AI Analysis: Employment insurance benefit characteristics by class of worker, monthly, unadjusted for seasonality

Executive Summary

Statistics Canada's Employment Insurance dataset (Table 14100007) spans nearly 30 years of monthly data (January 1997 – February 2026), tracking benefit payments and recipient counts across 18 income benefit types and 3 worker classes. Over this period, total EI benefit payments nearly doubled (+93.7%, from $1.27B to $2.46B), while benefit weeks paid out declined by 10.1%, signaling significantly higher average weekly benefit amounts driven by wage growth. The most dramatic anomaly in the dataset is the COVID-19 pandemic period (2020–2021), where emergency EI expansions caused benefit payments to spike far beyond historical norms.

Key Findings

• EI benefit payments grew by 93.7% over the study period, rising from $1.27 billion in January 1997 to $2.46 billion in February 2026, while benefit weeks simultaneously declined 10.1% from 4.9 million to 4.4 million weeks.
• The 'Regular and not a Part II Employment Benefit participant' income type dominates all categories with an average monthly value of $575 million — far exceeding all other benefit types.
• The dataset is heavily right-skewed: the mean value (~$168.9 million) is roughly 84 times higher than the median (~$2 million), driven by a few extremely large aggregate values reaching up to $5.1 billion.
• Self-employed workers account for a significantly smaller share of EI benefits, with an average value of ~$10.2 million compared to ~$159.2 million for employees and ~$242.7 million for all classes combined.
• The COVID-19 pandemic (2020–2021) produced the most extreme statistical outliers in the dataset's nearly 30-year history, with benefit payments spiking far beyond the 3×IQR threshold used to define anomalies.
• The dataset contains 58 unique time series vectors across 350 monthly periods, with correlation analysis across 18 benefit variables revealing clusters of highly correlated payment types that tend to rise and fall together over time.
• Seasonal patterns are consistently visible throughout the unadjusted data, with winter months historically associated with elevated EI claims — a recurring structural feature across nearly all worker categories and benefit types.