PDF Version [ 509 Kb / 14 pages | Download Adobe Reader ]

**See also:**

- Results of the Wage Gap Pilot Program Questions & Answers
- Wage Gap Program: Questions and Answers

The Pay Equity Office (PEO or the office) conducted the Wage Gap Pilot Program in 2011 with the following objectives:

- Outreach: to conduct outreach to private sector non-unionized employers with more than 250 employees
- Awareness: to raise awareness among these employers about their compensation practices and possible wage gaps
- Interest: to determine level of interest among these employers in their voluntary participation to examine compensation practices
- Opportunity to pilot PEO developed wage gap criteria: to field test PEO developed criteria for assessing the presence of an apparent wage gap

Apparent wage gaps were noted among employers who participated. The data submitted provides a snapshot of compensation practices at a point in time and is not representative of all private sector employers in Ontario.

The Pay Equity Office is responsible for enforcing the Pay Equity Act (R.S.O. 1990, c.P.7) (the Act). The office works to educate the public about the Act and investigates and resolves pay equity complaints. In addition, the office is empowered to conduct research into aspects of compensation practices.

The PEO commissioned a selection of private sector, non-unionized employers through Dun and Bradstreet. The following criteria were applied in the selection process:

- Location: Ontario
- Size: Corporate wide, 250+
- Industries: manufacturing, real estate, wholesale trade and finance insurance.

Employers whom the Office had monitored in the previous 11 years, those with open cases in the Office's database as well as federally regulated employers not covered by pay equity legislation in Ontario were filtered out of the final employer listing.

Surveys were mailed out to a total of 516 employers asking them to supply compensation data as of December 31, 2010:

- Job title/position
- Gender in that position
- Employees' current pay,
- Salary range (if applicable)
- Years of service

Five PEO developed criteria were applied to the incoming responses:

- Clustering of jobs: explored the grouping of women in stereotypical female jobs and alignment of jobs around wage/salary rates that might have been different depending on the gender of the incumbent;
- Natural breaks in clusters of jobs: examined the salary ranking of male and female jobs and how there were grouped within identified pay bands;
- Seniority vs. Job Rate: examined the job rate (highest rate of pay for identified jobs) against seniority (years of service) to identify regular patterns of women progressing at different rates compared to men;
- Distribution of jobs: examined the distribution of female jobs across the organization and how women were compensated compared to men within the same job title; and
- Job to job comparisons: compared male and female job titles that were estimated to possess similar skill, effort, responsibility and working conditions.

Of the 516 letters sent to employers requesting their participation, 420 (81%) responded with their compensation data. This high response rate underscores the importance of the PEO's outreach efforts and represents an encouraging level of engagement with respect to employers' interest in responding about their compensation practices.

Of the 420 employers who responded, 69 employers were excluded from further analysis for administrative reasons e.g. duplicate files. The remaining 351 respondents formed the sample for wage gap assessment.

Over two third of the respondents (68.5%) were from the manufacturing sector, reflective of the size of Ontario's largest employers.

Fifty-four percent (188 employers) were found to have an apparent wage gap based upon the review of their compensation data using the five PEO criteria (wage gap assessment).

The office conducted statistical analysis to examine the gender wage gap and to examine the influence of occupational tenure (years of service). To conduct this statistical analysis, the sample included all firms with an "apparent wage gap" and where there was missing information or miscoded observations, these firms were either adjusted or removed from the sample (for detailed information, please see Appendix 1 – Data). The resulting dataset is composed of 155 employers and 72,818 employees for analysis.

There are 30,402 (41.75%) female employees and 42,398 (58.22%) male employees in the sample. Employers have reported employee's current pay in multiple ways (hourly pay, weekly pay, annual pay or a combination of several pay frequencies).

Hourly pay is recorded for a total of 27,917 individuals. However, 10,649 of these also record an annual rate of pay, 17,268 individuals for whom only an hourly pay amount is recorded are used for the hourly analysis.

Figure 1 shows the distribution of hourly wages for male and female employees. As can be observed, female's hourly pay is concentrated at the lower range of the pay scale and male's hourly pay is concentrated at the higher range of the pay scale.

Approximately 20 percent of the male employees were paid less than $11 per hour while just over 30 percent of the female employees earned less than $11 per hour.

The average hourly pay of the 9,620 males in this group is $20.84 while the average hourly pay of the 7,648 females in this group is $16.49 (79.13% of average wage of males.)

