SMIF Portfolio Construction Lab

# SMIF Portfolio Construction Report

## Executive Summary

The portfolio is evaluated on whether its weights translate analyst views into efficient portfolio risk. Current expected return is 8.2%, annualized risk is 3.6%, tracking error is 0.3%, and the efficiency coefficient is 0.62.

## Data Assumptions

- Data Source: Built-in sample SMIF data
- Sample Prices: synthetic demo data, not market data
- Return Type: Log Returns
- Frequency: Weekly
- Annualization Factor: 52
- Holdings with missing price history are excluded from covariance and weights are normalized

## Portfolio Metrics

- Holdings Analyzed: 12
- Portfolio Variance V_p: 0.001303
- Portfolio Standard Deviation S_p: 3.6%
- Sharpe Ratio using a 2% risk-free rate: 1.71
- Top Risk Contributor: NVDA at 16.6% of variance

## Top Risk Contributors

- NVDA: Weight 10.0%, PCV 16.6%, beta_iP 1.66, Risk-heavy; reduce candidate if expected return does not justify risk.
- MSFT: Weight 12.0%, PCV 13.0%, beta_iP 1.09, Risk contribution roughly in line with weight.
- AAPL: Weight 11.0%, PCV 10.9%, beta_iP 0.99, Risk contribution roughly in line with weight.
- GOOGL: Weight 9.0%, PCV 10.7%, beta_iP 1.19, Risk contribution roughly in line with weight.
- AMZN: Weight 8.0%, PCV 10.7%, beta_iP 1.33, Watch: risk contribution is different from weight and needs a return view.

## Required Returns and Gaps

A stock with beta_iP above 1 must earn more than the portfolio average to justify its contribution to risk. Largest gaps:

- NVDA: Expected 13.5%, Required 9.4%, Gap 4.1%, Signal Positive Gap/Monitor Risk
- KO: Expected 5.0%, Required 7.5%, Gap -2.5%, Signal Negative Gap/Monitor
- PG: Expected 5.2%, Required 7.7%, Gap -2.5%, Signal Negative Gap/Monitor
- JNJ: Expected 5.5%, Required 7.8%, Gap -2.3%, Signal Negative Gap/Monitor
- COST: Expected 8.8%, Required 6.8%, Gap 2.0%, Signal Add/Keep

## Rebalance Proposal

Turnover estimate is 1.8%. Proposed risk is 3.5% versus current risk of 3.6%. Proposed beta dispersion is 0.37 versus current 0.37.

- Sell NVDA: 10.0% to 8.2%; Reduce high portfolio-risk contribution (beta_iP 1.66).
- Buy COST: 7.0% to 8.3%; Low beta_iP (0.21) with positive expected-return gap.
- Buy KO: 6.0% to 6.5%; Add lower-risk contribution candidate under equal-attractiveness logic.

## Tracking Error Summary

Tracking error equals sqrt((x - b)' C (x - b)). Current active expected return is 0.4%, tracking error is 0.3%, and information ratio is 1.45.

## Guaranteed Active Management

This analytical module reframes active management as an outperformance manufacturing process. In the traditional model, the client provides capital and bears most of the process-failure risk while the manager earns fixed or asset-based compensation. In the GAM structure, the LP is a debt-like investor, the GP is an equity-like investor, both capital sleeves are invested in the same portfolio, and GP equity backs the LP guarantee.

- LP/Debt Ending Value: $93,107,732
- GP/Equity Ending Value: $42,441,235
- Annualized Guaranteed Premium: 1.045%
- Ending Equity Cushion: 31.3%
- Liquidation Status: Operating
- Tracking-Error Policy: 23.5% maximum allowed by the cushion multiple assumption
- Active-Risk Intentionality: 95.0% intended, 3.2% unintended, 1.8% unknown
- Performance Analytics: fund annual return 7.9%, benchmark annual return -0.2%, tracking error 3.4%, information ratio 2.40

Professional Interpretation: the structure forces the manager to bear the risk of the outperformance process. The debt investor receives benchmark participation plus a fixed premium and pays no fees. The GP receives residual profits only after satisfying the LP guarantee. The equity cushion is not cosmetic; it is the collateral supporting the guarantee. The higher the tracking error, the faster a failed process can erode the cushion. The structure only works if active risk is mostly deliberate rather than accidental.

## Stock-Selection Evidence

- MSFT BUY: Stock 3.3%, Benchmark 2.1%, Relative 1.2%, Verdict Supported. Buy call outperformed the synthetic benchmark window.
- XOM SELL: Stock -0.7%, Benchmark 1.3%, Relative -2.0%, Verdict Supported. Sell call avoided relative underperformance in the demo sample.
- KO HOLD: Stock 0.9%, Benchmark 0.6%, Relative 0.3%, Verdict Neutral. Close enough to benchmark for a hold recommendation.

## Formula Appendix

- Portfolio Variance: V_p = w' C w
- Portfolio Standard Deviation: S_p = sqrt(V_p)
- C_iP: row_i(C) dot w
- ACV_iP: w_i * C_iP
- PCV_iP: 100 * ACV_iP / V_p
- beta_iP: C_iP / V_p, also PCV_iP / PCT_i when both are consistent
- Required Return: I(R_i) = E(R_p) + psi * S_p * (beta_iP - 1)
- Inefficiency Gap: ExpectedReturn_i - RequiredReturn_i
- UADE: Gap_i / Volatility_i
- Tracking Error: sqrt((x - b)' C (x - b))
- GAM Debt Return: Debt Return_t = Benchmark Return_t + Premium_t
- GAM Debt Capital: Debt Capital_t = Debt Capital_{t-1} * (1 + Debt Return_t)
- GAM Total Fund: Total Fund_t = Total Fund_{t-1} * (1 + Fund Return_t)
- GAM Equity Capital: Equity Capital_t = Total Fund_t - Debt Capital_t
- GAM Equity Cushion: Equity Cushion_t = Equity Capital_t / Total Fund_t
- GAM Liquidation Trigger: Equity Cushion_t < Minimum Cushion
- GAM Active Return: Fund Return_t - Benchmark Return_t
- GAM Information Ratio: Active Return / Tracking Error

Analytical simulation only. Not investment advice. Not a real guarantee, fund structure, legal agreement, or solicitation. Calculations depend on uploaded data quality.