ETF Due Diligence Framework

Direct Answer

ETF due diligence works through six questions in order: (1) Does the index methodology match the intended exposure? (2) What replication method is used and what structural risks does it introduce? (3) Is AUM large enough and liquidity sufficient for the position size? (4) Does the securities lending program introduce risk and who captures the revenue? (5) Is there material counterparty exposure from derivatives or lending? (6) Is the fund large and established enough that closure risk is minimal? Only after all six questions are answered satisfactorily does it make sense to compare expense ratios and tracking differences. An ETF that fails on index methodology is wrong regardless of its cost.

Key Takeaways

• Index methodology is the foundation. Two "technology ETFs" can have dramatically different portfolios depending on how each index defines and weights technology exposure.
• Physical full replication has the lowest structural risk. Sampling adds minor tracking error for large-index ETFs. Synthetic replication (swaps) reduces tracking error but introduces counterparty risk.
• AUM below $50 million is a closure risk warning sign. AUM below $100 million often means wider bid-ask spreads and higher transaction costs relative to larger alternatives covering the same index.
• Securities lending revenue can offset expense ratio cost and improve tracking difference — but the borrower-default risk sits with the ETF's investors, not the sponsor. Check collateral quality and indemnification.
• Counterparty exposure in synthetic ETFs is capped at 10% of NAV per counterparty under UCITS rules (EU); US ETFs have no equivalent regulatory cap — check the prospectus.
• Fund closure forces a sale, potentially triggering taxable events and losses from liquidating during an unfavorable period. Monitor AUM trend, not just current AUM level.
• Tracking difference (annual return gap vs. benchmark) is the best single all-in cost metric for long-term holders. Expense ratio alone understates cost for ETFs with poor securities lending or high-turnover indexes.

Step 1: Index Methodology

Before looking at any performance or cost data, read the index methodology document. This document — produced by the index provider (MSCI, S&P, FTSE Russell, Bloomberg, or a proprietary provider) — defines the rules the ETF is required to follow. Index methodology determines:

• Eligibility universe: Which securities can enter the index? US-listed only? Global? What minimum market cap, liquidity, or other screens?
• Weighting scheme: Market-cap weighted, equally weighted, fundamental-weighted (revenue, earnings, book value), factor-weighted (value score, quality score)? Weighting determines the portfolio's factor tilts and concentration.
• Reconstitution schedule: How often is the index rebalanced and constituent changes made? Quarterly reconstitution = higher turnover = higher transaction costs = potential tracking error relative to daily-priced alternatives.
• Buffer rules: Do departing constituents need to breach a lower threshold to be removed, or is removal immediate when they fall below the entry threshold? Buffer rules reduce turnover and reconstitution event price impact.
• Corporate action handling: How are spin-offs, mergers, delistings, and special dividends treated? Inconsistent treatment across competing indexes creates performance divergence.
• Total return vs. price return: Does the index reinvest dividends (total return) or ignore them (price return)? ETFs tracking price-return indexes will naturally appear to trail total-return benchmarks. Compare apples to apples.

Example divergence: Two "US small cap value" ETFs — one tracking S&P 600 Small Cap Value and one tracking Russell 2000 Value — have different size cutoffs, different value factor definitions, different constituent counts (400 vs. 1,400 stocks), and different rebalancing schedules. Over a 10-year period, these differences can produce meaningfully different returns with different factor exposures, even though both are described as "small cap value".

Red Flags in Index Methodology

• Proprietary index created by the ETF sponsor itself, with limited independent oversight or verifiable methodology rules
• Very small eligibility universe (fewer than 50 securities) that makes the index highly concentrated by design
• No published methodology document or one that is vague about weighting rules and reconstitution criteria
• Subjective inclusion criteria ("companies positioned to benefit from…") without quantitative screens that can be verified independently

Step 2: Replication Method

ETFs use one of three methods to track their index, each with different tracking error profiles and structural risks.

Full Physical Replication

The ETF holds every security in the index in proportion to its index weight. Most accurate for liquid, narrow indexes (S&P 500, NASDAQ 100). Tracking error comes primarily from dividend reinvestment timing, cash drag from creation/redemption settlement, and transaction costs during rebalancing. Securities lending is additive — revenue can more than offset these costs. No counterparty risk from the replication structure itself.

