Consultation Paper Australian equity market structure: Further proposals

Transcription

1 20 February 2012 Ms Antonia Fong Market and Participant Supervision Australian Securities and Investments Commission Level 5, 100 Market St SYDNEY NSW 2000 By Dear Ms Fong Consultation Paper Australian equity market structure: Further proposals The Australian Financial Markets Association (AFMA) welcomes the opportunity to comment on ASIC Consultation Paper 168 Australian equity market structure: Further proposals (CP 168). AFMA s initial comments on CP 168 are to be found in the attached commentary paper. CP 168 contains design decisions that will shape the Australian financial markets for many years to come. In this context, the regulatory framework should provide sufficient flexibility to avoid locking our markets into market practices and technology that prevail at one particular point in time. We would encourage ASIC to approach all of its decisions in regard to CP 168 with a primary motive of ensuring that Australia has fully effective capital markets to support the economy and achieves its full potential as a competitive international financial centre. AFMA and its members are keen to engage with ASIC in the further work that is required to finalise the Market Integrity Rules and Guidance. We are grateful for the assistance provided by ASIC staff in exploring the issues raised in consultation with our members and directly. Please contact me at djeffree@afma.com.au or on (02) if you would like to discuss any part of this submission further. Yours sincerely Damian Jeffree Director Policy

2 ASIC CONSULTATION PAPER 168 AUSTRALIAN EQUITY MARKET STRUCTURE: FURTHER PROPOSALS February 2012

3 TABLE OF CONTENTS 1. EXECUTIVE SUMMARY BACKGROUND Prioritisation The Flash Crash IOSCO and Other Regulators MARKET MANIPULATION AND DISORDERLY TRADING AUTOMATED TRADING ENVIRONMENT Algorithmic Trading Control Over Messages and Monitoring BCP and Annual Review of Systems and Connectivity AOP Attestation Changes Minimum Standards for Direct Electronic Access Legally Binding Agreement with AFS Licensees Market Operator Systems and Controls Market Makers ELP/HFT Licencing EXTREME PRICE MOVEMENTS S&P/ASX200 Stocks and ETFs Volatility Controls Cancellation Range Amendments Futures Volatility Controls Expansion of Scope ENHANCED DATA FOR MARKET SURVEILLANCE Additional Data Market Participants Synchronised Clocks Standard Record Format BEST EXECUTION Extension Public Reporting PRE-TRADE TRANSPARENCY AND PRICE FORMATION A Policy Issue for Treasury Review Background A Logical Framework Appropriate Regulatory Aims Targeting Commentary Response AFMA s Framework Example PROPOSED TIMING Page 1

4 1. EXECUTIVE SUMMARY AFMA members are concerned with the scope and pace of regulatory change at present in the Australian market place. The nature of some of the changes being contemplated may impair the efficiency of the equity market and its capacity to support Australian economic development. Moreover, the scale of the changes, as experienced in 2011, is putting substantial additional burdens on the industry when it is already facing difficult trading conditions and is contracting. AFMA argues for a review of ASIC s strategic approach to change in the markets and calls for a more measured approach and a renewed focus on the need for Australian businesses and the Australian capital markets to be internationally competitive. In our response to Consultation Paper 168 (CP 168) we call for the scaling back of the number and scope of the proposals given the well-functioning market that they address, and the prioritisation of the supported proposals over a two year time frame. Algorithmic Trading AFMA argues for a best practice response to the need to ensure consistency of standards in testing of algorithms rather than a strict liability policing style approach. We also support rules, revised from those proposed, to require automated order processing (AOP) kill-switches, and increased monitoring of AOP trades to address the evolution of automated trading. Business Continuity In relation to proposals for regulatory intervention in business continuity arrangements, AFMA argues that the case has not been made. Attestations Changes AFMA suggests modifications to the proposed changes to include cumulative changes but rejects yearly attestations. Market Making AFMA is supportive of reforms to the short selling regime but without micro-regulation around requirements to make a market. We reject requiring all clients who make a market under the wide Corporations Act 2001 definition to be licenced, when the benefits are not apparent and which will unnecessarily increase the regulatory burden. Page 2

