.Thomas 0.133 (0.238)
0.56 0.576 -0.333 0.599
D.W. Thomas-TBT Personnel Services 0.273 (0.267)
1.02 0.308 -0.251 0.796
Post-TBT Personnel Services 0.055 (0.180)
0.31 0.758 -0.297 0.408
Constant -0.132 (0.113)
-1.17 0.241 -0.355 0.089
As can be seen in Table 4, the statistically significant results are in the eras between Tremblay
and B-Pro Grooming. With 95% confidence and using the Pre-Wolf era as a base, between
Tremblay and Royal Winnipeg Ballet, the courts were between 5% and 74% more likely to find a
worker was an independent contractor. That range rose after Royal Winnipeg Ballet was decided.
35
Between Royal Winnipeg Ballet and B-Pro Grooming, the courts were between 1.5% and 79%
more likely to characterize a worker as an independent contractor.
Table 5: The Courts’ Emphasis on Intent by Era Era x intent (interaction variable)
Coefficient z P-value 95% Confidence Interval
Pre-Wolf 0.004 (0.201)
0.02 0.982 -0.390 0.399
Wolf-Tremblay 0.198 (0.266)
0.74 0.456 -0.323 0.720
Tremblay-Royal Winnipeg Ballet 0.269 (0.214)
1.26 0.208 -0.150 0.690
Royal Winnipeg Ballet-B-Pro Grooming 0.675 (0.248)
2.72 0.007 0.188 1.163
B-Pro Grooming-Kilbride 0.128 (0.323)
0.40 0.692 -0.506 0.762
Kilbride-D.W.Thomas 0.253 (0.300)
0.84 0.398 -0.335 0.842
D.W. Thomas-TBT Personnel Services 0.399 (0.324)
1.23 0.219 -0.237 1.036
Post-TBT Personnel Services 0.237 (0.232)
1.02 0.307 -0.218 0.693
constant -0.146 (0.124)
-1.18 0.238 -0.391 0.097
Using an interaction variable, any trends in each era between variables can be detected. Table 5
shows how intent interacted with each era described. Between Royal Winnipeg Ballet and B-Pro
Grooming, the courts’ finding was between 19% and 116% dependent on the common intent
held between the contracting parties.
Table 6: The Courts’ Emphasis on Financial Risk by Era Era x Financial Risk (interaction variable)
Coefficient z P-value 95% Confidence Interval
Pre-Wolf 1.864 (0.240)
7.75 0.000 1.393 2.336
Wolf-Tremblay 2.028 (0.267)
7.57 0.000 1.503 2.553
Tremblay-Royal Winnipeg Ballet 2.912 (0.367)
7.93 0.000 2.192 3.631
Royal Winnipeg Ballet-B-Pro Grooming 3.202 (0.510)
6.27 0.000 2.202 4.203
B-Pro Grooming-Kilbride 2.254 (0.448)
5.02 0.000 1.374 3.134
Kilbride-D.W.Thomas 2.554 (0.587)
4.34 0.000 1.401 3.706
D.W. Thomas-TBT Personnel Services 2.402 (0.488)
4.92 0.000 1.445 3.360
Post-TBT Personnel Services 2.367 (0.340)
6.95 0.000 1.700 3.035
constant -0.998 (0.093)
-10.71 0.000 -1.181 -0.815
36
Table 7: The Courts’ Emphasis on Control by Era Era x Control (interaction variable)
Coefficient z P-value 95% Confidence Interval
Pre-Wolf 2.578 (0.260)
9.89 0.000 2.067 3.089
Wolf-Tremblay 3.038 (0.406)
7.48 0.000 2.242 3.835
Tremblay-Royal Winnipeg Ballet 3.233 (0.376)
8.58 0.000 2.495 3.972
Royal Winnipeg Ballet-B-Pro Grooming 3.166 (0.452)
6.99 0.000 2.279 4.053
B-Pro Grooming-Kilbride 2.88 (0.502)
5.74 0.000 1.897 3.866
Kilbride-D.W.Thomas 2.453 (0.443)
5.53 0.000 1.583 3.322
D.W. Thomas-TBT Personnel Services 3.37 (0.778)
4.34 0.000 1.849 4.898
Post-TBT Personnel Services 2.815 (0.336)
8.37 0.000 2.156 3.475
constant -1.286 (0.102)
-12.57 0.000 -1.487 -1.086
