PolicyEngine Macro
model 02 — Bank of England structural VAR · boe-svar · UK · hosted
Explain UK growth and inflation.
Decompose UK GDP and inflation into six structural shocks, forecast with credible bands, and explain revisions between quarters.
Estimate, identify, decompose.
Four routes into the model, hosted to local: MCP tools, CLI mirrors, a pip install, and the development runner scripts. Each step's code follows its narrative.
Hosted MCP tools
No install, no credentials
The quickest route is the hosted MCP server. Three tools reach this model: forecast_uk (YoY GDP-growth and CPI-inflation fan charts with 68%/90% bands), latest_shocks (posterior sign probabilities of each identified shock in the latest quarter), and model_summary (headline replication results — instant, no estimation).
Where to read the results
What each tool tells you is on the overview; the estimation and identification choices behind them are on the methodology page.
CLI mirrors
Same tools from the terminal
The CLI mirrors the three MCP tools one for one: pe-macro forecast, pe-macro shocks, and pe-macro summary. Same outputs, no MCP client required.
Install locally
One pip install
The package installs straight from GitHub with the prepared quarterly dataset inside, so boe_var.data.load_data() works immediately.
All inputs are public
Every input series is public: ONS (CPI, CPI energy, real GDP), Bank of England (Bank Rate, sterling ERI), and FRED (oil, exchange rates, world components).
Development runner scripts
Clone to rebuild everything
The runner scripts are for development — clone the repository to rebuild the dataset or regenerate the paper outputs.
What lands where
outputs land in results/ — the replication figures (fig2–fig6), the forecast figures (fig1, fig7–fig9), and two summary reports. runtime scales with draw count; the importance weights are the expensive part — one numerical Jacobian per accepted draw.