CHANGING RELATIONSHIPS.

Silvano Stagni, Managing Director of Perpetual Motion Consulting and Research, asks whether we have the data to navigate the changes to the relationship between making the investment decision and executing it.

New rules, new reports, and changes to business processes stand out when financial regulations are implemented. They represent the visible part of the iceberg and we all know that the largest and most dangerous part lies beneath the surface. The separation of the investment decision from the execution decision was brought about by a combination of the unbundling of research from execution charges, best execution, market abuse regulation, an increase in transparency levels and the consequent empowerment of the client.

It does not matter if one person decides what to invest in and where to execute the relevant transaction. One must rely on publicly available information and the other must be based on the best possible combination of a) price, b) total cost of the transaction, c) likelihood of execution, and d) speed of execution. Research provides the information to make the decision to invest in a way that must not be connected with the decision of where and how to execute the relevant order.

MiFID II transaction report shows the code of the investment decision maker, the code of the trader, and the code of the venue. If there is an unusual market movement it will be possible to conduct an insider trading investigation. The trader identification code combined with the code of the execution venue, will feed into reports that assess the quality of the execution.

MiFID II and MAR strictly limit the inducement to promote order flow. This is the basis of the separation between the two decision-making processes. Broadly speaking the decision to invest is based on research and the decision on where to execute the relevant order is based on market data, venue charges and liquidity information.

What information needs to be captured to manage this separation of roles? And, 12 months into MiFID II, are we making use of the data that is available?

Nobody is asked to document why they have made an investment decision unless there is a doubt whether a decision to invest or disinvest was the result of publicly available information. If that is the case, it helps to show the relevant research. MiFID II rules on unbundling research have generated data on the usage of research for the whole of 2018, which is a large volume of data. The majority of available research portals either collect all the information relative to a research provider – everything used by a specific company, or research user, irrespective of the source – or everything available by a specific research provider, irrespective of who downloaded it.

In the summer of 2018, GreySpark published a paper called “Living with MiFID II: Accessing Unbundled Research” where it indicated that they only found one portal – Substantive Research – that “offers asset managers analytics that suggest a series of alternative pieces of research content related to the reports or notes that have already been consumed by the firms”. This is very valuable because it allows investment managers to narrow their field of vision to the specific content providers that ultimately meet their specific demands for specific pieces of research on a per-topic basis, and enable a way to document their investment decision.

MiFID II has potentially provided traders with reports that allow them to investigate all the possible execution venues for a specific instrument. The venue execution quality report (commonly known as the RTS 27 report, based on the number of the MiFID II regulatory technical standard that describes it) features too much data to be easily digested quickly. It consists of nine different tables, and some of them need to be repeated for every day in the quarter where transactions associated with a specific ISIN code were executed. The purpose of this report is to prove the ‘quality of execution’ of a specific venue, for a specific instrument in a specific trading mode (RFQ, auctions, etc). One of the general complaints is that the RTS 27 report is too big to be of any use. Execution venues and investment firms have a (smaller) yearly report (commonly known as RTS 28) that states the five most used execution venues for each instrument and explains the criteria used to choose execution venues.

So, do we have enough data to navigate the changes in the relationship between making the investment decision and executing the investment decision? Probably yes, but only when things go well and those decisions are not questioned.

The success factor of an investment decision is quite straightforward: either making a profit or stop a loss from getting worse; but it could be questioned (for instance) by:

•   The regulator during an insider trading investigation;

•   A client of the portfolio manager if the investment decision leads to a loss.

In both cases it would be convenient to have an audit trail documenting the sources used as a way to prove good faith, and that the decision to invest was reached based on publicly available information.

A client has five years to query the quality of execution of a specific transaction, but the evidence is buried in very large reports. There is also a mix of structured and unstructured information that is used to navigate both decisions and to show how one led to another. So, although technically the data is available, it may be unwieldy to use if one only relies on traditional query methods for sequential databases.

Software tools that capture and correlate both structured and unstructured data, possibly used in conjunction with big data software, could be used to substantiate those decisions when required. MiFID II is 12 months old, and making use of the vast volume of data generated by complying with its rules will be the next challenge. Navigating the relationship between making the investment decision and executing it in a way that allows you to stay away from ‘troubled waters’ is a prime example of the need to analyse large volumes of data in a straightforward way. There are tools available to do it but they require lateral thinking to start using them as there is no current ad hoc implementation of such tools at the moment.

©BestExecution 2019

[divider_to_top]