5 Volatility Controls Responding to the volatility control proposals, AFMA supports the careful and staged implementation of the limit-up and limit-down proposals with a small pilot project in the first instance moving to a slow and cautious timetable for the full implementation under the guidance of a consultative panel. Enhanced Data AFMA notes various points on the additional order message-fields proposal, and the importance of minimising implementation costs by avoiding partial implementation. We hold that the case has not been made for synchronising clocks at this point given the associated costs and technical difficulties. We note significant concerns that some of the proposals around additional data will increase participants workload, rather than decrease as intended, due to the difficulty in creating some of the envisaged systems but we note the industry s willingness to work with ASIC to find what additional data can be readily provided. We further suggest that ASIC may wish to undertake an internal review to reduce its regulatory impost in this area. Best Execution We reject the proposal that best execution requirements should be extended to other markets or products at this time given the already extensive loads on the industry and the lack of substantial benefit to investors that are currently possible. Pre-trade Transparency In relation to pre-trade transparency, AFMA considers good policy in this area as critical to the industry s future. We hold that the changes proposed constitute a very substantial policy change for the Government and that should only be the result of a Treasury-led review with Ministerial and Legislative oversight. AFMA proposes ASIC work to develop a clear framework and methodology for the assessment of perceived risks and that this be applied in future to determine whether regulatory attention is warranted. ASIC should then work to further develop its understanding of dark liquidity in both quantitative and qualitative terms in conjunction with a consultative panel. AFMA does not support the proposal to implement a $50,000 minimum if dark trading increases by 50% within the next three years as we argue the supporting arguments are Page 3

6 not sufficiently targeted. We instead provide an example of a logical basis for determining which stocks may benefit from the protection of their lit liquidity and an example of a targeted regime for intervention in dark venues for these stocks, noting this work is not definitive. 2. BACKGROUND The finance industry appreciates the need to ensure that the regulatory environment remains robust and current and welcomes the holistic approach being taken by ASIC in CP 168 that recognises the many connections and dependencies inherent in the existing structures. The equity market plays a central role in Australia s economic development by providing finance to assist the capital formation process for Australian business. The equity market also provides essential investment opportunities for superannuation and other investors. A fully efficient and well-regulated market is necessary to support our national aspiration for further economic development. Australia relies on significant capital inflows to support our economic development. One challenge for Australia in this regard is to maintain the attractiveness of our markets for foreign investors. A complementary but inter-related factor is the international competitiveness of Australia s financial markets. AFMA holds that the ultimate goal of regulation must be the creation of a globally competitive financial sector. Australia s most competitive financial centre is Sydney which is currently ranked 15 th in the Global Financial Centres Index, 1 behind all our main regional competitors including Hong Kong at third, Singapore at fourth, Shanghai at fifth and Seoul at 11 th. While this is a disappointing result, with the right combination of measures to ensure that the confidence in our markets is maintained and enhanced, and the competitiveness of the system as a whole is not compromised, Australia can aim to improve its competitive position over time. Many of the proposals in CP 168 go to both the confidence in the markets and to the competitiveness of Australia s financial system. AFMA considers that the effects on competitiveness of the proposals in CP 168, when considered as a whole, could create a significant regulatory burden and are at risk of creating significant disincentives for continued participation or establishment in Australia by international investors and intermediaries. This in turn would be to the detriment of the domestic offering that 1 Page 4

7 services the needs of local business and individual customers and would counter the aims of ASIC under the Corporations Act 2001 to maintain, facilitate and improve the performance of the financial system and entities in it. The domestic industry expects to face continued headwinds in 2012 and is already on a contractionary footing with redundancies reported in many areas and a recent roundtable reportedly estimating equities broking commissions to have been down as much as 15% in There have also been some concerning trends emerging in a reduction of participation of international firms in the domestic market and in the loss of around 800 operations jobs 3 over the last three years, many to offshore offices. Market participants report that the increases in regulatory burden since the Financial Crisis, combined with lower revenues, have already cost jobs in the sector and have increased pressures to offshore functions in order to remain competitive. These pressures are likely to increase with the ASIC cost recovery levy and the proposals in CP 168. There is a concern in the industry that while the thrust of policy should, in the present circumstances, be oriented to improving the industry s international competitiveness, it is instead focused on matters that risk potential damage to the industry s competitiveness. Particularly in the context of a contracting industry where the industry indicates that further regulatory loads may create increased job losses, it is important that ASIC adopts a conservative approach to proposals that may increase the regulatory burden. In this regard, best practice principles with respect to regulation such as those discussed in the Federal Government s Taskforce on Reducing Regulatory Burdens on Business Report 4, particularly Recommendation 7.1 and the discussion on targeting may be relevant. We are concerned that some proposals in CP 168 may not, as drafted, fully conform to these principles. AFMA s approach in responding to the consultation proposals seeks, as far as possible, to be indicative of the outcomes we believe are appropriate given these Taskforce principles and the International Council of Securities Associations (ICSA) Principles for Better Regulation 5 as guidelines AFMA Operations Survey Report data/assets/pdf_file/0007/69721/regulationtaskforce.pdf, see in particular Recommendation 7.1 on page 147 and the discussion on targeting on page Page 5

