Fama-French Analytics

ETFs vs. Theoretical Indexes

An index (like the S&P 500 or the Fama-French Small Value Research Index) is a mathematical abstraction. An Exchange Traded Fund (ETF) is a real legal entity that must buy stocks, pay lawyers, audit its books, and execute trades. This difference creates "Tracking Difference."

Sources of Performance Drag

1. Expense Ratios (The Fee)
This is the most obvious drag. If an index returns 10.0% and the ETF charges 0.40%, the theoretical maximum return for the investor is 9.6%.

2. Transaction Costs
Indexes don't pay commissions. ETFs do. When an ETF rebalances (buys/sells stocks to match the index changes), it pays bid-ask spreads. For illiquid Small Cap Value strategies, these costs can be substantial.

3. Cash Drag
ETFs hold a small amount of cash to handle daily redemptions or dividend payouts. In a rising market, this cash earns nothing, slightly hurting performance (in a falling market, it helps).

Sources of Performance Boost (Yes, really)

Securities Lending
ETFs often lend out the stocks they own to short sellers. The short sellers pay interest (borrow fees) to the ETF. The ETF manager typically keeps a cut (e.g., 20%), but passes the majority (80%) back to the fund.

In some niche categories like Small Caps or Cannabis, the borrow fees are so high that the Securities Lending income can exceed the Expense Ratio, effectively giving the ETF a negative cost!

Replication Methods: Full vs. Sampling

Full Replication: The ETF buys every single stock in the index.
Pro: Perfect tracking.
Con: Expensive for indexes with thousands of tiny stocks.

Sampling (Optimization): The ETF manager uses a computer model to buy a representative subset of the index. For example, instead of buying all 2,000 stocks in the Russell 2000, they might buy 800 stocks that statistically behave like the full 2,000.
Pro: Lower transaction costs.
Con: "Tracking Error" - the return might deviate day-to-day from the index.

Why This Matters for Fama-French Data

The data on this dashboard comes from the Kenneth French Research Data. These are Research Indexes. They assume:

  • Zero transaction costs
  • Holdings rebalanced annually (usually end of June)
  • No management fees

If you look at the "HML" (Value) factor return here, do not expect a Value ETF (like AVUV or VBR) to match it perfectly. The ETF is living in the real world; the Factor data is living in the lab. Both are useful, but they are not the same.