The median (2^{nd} Quartile) is $18.40 for males and $14.39 for females (78.2% of median wage of males.) In the case of hourly pay, there is an average wage gap of $4.34 per hour.

Annual pay is recorded for 55,537 individuals in the sample.

Figure 2 shows the distribution of annual wages for both male and female employees. Similar to the case for hourly wages, female wages are more concentrated towards the lower end of the pay scale and the male wages are more concentrated towards the higher end of the pay scale.

The average male annual wage is $74,943.42 and the average female annual wage is $58,324.51 (77.82% of male's wage). In the case of annual pay, there is an average wage gap of $16,618.91.

A majority of this dataset (81%) met the criteria for seniority versus job rate, indicating a pattern of women progressing in their compensation at different rates when compared to men.

In order to further investigate the gender wage gap in this dataset, a pooled regression was conducted using years of service as an explanatory variable. Generally, in this approach, wage gap factors include type of job, education, age, job tenure and other experience. In this pilot project, the collection of individual characteristics was limited to occupational tenure and detailed job characteristics was not requested (aside from job titles). As a result, applying a gender wage gap decomposition model with a single variable will explain a small portion of the gender wage gap. However, the model is illustrative of the suggested approach for future pay equity research.

In the case of hourly pay, there is an average wage gap of $4.34 per hour. The average years of service for males is 8.06 while the average for females is 7.34 years. For an additional year of service, males earn $0.54 more per hour while females only earn $0.42 more. Differences in years of service alone can explain 8.9 percent of this gender wage gap.

In the case of annual pay, there is an average wage gap of $16,618.91. The average years of service for males is 10.34 years and the average years of service for females is 9.41 years. For an additional year of service, males earn $982.21 more while females only earn $572.87 more. Differences in years of service alone can explain 5.5 percent of this gender wage gap.

The Pay Equity Office Wage Gap Pilot Program has wielded successful results in meeting its primary objectives.

From an outreach perspective, the program accessed Ontario's largest private sector non-union employers and increased awareness on the role of the office and drew attention to reviewing compensation practices for impacts on gender wage gaps. The PEO will continue to focus on activities that raise awareness regarding compensation practices.

From an interest perspective, the degree of participation in providing firm compensation data was unprecedented. This resulted in a meaningful compensation dataset for statistical analysis and an opportunity to note considerations for future pay equity research.

The statistical analysis presented in this report has provided insight into Ontario's largest private sector non-union firms. We can see that in terms of occupational tenure, for every additional year of service, there is a noted difference in pay. More research on causes of Ontario's wage gap can provide a clearer picture.

From 188 total firms that have an apparent wage gap, 33 firms were excluded from the analysis for the following reasons:

- Neither hourly wage nor annual wage was available
- Years of service were unavailable
- Hourly or annual pay rate was recorded as a range
- Years of service was recorded as a negative or impossibly large value (fewest years of service was – 100 while longest was over 159 years)
- Hourly or annual wage was unusually high (sales representative's hourly wage higher than a CEO's hourly wage / annual wage over $1,000,000)

A standard approach used in economics for explaining the wage gap is the Blinder–Oaxaca decomposition for linear regression models.

The Blinder–Oaxaca decomposition begins by fitting linear regression equations for male and female wages:

w_{m} = α_{m}+ β_{m} YEARS_{m}+ ε_{m}

w_{ƒ}= α_{ƒ} + β_{ƒ} YEARS_{ƒ}+ ε_{ƒ}

The decomposition proceeds to explain the difference in the average male and average female wage by introducing an intermediate hypothetical average female wage that uses estimated male parameters with female average years.

¯w_{m} − w¯_{f} = w¯_{m} − w¯_{f}^{∗}+w¯_{f}^{∗} − w¯_{f} = explained + unexplained gap

¯w_{m} = α_{m} + β_{m} YEARS_{m}

w¯_{f} = α_{f} + β_{f} (YEARS)_{f}

w¯_{f}^{∗} = α _{m} + β_{m} (YEARS)_{f}

Simplifying shows that the explained part is given by

¯w_{m} − w¯_{f}^{∗} = α _{m} + β_{m} (YEARS) _{m} − (α _{m} + β_{m} (YEARS) _{f} ) = β_{m} (YEARS_{m} − YEARS _{f} )

For convenience, we can nest the two separate male and female wage equations into a single equation to simplify estimation as follows:

w_{i} = α_{m} + β _{m} YEARS_{i} + α_{f}−_{m}SEX_{i} + β_{f}−_{m}SEX_{i} YEARS_{i} + ε_{i}

In the computer estimation results, y represents the wage, x1 the years of service and x2 the sex (0=Male and 1 otherwise) for each of the individuals in the dataset.

