Dataford
Interview QuestionsInterview GuidesExperiencesMock InterviewsPricing
Get started

Testing GDP Signal in Forecast Briefings

IMFEasyStatistics & Probability00:00
IMF
Your interviewer · Financial Analyst
In session
Interviewer

Welcome to your interview for the Financial Analyst role at IMF.

The question is on your right: Testing GDP Signal in Forecast Briefings. Take a moment with it first.

Talk your thinking through with me if you like - when you're confident, submit your answer and I'll grade it like a real screen. You have three graded attempts to score 7/10 or better.

Only Submit answer is graded - discussion is free practice.

Problem

Business Context

At MacroPulse, a financial intelligence platform, analysts publish a weekly “global outlook” note. Management wants to know whether the team’s briefing process is actually aligned with real global economic movement, rather than just reacting to noise.

Problem Statement

You are given 12 recent months of data comparing the platform’s monthly economic sentiment score with the actual monthly global GDP growth surprise index. Test whether the average difference between the two series is zero, and quantify the uncertainty around that difference.

Given Data

The sentiment score and GDP surprise index are both standardized to comparable monthly units.

MonthSentiment ScoreGDP Surprise IndexDifference (Sentiment - GDP)
10.420.350.07
20.310.280.03
30.150.22-0.07
4-0.050.01-0.06
5-0.18-0.12-0.06
6-0.12-0.08-0.04
70.080.020.06
80.210.170.04
90.270.190.08
100.110.090.02
11-0.09-0.03-0.06
120.050.010.04

Use a significance level of 0.05.

Requirements

  1. State the null and alternative hypotheses for the mean paired difference.
  2. Compute the sample mean and sample standard deviation of the monthly differences.
  3. Perform a paired t-test.
  4. Construct a 95% confidence interval for the mean difference.
  5. Decide whether the briefing process appears systematically biased high or low.
  6. Briefly explain what this implies about staying current with global economic trends.

Assumptions

  • Monthly paired observations are independent enough for a simple paired analysis.
  • The distribution of paired differences is approximately normal.
  • The 12 months are representative of the recent operating environment.