-!b: Jtlt Ms. Maninder Cheema Deputy General Manager, Market Regulation Department - Division of Policy, Securities and Exchange Board of India, SEBI Bhavan, Plot No. C4-A, "G" Block, Bandra Kurla Complex, Bandra (East), Mumbai - 400 051 'o]3, r,, / "-Yl-.,. v!/.:..: I, 20ll Sub: Comments on Proposal on Co-location / Proximity Stock Exchanges. offered try the We write with reference to SEBI discussion paper on the above subject and give our comments on the same. Background: We believe that Co-location facility provides opportunity for users to achieve low latency access to a liquidity venue thus increases a platforms' competitiveness in attracting market users who have trading strategies that are time sensitive. Further, the effect of Co-location on the wider market include the opportunities that stem from a new service offering, for example technology companies and data centre companies that provide the infrastructure required to support colocation. In addition, generally by offering co-location service execution venues are able to attract new participants to the Indian market which enhances the benefits to the wider market with increased trading activity and available liquidity. About the document:. Analyses SEBI proposals associated with the trading from co-location and regular non Co-location facility. Outlines pros and cons of the proposals in the discussion paper o Provides reference about global practices and trends in the co-location space' NATIONAL SECRETARIAT 304-A, Veena Chamber, opp. BSE' 21, Datal Su""t''on' - o* *ii i" "'no"' " sorrrrnn,*o,o *ru,onnl courlcll (T.K.x. 0HAPTER) Cout{clL (A P CHAPTER) lgul!$ 1IDll[9ONAt ;",.*i,:,n1,fu:a "; Ta"flffi,iJ;ilI"ffH**,oooo, fi:13:"".lle;;fi#il'il". rereraxi ori-23715r31 /32 reretax: oi!'-)iiiriiiq *--- ctrennar--aoo os; - - Naravanaouda HYderabad'50002e I:t onoo6zzssza Fax 14o 27562182 i;,ui:inmin,omrn",n.,",,*i"#6,'i,.'*, f**l"ia,"rmg:l:"""", Yi?l,i"ll#333i3f"uo**.,."", Website : www.an mi in
Response to the SEBI proposal in the discussion paper Proposal #6.1 Stock exchange shall provide colocation i proximity hosting in a fair, transparent and equitable manner, including ensuring that all stock brokers and data vendors located at co-location / proximity hosting experience similar latency between their systems and the stock exchange's trading platform. Our view: 1. We welcome SEBI's proposal of providing hosting facilities in a fair and transparent manner as it addresses normalized latency for all market participants in co-location facility. Going by the global standards listed in subsequent section, providing an equal-cable-length from every rack space within the co-location to the matching engine of trading venue will standardize the transmission latency and thus will not give any participant an unfair advantage within cojocation space. 2. Global reference: As per Australian Securities and Investment Commission (ASIC) report 123-May-2013)t, cable length and network infrastructure are provided in the same way for all users within each service in cojocation facility. Similarly Hong Kong Exchange also provides equal length cable to all hosting users regardless ofrack location. b. In Tokyo Stock Exchange (TSE) Co-location facility, rack space is priced equally with no preferential treatment to any market participant. 1 Market Assessment Report: ASX Group by ASIC dated May 2013.