Sampling / Optimized Replication

The ETF holds a subset of index constituents chosen via optimization to match the index's risk factor exposures while reducing transaction costs. Common for large, illiquid indexes (Russell 2000 — 2,000 stocks, many micro-caps with wide spreads). Tracking error is higher than full replication but often acceptable if the optimization is well-designed. Check actual tracking difference over multiple years to verify the optimization works in practice, not just in theory.

Synthetic Replication

The ETF enters a total return swap with a counterparty bank: the ETF delivers the return of a substitute basket (often highly liquid securities in different sectors/geographies), and the bank delivers the index total return. Used for hard-to-replicate exposures: commodity indexes requiring futures, certain emerging markets with capital controls, and some fixed-income indexes. Tracking error is typically very low — the bank guarantees the index return within the swap terms. However, the counterparty risk is real: if the bank defaults between swap payments, the ETF has not received the index return for that period. UCITS rules cap per-counterparty exposure at 10% of NAV; check the prospectus for specific collateral and netting arrangements.

Synthetic ETFs are common in Europe; less common in the US. They are not inherently dangerous — many large, reputable providers use them effectively — but they require understanding the counterparty arrangements, collateral quality, and what happens operationally in a counterparty default scenario.

Step 3: AUM and Liquidity

AUM (assets under management) and liquidity are related but distinct concerns.

AUM and Fund Viability

A fund sponsor incurs fixed and variable costs to operate an ETF. Below a certain AUM threshold (typically $50–100 million), the fund generates insufficient revenue from the expense ratio to cover these costs profitably. Sponsors may cross-subsidize small funds within a large lineup, but niche or thematic ETFs that don't attract AUM are candidates for closure or merger. Signs of closure risk: AUM below $50 million, AUM declining over 12+ months, niche category, sole product from a small issuer, recent management fee cuts (sometimes a desperate attempt to attract flows).

Liquidity and Transaction Cost

ETF liquidity has two layers. Primary market liquidity is the liquidity of the underlying securities — authorized participants can create and redeem shares in large blocks, making the ETF as liquid as its holdings. Secondary market liquidity is the bid-ask spread and depth at which the ETF itself trades on the exchange. Wide spreads occur when the underlying is illiquid (market makers can't hedge efficiently), when trading volume is low, or when the ETF is very small. For position sizes below $50,000, the bid-ask spread is the dominant transaction cost — no commission savings offset a 0.5% spread paid on every round trip. For large institutional positions, impact cost (moving the price by buying a large block) becomes relevant and scales with position size relative to daily volume.

Rule of thumb: bid-ask spread in basis points should be no more than 1–2x the expected holding period in months for the spread cost to be a small fraction of the total expected return. A 20 bp spread on a 1-month hold is 20 bp/month transaction cost. A 20 bp spread on a 10-year hold is 0.2 bp/month — negligible.

Step 4: Securities Lending Policy

Most large ETFs lend some of their holdings to short sellers in exchange for a fee. The lending mechanics and revenue sharing vary significantly across sponsors.

Revenue Sharing

Lending revenue split between the ETF and the sponsor is disclosed in the prospectus and annual report. iShares returns 100% of securities lending revenue to the fund. Vanguard returns 100% (Vanguard's structure means the fund company is owned by the funds). Other sponsors may keep 20–40% as a management fee, reducing the benefit to investors. Compare the stated securities lending income in the annual report to the expense ratio to understand whether lending more than offsets costs (tracking difference better than expense ratio) or only partially offsets them.

Collateral and Indemnification

When the ETF lends securities, the borrower posts collateral — typically cash or high-quality government securities — worth 102–105% of the loaned securities' value. If the borrower defaults, the ETF uses the collateral to repurchase the loaned shares in the open market. If the market price has risen above the collateral value, the ETF incurs a loss. Most large sponsors indemnify the fund against this shortfall from their own balance sheet, making collateral management risk effectively transferred to the sponsor. Smaller or less-established ETF sponsors may not offer indemnification — check the prospectus lending program section explicitly.

Step 5: Counterparty Risk Assessment

Counterparty risk exists in two forms in ETFs:

Securities Lending Counterparty Risk

Assessed in Step 4. With full sponsor indemnification and overcollateralization, residual risk is primarily the credit quality of the sponsor's own balance sheet (their ability to make good on the indemnity). For large, highly rated sponsors (BlackRock, Vanguard, State Street), this is minimal but non-zero systemic risk.

