Finance & Economics Methods Primer: Simple Identification Strategies

In finance and economics, the central challenge is identification: separating true effects from confounding noise.

You do not need advanced econometrics to start. You do need clean design logic.

Why identification matters

Correlation alone is often misleading because:

  • omitted variables influence both X and Y,
  • selection effects bias comparisons,
  • and reverse causality can flip interpretations.

Identification strategies reduce these risks.

1) Before–after (same unit over time)

Compare outcomes before and after an event for the same unit.

Use when: the event is sharp and no major concurrent shocks are likely.

Risk: other time-varying factors may drive changes.

2) Difference-in-differences (DiD) intuition

Compare change over time in a treated group vs control group:

`(Post−Pre in Treated) − (Post−Pre in Control)`

Use when: treatment affects one group but not another.

Key assumption: parallel trends (without treatment, both groups would move similarly).

3) Instrumental variables (IV) logic

Use an instrument Z that shifts X but affects Y only through X.

Use when: X is endogenous (e.g., choice-related).

Challenge: valid instruments are hard to justify; weak instruments create unstable estimates.

4) Regression discontinuity (RD) intuition

Exploit cutoff-based assignment (e.g., eligibility threshold). Compare units just above vs just below cutoff.

Use when: assignment rule is strict and manipulability near cutoff is low.

5) Event studies (finance workhorse)

Estimate abnormal returns around information events (earnings, policy announcements).

Use when: event timing is clear and market data is available.

Risk: overlapping events and market-wide shocks.

Practical method selection for students

Ask three questions:

  1. What is the source of variation identifying effect?
  2. What confounders remain plausible?
  3. What assumptions am I willing to defend in writing?

Choose the simplest design that answers these credibly.

Reporting standards (even for beginner work)

  • State identification assumption explicitly
  • Show robustness checks where feasible
  • Discuss threats to validity honestly
  • Avoid causal language when design does not support it

Good research is often conservative in claims and precise in assumptions.

---

If you want a step-by-step path to choose methods you can actually execute, grab the free '8‑Week Research Roadmap + Proposal Template'. For structured mentorship on question design, methods, and writing, you can also explore the Core Research Fellowship.

Previous
Previous

AP Research Paper Format: Structure, Sections, and Common Errors

Next
Next

Finance Research Topics for Students: 30 Smart Ideas for Strong Projects