When does input uncertainty (the proxy problem) actually matter for CME, and when can I safely ignore it?
The curriculum says input uncertainty depends on context — it's a problem for testing theory but less of an issue for practical use. Can you give concrete examples of when I should worry about it vs. when it's fine to use imperfect proxies?
The distinction the curriculum draws is between theoretical validation and practical application — and it makes a real difference for how you approach CME.
When Input Uncertainty IS a Serious Problem:
If your goal is to test whether a theoretical model is correct or to identify anomalies relative to the model, proxy errors can invalidate your entire analysis.
Example — Ashworth Academic Research:
Professor Ashworth discovers that stocks with high book-to-market ratios earn returns 3% higher than the CAPM predicts when using the S&P 500 as the market proxy. She concludes this is a CAPM anomaly (the 'value premium').
But this conclusion rests entirely on the S&P 500 being a good proxy for the true market portfolio. If the true market portfolio includes real estate, private equity, and human capital — assets correlated with value stocks — the apparent anomaly might disappear. The input uncertainty makes it impossible to determine whether the value premium is real or an artifact of the wrong benchmark.
When Input Uncertainty Is LESS of a Problem:
If your goal is to generate useful empirical estimates for portfolio construction — not to prove a theory — imperfect proxies can be perfectly adequate.
Example — Crestfield Pension Fund:
Crestfield's analyst uses the CAPM with the MSCI All Country World Index as the market proxy to estimate expected returns for 10 equity markets. She knows the MSCI ACWI isn't the theoretical market portfolio. But she's not trying to prove CAPM — she's trying to produce reasonable cross-sectional return estimates that are internally consistent.
For this purpose, the proxy works fine because:
- The relative ranking of expected returns across markets is likely correct even if the absolute levels are off
- The estimates are cross-sectionally consistent (all derived from the same framework)
- The errors from using an imperfect proxy are systematic and therefore less likely to produce absurd allocations
The Practical Framework:
| Goal | Input Uncertainty Impact | Action |
|---|---|---|
| Test whether CAPM/APT is theoretically valid | 🔴 Critical — cannot be resolved | Acknowledge limitation |
| Identify pricing anomalies relative to a model | 🔴 Serious — 'anomaly' may be proxy artifact | Use multiple benchmarks |
| Generate cross-sectional return estimates for allocation | 🟡 Manageable | Use best available proxy |
| Rank asset classes by expected return | 🟢 Low impact | Relative rankings are robust to proxy choice |
| Estimate risk premiums for valuation | 🟡 Moderate | Sensitivity test with different proxies |
Key Exam Insight: The CFA exam often presents a scenario where an analyst uses a stock index as the market portfolio. If the question asks about testing theory or identifying anomalies, input uncertainty is the key concern. If the question is about practical allocation, input uncertainty is secondary.
Explore more uncertainty questions in our CFA Level III question bank.
Master Level III with our CFA Course
107 lessons · 200+ hours· Expert instruction
Related Questions
What exactly is the Capital Market Expectations (CME) framework and why does it matter for asset allocation?
How do business cycle phases affect asset class return expectations?
Can someone explain the Grinold–Kroner model step by step with numbers?
How do you forecast fixed-income returns using the building-blocks approach?
PPP vs Interest Rate Parity for forecasting exchange rates — when do I use which?
Related Articles
Join the Discussion
Ask questions and get expert answers.