8 2.1. Prioritisation AFMA s primary submission in relation to ASIC s Consultation Paper 145 (CP 145) was that a process of reprioritisation and a more pragmatic timetable for implementation were essential to ensuring the Government s key policy objective of enabling market operator competition to commence was achieved. Given the large amount of work that ASIC s careful and ultimately successful implementation of this limited subset of the original proposal entailed for both ASIC, market operators and market participants, AFMA would contend that ASIC s responsiveness to this concern for a more practical scale and timing was critical for the successful commencement of market operator competition. As with CP 145, AFMA s key concern in responding to CP 168 is to stress the need to scale back the extent of reforms, prioritise and stage the introduction of a more limited set of reforms over a longer timeframe in recognition of the already busy reform agenda and difficult operating environment. This year already has a heavy workload for participants even without the proposed changes in CP 168. The implementation of the Chi-X commencement is not yet complete, with those connected continuing to switch order flow for national best bid and offer (NBBO) compliance on to the new market, and the majority of market participants not yet connected to either PureMatch or Chi-X. The market is also expecting a number of mandatory market operator upgrades and preparation for Cash Market Margining in Q ASIC has indicated there is also the potential for short-sale tagging to be introduced. Market participants indicate that depending on the implementation details this could cost from $500,000 to over $1 million per large broker given the number of systems that would need to be connected to ensure accurate tagging. Members have indicated it would also have a significant ongoing maintenance cost. In passing we note our strong opposition to short-sale tagging which AFMA holds will offer limited improvement to the existing triple-layered short selling disclosure regime at significant cost to the industry and we would encourage ASIC to again reconsider the merits of proceeding with this proposal, particularly at this time. There is also the wider context of other regulatory changes underway in the sector including the Future of Advice (FOFA) reforms, Basel liquidity changes, the US Volcker Rule, the US Foreign Account Tax Compliance Act (FATCA) and global reforms on OTC clearing. Page 6

9 AFMA recommends that a number of proposals, with particular reference to the expansion of the applicability of Best Execution and clock synchronisation, be removed from consideration for now. Of those proposals that we support in their original or a modified form, we propose that they be scheduled, in a staged way, over a two year period commencing from the finalisation of the relevant guidance. We hold that for many of the proposals, while a six or twelve month implementation period may be reasonable in isolation, when considered in the context of the many other changes that are proposed in CP 168 and that are likely to occur in the market outside CP 168 these timeframes become unrealistic. Given the very ambitious scope of CP 145 and the concerns raised on the scale of CP 168, more generally we would support a review of ASIC s strategy for industry change over the coming years with a view to decreasing the rate of change to ensure optimal orderliness of the process. The high rate of change over the last two years has been costly and difficult for market participants and which, given the longstanding record of stability and confidence in the domestic markets, may not have been necessary. The risks associated with a high-change agenda include excessively high costs, uncertainty, an unavoidable amount of disruption, decreases in process predictability and orderliness, and a lack of clarity. A slower and more deliberate approach can assist, for example, in minimising the need for secondary documentation to explicate rules from primary documentation and the risks of oversights or unintended consequences associated with introducing too many changes at the one time The Flash Crash AFMA understands that one of the main motivations for many of the proposals is to prevent a local occurrence of the flash crash that occurred on 6 th May 2010 in the US. Flash crashes and uncontrolled algorithms are a risk to the market that has emerged with the increased use of algorithms and have many parallels with the fat finger errors associated with the introduction of screen trading. While both the advent of screen trading and algorithmic trading are overwhelmingly positive developments that have led to large increases in market efficiency, updating regulation to mitigate their associated risks is appropriate. In the case of algorithmic trading a measured approach to reducing the risks of flash crashes and other associated issues is warranted on an incremental basis and is supported by AFMA. Page 7

10 While a significant flash crash locally would be highly undesirable and detract from market confidence, it is important that measures to minimise the risk of one occurring should not themselves do more damage. In responding to risks (such as flash crashes) the Taskforce on Reducing the Regulatory Burdens on Business noted: Principles laid down to guide regulatory approaches should require regulators to use a risk-based approach, with any measures to be targeted at specified problem areas, and not designed to eliminate the risk of an event occurring. 6 [Original emphasis] As drafted, the testing requirement proposals come close to requiring perfection from market participants in relation to their algorithms. This is not merely a drafting issue over the choice of words. Even if a reasonability test is introduced we do not consider the approach being taken is the appropriate or most effective approach as it is predicated on an approach that attempts to eliminate risk regardless of cost and practicality. There needs to be greater recognition given to the costs and reasonableness of measures for market players in proposed measures. Given the potential for damage to the industry s competitiveness resulting from the potential increases in the cost burden placed on market users associated with the approach adopted, we are of the view that the approach to flash crash risk should be reformulated IOSCO and Other Regulators The International Organisation of Securities Commissions (IOSCO) plays an important role by providing guidance to national regulators across the globe in relation to international regulatory standards. The high level co-ordination of regulation provided by IOSCO, when used effectively, assists regulated firms by providing more harmonious responses by national regulators to regulatory issues that IOSCO provides guidance on. However, IOSCO is not a policy making body nor is it an international regulator and it does not have accountability to government or the usual parliamentary and other governance mechanisms that exist to assist national regulators. For instance, by the nature of its design, IOSCO cannot perform the cost-benefit analysis that is typically expected to justify any significant regulatory measures in a developed economy data/assets/pdf_file/0007/69721/regulationtaskforce.pdf page In contrast, the Bank for International Settlements has conducted quantitative impact assessments for the Basel 3 reforms to bank regulation. Page 8

