LAISHA
Open Laisha

ICAR AICE-JRF/SRF Econometrics: practice questions

ICAR AICE-JRF/SRF (PhD admission and fellowship) · India · National Testing Agency for ICAR

Original practice questions written to the published syllabus · not past papers
10worked here, free
125written to this exam in this subject
15syllabus units

About ICAR AICE-JRF/SRF

ICAR AICE-JRF/SRF (PhD admission and fellowship) is a PhD, fellowship and lectureship exam run by National Testing Agency for ICAR, for candidates in India. LAN-based CBT, 2 hours, 120 MCQs, 480 marks, +4 / -1. 2026 exam on 4 July 2026. 72 subjects at 122 cities.

Source: cdnbbsr.s3waas.gov.in.

Checked against the exam body's published information. Formats and fees change; confirm the current edition there before you book.

Econometrics in ICAR AICE-JRF/SRF

CLRM assumptions and which one each violation breaks; BLUE and the Gauss-Markov theorem; consequences, detection and remedies for multicollinearity, heteroscedasticity and autocorrelation; Durbin-Watson decision rules and Durbin's h; dummy trap; Koyck, adaptive expectations and Nerlovian partial adjustment; order and rank conditions; ILS and 2SLS; DF/ADF, cointegration and ECM; LPM, logit and probit. Short numericals: slope from sums, R-squared, adjusted R-squared, F from R-squared, t, DW, VIF, elasticities from log models, long-run elasticities.

Regression interpretation for business data: dummy variables, logistic regression for choice and adoption, multicollinearity among marketing variables, reading regression output (coefficients, t, p-values, R-squared, F).

Estimating economic relationships from data and knowing when the estimates can be trusted: OLS and its assumptions, their violations, simultaneous equations, time series and binary-choice models.

Our 125 ICAR AICE-JRF/SRF questions in this subject: 29 easy, 65 medium, 31 hard, across 15 units.

10 practice questions, worked

Choose an answer, then open the working. Each explanation says why the right answer is right and why the others are not.

These 10 were written to ICAR AICE-JRF/SRF and are not shown on any other exam's page.

Question 1 · medium

A linear yield response Y = 21 + 0.12N (Y in q/ha, N in kg/ha) is estimated. At the sample means N = 100 kg/ha and Y = 33 q/ha, the elasticity of yield with respect to nitrogen is

  1. 0.364
  2. 0.120
  3. 0.040
  4. 3.030
Answer and working

A. 0.364

In a linear model the elasticity is b x (X-bar/Y-bar) = 0.12 x 100/33 = 0.364. The slope 0.12 is a marginal effect in q per kg, not an elasticity; 0.040 inverts the ratio of means. 3.030 is the ratio of means (100/33) without the slope.

Question 2 · easy

Which assumption is NOT needed for the Gauss-Markov theorem that OLS is BLUE?

  1. Zero conditional mean of the error
  2. Homoscedastic errors
  3. Normal distribution of the error term
  4. No autocorrelation between errors
Answer and working

C. Normal distribution of the error term

The Gauss-Markov theorem needs linearity in parameters, zero conditional mean, homoscedasticity, no autocorrelation and no perfect multicollinearity. Normality is added only to derive exact t and F distributions for inference in small samples.

Question 3 · easy

Which pattern in regression output most strongly suggests multicollinearity?

  1. High R-squared and significant F, but few individually significant t ratios
  2. Low R-squared and a Durbin-Watson statistic near 2
  3. Residuals that fan out as X increases
  4. A significant RESET test
Answer and working

A. High R-squared and significant F, but few individually significant t ratios

Collinear regressors jointly explain Y well (high R-squared, significant F), but their separate effects cannot be isolated, so individual standard errors are large. A fan of residuals points to heteroscedasticity; RESET to misspecification.

Question 4 · hard

With the same bounds (dL = 1.284, dU = 1.567), a computed d of 2.90 indicates

  1. positive autocorrelation
  2. no autocorrelation
  3. an inconclusive result
  4. negative autocorrelation
Answer and working

D. negative autocorrelation

For negative autocorrelation compare d with 4 - dL = 2.716 and 4 - dU = 2.433. Since 2.90 > 2.716, negative autocorrelation is indicated. Between 2.433 and 2.716 would be inconclusive.

Question 5 · medium

True model: yield = b1 + 0.3 fertiliser + 5.0 soil index + u. Soil index is omitted, and the slope from regressing soil index on fertiliser is 0.02. The expected value of the estimated fertiliser coefficient is

  1. 0.30
  2. 0.40
  3. 0.20
  4. 0.10
Answer and working

B. 0.40

E(b2) = b2 + b3 x b32 = 0.3 + 5.0 x 0.02 = 0.40. The fertiliser effect is overstated because better soils receive more fertiliser and also yield more. 0.10 is the bias alone. 0.20 subtracts the bias instead of adding it.

