This repository contains a local-data Python pipeline testing whether Canadian wage growth behaves more like a delayed response to housing and shelter-driven cost pressure than an initiating wage-led inflation channel.
The empirical chain tested is:
mortgage rates / credit conditions
-> house prices / land values
-> CPI shelter / rent / mortgage-interest pressure
-> delayed wage adjustment
land_housing_wage_pipeline.py - end-to-end extraction, cleaning, analysis, plotting, and report generation.data_clean/ - cleaned monthly and quarterly analysis panels.tables/ - lead-lag scans, Granger direction tests, regressions, robustness checks, and file audit.figures/ - generated matplotlib figures.report/land_housing_wage_channel_report.md - full markdown report.report/land_housing_wage_channel_article.html - screenshot-friendly article version for social posts.The raw StatCan ZIP exports and Bank of Canada CSV files are intentionally not committed because they are large local downloads. Place them in either ./data/ or ./.data/ before rerunning the pipeline.
python .\land_housing_wage_pipeline.py
The script creates or updates:
data_raw/data_clean/figures/tables/report/The reported results are about predictive content, forecasting sense, and lead-lag structure. They should not be read as structural causality.
In the generated run, the evidence is more consistent with a land/housing-cost transmission channel than with a wage-price spiral. Wages appear downstream of prior shelter and CPI pressure rather than as the initiating source of inflation.