Securities Lending

The mechanics, without mystique

A short seller wants to sell a stock they don't own; settlement systems require actual shares to deliver. Index funds are the world's ideal lenders: enormous, stable inventories they won't suddenly need. So the fund's lending agent (often an arm of the provider or its custodian) lends specific holdings to vetted borrowers, who post collateral of typically 102–110% of the loan's value, marked to market daily, in cash or high-grade securities. The borrower pays a fee; the fund can also earn on reinvesting cash collateral. Loans are recallable: the fund can demand its shares back (for sales, index changes, or, in some policies, proxy voting).

The economics: for broad developed-market equities, lending demand is modest and fees are thin, adding perhaps 1–5 basis points per year to a world or S&P 500 fund. For small caps, emerging markets and "specials" (heavily shorted names), demand is hot and income can reach 10–30+ bp: which is why some small-cap ETFs persistently out-track their own TERs (Chapter 35's positive TDs, explained).

The two real risks, and their mitigations

1. Borrower default. The borrower fails and doesn't return the shares. The fund keeps the collateral and rebuys the shares in the market; the risk is the gap if the shares jumped faster than the collateral covered. Mitigations: over-collateralization with daily marking, borrower vetting, and, at several large providers, a borrower-default indemnification: the lending agent contractually makes the fund whole. Historical record: actual investor losses from equity-lending defaults in mainstream funds are, on the public record, essentially nil; the mechanism has absorbed defaults (including 2008's) via collateral as designed.

2. Cash-collateral reinvestment. The sneaky one, and the source of the only real historical damage. Cash collateral must be parked somewhere; pre-2008, some lending programs (notably outside the ETF world, in securities-lending pools run for institutions) reached for yield in that parking and got burned when "safe" paper froze. The lesson was learned in rules: UCITS and post-crisis US practice confine reinvestment to money-market-grade instruments. When reading disclosures, this is the paragraph to actually read: conservative reinvestment (government MMFs, overnight repo) is the correct answer; anything creative is a flag.

A third, softer issue: lent shares and voting. Shares on loan generally can't be voted; providers balance lending income against stewardship policies, and some recall for contentious votes. A governance detail worth knowing exists, rarely decision-relevant for index investors.

The revenue split: where providers differ most

The fee pie divides between the fund (your return) and the lending agent/manager (their revenue), and the split is a published number that varies meaningfully by provider. The mid-2020s landscape, as dated examples: iShares returns a stated majority share to its funds (historically around 62.5% in equity programs, with BlackRock's agent taking the rest and providing indemnification); Vanguard has long returned essentially all net revenue to the fund, running lending as a service rather than a profit center; European houses (Xtrackers, Amundi, UBS) publish splits commonly in the 60–75%-to-fund range with lending caps. UCITS rules (ESMA guidelines) require that all revenue net of direct and indirect operational costs be returned to the fund: the operative word being the cost definition, which is where the agent's share legally lives.

The evaluative frame this book endorses: judge the package, not the split alone. A 62.5% split with indemnification and conservative collateral can serve investors better than 100% with weaker protections; and all of it is ultimately audited by one number you already know how to read, the multi-year tracking difference, where lending income either shows up or doesn't.

Reading the disclosures in four minutes

Where the three numbers live (Chapter 25's pyramid applied):

1. Maximum and expected on-loan percentages: prospectus (efficient-portfolio-management section) for the cap (UCITS funds often cap at 20–50% of assets; some famously lend up to 100% of lendable assets within limits), factsheet or provider lending page for current/average on-loan.
2. The revenue split and indemnification: prospectus and the provider's securities-lending disclosure page.
3. What actually happened: the annual report (Chapter 25, Layer 3): realized lending income, collateral held, largest borrowers. Policy versus practice, reconciled yearly.

A worked example

Two MSCI World funds, Chapter 35's Fund A and Fund C revisited with the hood open:

• Fund A: lends up to 15% of assets, currently 4% on loan, 70% of revenue to fund, government-MMF cash reinvestment, indemnified. Lending contribution: ~2 bp/yr. TD −0.02%.
• Fund C: lends up to 50%, currently 18% on loan, 62.5% to fund, indemnified, same conservative reinvestment. Contribution: ~7 bp/yr. TD +0.04%.

Both are inside good practice; C simply runs the side business harder and pays you for it. The investor who chooses A "because lending is risky" without reading either disclosure, and the one who chooses C "for the yield" without checking the collateral paragraph, are making the same mistake in opposite directions. The four-minute read converts both into a decision.