Question 6 · medium

A researcher must choose between two non-nested models of farm credit demand, neither a special case of the other. A suitable test is

  1. the Chow test
  2. the Davidson-MacKinnon J test
  3. the Durbin-Watson test
  4. the Goldfeld-Quandt test
Answer and working

B. the Davidson-MacKinnon J test

The J test adds the fitted values of the rival model as a regressor and tests their significance, in both directions. It can reject both models or neither. The other tests address structural change, autocorrelation and heteroscedasticity.

Question 7 · medium

In a Koyck distributed lag with impact coefficient b0 = 0.25 and decay rate lambda = 0.6, the long-run multiplier and mean lag are

  1. 0.417 and 1.5 periods
  2. 0.625 and 1.5 periods
  3. 0.625 and 0.67 periods
  4. 0.150 and 1.36 periods
Answer and working

B. 0.625 and 1.5 periods

Long run = b0/(1 - lambda) = 0.25/0.4 = 0.625. Mean lag = lambda/(1 - lambda) = 0.6/0.4 = 1.5. 1.36 is the median lag (-ln 2/ln lambda), a different measure.

Question 8 · easy

Applying OLS to a structural demand equation in which price is endogenous gives estimates that are

  1. unbiased but inefficient
  2. biased and inconsistent
  3. BLUE
  4. consistent but biased in small samples
Answer and working

B. biased and inconsistent

Price is determined jointly with quantity and so is correlated with the demand error. That is simultaneity bias, which does not disappear as the sample grows.

Question 9 · medium

In the Dickey-Fuller regression Delta Y_t = delta Y_(t-1) + u_t, the null and alternative hypotheses are

  1. H0: delta = 0 (unit root); H1: delta < 0 (stationary)
  2. H0: delta < 0 (stationary); H1: delta = 0 (unit root)
  3. H0: delta = 1; H1: delta not equal to 1
  4. H0: delta = 0 (stationary); H1: delta > 0 (explosive)
Answer and working

A. H0: delta = 0 (unit root); H1: delta < 0 (stationary)

delta = rho - 1, so delta = 0 means rho = 1, a unit root. The alternative is one-sided stationarity. KPSS reverses the roles, taking stationarity as the null.

Question 10 · medium

A regression of one district's wheat price on another's, both in levels over 25 years, gives R-squared 0.96 and Durbin-Watson 0.31. This should first raise the suspicion of

  1. perfect multicollinearity
  2. a well-specified cointegrating relationship, proven
  3. spurious regression between non-stationary series
  4. negative autocorrelation
Answer and working

C. spurious regression between non-stationary series

Granger and Newbold's warning sign is R-squared above d. The series may be cointegrated, but that must be tested (residual ADF with Engle-Granger critical values); a high R-squared alone proves nothing. d near 0 means strong positive, not negative, autocorrelation.

A study plan for Econometrics in ICAR AICE-JRF/SRF

This plan follows where our questions actually fall, so the time goes where the marks are likely to be. It is our reading of the syllabus, not the exam body's weighting.

  1. First, the heavy units (40% of the 125 questions): Linear probability, logit and probit models (12, 10%); Simultaneous equations and identification (11, 9%); ARIMA models (9, 7%); Classical linear regression model and OLS (9, 7%); Cointegration and error-correction models (9, 7%).

    Read the manual chapters for these first and do every worked question you can find on them.

  2. Then the middle band: Hypothesis testing in regression (9, 7%); Stationarity and unit roots (9, 7%); Autocorrelation (8, 6%); Multicollinearity (8, 6%); Specification error (8, 6%).

    Aim for one sitting per unit, then mixed questions across them.

  3. Then the lighter units: BLUE properties and the Gauss-Markov theorem (7, 6%); Distributed lag models (7, 6%); Heteroscedasticity (7, 6%); Dummy variables (6, 5%); Maximum likelihood (6, 5%).

    Learn the definitions, classifications and standard numbers; these units are more often recalled than worked.

  4. Leave time for the hard ones: Heteroscedasticity (3 hard of 7), Autocorrelation (3 hard of 8), Cointegration and error-correction models (3 hard of 9).

    Come back to these after a first pass, untimed, and read every explanation, including for the options you rejected.

  5. Bank the easy marks: ARIMA models (3 easy of 9); Classical linear regression model and OLS (3 easy of 9).

    Do these early in revision so they are fast on the day.

  6. Finish timed. Sit the subject as a timed set without notes, mark it, and send each wrong answer back to its unit above.

Last updated · How Laisha writes and checks its pages · Report a mistake

Ask Laisha about this

Opens Laisha's chat with your question ready to send. Sign in or start free.

Unlock every question — ICAR AICE-JRF/SRF Econometrics

US$10 a month · one price worldwide · cancel any time

  • Every practice question in this subject, with full working
  • Timed exam papers and progress by unit
  • Ask Laisha to explain any answer — up to 1,000 questions a month
Free or Smarter
FreeSmarter
Grow guides, pest pages, articlesRead allRead all
Ask Laisha5 a dayUp to 1,000 a month
Exam practiceThe sample on each pageEvery question, timed papers
Crop calendar and pest alerts—For your own fields
Calculators and tools—All
Uploads and reports—100 a month, real files
10 days free for every new account, no card.