RBI Grade B (DSIM) - Department of Statistics and Information Management · India · Reserve Bank of India Services Board
Original practice questions written to the published syllabus · not past papersDid the scheme, technology or variety cause the change? Potential outcomes, matching and weighting, difference-in-differences, instrumental variables, regression discontinuity and synthetic control, panel data models, quantile regression, and stochastic frontier and DEA efficiency analysis. Classical econometrics (OLS violations, simultaneous equations) is in the Agribusiness track and is assumed here.
In the exams: Not an explicit unit in AICE-JRF Ag Stat; AICE-JRF Ag Econ research methodology (present); RBI Grade B DSIM (reported); research and evaluation roles (applied)
RBI Grade B (DSIM) and Indian Statistical Service (reported)
Econometric identification: endogeneity and instrumental variables, panel data fixed and random effects with the Hausman test, and basic programme-evaluation designs.
How our questions for RBI Grade B spread across the units of Causal Inference, Panel Data & Efficiency Analysis. Difficulty: 21 easy, 40 medium, 16 hard.
| Randomised evaluation in agriculture and its limits | 9 | |
| Propensity score matching and weighting | 9 | |
| Difference-in-differences and event studies | 8 | |
| Instrumental variables and local average treatment effects | 8 | |
| Panel data: fixed and random effects and the Hausman test | 8 | |
| Dynamic panels (introduction to GMM) | 8 | |
| Potential outcomes and the identification problem | 7 | |
| Regression discontinuity designs | 7 | |
| Quantile regression | 7 | |
| Synthetic control | 6 |
Choose an answer, then open the working. Each explanation says why the right answer is right and why the others are not.
In a causal graph, both adopting a technology and high yield cause a farm to be selected for a media award. Analysing only award-winning farms will:
B. It conditions on a collider
The award is a common effect (collider) of adoption and yield; conditioning on it induces an association between them even if none exists in the population.
Further reading: Imbens and Rubin, Causal Inference; Angrist and Pischke, Mostly Harmless Econometrics
What is the propensity score?
B. The probability of treatment given covariates
e(X) = P(D = 1 | X); Rosenbaum and Rubin showed it balances observed covariates between treated and untreated units.
Further reading: Imbens and Rubin, Causal Inference; Rosenbaum and Rubin (1983)
A matched evaluation of a credit scheme shows excellent balance on all observed covariates. What bias can remain?
C. Bias from unobserved confounding factors
Matching relies on selection on observables (unobserved confounders such as managerial ability are not handled); unobserved factors affecting both take-up and outcome still bias the estimate. Sensitivity analysis gauges how strong they would need to be.
Further reading: Imbens and Rubin, Causal Inference; Rosenbaum and Rubin (1983)
What does imposing common support do in propensity-score matching?
B. Drops treated units without matches
Treated units whose scores exceed the controls' range have no counterparts and are dropped; the estimate then refers to the remaining units, which should be reported.
Further reading: Imbens and Rubin, Causal Inference; Rosenbaum and Rubin (1983)
In an event study, coefficients for the three years before a scheme are large and significant. What does this suggest?
A. Pre-trends or anticipation
Leads should be near zero under parallel trends; significant leads cast doubt on the design.
Further reading: Angrist and Pischke; Cunningham, Causal Inference: The Mixtape
At what level should standard errors be clustered in a DiD study of a district-level scheme with yearly farm data?
C. The district
Cluster at the level of treatment assignment to allow for serial and within-district correlation; Bertrand, Duflo and Mullainathan showed unclustered DiD errors are far too small.
Further reading: Angrist and Pischke; Cunningham, Causal Inference: The Mixtape
A lottery offering subsidised tractors raises mean farm income by 6,000 currency units and tractor ownership by 30 percentage points. What is the Wald IV estimate of the effect of tractor ownership?
D. 20,000 currency units
Wald = reduced form / first stage = 6,000/0.30 = 20,000. 1,800 multiplies; 6,000 is the reduced form (ITT).
Further reading: Angrist and Pischke; Imbens and Angrist (1994)
Distance to the nearest input dealer is used as an instrument for fertiliser use. Which threat to exclusion is most plausible?
A. Distance also shapes prices and access
Remote farms may also face lower output prices, less credit and fewer services, which affect income directly, violating exclusion.
Further reading: Angrist and Pischke; Imbens and Angrist (1994)
A synthetic control tracks the treated state poorly before the policy, with a large pre-treatment RMSPE. What follows?
A. The gap is not a credible effect
A credible counterfactual must reproduce the pre-treatment path; with poor fit the gap may reflect mismatch rather than the policy.
Further reading: Abadie (2021), Using synthetic controls
A treatment's conditional quantile effect at tau = 0.1 is large. Does this mean the farms that were poorest before treatment gained most?
A. No; ranks need not be preserved
Quantile effects compare distributions, not individuals; without rank invariance, the farms at the 10th percentile after treatment need not be those that were there before.
Further reading: Koenker, Quantile Regression
Ten here, 77 in the app. The Laisha app has every practice question we have written to RBI Grade B in this subject, marked as you go, with timed papers, a jump-to-unanswered list and a certificate when you pass at 70 per cent.
Free for ten days, then one subscription of US$10 a month. We would rather say so here than have you find it later.
An AI operating system for agriculture: free guides, manuals and tools written for every country.