Table 8: The Courts’ Emphasis on Integration by Era Era x Integration (interaction variable)
Coefficient z P-value 95% Confidence Interval
Pre-Wolf 2.382 (0.300)
7.94 0.000 1.794 2.970
Wolf-Tremblay 2.413 (0.351)
6.86 0.000 1.724 3.103
Tremblay-Royal Winnipeg Ballet 3.178 (0.391)
8.11 0.000 2.410 3.946
Royal Winnipeg Ballet-B-Pro Grooming 3.124 (0.437)
7.14 0.000 2.267 3.981
B-Pro Grooming-Kilbride 2.240 (0.479)
4.67 0.000 1.301 3.180
Kilbride-D.W.Thomas 2.313 (0.501)
4.61 0.000 1.329 3.296
D.W. Thomas-TBT Personnel Services 2.755 (0.578)
4.76 0.000 1.620 3.889
Post-TBT Personnel Services 2.397 (0.380)
6.30 0.000 1.651 3.143
constant -1.167 (0.131)
-8.89 0.000 -1.424 -0.909
37
Table 9: The Courts’ Emphasis on Ownership of Tools by Era Era x Ownership of Tools (interaction variable)
Coefficient z P-value 95% Confidence Interval
Pre-Wolf 1.643 (0.248)
6.62 0.000 1.157 2.130
Wolf-Tremblay 2.036 (0.304)
6.69 0.000 1.439 2.633
Tremblay-Royal Winnipeg Ballet 2.299 (0.313)
7.33 0.000 1.684 2.914
Royal Winnipeg Ballet-B-Pro Grooming 2.379 (0.377)
6.30 0.000 1.639 3.119
B-Pro Grooming-Kilbride 1.765 (0.439)
4.02 0.000 0.904 2.626
Kilbride-D.W.Thomas 1.885 (0.535)
3.52 0.000 0.835 2.934
D.W. Thomas-TBT Personnel Services 2.785 (0.702)
3.97 0.000 1.408 4.162
Post-TBT Personnel Services 1.478 (0.268)
5.50 0.000 0.951 2.004
constant -0.808 (0.091)
-8.88 0.000 -0.986 -0.629
As seen in Tables 6 to 9, the courts in all the eras relied heavily on Control, Financial Risk,
Integration, and Ownership of Tools with statistical significance. These indicators are very
powerful given the P-values are 0.
Table 10: The Courts’ Emphasis in the Pre-Wolf Era Pre-Wolf Era (interaction variable)
Coefficient z P-value 95% Confidence Interval
Control 1.950 (0.361)
5.40 0.000 1.242 2.658
Intent -1.911 (0.344)
-5.56 0.000 -2.585 -1.237
Integration 0.869 (0.470)
1.85 0.064 -0.052 1.790
Financial risk 0.204 (0.396)
0.52 0.606 -0.572 0.982
Ownership of tools 0.290 (0.430)
0.67 0.500 -0.552 1.133
constant 0.007 (0.060)
0.12 0.906 -0.111 0.125
Though the courts in the Pre-Wolf era heavily relied upon Control when characterizing a worker,
the court hardly relied upon Intent. In Table 10, the coefficient of Intent is negative, which
means an increase in Intent reduced the likelihood of the court finding the worker was an
independent contractors.
38
Table 11: The Courts’ Emphasis in the Wolf to Tremblay Era Wolf to Tremblay Era (interaction variable)
Coefficient z P-value 95% Confidence Interval
Control 2.257 (0.571)
3.95 0.000 1.138 3.377
Ownership of tools 1.912 (0.685)
2.79 0.005 0.570 3.255
Financial risk 1.625 (0.550)
2.95 0.003 0.546 2.705
Intent -2.699 (0.627)
-4.30 0.000 -3.929 -1.468
Integration -1.057 (0.839)
-1.26 0.208 -2.703 0.588
constant 0.005 (0.058)
0.09 0.927 -0.108 0.119
As seen in Table 11, the courts have established a priority of factors they rely on in the
characterization process. The courts rely on Control, then Ownership of Tools, then Financial
Risk. The courts place the reliance on Intent and that factor actually has a negative impact on the
outcome of the characterization process.
