Skip to main content

Interest rate – inflation – output interactions (VAR)

This illustrates VAR: analyzing the macro system dynamics among the policy rate, inflation and output — without imposing a priori causal direction. Figures are illustrative.

Summary: estimate a 3-variable VAR, read the impulse responses (IRF) and variance decomposition (FEVD) to see how an interest-rate shock propagates.


Step 1 — Ideation

  • Question: how does an interest-rate hike affect inflation and output over time?

Step 2 — Literature Review

Monetary policy, the transmission mechanism; macro VAR/SVAR.

Step 3 — Data Collection

VariableSymbolMeasurementSource
Interest rateratepolicy rate (%)central bank; IMF
Inflationinf% change in CPIGSO; World Bank
Outputlnylog real GDP / production indexGSO

Step 4 — Modeling

Check stationarity (difference if needed). Choose the Multivariate time series family → VAR; select lags by AIC/BIC:

Yt=c+A1Yt1++ApYtp+εt,Yt=(ratet,inft,lnyt)Y_t = c + A_1 Y_{t-1} + \dots + A_p Y_{t-p} + \varepsilon_t, \quad Y_t = (rate_t, inf_t, lny_t)'

Illustrative results (format — not real results):

AnalysisSample result
Optimal lag2 (by AIC)
IRF: rate ↑ shockinf falls after 2–4 periods; lny falls temporarily
Grangerrateinf (p < 0.05)
FEVD (lny, 10 periods)~25% due to interest-rate shock

Sample interpretation: a tightening shock (rate up) lowers inflation after a few periods and temporarily reduces output — consistent with monetary transmission. To interpret structural shocks, use SVAR.

* --- Macro VAR example ---
tsset time

* Estimate VAR(4)
var rate inflation output, lags(1/4)

* Granger causality
vargranger

* Impulse Response Functions
irf create myirf, set(irf_macro) step(20)
irf graph oirf, impulse(rate) response(inflation output)

* Variance Decomposition
irf graph fevd

Step 5 — Reporting

Export a report + IRF/FEVD plots + replication code.

Video tutorial

Video Tutorial: Running VAR in EcoLab

See also