This is another in my series of posts on the SEC’s proposal to require money market funds with floating net asset values (“institutional money funds”) to implement swing pricing during any pricing period in which the fund has net redemptions. This post continues the analysis of the estimated costs that an institutional money fund “must include, for each security in the fund’s portfolio” when determining any swing price. These costs are:
- Spread costs,
- Brokerage commissions,
- Custody fees, and
- Any other charges, fees, and taxes associated with portfolio security sales.
This post considers custody fees.
What Are Custody Fees?
The proposing release does not explain what would be included in “custody fees.” In my experience, custodians typically charge a fee based on the value of the assets held in custody, subject to a minimum fee or in addition to a base fee. Selling a vertical slice of a fund’s portfolio would reduce an asset-based fee, so the SEC could not have been referring to these custody fees in the proposed amendments.
Custodians also charge “ticket” or transaction fees for processing portfolio transactions. These fees are charged for each trade ticket submitted by a fund. Sale of a vertical slice of a fund’s portfolio would result in numerous trade tickets and significant transaction fees.
For example, the two major U.S. custodians charge between $3.75 and $5 per ticket, depending on whether the investment is cleared through DTC or the Fed Securities Wire and other factors.
The $600 million institutional money fund I used as an example in my earlier posts is based on a real prime fund. The fund’s year-end portfolio holdings report shows it held 118 individual investments, mostly commercial paper. If the fund’s custodian charged $4.50 per ticket, the sale of a vertical slice would result in ticket charges of $531 (assuming multiple trades are not required to sell an investment).
The Perverse Consequences of Including Ticket Fees
Readers may recall that, in my earlier example, estimated costs of $301 were sufficient to produce a swing price that rounded down to $0.9999 from a net asset value per share (“NAV”) of $1.0000. This suggest that combining $531 of transaction fees with estimated spread costs could reduce the swing price to $0.9998 or lower.
Perversely, because these transaction fees are based on the number of tickets, rather than the dollar amount of the trades, smaller net redemptions will result in larger swing factors and lower swing prices. Using the swing factor formula (Swing Factor = Estimated Costs ÷ (Net Redemptions + Estimated Costs)) and holding transaction fees constant at $531, the following chart illustrates this relationship between net redemptions and swing prices.
Passing $531 of transaction fees onto redeeming shareholders requires a swing price 5 basis points lower than the NAV when the net redemptions are $1 million, but only 1 basis point lower when the net redemptions are $5 million. Rounding the swing price to four digits keeps the swing price at $0.9999 through $10 million in net redemptions. At $11 million, the transaction fees by themselves would not result in a swing price different from the NAV.
Thus, under the SEC’s swing pricing proposal, the effect of transaction fees on swing prices will be highest when the net redemptions are lowest, and thus least likely to require the fund to incur these fees.
Unrealistic Assumptions
No competent portfolio manager would sell 118 securities to raise $1 million for net redemptions, much less do so during each pricing period. Indeed, the proposing release acknowledges that
Using a swing price to pass through phantom transaction fees the fund never incurs will dilute shareholders redeeming at the swing price for the benefit of the remaining shareholders. Lowering the swing price more when net redemptions are smaller will be contrary to the SEC’s aim that “swing pricing result in fairer, non-dilutive pricing particularly when there are heavy redemptions.”
Rather than forcing institutional money funds to assume they will sell vertical slices of their portfolios, it would be better to base swing pricing on the estimated expenses they are expected to incur, if any, because of the net redemptions. Otherwise, the SEC should not be surprised that an unrealistic assumption produces counterproductive results.