Swap Counterparty Risk (Synthetic ETFs)

The bank providing the total return swap is the ETF's primary counterparty. Key questions: Who is the swap counterparty? What is their credit rating? What collateral does the ETF hold in case of counterparty default? How frequently is the swap marked to market and margined? Under UCITS, counterparty exposure cannot exceed 10% of NAV at any time — any excess must be reset by the counterparty posting additional collateral. US ETFs have no equivalent rule; some synthetic US ETFs have negotiated similar protections voluntarily. Review the prospectus section on "Counterparty Risk" and "Collateral" explicitly for synthetic ETFs.

Step 6: Fund Closure Risk

When an ETF closes, investors receive cash equal to NAV at liquidation. This seems benign — NAV is what your shares are worth — but closure has several adverse practical effects.

• Forced taxable event: In a taxable account with embedded unrealized gains, liquidation forces recognition of those gains, triggering a tax bill regardless of your timing preferences.
• Liquidation discount: As the fund liquidates its portfolio in the weeks before closure, it sells holdings at market prices. For illiquid underlying securities, this may push prices down — a liquidation discount that shows up as slightly lower NAV vs. fair value during the wind-down period.
• Disruption to strategy: If the ETF is a key component of a systematic allocation or a specific exposure you need, closure forces you to find a replacement — potentially at a disadvantageous time or without an exact equivalent.

Monitor: AUM trajectory (12-month change, not just current level), any sponsor announcements, news about the issuer's ETF lineup rationalization. When a small ETF from a large sponsor is merged into a similar, larger fund (the more common outcome vs. outright liquidation), the transition is usually more investor-friendly — the swap is often done as an in-kind exchange, with no forced tax event.

Worked Scenario: Due Diligence on Two "Tech ETFs"

1. ETF A: Tracks a proprietary sponsor-created "Innovation Technology Index." 45 holdings. Expense ratio: 0.50%. AUM: $80M. Physical replication. Methodology: "companies with high revenue growth and R&D intensity in technology and technology-adjacent industries." No quantitative thresholds specified.
2. ETF B: Tracks MSCI USA Information Technology Index. 87 holdings. Expense ratio: 0.13%. AUM: $15B. Physical full replication. iShares structure — 100% securities lending revenue returned to fund. Clear quantitative GICS methodology.
3. Step 1 — Index: ETF A has a vague proprietary methodology with subjective criteria. ETF B has a published, quantitative, GICS-based methodology independently maintained by MSCI. ETF B wins on methodology clarity and oversight. ETF A's construction is opaque.
4. Step 2 — Replication: Both physical. No differentiation on structural risk from replication method.
5. Step 3 — AUM/Liquidity: ETF A at $80M is viable but approaching the warning zone. It has a narrow, thinly-traded underlying and wider bid-ask spreads. ETF B at $15B has tight spreads, deep secondary market, no liquidity concern.
6. Step 4 — Securities Lending: ETF B returns 100% to fund; ETF A prospectus reveals sponsor keeps 40% of lending income. At a 0.20% securities lending rate, ETF B earns 0.20% additional offset; ETF A earns only 0.12%. ETF B's effective all-in cost (tracking difference) may be better than its expense ratio. ETF A's lending terms reduce but don't eliminate this advantage.
7. Step 5 — Counterparty: Both physical ETFs. Securities lending counterparty risk exists at both, with ETF B (iShares/BlackRock) having the stronger indemnification and balance sheet.
8. Step 6 — Closure Risk: ETF A at $80M with declining AUM in a thematic niche has meaningfully higher closure risk. ETF B at $15B is one of the market's core benchmark products — closure is implausible under any realistic scenario.
9. Verdict: ETF A fails on methodology clarity, relative AUM, lending terms, and closure risk. The 0.37% expense ratio difference does not compensate for these structural deficiencies. ETF B is the appropriate choice unless the investor specifically requires the proprietary index's different exposure — in which case, they should first confirm that exposure is actually what they want and is not achievable through ETF B plus a satellite position.

Summary Checklist

FAQ

Sources

• SEC ETF Rule (Rule 6c-11) — Modern framework for ETF operations and disclosure
• ESMA UCITS Guidelines on ETFs — Counterparty risk, collateral, and synthetic replication rules
• ETF.com: How to Evaluate an ETF's Securities Lending Program
• Vanguard: ETF Due Diligence — A Framework for Evaluating ETFs