AI Analysis: Historical (real-time) releases of monthly retail sales, price, and volume

Executive Summary

Statistics Canada's Table 20100082 tracks monthly Canadian retail sales across 30 NAICS categories from January 2017 to December 2025, revealing a 38.9% surge in nominal retail sales over the period — though roughly half of that growth was driven by price inflation rather than real volume increases. The dataset's 36 release vintages enable revision analysis, while four distinct measures (current prices, constant prices, price index, and volume index) together confirm genuine real retail growth of approximately 18–19% alongside significant inflationary pressure. Only 4 months recorded extreme volatility (>10% month-over-month swings), most likely tied to COVID-19 disruptions in 2020.

Key Findings

• Nominal retail sales in current prices grew 38.9% from $50,417M in January 2017 to $70,013M in December 2025, but real volume growth was only ~18–19%, highlighting that roughly half the nominal gain was inflation-driven.
• The Fisher Chained Price Index rose 17.4% (from 100.6 to 118.1) over the period, confirming meaningful retail price inflation well above the 2012 base year of 100.
• The dataset contains 241,056 rows spanning 108 reference months and 36 historical release vintages (2023–2026), making it uniquely suited for studying how Statistics Canada revises its retail estimates over time.
• The value distribution is heavily right-skewed: the mean of 6,429 is nearly 3x the median of 2,270, driven by high-value aggregate categories pulling averages upward across the 200,484 usable records.
• No monthly retail sales totals fell outside the IQR-based outlier bounds ($34,796M–$84,887M), but 4 months recorded extreme month-over-month swings exceeding ±10%, most likely corresponding to COVID-19 disruptions in 2020.
• Across 30 NAICS classifications, there is a vast spread in average sales values — broad aggregates like 'Retail trade [44-45]' dominate at the top, while narrow sub-sectors like 'New car dealers [44111]' represent the lower-ranked categories.
• The Fisher Chained Price Index is the most stable measure in the dataset, with a standard deviation of just 6.19 and values tightly clustered between 98.7 and 119.8, contrasting sharply with current- and constant-price series that span ranges exceeding 70,000 million dollars.