Anyone familiar with even a snippet of MiFIDII will have come across threshold tests. On 28th September ESMA published its ” Regulatory technical and implementing standards – Annex I” (RTS), a document that uses the word ”threshold” no less than 439 times. Few of us will have had occasion to look into detailed calculation requirements of even one of these calculations, so I hope it will be useful to share some first impressions.
I have had occasion to look at one instance in some level of detail, the ‘Main Business Activity’ exemption threshold which intends to specify the conditions under which non-financial firms who engage in commodity trading may be exempted from applying for a MiFIDII license, thus sidestepping all the obligations that would entail.
Even a relatively short dip into the subject showed very clearly there is still a lot of uncertainty in the methodology to resolve, potentially unintended consequences to avoid, and some potentially perverse incentives that are introduced. The uncertainty is amplified to a large extent by a combination of inconsistency between the RTS text and the accompanying consultation feedback commentary, and silence in the RTS on some key computational assumptions.
Main Business Threshold
The Main Business Threshold (MBT) compares a 10% threshold against a ratio calculated (roughly speaking) a group’s European speculative derivatives trading activity divided by a proxy measure of the size of its main business. If the result is <10% then the group does not need to register a MiFID entity. If the result is >10%<50% a further test is required to establish exemption. I won’t go into that second test today.
Progress has been Made
First let’s recognise some progress made in this iteration. ESMA has made several sensible (N.B. Not every industry commentator/body agrees with me!) decisions since its first attempt at how this ratio should be calculated.
- ESMA has withdrawn the suggestion that accounting measures should be used, apparently acknowledging that the cost of overhauling the double entry book-keeping for MiFID derivatives of an entire firm globally might actually be quite expensive, would produce a subjective result due to wide variances in accounting policy and practice, and most likely deliver the number too late, due to year-end close and audit timetables not aligning with the timeframe in the Level 1 text for exemption applications.
- Instead ESMA has proposed a new calculation based on filtering and aggregating and comparing values that can be derived from individual trades and positions. This should be welcomed by the market participants at least in terms of simplicity, not least because it is a similar approach to the one they will recently have implemented for the EMIR NFC/NFC+ Threshold calculation.
The MBT compares speculative commodity derivative transactions of the group with third parties (sum of gross notional, or “GNV”) with the aggregate gross notional of all Financial Instruments of the group, speculative or hedged.
I should declare some simple but important interpretations on which my analysis relies:
- “Financial Instrument” means all asset classes, unless otherwise qualified. This means that the denominator includes tradeable securities, i.e. Equities and Bonds as well as derivative instruments in interest rates, FX and commodities. This is basic explicit MiFID Level 1 and RTS text so shouldn’t be too controversial.
- The calculation is performed on a ‘trading turnover’ or trade date window basis, not on contracts liver in the window as is the case with the EMIR NFC Threshold. The RTS is silent on this subject, but the momentum of the consultation commentary is consistent with the trade date window assumption.
- ‘Group’ means the whole group of companies globally, unless qualified by the term ‘European’. Again, this should not be controversial, but is easily confused when reading the text and various commentaries on various threshold calculations at once.
With these straightforward interpretations declared thee are some very problematic outcomes of the calculation as it is currently specified, outlined below:
Funding and Hedging Frequency Bias
The same firm may produce a different MBT ratio depending on how and when it last raised funds from the market, and how it chose to hedge the financing transactions. This is because the trade date window is a maximum of three years, and initially will be only one year.
- Unlike many commodities the tenor of financial hedging instruments has no influence on Notional Value expressed in currency, so a 1 year EUR1m Interest Rate Swap has the same gross notional as a 10 year EUR1m IRS, and conversely ten 1 year swaps has ten times the gross notional of 1 ten year swap with the same terms and cash flows.
- Regularly extended Short term finance (<3 years), by way of equity or bonds will have more impact on the ratio than long term hedging of the full investment payback of the underlying assets (often >10 or even >20 years)
- Foreign Currency Bond finance plus Interest Rate hedge plus Foreign Currency hedge might count three times in comparison to the equivalent local currency bond with no hedges.
Reporting/Operating Currency Bias
- A European firm of a particular business size will produce a different MBT ratio if its geographic profile means it needs to hedge foreign currency flows using Financial Instruments (which count in the denominator), as opposed to a single country/single currency entity with the same assets, and turnover.
Funding Instrument Bias
- Equity is a significant portion of the balance sheet of most real-economy firms that use the financial markets. It should definitely be recognised in measuring the size of the main business. However the notional value of equity issued bears no resemblance to the size of most large public firms. I checked a sample of balance sheets and found that shares issued at 10 Euro cents often raise EUR10 funding for the business. Measuring Equity Financial Instruments at notional just doesn’t make sense.
Anomalies like this lead to uncertain or arbitrary regulatory status, wasted resources, potentially provide perverse economic incentives, and at worst incentive to game the system. More work is required. ESMA might start with the following:
- Adopt the EMIR approach of measuring outstanding contracts in the measurement window rather than new trades done thus avoiding spurious bumpiness of the MBT denominator.
- Choose to measure the GNV of the bonds not their related Interest Rate/Currency derivatives, thus avoiding the double counting and potential gaming of the calculation.
- Measure Equity as Shareholders Funds rather than notional value thus representing the true size of the business that is funded.