15  Session 6: IRR · The Internal Rate of Return Problem

Unit 2 — Measurement and Theory
Book Chapter 4 (sections 4.2–4.4)
Track Common core (both tracks)

15.1 Learning Objectives

By the end of this session, students will be able to:

  1. Define IRR precisely and state its mathematical relationship to NPV.
  2. Identify the two structural biases of IRR (reinvestment assumption + sign-pattern dependence).
  3. Distinguish money-weighted from time-weighted returns and explain when each is appropriate.
  4. Compute IRR for a simple cash flow pattern and identify when multiple IRRs exist.
  5. Explain why a constant-rate metric cannot correctly summarize performance when discount rates are stochastic.
  6. Prepare for Session 7’s PME discussion by understanding why IRR’s failures motivated PME’s development.

15.2 Pre-Class Assignment

  • Read: Book Chapter 4, sections 4.2–4.4 (~10 pages)
  • Optional: Phalippou (2008), “The Hazards of Using IRR to Measure Performance: The Case of Private Equity,” Journal of Performance Measurement

15.3 In-Class Outline (75 minutes)

Time Segment Format
0:00–0:05 Why measurement matters: returns based on bad valuations are bad returns Lecture
0:05–0:20 What IRR is, mathematically and conceptually Lecture + math
0:20–0:35 Money-weighted vs. time-weighted returns Lecture + worked example
0:35–0:50 Two IRR biases: reinvestment + sign-pattern Lecture + worked examples
0:50–1:05 Multiple IRRs and the descending-discount problem Lecture + numerical examples
1:05–1:15 The constant-rate fallacy: IRR under stochastic discount rates Lecture

15.4 Discussion Questions

  1. The GP-LP relationship has IRR baked into the carried interest formula. If GE-LAV becomes adopted methodology, does the carry formula need to change? How?
  2. Public mutual funds use TWR. Why has PE never moved to TWR despite known IRR flaws?
  3. Two PE funds have identical IRRs of 18% but very different cash flow timing patterns. How would you distinguish them from each other in selecting which to commit capital to?

15.5 Worked Numerical Example: When IRR Misleads

Setup: Two PE funds, both reporting IRR = 18%.

Fund A (front-loaded distributions): - \(t=0\): \(-\$100M\) - \(t=2\): \(+\$80M\) - \(t=4\): \(+\$40M\) - \(t=6\): \(+\$20M\) - IRR: 18% (capital out quickly)

Fund B (back-loaded distributions): - \(t=0\): \(-\$100M\) - \(t=2\): \(+\$0M\) - \(t=4\): \(+\$0M\) - \(t=8\): \(+\$200M\) - IRR: 18% (capital out only at end)

The reinvestment problem:

If LP can reinvest at 7% public market rate:

  • Fund A realized return: $80M from year 2 grows at 7% to \((80 \cdot 1.07^6) + (40 \cdot 1.07^4) + 20 = \$120M + \$52M + \$20M = \$192M\) at year 8 — total return 92% over 8 years = 8.4% annualized
  • Fund B realized return: $200M at year 8 — total return 100% = 9.05% annualized

Interpretation: Same reported IRR (18%), but Fund B’s realized return to the LP is higher because the IRR overstatement is larger for Fund A.

Lesson: Reported IRR can rank funds in the wrong order when reinvestment assumption matters.

15.6 What to Expect Next Session

Session 7 dives into PME (Public Market Equivalent) and its variants. We’ll cover: - The Kaplan-Schoar PME and what it computes - Long-Nickels PME (with implicit reinvestment in the benchmark) - Direct Alpha (Gredil-Griffiths-Stucke) - LA-IRR and LA-PME — the GE-LAV-corrected versions - Which method to use in which context

Reading: Book Chapter 4, sections 4.5–4.10 (~12 pages).


← Session 5 | Schedule | Next: Session 7 →