Risk Forum 2025 plenary session
Post Announcement Drift and Neo-Brokers
Risk Forum 2025 plenary session (Tuesday March 18, afternoon)
This year is the last time ILB organise its flagship conference, the risk forum, at the CCI of Paris, that has been bought by a luxury brand and will t urn to be private.
This blog post is about the last plenary session of this event, hold in partnership with Euronext.
The format is a plenary speaker followed by a round table. ILB organising committee and Euronext proposed me to moderate the roundtable. I accepted with pleasure: listening carefully research on the post earning drift, mixed with graphs, it of deep insurer interest to anyone who, like me, worked 10 years in asset management (hedge fund and sovereign wealth fund).
Moreover, the topic of the roundtable: new models for retail trading, and the recent surge in the Payment For Order Flow (PFOF) in several countries (not in Europe, putting aside temporary exceptions lile Germany) is of interest since it sets an interesting debate between proponents of centralised trading venues and believers in the virtues of bilateral market making.
Explaining the Post Announcement Drift with social networks
- Chairwoman: Emilie RIEUPEYROUX, Euronext.
- Guest speaker: David HIRSHLEIFER, Professor of Finance, Marshall School of Business, University of Southern California “News Diffusion in Social Networks and Stock Market Reactions”
Pr Hirshleifer gave a convincing presentation on the way social networks convey information around Earning Dates.
The Post-Earning Announcement Drift is a well documented effect1 describing the way information do not propagate immediately after the earning announcements of companies.
In his paper, Hirshleifer uses the graph of the connection of social networks (FB, s, etc) and Google Searches, with the assumption that the information starts in the node where the HQ of companies are located. The spread between the first and last deciles of his indicator seems to explain a surge of 28% on the announcement date and -29% in the subsequent drift. Strangely the volumes are indeed more intense, even after the announcement, along this graph.
This is one more way to understand the rich debate between price formation (market participants push the price via a liquidity consuming mechanism) and price discovery (price follow a perpetual Brownian bridge, targeting the value of having the instrument in my portfolio at a future date).
New Retail Trading Models: can we reconcile free-commission trading and best execution?
The second party of this plenary session was a roundtable. Despite its lack of diversity
- Chairman: Charles-Albert LEHALLE, Ecole Polytechnique.
- Paul BESSON, Head of Quant Research, Euronext.
- Philippe GUILLOT, Secrétaire général adjoint, en charge de la Direction des données et marchés.
- Eric RETTIEN, Head of Retail Services, Kepler Cheuvreux.
Information is an essential component of the price formation process: it triggers decisions, and in case of consensus, these decisions will move prices. In case of disagreement, the asymmetrical behaviours of liquidity consumers and liquidity providers will temporary push the price. A diversity of profiles of market participants makes this consensus or disagreement meaningful. More participants means more ways to understand information, at different time scales. This is a central arguments in favour of centralised trading places. Seen from the viewpoint of a market maker, the situation is nevertheless different: the better you know your clients / counterpart the more adapted to its habits in terms of liquidity consumption you can be. It speaks in favour of discrimination of flows, because in principle we are convinced that we need market makers.
On the one hand you can focus on the instantaneous reaction to information, as if market participants have no prior. On the other hand you can focus on the habits, on the way market makers, informed investors, or noise traders take profit of long term relations with liquidity providers. They are ready to pay to have the guarantee they you be present in the future. You accept to not be fully optimal at each instant in exchange of guarantees for the future.
The roundtable focuses on Payment for Orderflows and new forms of trading mechanisms in place by Neo-brokers. This is not new it has existed since at least 1984. This practice, in which market makers pay brokers for routing the orders of their retail customers, involves a clear segmentation of the retail flow, which is specifically directed to certain market makers in exchange for compensation. Since the pandemic, and indeed it started a little earlier, low-cost retail brokers appeared (Trade Republic, Robinshood, etc). The way they operate is that (to spare the costly money of directing their clients’ orders to exchanges) they sell this flow to intermediates.
What does it means for price formation? what does it means in terms of best execution for these retail clients? It is the kind of questions we will address in this roundtable.
Over the past few years, we have seen a growing participation of retail investors in the financial markets. From your perspective, what have been the most significant factors contributing to this boom?
From the regulator’s perspective: The regulator put the emphasis on using the term “bilateral price formation process” instead of “bilateral trading”. Indeed, the virtue of flow segregation is not natural, since the more concentration, the more participants bring their bit of information to price formation. The analogy of having a two stages market and having access to a (more expensive, for open 7/24) convenient store on the top of a big market has been made. That being said, the exact value you get from the margin you pay should be transparent. The introduction of circuit breakers for retail investors, if they have a specific channel to go to markets, was mentioned as a protective measure. The question of “trajectory crossing orders” has been motion: what is the level of information of an order that accepted to be executed at the mid for the next hours whatever happen in the future? It has to be zero, and hence it offers at zero cost a liquidity option. Another important remark is the change of practices of retail flow attracted by Neo-brokers: they trade more frequently and with smaller size than other retail flow.
From the broker’s perspective: The issue of price improvement was raised by the broker. Price improvement, where an order is executed at a more favorable price than the best available price, is a key aspect of the best execution obligation. A price improvement on the bid-ask spread for a retail investor is not an important component of the costs that are made of fees. Especially once you take into account the frequency of trading and the size of such investors.
From the exchange’s perspective: The exchange emphasised the need to better define “best execution”. This concept is central to the European Markets in Financial Instruments Directive (MiFID) and aims to ensure that brokers obtain the best possible price for their clients’ orders. Nevertheless, it is clear that
- the gain associated to a price improvement does not have the same effect of a portfolio of a retail client of a Neo-broker (small size, low frequency), and an institutional investor
- usually one accepts to pay a premium to a market maker (with whom one trade in a bilateral venue) to be sure this market maker will be there in the future in case of liquidity. Nevertheless “nobody catches a falling knife”: liquidity tentions like the Flash Crash have shown that these market makers do not keep the liquidity but turn it back very quickly to markets.
From an academic perspective: It is fair to say that offering zero (neo-)brokerage fees requires payment for order flow (PFOF). It opens trading to a category of retail investors that would not have been confortable with larger fees (implicitly requiring to trade less and for larger size).
Footnotes
Marais, M. Laurentius. “Discussion of post-earnings-announcement drift: Delayed price response or risk premium?” Journal of Accounting Research 27 (1989): 37-48.↩︎