Table 12: The Courts’ Emphasis in the Tremblay to Royal Winnipeg Era Tremblay to RWB Era (interaction variable)
Coefficient z P-value 95% Confidence Interval
Control 2.098 (0.577)
3.63 0.000 0.965 3.230
Financial risk 1.587 (0.774)
2.05 0.040 0.070 3.105
Intent -1.565 (0.497)
-3.15 0.002 -2.541 -0.590
Integration -0.453 (0.941)
-0.48 0.630 -2.299 1.392
Ownership of tools 0.487 (0.798)
0.61 0.542 -1.078 2.053
constant -0.052 (0.058)
-0.90 0.368 -0.167 0.062
As in the previous era, the courts prior reliance on Control is maintained in the Tremblay to
Royal Winnipeg Ballet era. The courts place a reduced emphasis on Financial Risk than in the
previous era. Finally, the courts still place very little reliance on intent, and the coefficient is still
negative. This means the court is more likely to characterize a worker in a manner opposite that
intended by the contracting parties. However, the coefficient on Intent is weaker in this era than
in the previous era.
39
Table 13: The Courts’ Emphasis in the Royal Winnipeg to B-Pro Grooming Era RWB to B-Pro Grooming Era (interaction variable)
Coefficient z P-value 95% Confidence Interval
Control 2.114 (0.666)
3.17 0.002 0.808 3.421
Financial risk 1.667 (0.801)
2.08 0.037 0.096 3.238
Integration -1.390 (0.861)
-1.61 0.106 -3.079 0.297
Ownership of tools -0.509 (0.807)
-0.63 0.528 -2.091 1.072
Intent -0.294 (0.432)
-0.68 0.496 -1.141 0.552
constant -0.037 (0.056)
-0.65 0.516 -0.148 0.074
As seen in Table 13, the only statistically significant results found are the courts’ reliance on
Control and then Financial Risk. Using these interaction variables, no statistically significant
reliance can be found on Intent in this era. However, if we look back at Table 5, the only
statistically significant reliance on Intent amongst all the eras, was found in the Royal Winnipeg
Ballet to B-Pro Grooming era. Thus, although Intent was relied upon most heavily in that era,
that reliance is outweighed by the courts’ reliance on Control and Financial Risk.
Table 14: The Courts’ Emphasis in the B-Pro Grooming to Kilbride Era B-Pro Grooming to Kilbride Era (interaction variable)
Coefficient z P-value 95% Confidence Interval
Control 3.791 (1.165)
3.25 0.001 1.507 6.076
Financial risk 2.234 (1.132)
1.97 0.048 0.015 4.453
Integration -1.950 (1.430)
-1.36 0.173 -4.754 0.853
Ownership of tools -1.436 (1.230)
-1.17 0.243 -3.848 0.975
Intent -1.684 (0.997)
-1.69 0.091 -3.639 0.269
constant 0.016 (0.055)
0.29 0.770 -0.093 0.125
As seen in Table 14, the courts have maintained their reliance on Control and Financial Risk. No
other statistically significant results were found regarding the courts’ reliance on the other
objective factors or Intent.