11 IOSCO s regulatory standards are not designed for any particular economy and it readily accepts, and indeed expects, that national authorities should consider the relevance of specific IOSCO guidance to their financial markets and economy, having regard to national law and the structure, traditions and practices in their markets. Thus, ASIC (and Treasury as the relevant policy department) must consider the relevance and appropriateness of specific IOSCO regulatory guidance to the Australian market. While it is appropriate for ASIC to be cognisant of recommendations of IOSCO and developments in other jurisdictions, we think ASIC should more carefully question the relevance of IOSCO recommendations to the financial services industry in Australia and avoid the risk to our markets of accepting IOSCO recommendations as unassailable justification for local regulatory measures. AFMA acknowledges that where domestic policy independently and broadly aligns with the approach IOSCO recommends then there may be a case for further alignment with the recommended IOSCO approach in order to coordinate with other jurisdictions. IOSCO may also provide guidance as to the topics that other regulators are considering but this should not be considered to provide evidence that these topics are in fact real issues for the market or are relevant to the domestic circumstances. Further, we note the need to increase the balance given to what our regional competitor jurisdictions such as Hong Kong are doing, given that it is in comparison to these jurisdictions rather than Canada, or the US, that the Australian market must find areas of competitive advantage. Currently there is very limited consideration given to these jurisdictions in comparison to, for example, Canada, which while similar to Australia in some respects is different in others. For example it has competitive advantages in being so close to the US markets and the liquidity boost that provides. 3. MARKET MANIPULATION AND DISORDERLY TRADING With particular reference to inadvertent crossings, AFMA continues to strongly support the reintroduction of an 'intention' element into the civil penalty provisions of the Corporations Act, in respect of market manipulation. We note ASIC has signalled concern around the growth in inadvertent crossings that has accompanied the growth of automated trading. Given the potential for unintentional crossings to be subject to civil penalties ASIC should focus on resolving these framework issues. There are models and technological practices in overseas jurisdictions that may provide assistance in addressing the issue. AFMA would be pleased to work with ASIC towards legislative reform in this area. Page 9

12 4. AUTOMATED TRADING ENVIRONMENT 4.1. Algorithmic Trading As noted, AFMA supports an evolution of the rules around automated trading to reflect its increased presence in the market and our knowledge of the risks associated with this market development from international and domestic observations. Market participants report that trading systems for both agency and proprietary execution are now extremely complex and a single client order can result in multiple algorithmic systems creating market orders through a number of channels to multiple markets and crossing venues. Market participants, recognising their duties to create an orderly market, have, without further regulatory intervention, developed and applied sophisticated processes and standards to algorithm development that would not have been possible for regulators to develop. In other words, the approach to developing targeted and proportionate regulation recommended under better regulation principles is working reasonably effectively in this instance. As a result of this large amount of highly technical work, despite their complexity, in the vast majority of the time these systems work efficiently without incident to the benefit of investors and in accordance with the wishes of clients and with no detriment to market integrity. However, from time to time incidents do occur and the industry supports an alternative approach to reduce their frequency and severity. The approach taken in relation to the automated trading environment by CP 168 in relation to algorithms is to: Create a legal liability that attaches to users of algorithms that participate in incidents including deemed fault provisions that would not apply to manually entered orders 8. (MIR 5.6.3B (1)) Create a legal liability that attaches to brokers when their client s algorithms, that are typically beyond their direct control, participate in incidents.(mir 5.6.3B (2)) AFMA holds that the approach proposed in CP 168 does not sufficiently acknowledge the complexity and limitations of the software testing process, and as a result creates 8 ASIC specifically notes in its rationale that the testing regime should cover the potential flow-on effects, such as where the algorithm generates orders that trigger other algorithms to submit orders, resulting in prices cascading away from the fair value of the products. Manual orders are not subject to this category of cause. Page 10

