Models
Start with the macro question, then open the lightest model route that can answer it.
HelpQuestions first
A model approach is only useful after the question is clear. Use the cards below to route straight into the lightest model path that fits the macro problem you are actually trying to answer.
“Question first. Approach second. Job last.”
Recent runs
Recent empirical and DSGE work now carries across the model routes, so you can reopen the last lab instead of starting cold.
No model runs yet. Start an empirical or DSGE lab and the latest run will show up here.
Start with the lightest model that separates long-run capacity from short-run demand, then move to empirical forecasts if the question becomes near-run direction.
Use a structural inflation route when persistence, expectations, and policy transmission are the real question; keep a simpler framework nearby for first-pass intuition.
Labor questions usually need one model for slack and one model for timing. Start with a labor-macro framework, then move to forecast workflows if the signal is the job rather than the mechanism.
If the question is timing, expectations, and counterfactuals, move into a policy-capable structural route early rather than staying in a generic model chooser.
Use richer model routes when balance sheets, credit frictions, and feedback loops stop looking like side channels and start driving the macro outcome.
Approaches
Compact macro frameworks for tracing one mechanism at a time, with browse, explore, proof, and compare views built for teaching-scale analysis.
Data-Driven Models start from observed macro data: bring in a provider series or uploaded dataset, choose a model class, keep diagnostics visible, and move into the live baseline or broader guided routes.
Forward-looking structural systems used to formalize policy analysis, expectations, and intertemporal macro adjustment.
Bottom-up macro simulations for studying what emerges when heterogeneous households, firms, banks, and policymakers follow local rules and interact over time.
Jobs
Theory-Based Models
Stripped-down frameworks isolate one mechanism cleanly before richer estimation enters the picture.
Data-Driven Models
Data-driven routes are most useful when the task is baseline direction, live monitoring, diagnostics, and forecast comparison.
DSGE
Move toward models that can separate mechanism from counterfactual and make policy assumptions explicit.
Agent-based
Richer computational or heterogeneous-agent models matter when averages hide who is actually taking the shock.
