Josh's StudyWiki

SelectionBias-SecuritiesMarkets

3 min read

Connection: Selection Bias ↔ Securities Markets

Securities markets are one of the richest real-world environments for selection bias. The funds you can compare all exist today — the ones that failed closed and disappeared from the record. The track records you see are pre-screened. PHIL 252’s taxonomy of selection bias variants maps precisely onto the ways financial performance data misleads investors.

graph LR
    subgraph PHIL252
        SB[Selection Bias Variants]
        RC[Right-Censoring]
        OSE[Observation Selection Effect]
        EX[Extrapolation]
    end
    subgraph ADMN201
        MF[Mutual Funds & ETFs]
        HF[Hedge Funds]
        SM[Securities Markets
Bull & Bear]
        IDX[Indexes & Benchmarks]
    end
    RC -->|survivorship bias in fund data| MF
    RC -->|hedge fund track records| HF
    OSE -->|only visible funds are surviving ones| IDX
    EX -->|past returns ≠ future results| SM
    SB -->|corrupts all performance comparisons| MF

From PHIL 252

SelectionBiasVariants introduces six distinct ways sample design can corrupt conclusions before reasoning begins. The most investment-relevant:

  • Right-Censoring / Survivorship Bias: A study ends (or a dataset is compiled) before all outcomes have occurred, systematically dropping the failures. In fund data: only surviving funds appear in performance rankings.
  • Observation Selection Effect: The observer’s position is linked to the variable measured. You can only observe a fund if it still exists — you never see the ones that closed.
  • Extrapolation: Applying results from one group to a different one. A fund’s 5-year track record during a bull market doesn’t extrapolate to bear conditions.
  • WEIRD Populations: “Historical returns” may be drawn from an unusually favorable period — not representative of all market environments.

From ADMN 201

InvestmentVehicles covers mutual funds, ETFs, hedge funds, and commodities. SecuritiesMarkets covers primary vs. secondary markets, bull/bear cycles, margin trading, short selling, and market indexes. Key touch points:

  • Fund performance rankings: Presented as if they represent the industry, they actually represent only survivors
  • Index composition: Indexes drop underperformers and add outperformers — making the index look better than most individual investors do
  • Hedge fund marketing: Average returns of existing funds ignore the many that closed with losses
  • “Past performance does not guarantee future results” — the legal disclaimer is a direct acknowledgment of extrapolation bias

Why This Matters

An investor who ignores selection bias will systematically overestimate the odds of success when choosing investments. The funds that made the brochure are the ones that happened to win — not necessarily the best-managed. PHIL 252’s framework lets you ask the right questions: Who is in this sample? What happened to those who didn’t make it in?

SelectionBiasVariants, Bias, DataVisualization, SecuritiesMarkets, InvestmentVehicles, RiskManagement

Graph View

Backlinks