13 liabilities which are not justifiable or warranted. It creates an undesirable strict liability approach to issues which does not work with the industry to improve outcomes and may result in an inefficient and less effective pro-forma defensive approach to system testing. We do not view the rules as currently drafted as the most efficient evolution of the existing rules and guidance. Limitations of Testing As a starting point it should be acknowledged that, in common with other areas of market activity subject to regulatory scrutiny, some incidents are not reasonably foreseeable or preventable given the complexity of the systems and the impossibility of testing all possible scenarios for all possible market conditions. This is important to understand as this has implications for how best to approach the issue. For example, given this fact it would be unfair and unreasonable to deem fault for incidents. The flash crash itself presents a good example of a situation that would have been very difficult to foresee or test for prior to it occurring. With the benefit of experience the market can establish processes such as volatility controls and testing scenarios that should reduce the risk of recurrence. A typical algorithmic system may contain dozens of operating parameters which would need to be tested against thousands of market conditions and against an unknown number of unknowable other algorithms. The creation of a strict liability to ensure algorithms comply with the Market Integrity and Market Operating Rules is inappropriate given that best practice efforts may not avoid an algorithm participating in an event that is later deemed disorderly. However, changing the language away from ensure and including a reasonability 9 provision would not correct the fundamental problems with this approach. While such a change may make the Rule no longer impossible to fully comply with, it would maintain an unhelpfully negative policing style approach to participation in disorderly events that would lead to inappropriate and inefficient outcomes. For example, under this approach a simple time slice (TWAP) algorithm that was limited to 5% of volume and could reasonably be expected not to cause an incident would need a complex and otherwise unnecessary system to estimate whether an incident was already underway or could be expected to begin if another slice was sent to market. 9 We note a number of instances in CP 168 where the inclusion of reasonability tests through the drafting process would have been helpful. Page 11

14 Further, such an algorithm would require an inefficient and inappropriate testing regime to determine for all market conditions and all possible interactions with other algorithms whether it could cause a market issue. Definition Issues We also note that it may be inappropriate to deem algorithmically generated orders to be at fault merely due to their participation in a market disruption event when manually entered orders need to be proven to be at fault to be deemed at fault. This may discriminate against a particular class of trader. As an example of the inconsistency (B) would presumably not apply to a trading system that generated suggested trades for DTR approval but did not actually generate trading messages. AFMA holds that the rules should apply to both algorithmically generated orders and manually entered orders in a manner that provides regulatory neutral treatment of competing delivery channels. Appropriate testing regimes should form a part of a firm s compliance regime at the macro level but micro-regulation at the algorithm level is problematic given the kinds of discrepancies noted and ultimately non-optimal. Pro-forma Testing and Other Issues The approach proposed by ASIC in Market Integrity Rule (MIR) 5.6.3B (1) and (2) could also have a number of undesired effects. It could, for example, promote a pro-forma approach to testing where the primary purpose of testing becomes producing a legally sound record trail that ticks the boxes rather than finding all possible issues. This approach may also result in ASIC needing to become expert or engage the services on an ongoing basis, at significant expense, of experts in reviewing and formulating software testing procedures and at software quality control and debugging. ASIC may find itself needing to make difficult judgement calls in an area in which it is not specialised. It is not appropriate for a regulator to involve itself in this level of detail in making decisions about business operational matters. The need to ensure that testing procedures form a legally defensible audit trail may add significantly to the cost of testing. AOP Clients All the concerns outlined above also apply from the client s perspective in relation to proposed MIR 5.6.3B (2). In addition, from the market participant s perspective in Page 12

15 relation to the reasonability test it is unclear what reasonable steps to ensure a client s compliance would be. More fundamentally, AFMA would question the utility of using market participants as regulatory agents in matters regarding client actions. Under the previous regime this second-best approach was adopted to recognise the limitations of ASIC s role in interacting with clients of market participants. However, one of the benefits of ASIC taking over market supervision is that ASIC is now able to deal directly with the clients of market participants in relation to matters such as this. The continued use of market participants as proxies is deficient and creates unnecessary liabilities for market participants for the actions of parties over which they have limited control. Best Practice Alternative Rather than an approach of creating a legal hook that attaches automatically to a participant in the event of participation in a market incident by an algorithm (either theirs or their client s), regardless of whether the algorithm was a cause of the incident or whether the market conditions were reasonably foreseeable or should or could have been tested for, AFMA supports instead an approach of working with the industry to ensure testing standards across market participants are consistent and are at an optimal level through the development of a best practice regime for algorithm testing. Such an approach would enable the evolution of testing practice to continue without a focus on creating legal defences and heavy regulatory oversight. It would ensure the regulation and standards are maintained at the macro level and that micro-regulation, and its associated costs and inefficiencies, is avoided. This would also assist ASIC in its education role while providing a benchmark for internal reviews. An advantage of this approach is that a single incident would not be a deemed breach of the Market Integrity Rules. For algorithms that unwittingly participated in a market incident driven by another algorithm, participants would not be required to self-report and ASIC would not be required to investigate. Where an incident or pattern of incidents did give rise to concern, ASIC would have the flexibility to assess the firm s practices in the context of a clearly defined benchmark and determine whether a licence breach had occurred. Page 13

