Part I. Foundations
The decision framework and the research workflow.
Why this handbook. Applied economists rarely lack methods; they lack a defensible way to choose among them. This handbook starts from the question, not the estimator: which test or diagnostic tells you which model to use, what to do when the evidence is mixed, and how to write the result up. Every chapter takes one real dataset through that path (tests, estimation, diagnostics, robustness, and the text of the paper) with the code and its output shown, so you can reproduce each step.
The decision paths, the choice of tests and the design of the inference are based on my own learning, knowledge and implementation. Interactive versions of the guides are on the main website. The handbook is free to read and share.
Work in progress. This handbook is still being revised. Please check numerical results and references against the original sources before citing. If you find a mistake, a typo or an unclear passage, please report it to tuhin.hnu@gmail.com. Corrections are welcome and will be credited in later versions.
Part II. Time-Series Methods
OLS, VAR, VECM, ARDL, NARDL, SVAR, local projections, threshold and smooth-transition models, structural breaks, 2SLS and GMM.
Part III. Panel-Data Methods
Pooled OLS, fixed and random effects, GMM, Driscoll–Kraay, mean group, PMG, CS-ARDL, panel local projections and panel thresholds.
Part IV. Causal-Inference Designs
Instrumental variables, difference-in-differences, staggered DiD, synthetic control, generalized synthetic control, interrupted time series, regression discontinuity.
Part V. Machine Learning
LASSO, ridge, elastic net, random forest, gradient boosting, double machine learning and causal forests.
Part VI. Bayesian Methods
Bayesian regression, BVAR, steady-state VAR, FAVAR, mixed-frequency, panel VAR, stochastic-volatility, TVP and threshold VARs.
Part VII. Markov-Switching Models
MS-AR, MS regression, MS-VAR and MS panel VAR.
Part VIII. DSGE with Dynare
RBC, New Keynesian, small open economy, optimal policy, identification, Bayesian estimation, medium-scale model, zero lower bound, higher-order solutions.