27 Session 18: Platform Demo · Historical Validation
| Unit | 3 — Decision and Application |
| Book Chapter | 9 (sections 9.6–9.10) |
| Track | Common core (both tracks) |
27.1 Learning Objectives
By the end of this session, students will be able to:
- Execute a complete portfolio valuation workflow on the GE-LAV® platform — from raw data to investment committee output.
- Validate platform outputs against historical episodes and identify what the platform would have recommended at each turning point.
- Identify known limitations of the platform: data gaps, calibration sensitivity, regime-shift assumption violations.
- Translate platform outputs into business language suitable for IC memos, board decks, or regulatory disclosure.
- Diagnose typical errors in platform usage (wrong asset class selection, stale calibration, ignoring confidence intervals).
27.2 Pre-Class Assignment
- Read: Book Chapter 9, sections 9.6–9.10 (~10 pages)
- Try: Run the platform tutorial walkthrough (link on course site, 30 min)
27.3 In-Class Outline (75 minutes)
| Time | Segment | Format |
|---|---|---|
| 0:00–0:05 | Recap Session 17 · Today: full pipeline + validation | Lecture |
| 0:05–0:25 | Live demo: end-to-end portfolio valuation | Live demo |
| 0:25–0:45 | Validation: GFC, COVID, 2022 backtest | Lecture + chart deep-dive |
| 0:45–0:55 | Platform limitations | Lecture |
| 0:55–1:10 | Translating outputs to IC language | Activity |
| 1:10–1:15 | Wrap-up Unit 3 · Bridge to math intuition (Sessions 19–24) | Lecture |
27.4 Discussion Questions
- The platform improved hypothetical 2008–2010 returns by 5–8 percentage points. Why didn’t the GE-LAV® framework get adopted earlier, given the empirical success?
- The 2022 rate shock revealed a \(\bar{L}\) stability assumption that may be wrong in the “higher for longer” environment. If you were running the platform, how would you decide whether \(\bar{L}\) has shifted vs. just experienced a long stress excursion?
- The platform produces a “GE-LAV mark” that differs from GP-reported NAV. As an LP, do you report the platform mark to your board, or the GP mark? Why?
27.5 Worked Example: Translating Platform Output to IC Memo
Platform output for Fund A (PE Buyout, vintage 2020, NAV $75M, 5 yr remaining):
Reported NAV: $75.0M
LAV Mark (Jensen-adjusted): $75.6M (+0.8% Jensen bias)
GE-LAV Mark (current state L = -0.5): $73.1M (-2.5% equilibrium correction)
Exit Boundary Analysis:
Current L_t: -0.5
Critical L*(t=5yr): -1.2
Recommendation: HOLD (L > L*)
Stress Test:
GFC-Scenario Value: $34.5M (-54%)
COVID-Scenario Value: $63.8M (-15%)
Pigouvian Tax (if forced exit): 0.8% on transaction valueTranslation to IC memo language:
“Fund A (PE Buyout vintage 2020): The current GE-LAV-adjusted fair value is $73.1M, slightly below the GP-reported $75.0M (a 2.5% downward adjustment reflecting the current mild liquidity stress). Standard Jensen convexity bias adds 0.8% to value when properly accounted for; the equilibrium correction (-2.5%) reflects current secondary market conditions. The exit boundary analysis confirms hold — current liquidity state is above the critical exit threshold for the remaining horizon. Under a GFC-depth stress scenario, fair value would decline to $34.5M (a 54% drawdown). The position is appropriately retained at current parameters but should be monitored as liquidity conditions evolve.”
27.6 What to Expect Next Session
Session 19 begins Unit 4: Math Intuition Bridges. This is where the framework’s mathematical machinery gets explained — at the intuition level, not the proof level.
Sessions 19–24 cover: - 19: Brownian motion, Itô, OU (concepts) - 20: Stochastic control, HJB (concepts) - 21: Mean-field games (concepts) - 22: Fokker-Planck, master equation (concepts) - 23: LAV operator, GE equilibrium (concepts) - 24: Jensen bias, Pigouvian tax, welfare (results)
Track 1 students: These sessions give you the meaning of the math without the algebra. You’ll be assessed on conceptual understanding, not derivations.
Track 2 students: These sessions are the first pass. You’ll revisit each topic with full derivations in Sessions 25–29.
Reading: Book Chapter 10 (probability, Brownian motion, Itô) — read for concepts, not formulas. ~20 pages but skim heavily.