16 Firms that are doing the right thing in terms of testing their algorithms would not have their costs undeservedly increased. ASIC has indicated its support in general for utilising best practice where appropriate and AFMA would be pleased to work with the industry to develop guidelines for application in this area. For AOP Clients the market participant best practice guidelines may be directly appropriate or may assist in informing a similar best practice regime, and again, AFMA would be pleased to contribute our relevant work from the market participant side. Extension to Other Markets We understand that ASIC has no immediate intention to extend these proposals to other markets and would consult further in the event that it did. While there may be some scope to extend a best practice regime for testing, we would need to assess each proposal on a case by case basis as it is difficult to make generalisations without more detailed investigations. AFMA would not support the extension of the proposals as currently drafted or with reasonability tests included Control Over Messages and Monitoring AFMA holds that, as drafted, 5.6.3A imposes an unreasonable and impractical standard of monitoring on market participants. The level of control prescribed in the proposed rules goes beyond what is currently available in commercial monitoring software and may not be possible. This issue should be directly addressed with redrafting; interpretive responses would not give sufficient comfort to the industry. Further, there is much redundancy and overlap in the proposed rules with the existing structures. Here, as in many areas of the market s regulation and guidance the need for a consolidation and updating of existing regulation into a more harmonised package has become more pressing. AFMA members are supportive of a prioritisation of this work to increase certainty and reduce compliance costs before the adoption of any proposals that may further compound the issue. It may be the case that the inclusion of reasonability tests and some redrafting and refinement of scope may achieve the outcomes that ASIC is aiming for without unduly burdening market participants. Page 14

17 Direct and Immediate Control AFMA supports the requirement for market participants to have direct and immediate control over orders sent by AOP Clients a kill switch. This is a logical extension of the requirement to be able to monitor AOP client orders to ensure their compliance with the Market Integrity Rules and Operating Rules. However, as drafted, 5.6.3A (a) requires market participants to inspect the stream of incoming orders and prevent an order or series of orders that may interfere with the efficiency and integrity of the market or proper functioning of the trading platform from entering the market by automated filters and controls. This is a much tighter requirement than for the existence of a kill switch. It also goes much further than the accompanying commentary suggests, which claims it is an extension of MIR The relevant section of MIR requires: (a) having appropriate automated filters to enable Trading Messages to be submitted into the Trading Platform without interfering with the efficiency and integrity of the Market or the proper functioning of the Trading Platform. The existing MIR is focussed on having filters which are, in general, fairly straightforward systems to reject aberrant orders. We note, in passing, that there is a lack of specificity in 5.6.3(a) but that informal good practice standards have emerged over time A(a), in contrast, requires determination of which orders or series of orders (a significantly more difficult exercise) and that these be automatically prevented from entering the market by filters and controls. It would place a large responsibility and unreasonable liability on market participants. It may be impossible to fully implement as it may be impossible to determine which orders or series of orders may have a detrimental effect on the integrity of the market. It would also create a large drag on the market's speed and responsiveness particularly for providers of access to third parties. By definition any order that does interfere with the efficiency or integrity of the market or the proper functioning of a trading platform immediately results in a breach of the Market Integrity Rules for the Trading Participant, regardless of whether such an order or series of orders (over an indefinite timeframe) would have been reasonably possible to detect. A broker may receive up to several thousand orders per second across thousands of products. The current regime allows for filters to make appropriate checks before allowing each on to the market. To attempt to define a regime that will extract each Page 15

18 single order that may cause a problem is in practical terms technically impossible given the amount of analysis required for each of these orders and the markets to which they are destined. To attempt to extract all "series of orders" across an undefined timeframe is a computing problem of the highest order. If it is ASIC s intention to limit the enforcement of this rule to an acceptable level of risk, this would in turn create the issue that participants could not know with certainty what ASIC s view of an acceptable level of risk would be in a particular set of circumstances. Participants are keen to avoid, as far as possible, these types of variances being discovered or determined ex post facto. As drafted, this provision does not fully reflect the realities of modern order flow, the difficulties of determining the potential for market impact across orders, or the speed of modern trading. A workable version of this provision would require brokers to have the ability to disconnect a client connection to the market and withdraw all the associated orders within a reasonable timeframe. We would propose the following drafting which is a clear requirement for a kill switch : (a) having a control for each client connection to the market which enables the timely detection and disconnection of order flow through that connection and removal of orders in the market from that source within a reasonably expeditious timeframe. As currently drafted, there appears to be a significant amount of overlap and redundancy between 5.6.3A (a) and 5.6.3A (b) and between both 5.6.3A (a) and 5.6.3A (b) and (a). We argue that the existing 5.6.3(a) deals sufficiently well with the requirement to have filters to automatically reject orders (or series of orders) that are detected by these systems. If ASIC would like the filters to reject a particular series of orders (which may have dominated a price move for example) then these criteria should be clearly specified. Filter design is best approached on a specification basis rather than a less specific reasonableness approach in contrast to algorithm testing. Consideration should then be given to whether the most efficient venue would be in market participant or exchange-side systems. Exchange-side placement often has benefits not just in efficiency of implementation but also in removing less desirable incentives for competition in filter design by market participants. Page 16