40
Table 15: The Courts’ Emphasis in the Kilbride to D.W. Thomas Era Kilbride to D.W. Thomas Era (interaction variable)
Coefficient z P-value 95% Confidence Interval
Control 1.620 (0.671)
2.41 0.016 0.304 2.937
Integration -1.348 (0.975)
-1.38 0.167 -3.259 0.563
Financial risk 1.083 (0.846)
1.28 0.201 -0.576 2.743
Ownership of tools 0.792 (0.904)
0.88 0.381 -0.980 2.565
Intent -0.676 (0.485)
-1.39 0.164 -1.628 0.275
constant 0.011 (0.055)
0.20 0.838 -0.097 0.120
Table 16: The Courts’ Emphasis in the D.W. Thomas to TBT Personnel Services Era D.W. Thomas to TBT Era (interaction variable)
Coefficient z P-value 95% Confidence Interval
Control 2.986 (1.516)
1.97 0.049 0.014 5.958
Integration -2.583 (1.689)
-1.53 0.126 -5.894 0.726
Financial risk 0.649 (0.891)
0.73 0.466 -1.097 2.397
Ownership of tools 2.057 (1.639)
1.25 0.209 -1.155 5.271
Intent -1.476 (1.265)
-1.17 0.243 -3.957 1.004
constant 0.005 (0.055)
0.11 0.915 -0.102 0.114
In the Kilbride to D.W. Thomas and the D.W. Thomas to TBT Personnel Services eras, the only
significant result was the courts’ reliance on Control.
Table 17: The Courts’ Emphasis in the Post-TBT Personnel Services Era Post-TBT Era (interaction variable)
Coefficient z P-value 95% Confidence Interval
Control 1.950 (0.361)
5.40 0.000 1.242 2.658
Intent -1.911 (0.344)
-5.56 0.000 -2.585 -1.237
Integration 0.869 (0.470)
1.85 0.064 -0.052 1.790
Financial risk 0.204 (0.396)
0.52 0.606 -0.572 0.982
Ownership of tools 0.290 (0.430)
0.67 0.500 -0.552 1.133
constant 0.007 (0.060)
0.12 0.906 -0.111 0.125
41
Interestingly, the significant results show the courts’ reduced reliance on Control as compared to
other eras. Also, the courts’ do not rely on Intent in this era. The results show that in this era, the
courts’ characterization of a worker is opposite that intended by the contracting parties. This is a
similar trend found in the eras before Royal Winnipeg Ballet was decided.
The remaining tables can be found in the Appendix. After a review of the industries and
according to Table 18, the only statistically significant result found was that for the Property
management industry. The results show that in that industry, a worker is between 1.32% and
147% likely to be found an independent contractor than an employee.
Taking all the judges into account in Table 19, no statistically significant results can be found.
Thus, no one judge is more likely to characterize a worker as either an independent contractor or
an employee. Separating the judges by their court, we see in Table 20, that no one judge from the
FCA characterizes a worker one way or the other in a statistically significant way. Examining the
TCC however, reveals an interesting result. According to Table 21, Justice Sheridan was found
more likely to find a worker as an independent contractor. With all due respect to Justice
Sheridan, the results show that she was approximately 80% more likely to characterize a worker
as an independent contractor than an employee than the omitted justices.
As seen in Table 22, the province of the hearing has no statistically significant bearing on the
courts’ characterization of a worker.
However, a very strong trend observed is where the court reverses. Either the TCC reverses the
characterization of the Minister, or the FCA reverses the characterization made by the TCC. In
either circumstance, when the court reverses the decision below, it is more likely to characterize
the worker as an independent contractor. According to Table 23, with 95% confidence, I can say
when the court reverses, it is between 146% and 195% more likely to characterize a worker as an
independent contractor than an employee.
Table 24 reveals a very interesting trend. Though the gender of the worker has no statistically
significant bearing on the characterization finding, the gender of the judge does. It seems that a
42
female judge between 5.5% and 56% more likely to find a worker is an independent contractor
than a male judge. Tables 25 and 26 confirm this.
According to Table 25, after isolating for the gender of the judge, we see that indeed, with 95%
confidence, a female judge is between 5.5% and 56% more likely to characterize a worker as an
independent contractor. The opposite is true for a male judge. According to Table 26, we also see
that with 95% confidence, a male judge is between 5.5% and 56% less likely to find a worker is
an independent contractor. Instead, a male judge is more likely to find a worker is an employee.
I wanted to see in which era the gender of the judge had the greatest effect on the
characterization process. In Table 27, isolating for each era and the gender of the judge, we find
that the gender of the judge had a statistically significant affect on the characterization of a
worker in the Tremblay to Royal Winnipeg Ballet era. With 95% confidence, a female judge was
between 1.55% and 111% more likely to find a worker was an independent contractor than an
employee.