The regression results for individuals with hourly wages are:

Table 1: Call |
---|

lm(formula = y ∼ x1 + x2 + x1 * x2) |

Table 2: Residuals | |
---|---|

Min | −21.939 |

1Q | −5.156 |

Median | −2.555 |

3Q | 2.061 |

Max | 264.224 |

Table 2a: Coefficients | ||||
---|---|---|---|---|

| Estimate | Std. Error | t value | Pr(>|t|) |

(Intercept) | 16.50135 | 0.14428 | 114.367 | <2e-16 *** |

x1 | −3.10172 | 0.21837 | −14.204 | <2e-16 *** |

x1:x2 | −0.11638 | 0.01985 | −5.862 | 4.66e-09 *** |

Signif. codes: 0 '***' 0.001 '*' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 10.27 on 17,264 degrees of freedom Multiple R-squared: 0.1646, Adjusted R-squared: 0.1645 F-statistic: 1,134 on 3 and 17,264 DF, p-value: < 2.2e-16 | | | | |

These results imply that:

>α_{m} = 16.50135,β_{m} = 0.53783,α_{f} − α_{m} = - 3.10172,β_{f}−β_{m} = −0.11638

α_{f} = α_{m}−3.10172 = 13.39963

β_{f} = β_{m}−0.11638 = 0.42145

Table 2b: Hourly Summary | |||
---|---|---|---|

Group | Size | Avg.Wage | Avg.Years |

Female | 7,648 | 16.49 | 7.34 |

Male | 9,620 | 20.84 | 8.06 |

Combined | 17,268 | 18.91 | 7.74 |

Thus, the explained portion of the hourly wage gap is given by:

β_{m} (YEARS_{m} − (YEARS) _{f} ) = 0.42145(8.06-7.34) = $0.30

(Note: This value of $0.30 is less than the $0.38 used in the text due to rounding of years of service.)

The regression results for individuals with annual wages are:

Table 3: Call |
---|

lm(formula = y ∼ x1 + x2 + x1 * x2) |

Table 3a: Residuals | |
---|---|

Min | −79,180 |

1Q | −20,904 |

Median | −7,881 |

3Q | 12,321 |

Max | 90,566.3 |

Table 3b: Coefficients | ||||
---|---|---|---|---|

| Estimate | Std. Error | t value | Pr(>|t|) |

(Intercept) | 64,782.67 | 294.29 | 220.13 | <2e-16 *** |

x1 | 982.21 | 21.77 | 45.12 | <2e-16 *** |

x2 | −11,850.51 | 458.26 | −25.86 | <2e-16 *** |

Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 34,300 on 55,533 degrees of freedom Multiple R-squared: 0.09153, Adjusted R-squared: 0.09148 F-statistic: 1,865 on 3 and 55,533 DF, p-value: < 2.2e-16 | | | | |

Table 4: Annual Pay Summary | ||||
---|---|---|---|---|

Group | Size | Avg.Wage | Average Years | |

Female | 22,766 | 58,324.51 | 9.41 | |

Male | 32,771 | 74,943.42 | 10.34 | |

Combined | 55,537 | 68,130.91 | 9.96 | |

Average annual wage difference of $16,618.91 | | | |

Table 5: Parameter Estimates Based on Regression Results | ||||
---|---|---|---|---|

Gender | α | β | Average Years | |

Male | 64,782.67 | 982.21 | 10.34 | |

Female | 52,932.16 | 572.87 | 9.41 |

β_{m} ((YEARS) _{m} − (YEARS) _{f} ) = 982.21(10.34-9.41) = $913.46

(Note: This value of $913.46 is less than the $915.34 used in the text due to rounding of years of service.)

[1] Statistical analysis of data and statistical findings provided by Dr. Amy Peng, Associate Professor, Department of Economics, Ryerson University.

ISBN 978-1-4606-5196-4 (HTML)

ISBN 978-1-4606-5197-1 (PDF)

ISBN 978-1-4606-5195-7 (PRINT)