19 The Australian markets have lagged international markets in exchange-side filter implementation and are only now implementing responses to AOP challenges. We understand that market-maker protection systems such as are available on Eurex are not yet contemplated locally. The requirement to automatically remove orders beyond what is required by the automated filter rejections will be dealt with in our revised proposed drafting for 5.6.3A (b). We note the commentary mentions (at 87) the ability to remove connectivity to an exchange but this is not reflected in the draft rules. We would not support including this requirement on grounds of redundancy. Pre-trade Controls 5.6.3A (b) requires monitoring of AOP messages and preventing them from entering the market (presumably using automated systems) or cancelling the resulting orders 10 (presumably by automated or manual means or a combination). Noting our position that (a) handles the automated rejection of pre-market Trading Messages by filters well enough, and that no manual pre-market rejection of AOP orders is technically possible, our redrafting of 5.6.3A (b) focusses on the proposal to require additional monitoring of orders in the market and to cancel these orders. As currently drafted 5.6.3A (b) does not have reasonability tests and may be read tightly to require immediate withdrawal of orders which is a standard that even commercial order-monitoring software may not be able to readily meet. We propose the following drafting: (b) monitoring orders and series of orders and cancelling those orders or series of orders in a reasonably expeditious timeframe where they would be likely to interfere with the efficiency or integrity of the market. Real-time Monitoring and Post-trade Reporting and Analysis As ASIC notes in CP 168 (at 92) there are existing Market Integrity Rules requiring market participants to monitor all of their own and their client s trading. We do not view the rules referred to (5.5.3 and 5.6.3) as being clarified by the proposed 5.6.3A(c) which appears instead to be a significant new requirement. 10 Drafted as Trading Messages but we presume this is a drafting error. Page 17

20 In view of the need to coordinate the proposed rule and its interaction the suspicious activity reporting requirements we propose that the discussion of the rule be deferred until that time BCP and Annual Review of Systems and Connectivity AFMA questions the necessity and appropriateness of regulatory intervention in business continuity planning. Market participants view the proposal to mandate aspects of business continuity planning and subject them to regulatory review as wholly unwarranted, unnecessary and likely to contribute to costs and barriers to entry to market participation without benefit to the integrity of the market. In the alternate, AFMA holds a number of concerns with the proposed drafting of proposed MIR 5.6.3C. Business Continuity General Objection Business continuity planning is an important and technically complicated aspect of modern broking. Particularly for AOP, most large firms have multiple fall back options for their clients for many, or most of their key, systems that they have developed in the absence of regulatory intervention due to the interest of their clients in having these facilities and the associated competitive imperative. However, while back-up options are essential for market operators for the confidence in the integrity of the domestic market (particularly where there is only one execution venue for a particular product) the same is not true at the market participant level as there are many competing brokers. Clients of market participants may choose at minimal cost to have two or more broker connections (as many already do) as part of ensuring their connection to the market. Clients of market participants may also choose to connect their AOP systems via market participants with extensive redundancy systems. Participants with extensive redundancy may use this as a point of marketing difference. As system redundancy (and it is important to note there are many levels of redundancy) necessarily adds to costs, extensive system redundancy measures may not be appropriate for all market participant business models or all systems within a broker. Some small participants may value the decrease in trading costs that may be offered by a broker with more limited back-up systems and be content to use another broker or phone-based arrangements in the event of an outage. Page 18