As previously mentioned, the gender of the worker had no statistically significant influence in
the characterization process. For completeness, I checked that result against each era. According
to Table 28, the gender of the worker is not a statistically significant variable in any of the eras.
According to Table 29, the type of representation the taxpayer has, whether it by a lawyer or an
agent, or if the taxpayer is self-represented, has no statistically significant bearing on how the
worker will be characterized.
In the next part, I outline the implications of the observations described above.
43
Part 4 Implications
1 Analysis is Mostly Objective As we saw in Tables 6 through 9, there are very powerful statistical results that indicate the
courts’ reliance on the four objective factors. Control is also the most heavily relied upon factor
in the courts’ analysis. Interestingly enough however, is the courts’ reliance on integration
despite Chief Justice Bowman’s comments in Lang. There is still a powerful reliance on
integration found in all the eras over the last 15 years. Therefore, regardless of the fluctuation of
intent, the courts’ have maintained the characterization analysis as mostly an objective process.
2 Emphasis and De-emphasis of Intent Intent was negative in the eras before Royal Winnipeg Ballet. Also, the courts’ reliance on Intent
was only statistically significant in the Royal Winnipeg Ballet to B-Pro Grooming era. After this
era, no statistically significant reliance on intent was observed until the Post-TBT Personnel
Services era. Here we saw the coefficient of intent drop back down into the negative value. This
indicates that the courts have returned to a similar reliance on intent as observed before Royal
Winnipeg Ballet. This also means that the courts’ have returned to a more objective
characterization process. Intent was not significantly relied upon before Royal Winnipeg Ballet.
After Royal Winnipeg Ballet, intent was significantly relied upon and we also observed the
“more strongly” standard adopted by the courts. This was the case until TBT Personnel Services
was decided, after which we observed an abandonment of the “more strongly” standard and the
reduced reliance on intent. Thus, the emphasis on intent has more or less come and gone, and the
courts have returned to a more objective characterization process. Intent still remains a
consideration, though the amount of weight given to it has dropped dramatically in the Post-TBT
Personnel Services era.
44
3 Property Management None of the industries revealed any statistically significant trend, except for Property
management. The court is more likely to find a worker in Property management is an
independent contractor, rather than an employee. This means that unless the worker is in
Property management, the industry will have no statistically significant bearing on the
probability of the courts’ characterization of a worker.
4 Reversed Decisions As seen in Table 23, if the court reverses the decision made below it, the court is likely to find a
worker is an independent contractor, rather than an employee. This means that the previous
characterization was that of an employee. Thus, on appeal, if the court reverses, the
characterization is then that of an independent contractor.
5 Gender of Judge Amongst all the factors I considered, such as the province, adjudicative forum, type of
representation, gender of the judge, gender of the worker, consideration of PSBs, and whether
the taxpayer was in the profit or non-profit sector, only the gender of the judge showed a
statistically significant result. The gender of the worker does not yield a statistically significant
result, but the gender of the judge does.
As found in Tables 24 to 27, female judges are more likely to find a worker is an independent
contractor and male judges are more likely to find a worker is an employee. Also, as seen in
Table 27, this trend is only statistically significant in the Tremblay to Royal Winnipeg Ballet era.
After the courts decided Royal Winnipeg Ballet and said intent was determinative of the nature of
the relationship between contracting parties, no statistically significant results regarding the
gender of the judge were observed.
45
I can only say with 95% confidence that a female judge is between 5% and 56% more likely to
find a worker is an independent contractor. However, it was an interesting trend I found in the
540 cases I analyzed.
46
Part 5 Conclusions
After conducting my study, my ultimate conclusion is that intent was prevalent in the courts as a
factor heavily relied upon in the characterization process after Royal Winnipeg Ballet was
decided. My study also revealed many other significant trends. To my surprise, the process over
the last 15 years has remained relatively objective. Only between Royal Winnipeg Ballet and B-
Pro Grooming was a spike in the courts’ reliance on intent observed. Observing the courts
keeping the analysis mainly objective was comforting.
