Institutional Trading Best Practices: RFQ, NAV Trading and Blocks

The four channels

1. Risk price (the block trade). You ask several dealers to quote a firm price for the entire size, immediately. The winning dealer buys or sells the whole block onto its own book and manages the hedge afterwards. You get certainty of price and completion in one print; the dealer charges for the risk it absorbs, priced from the basket's cost plus a margin. Best when the position must be established now, when the basket is liquid enough for the dealer to hedge, and when you value certainty over shaving basis points.

2. NAV trading — the quiet workhorse. You agree to transact at the fund's official NAV struck at the underlying market's close, plus or minus an agreed spread in basis points. This removes intraday tracking risk entirely, scales to any size the fund itself can absorb, and is normally the cheapest all-in route: the natural default whenever today’s closing level is an acceptable price, and the only sensible choice when your benchmark is NAV — the classic case being a fund manager transitioning a portfolio, or anyone whose performance is measured against the index's closing level. Common in fixed income and international equity, where intraday fair value is hardest to pin down.

3. Agency execution — for particular situations. A broker works the order through the session, participating in displayed volume, sourcing hidden liquidity and possibly using the closing auction (Chapter 73). You pay a commission and bear the market's movement during execution. A specialist’s route rather than a default: for orders outsized relative to the fund’s normal activity where patience genuinely reduces cost — though even there, a NAV trade handles the same size without working the market at all, so agency earns its keep only when no NAV or risk quote is attractive.

4. Primary market (create or redeem) — usually someone else’s job. For the largest orders, you go through an AP to the creation window: deliver the basket or cash, receive shares at NAV plus a transaction fee. You pay the basket's trading cost rather than the ETF's spread, which for very large size is usually cheaper (Chapter 71's Layer 3, accessed directly). Best when the size is large relative to secondary liquidity and you have time to assemble or distribute the underlying.

The RFQ process, done properly

Most European institutional ETF flow runs through RFQ platforms (Tradeweb and Bloomberg being the dominant venues), where a buy-side desk requests quotes from multiple liquidity providers simultaneously and executes against the best.

The practical discipline:

• Compete the quote. Three to five dealers is the working standard: enough for genuine competition, few enough to limit information spread. Single-dealer execution on size is the most common avoidable cost in institutional ETF trading.
• Control information. Every dealer you ask learns something about supply and demand. Requesting quotes from a dozen desks on an illiquid fund leaks your intention into the market before you have traded. Size the panel to the trade.
• Quote both ways where sensible. Asking for a two-sided market conceals your direction, at the cost of slightly wider quotes.
• Know your fair value before you ask. Compute the basket's value independently. A quote is only assessable against a fair-value estimate you trust; without one, you are accepting the dealer's number on the dealer's word.
• Record the alternatives. Best-execution obligations under MiFID II and comparable regimes require demonstrating process, not merely outcome. Keep the competing quotes.

NAV trading, and why it is not free

NAV trading feels like a free lunch, you get the official price with no intraday risk, and it is not, for two reasons worth stating plainly:

You pay a spread to NAV. The dealer quotes NAV plus x basis points, and x embeds their cost of hedging from the moment of agreement to the close plus their margin. Competing that spread matters exactly as much as competing a risk price.

You inherit the fund's own execution. Transacting at NAV means your entry price is whatever the fund's closing valuation was, including any staleness in evaluated pricing for bonds (Chapter 33). In stressed markets, that can be a materially different number from where the ETF was actually trading: sometimes to your advantage, sometimes not.

The correct use is alignment: if your performance is measured against the index close, NAV trading eliminates a tracking error you would otherwise carry. If it is not, you are paying for certainty you do not need.

Cost measurement: what "good execution" means

Institutional practice measures execution rather than asserting it. The standard framework:

• Choose the benchmark before trading: arrival price (the market at the moment of decision), VWAP, or NAV.
• Measure implementation shortfall: the gap between the decision price and the achieved price, including commission, spread, market impact and delay cost.
• Attribute it: how much was spread, how much was impact, how much was the market moving while you waited.
• Compare channels retrospectively: the same order size in the same fund executed via RFQ, agency and creation over time builds a real cost curve for your flow.

The discipline matters because the cheapest-looking channel is frequently not the cheapest. A tight risk price on a difficult basket may embed more cost than a patiently worked agency order; a NAV trade may look clean while the fund's closing valuation carried a discount you absorbed.

Why several of these are wrong for retail

Stated explicitly, because our standards forbid blending the playbooks:

• RFQ and risk pricing are unavailable at retail size and would be worse if they were: the standing quote already prices small orders efficiently.
• NAV trading removes intraday risk that a long-term private investor does not meaningfully carry, at a spread they would pay for nothing.
• Working an order across a session multiplies commissions and attention for a private investor whose entire order fits inside the displayed quote.
• Creation and redemption require AP relationships and creation-unit-scale capital.

The retail equivalent of every technique here is Chapter 58's six habits, which achieve near-zero execution cost precisely because retail size is the easy case. Sophistication is not transferable downward.

A worked example

A pension fund must invest €120m into a European equity UCITS ETF as part of a transition, benchmarked to the index close.

• Fair value first. The transition manager computes the basket's value independently and estimates hedging cost.
• Channel choice. Because performance is measured to the close, a NAV trade aligns benchmark and execution; the alternative, working it intraday, would introduce tracking error against the very benchmark being measured.
• Competition. Five dealers are asked for NAV-plus quotes on the full size. Range: +6bp to +14bp. Winning quote +6.5bp.
• Cross-check. Against the alternative of creating directly (basket assembly estimated at +9bp all-in, plus two days of exposure risk), the NAV trade is both cheaper and cleaner.
• Measurement. Implementation shortfall recorded against the close: 6.5bp, entirely spread, zero delay cost by construction.

Total cost on €120m: about €78,000, versus a plausible €170,000 had the desk taken the first quote it was offered. The competition, not the cleverness, was worth €90,000.