Problem sets & exercises

Problem sets

Two tracks of practice. The analytical problem sets are pencil-and-paper: derivations, probability, reading regression output. The empirical exercises follow Stock & Watson's own numbering and come with a dataset and a Stata .do solution. Work the question first; solutions open after each deadline.

All problem sets and exercises are open now. Every question and dataset below is available for the whole course, not just the current week, so you can work ahead. Solutions and .do files are released as each lecture is reached.

Analytical problem sets

Problem Set 1 · S&W 1–3 · Lectures 1–2

Review of Probability & Statistics

Random variables, expectation and variance, conditional means, the sampling distribution, estimators and hypothesis tests.

Questions pdf Solutions

Problem Set 2 · S&W 4 · Lecture 3

Linear Regression with One Regressor

The OLS estimator, fitted values and residuals, R² and the standard error of the regression, the least-squares assumptions.

Questions pdf Solutions

Problem Set 3 · S&W 5 · Lecture 4

Hypothesis Tests and Confidence Intervals

Inference on the slope, the binary regressor as a difference in means, heteroskedasticity-robust standard errors, Gauss–Markov.

Questions pdf Solutions

Problem Set 4 · S&W 6 · Lecture 5

Multiple Regression and Omitted-Variable Bias

The two conditions and sign of omitted-variable bias, ceteris paribus, adjusted R², multicollinearity and control variables.

Questions pdf Solutions

Problem Set 5 · S&W 7 · Lecture 6

Hypothesis Tests in Multiple Regression

Tests on one coefficient and joint tests, the F-statistic, restrictions across coefficients, specification and confidence sets.

Questions pdf Solutions

Problem Set 6 · S&W 10 · Lecture 7

Panel Data and Fixed Effects

Between and within variation, entity and time fixed effects, what fixed effects can and cannot remove, clustered standard errors.

Questions pdf Solutions

Problem Set 7 · S&W 15 · Lecture 9

Dynamic Causal Effects & Time-Series Regression

Distributed lags and the impulse response, dynamic and cumulative multipliers, HAC standard errors, exogeneity.

Questions pdf Solutions

Problem Set 8 · S&W 16 · Lecture 10

Volatility, VAR and Cointegration

Volatility clustering and ARCH/GARCH, Value-at-Risk and the leverage effect, vector autoregressions, unit roots and cointegration.

Questions pdf Solutions

All problem sets in one file. Prefer a single booklet? the combined booklet is released alongside the lectures. The combined solutions are released at the end of the course.

Empirical exercises · Stata

These are Stock & Watson's own empirical exercises (numbered Echapter.exercise). Each names a dataset and asks you to run the analysis; the Stata .do file is the worked solution and opens after the deadline. The same data also drives the Python and Stata labs on the Modules page.

E3.1 · S&W 3 · Lecture 2

Distribution of average hourly earnings

Data: CPS96_15.csv — means, dispersion and the gender gap across years.

Exercise Data Solutions .do

Lab 2 · S&W 4.1, 4.2, 5.1, 5.2 · Lectures 3–4

Regression: estimation and inference (all four in one file)

Data: Growth.csv and Earnings_and_Height.csv. Worked solutions are the Lab 2 Stata do-file and Python notebook on the Modules page.

Exercises Growth Earnings & height

E4.1 · S&W 4 · Lecture 3

Economic growth and trade

Data: Growth.csv — the growth–trade slope and the Malta outlier.

Exercise Data Solutions .do

E4.2 · S&W 4 · Lecture 3

Earnings and height

Data: Earnings_and_Height.csv — does height predict earnings, and why.

Exercise Data Solutions .do

E5.1 · S&W 5 · Lecture 4

Earnings and height — inference

Data: Earnings_and_Height.csv — confidence intervals and robust standard errors.

Exercise Data Solutions .do

E5.2 · S&W 5 · Lecture 4

Growth and trade — inference

Data: Growth.csv — testing the trade-share coefficient.

Exercise Data Solutions .do

E6.1 · S&W 6 · Lecture 5

Birthweight and smoking

Data: birthweight_smoking.csv — adding controls and signing the bias.

Exercise Data Solutions .do

E6.2 · S&W 6 · Lecture 5

Growth and trade — multiple regressors

Data: Growth.csv — controls and the ceteris paribus slope.

Exercise Data Solutions .do

E7.1 · S&W 7 · Lecture 6

Birthweight and smoking — joint tests

Data: birthweight_smoking.csv — the F-test on a group of controls.

Exercise Data Solutions .do

E7.2 · S&W 7 · Lecture 6

Earnings and height — joint tests

Data: Earnings_and_Height.csv — joint significance and the overall F.

Exercise Data Solutions .do

Time-series & panel exercises · Stata & Python

Stock & Watson's empirical exercises for the panel and time-series chapters (S&W 10, 14–16), reproduced on the course datasets. The S&W variable names map to the FRED columns in our files: the PCE price index PCEP is PCECTPI, the CPI is CPIAUCSL, industrial production IP is INDPRO, the oil price for the shock Oₜ is WPU0561, and real GDP is GDPC1.

E10.1 · S&W 10 · Lecture 7

Concealed-weapons laws and violent crime

Data: Guns.csv — a state panel; fixed effects and clustered standard errors on the "shall-issue" effect.

Exercise Data Solutions

E10.2 · S&W 10 · Lecture 7

Income and democracy

Data: income_democracy.csv — is democracy a normal good? Country panel with clustered SEs and fixed effects.

Exercise Data Solutions

E14.1 · S&W 14 · Lecture 8

Forecasting the rate of inflation

Data: us_macro_quarterly.csv (PCEP=PCECTPI) — AR models, BIC vs AIC, the ADF unit-root test, pseudo out-of-sample forecasts.

Exercise Data Solutions

E14.2 · S&W 14 · Lecture 8

Can you beat the market?

Data: Stock_Returns_1931_2002.csv — return predictability and the dividend yield; replicate Tables 14.2 and 14.6.

Exercise Data Solutions

E15.1 · S&W 15 · Lecture 9

Oil prices and industrial production

Data: us_macro_monthly.csv (IP=INDPRO, oil WPU0561) — a distributed-lag model with HAC standard errors; dynamic and cumulative multipliers.

Exercise Data Solutions

E15.2 · S&W 15 · Lecture 9

CPI versus PCEP inflation

Data: us_macro_quarterly.csv (CPI=CPIAUCSL, PCEP=PCECTPI) — substitution bias, a constant-only regression and HAC inference.

Exercise Data Solutions

E16.1 · S&W 16 · Lecture 10

Iterated and direct multiperiod forecasts

Data: us_macro_quarterly.csv — iterated vs direct inflation forecasts and the DF-GLS unit-root test (extends E14.1).

Exercise Data Solutions

E16.2 · S&W 16 · Lecture 10

GDP-growth volatility and the Great Moderation

Data: us_macro_quarterly.csv (GDP=GDPC1) — an AR(2) model with GARCH(1,1) errors and the post-1983 fall in volatility.

Exercise Data Solutions

Shanghai · China Course home Teaching CV