Also comforting is the finding that most of the non-legal factors, such as province, forum,
industry, gender of the worker, and type of representation the taxpayer had, all yielded
statistically insignificant results. The courts are not influenced by such matters, nor should they
be. The courts should make their findings based on the merits of each case. And in large part
they do. The only observable trend I found that is disconcerting is that female judges are more
likely to find a worker is an independent contractor, and male judges are more likely to find a
worker is an employee. I cannot explain this trend. All I can say is that I observed this trend and
that the trend was statistically significant.
However, the most important part of this study was to discover the magnitude of the fluctuation
of the courts’ reliance on intent over the last 15 years of characterizing a worker. I found the
objective factors have been most heavily relied upon, with powerful statistical descriptors. I have
also found relatively few statistically significant trends with respect to intent. Other than the
spike after Royal Winnipeg Ballet and the decline after TBT Personnel Services, no other trends
regarding intent were observed. Therefore, it seems that while the courts placed a heavy reliance
on intent after Royal Winnipeg Ballet, the present era shows the courts have reduced their
reliance on intent to that of the Pre-Wolf era, before intent was introduced into the
characterization analysis.
In conclusion, I have observed the characterization process has evolved several times over the
last 15 years. Precedential case law has changed and the courts below have applied the law as it
47
has transformed. Different standards and trends have emerged and disappeared while others have
remained. This study has led to several interesting deductions regarding the role of intent in the
characterization process. It will be interesting to see how the courts move forward and establish
new eras in the case law for further studies to be conducted.
48
Bibliography Statutes
1. Income Tax Act, RSC 1985, c 1 (5th Supp), as amended.
25. Nabeel Peermohamed, “The Role of Intent in Characterizing a Worker” (2013) 3:1 Canadian Tax Focus.
26. Kurt G Wintermute, “A Worker’s Status as Employee or Independent Contractor: Recent Case Law, Trends, and Planning,” Report of Proceedings of Fifty-Ninth Tax Conference, 2007 Tax Conference (2008) 34:1.
50
Appendix: Additional STATA 12.0 Date Output Tables
51
Table 18: The Courts’ Characterization based on Industry Industry (trucking, 59, as a base) Coefficient z P-value 95% Confidence Interval Accounting 0.274
(0.590) 0.47 0.642 -0.881 1.431
Arts and entertainment 0.251 (0.315)
0.83 0.426 -0.366 0.869
Auto repair -0.653 (0.425)
-1.53 0.125 -1.487 0.181
Broadcasting and marketing 0.178 (0.354)
0.50 0.615 -0.516 0.873
Church and charity -0.092 (0.412)
0.27 0.822 -0.901 0.715
Cleaning 0.095 (0.345)
0.28 0.783 -0.581 0.771
Construction and landscaping -0.007 (0.251)
-0.03 0.975 -0.500 0.484
Consulting 0.160 (0.338)
0.48 0.634 -0.502 0.824
Education 0.328 (0.287)
1.29 0.253 -0.235 0.892
Employment agency -0.544 (0.528)
-0.97 0.302 -1.579 0.490
Engineering 0.021 (0.428)
0.05 0.960 -0.818 0.861
Finance -0.232 (0.590)
-0.39 0.694 -1.388 0.924
Fitness -0.118 (0.338)
-0.31 0.726 -0.781 0.544
Food and farming 0.021 (0.337)
0.06 0.950 -0.640 0.682
Healthcare -0.005 (0.245)
0.10 0.982 -0.485 0.474
Information technology 0.303 (0.341)
