# Size-discovery protocols are not on the efficient frontier

## Practice improves allocations but more can be done, says Darrell Duffie

Darrell Duffie is the Dean Witter distinguished professor of finance at Stanford University’s Graduate School of Business, and professor by courtesy in the university’s economics department. He has recently been an adviser on financial markets to a range of bodies, including the Financial Stability Board and several Federal Reserve banks.

Most basic economics courses teach that a market only works efficiently when the price is adjusted to match supply and demand. Modern trade platform operators know better.

When the price adjusts to clear the market, as on a central limit order book, large traders are inhibited by their own price impact. They hold back from expressing their full demand or supply. For example, the seller of a large block submits sell orders gradually, given the limited demand available at or near the best bid price. This can lead to costly delays in liquidating or acquiring large positions.

Platform operators have responded to price-impact concerns by adding special trade protocols, collectively known as size discovery.

One of the earliest examples of size discovery is the so-called workup protocol used in the interdealer market for US Treasuries. When a workup session is triggered on BrokerTec, for example, the last trade price is frozen, and central-limit-order-book trading is suspended. Buyers and sellers queue for the opportunity to trade at the frozen workup price.

At the end of the workup session, there is almost inevitably a residual order imbalance on one side of the market. The workup price does not clear the market. But because the workup price is frozen, buyers and sellers face no price impact and are therefore much less concerned about placing block-sized orders. Large inventory imbalances of buyers and sellers can thus quickly be offset, with a potentially large saving in inventory costs.

###### By shading down the magnitude of their workup orders, market participants have higher inventory costs than they might have with further improvements in market design

Fleming and Nguyen (2015) estimate that workup accounts for 43–56% of total BrokerTec trade volume in on-the-run Treasuries.

Workup and other size-discovery mechanisms, including matching sessions and dark pools, are popular in a wide range of other markets.

For example, on GFI’s platform for trading credit default swaps, Collin-Dufresne, Junge, and Trolle (2016) find that matching sessions and workups account for over 70% of trades in popular credit default swap (CDS) index products such as CDX.NA.IG.5YR – a composite of five-year CDSs referencing 125 investment-grade firms. Trade platforms for interest rate swaps also make heavy use of workup or matching sessions. Size discovery also shows up in equity markets, in the form of block-crossing dark pools such as Liquidnet and Posit.

My research with Haoxiang Zhu of the Massachusetts Institute of Technology shows that augmenting a price-discovery trading protocol, such as a central limit order book or double-auction market, with size discovery leads to a substantial improvement in market efficiency. But we also find that there are further, untapped potential efficiency gains.

### Winners’ curse

#### Risk Management

###### The bank, the vendor, the turrets and the golf day

After DBS switched supplier, a row broke out, raising questions about entertainment and influence