21 AFMA holds that the case for action has not been clearly established by ASIC for this level of intrusive micro-regulation and ASIC cannot reasonably claim to be promoting confident and informed participation by investors and consumers in the financial system by requiring all market participants to have full redundancy for AOP offerings when: Many brokers already offer full redundancy for AOP; Clients of market participants can and do use a number of brokers and achieve high redundancy through this means; and A single broker s AOP system is not a single point of failure for the system as a whole. The rationale to justify the requirement of business continuity plans in the Market Integrity Rules that: As a matter of good practice and governance, we expect that a market participant should have reasonable business continuity and disaster recovery plans for its AOP systems. This will ensure that any failures can be quickly identified and rectified. This expectation is outlined within ASX Market Rules Guidance Note 22 (which is taken into account by ASIC since the transfer of market supervision). We propose to incorporate this guidance into a market integrity rule. 11 Appears to rely on this line from Guidance Note 22 from ASX: Trading Participants must adopt and enforce written procedures reasonably designed to maintain adequate personnel and facilities for the timely and efficient delivery of customer orders and reporting of executions. In this regard, the procedures should cover capacity reviews, disaster recovery and redundancies, and advance disclosure to customers of both potential systems problems and alternative procedures for customers to use if problems occur. [Emphasis added] The ASX Guidance Note is not equivalent to a MIR and care must be taken in their transition from ASX to ASIC. AFMA holds that the Guidance Note indicating what measures are in keeping with the MIR in terms of business continuity would be better transitioned into an ASIC Regulatory Guide which, as with ASX Guidance Notes, is not an enforceable standard. The proposed translation of the Guidance Note into a MIR significantly changes the requirements on market participants. The Guidance Note suggests 12 having written procedures which cover disaster recovery and redundancies and alternate 11 CP 168 page Being the nature of Guidance Notes. Page 19

22 procedures. The MIR requires maintaining a connection between the AOP system and the Trading Platform at all times. This is a very different formulation. There is an exception where the Trading Participant is unable to maintain the connection, however it is unclear how this exception would work in practice, i.e. what causes of being unable to maintain a connection would excuse a participant from the requirement to maintain the connection at all times. The ASX Guidance Note suggests market participants should aim to ensure clients are aware of redundancies, disaster recovery and alternative arrangements such as they are, while in contrast the proposed MIR requires participants to be able to recover its normal business operations, including the connection between the AOP system and the Trading Platform as soon as practicable after an emergency or other significant disruption which is a very different matter. The Guidance Note cannot reasonably said to translate into the MIR as drafted. AFMA would instead support an ASIC Regulatory Guide that, in line with the ASX Guidance Note suggests that clients should be informed of disaster recovery and redundancy arrangements. The case has not been made for an expansion of the requirement by ASIC at this time, while AFMA holds that the case against expansion is clear. Business Continuity Alternate Argument AFMA is strongly of the view that 5.6.3C should no longer be considered as it is unjustified and will lead to increased costs without a net benefit to the market. In the event that this argument is not accepted we also draw attention to the discussed issues in the current drafting with (1)(a) resulting in any disconnection of an AOP system from the Trading Platform automatically creating a breach in the rules (noting that even many fail-over systems will have a small disconnection) unless the Trading Participant is unable to maintain the connection. Noting that unable may be read narrowly as when it is impossible due to reasons beyond your control for example the exchange was down or a network service provider has a major outage. Any such rule should include a reasonability test that avoids automatic breaches in the event of an outage. The drafting should not require as soon as practicable restoration of service but focus on ensuring that reasonable measures are in place that are designed to allow the restoration of service in a reasonable timeframe. This is to avoid situations where a market participant, operating in good faith and with reasonable measures in place, would be in breach of their licence in the event that they were unable to quickly recover connections dues to outages at both their main data centre and disaster recovery centre. This situation may occur, for example, due to unrelated communications outages. Page 20

23 4.4. AOP Attestation Changes AFMA does not support the proposed changes to AOP attestation systems as they are insufficiently targeted. Members report that, given the number of systems affected, for some members it would create a need to continuously certify systems throughout the year to meet the yearly requirement. Rather than requiring all systems to be certified annually, in addition to the current regime, participants should be required to have a certification for the system if there have been significant cumulative changes in the system or the environment in which the system operates since the last attestation. This would avoid unnecessary annual certifications of systems that may have been recently recertified or operate in an environment that has not changed in the past year. AFMA would be pleased to work further with ASIC on refining this revision to the proposal Minimum Standards for Direct Electronic Access Overall, AFMA considers the current Market Integrity Rules, which refer to Authorised Persons and AOP, are sufficient to ensure that direct market access and resultant order flow is properly controlled. The proposal supplements the current adequate principlesbased requirements which ensure that market participants prudently control their clients access to the market. There are expectations in the proposals of more intrusive due diligence on a client control framework before enabling a client s access to the market. The case justifying this level of regulatory intervention has not been made out. The proposal introduces significant additional requirements for direct access clients that do not apply to clients not using direct electronic access (DEA). This creates an unbalanced regulatory environment which discriminates against one class of investors. In effect the MIR is imposing on the market participant a due diligence exercise which is equivalent to ASIC s decision making process around granting financial services licenses. If this is ASIC s regulatory objective the more straightforward approach would be for ASIC to impose obligations on DEA clients directly. This proposal places unrealistic expectations on market participants regarding assessment of clients, because of the vague and subjective nature of the criteria. To make this proposal workable ASIC would need to provide detailed guidance to market participants on how they are to conduct assessments against the criteria. Page 21