0.89 0.374 -0.365 0.972
Insurance -0.409 (0.461)
-0.89 0.375 -1.314 0.495
Legal -0.503 (0.447)
-1.12 0.261 -1.380 0.374
Manufacturing -0.129 (0.299)
-0.43 0.665 -0.717 0.458
Native affairs -0.946 (0.630)
-0.99 0.133 -2.181 0.289
Natural resources 0.510 (0.365)
1.44 0.163 -0.206 1.226
Property management 0.742 (0.372)
2.00 0.046 0.013 1.472
Restaurant and bar -0.409 (0.461)
-0.89 0.375 -1.314 0.495
Retail 0.263 (0.246)
1.07 0.285 -0.219 0.746
Safety and security -0.158 (0.503)
-0.32 0.753 -1.145 0.828
Salon and spa -0.544 (0.390)
-1.39 0.163 -1.310 0.221
Taxi 0.757 (0.417)
1.82 0.069 -0.060 1.575
Technician -0.481 (0.398)
-1.21 0.228 -1.263 0.300
Trucking 0 (omitted) constant -0.021
(0.163) -0.13 0.896 -0.341 0.298
52
Table 19: The Courts’ Characterization based on the Judges Judge Coefficient z P-value 95% Confidence Interval Angers 0.475
Table 21: The Courts’ Characterization based on the TCC Judges TCC Judge Coefficient z P-value 95% Confidence Interval Archambault 0 (omitted) Beaubier 0.706
Table 22: The Courts’ Characterization based on the Provinces/Territories Province/Territory Coefficient z P-value 95% Confidence Interval Alberta -0.307
Table 24: The Courts’ Characterization based on Gender Gender Coefficient z P-value 95% Confidence Interval Judge 0.308
(0.129) 2.39 0.017 0.055 0.561
Worker -0.044 (0.121)
-0.37 0.715 -0.283 0.194
constant -0.039 (0.073)
-0.54 0.586 -0.183 0.103
Table 25: The Courts’ Characterization if a Female Judge Presides Gender Coefficient z P-value 95% Confidence Interval Female Judge 0.308
(0.129) 2.38 0.017 0.054 0.560
constant -0.054 (0.061)
-0.88 0.378 -0.174 0.066
Table 26: The Courts’ Characterization if a Male Judge Presides Gender Coefficient z P-value 95% Confidence Interval Male Judge -0.308
(0.129) -2.38 0.017 -0.560 -0.054
constant 0.253 (0.113)
2.23 0.025 0.031 0.475
57
Table 27: The Courts’ Characterization based on the Gender of the Judge by Era Era (interaction variable with judge gender)
Coefficient z P-value 95% Confidence Interval
Pre-Wolf -0.085 (0.423)
-0.20 0.840 -0.915 0.745
Wolf-Tremblay -0.059 (0.383)
-0.16 0.876 -0.812 0.692
Tremblay-Royal Winnipeg Ballet 0.566 (0.280)
2.02 0.044 0.015 1.116
Royal Winnipeg Ballet-B-Pro Grooming 0.527 (0.285)
1.85 0.064 -0.031 1.085
B-Pro Grooming-Kilbride 0.620 (0.506)
1.23 0.220 -0.371 1.612
Kilbride-D.W.Thomas 0.484 (0.379)
1.28 0.201 -0.258 1.228
D.W. Thomas-TBT Personnel Services 0.620 (0.506)
1.23 0.220 -0.371 1.612
Post-TBT Personnel Services 0.054 (0.223)
0.24 0.808 -0.383 0.492
constant -0.054 (0.061)
-0.88 0.378 -0.174 0.066
Table 28: The Courts’ Characterization based on the Gender of Worker by Era Era (interaction variable with judge gender)
Coefficient z P-value 95% Confidence Interval
Pre-Wolf -0.115 (0.218)
-0.53 0.598 -0.542 0.312
Wolf-Tremblay -0.117 (0.245)
-0.48 0.631 -0.598 0.363
Tremblay-Royal Winnipeg Ballet 0.112 (0.279)
0.40 0.688 -0.436 0.660
Royal Winnipeg Ballet-B-Pro Grooming 0.518 (0.293)
1.77 0.077 -0.056 1.093
B-Pro Grooming-Kilbride -0.349 (0.327)
-1.07 0.286 -0.992 0.293
Kilbride-D.W.Thomas 0.163 (0.409)
0.40 0.691 -0.640 0.966
D.W. Thomas-TBT Personnel Services -0.554 (0.486)
-1.14 0.254 -1.508 0.398
Post-TBT Personnel Services -0.169 (0.273)
-0.62 0.537 -0.706 0.367
constant 0.028 (0.067)
0.42 0.672 -0.103 0.159
Table 29: The Courts’ Characterization based on Type of Representation Coefficient z P-value 95% Confidence Interval Representation (Lawyer, Self, or Agent) 0.025