Proposal #6.2 Stock exchange shall ensure that the size of the co-rocated / proximity hosting space is sufficient to accommodate ai the stock brokers and data vendors that are desirous of availing the facility. Stock exchange shall also avoid situation of monopolizing of rack space by certain stock brokers or data vendors. Oirr view: L we concur and appreciate this initiative and believe that this wi, surely address the_faimess and equality objective of this proposal. We firmly believe that it is up to the individual member to decide whether the incrementar benelit ofco-location outweighs the incremental cost ofthe incremental infrastructure. 2. we also recommend that within co-rocation facility different sizes of rack should be made available for members and participants may elect to use the services suitable to their trading needs. 3. Global Reference: a.to accommodate more trading members, TSE recently moved from their old co_ location space to the new facilities (JpX)2 which ha. more number of available racks. b.as per ASIC reportl, access to market data is available to all users at the Co_ location facility through different services. Each service has a diil-erent latency profile, which is provided the same way for all users within each service. 2 Japan Exchange Group (lpx) was established via the business combination between Tokyo stock Exchange Group and osaka Securities Exchange on January 1, 2013
SEBI Proposal #6.3.1 Stock exchange shall identif, and categorize orders as (a) orders emanating from servers ofthe stock broker placed at the co-location / proximity hosting facility, and, (b) orders emanating from other terminals / servers ofthe stock brokers. Our yiew: 1. We believe that tagging of order based generated either from Co_ location/proximity or from Non-CoJocation facility will be useful for statistical reasons. This can always pave way to study correlation of various investor volumes and nature oftrading. SEBI Proposal #6.3.2 Separate order-validation mechanism and a separate queue shall be maintained for each of the aforementioned categories of orders. Our view: l we would like to seek more crarity about the separate order validation mechanism proposed in the paper. Does separate means same set of risk validations separately applied on both the queues or creating different risk validation for cojocation queue? 2. we would arso like to bring attention that cunently algorithmic orders are already sub-lected to extensive risk validation as prescribed by SEBI/Stock e*charges besides standard risk validations applicabte for stock brokers.
SEBI Proposal #6.3.3 and 6.3.4 A round-robin methodology shall be used to time-stamp and forward validated orders from the tr,vo order-queues to the order-book, i.e., if an order is taken from the queue of orders emanating from co-location i proximity hosting facility, then the next order shall be lrom the other queue. In the event any of the order-queues are empty, orders can be sequentially taken from the other queue till a valid order arrives in the empty queue. 'The time-stamp given as per paragraph 6.3.3 above shall be used to determine the time priority during matching of orders. Our view: 1. We fully understand and appreciate the intent of having two separate queue mechanism, however we believe further empirical study is required to ensure there are no unintended consequences of this proposal. Besides orders placed by non co- location intermediary and retail investor should not suffer in the bargain. 2. One ofthe consequences that we think and would tike to highlight with advent of two queue mechanism is that clients can run their algo by ptacing multiple orders from both co-location and non cojocation facilities. Once either order is executed, the remaining order would be canceled. This would only serve to add "noise" to both trading and market data. 3. Given that separate queue system may potentially benefit non-co-location orders, we believe that this could pave way for new set of predatory algorithms3 from participants to grab lost opportunity from cojocation queue. 3 Fleeting is one such predatory practice. As per ASIC report in May 2013, an order may be described as fleeling' if it fails to rest, within a market, for a meaningful period of time. This liquidity, although posted, is effectively inaccessible because investors are unable to trade purposefully against it.
a As per study conducted by one of the of the participants it is 3.3 bps in the index universe @anmi 4' cunently, exchanges use price / time priority principre to determine the sequence in which submitted orders must be executed i.e. orders at same price is executed at first-come-first serve basis. our assumption from the proposar is that exchanges may introduce an intermediary layer between queue and order book for roundrobin allocation of orders from both the queues; subsequently regular frequency matching priority will be assigned on same price order / quotes., Therefore, routing orders altematively may arso not be just and equitable. It is feared that it may affect the price discovery system adversely. System may buy at higher price and may sell at lower because of two different order streams where complete depth ofmarket is not known. we will await more crarity from exchanges on the exact implementation of this proposal and arso impact of new matching logic on existing price/time priority principle. Conclusion: 5. With respect to above point, we also think that a sub-optimal frequency of assigning matching priority to orders may lead to disruptive irad ing pattern. This in particular may lead to low participation of high frequency o.d".. f.o,, "olocation facility and thus have potential to impact the tight bid-ask spread aof liquid stocks (which is at present most competitive urnong elec."gion.) creating separate queues may lead to predatory practices ofsending orders into both the queues which may lead to more noise in messages and market data At this moment, we are also unclear about the implementation detairs of the two queue methodology such as the periods on time stamp and mechanism on matching during each cycle.
o Rack space should be uniformly priced with no preferential treatment to any clients. Latency from rack space should be uniform by maintaining same cable length Therefore, routing orders altematively may also not bejust and equitable. It is feared that it will affect the price discovery system adversely. System may buy at higher price and may sell at lower because oftwo different order streams where complete depth of market is not known. We therefore request SEBI to consider the above issues and defer the proposal of paragraph 6.3 till the matter is reviewed in greater depth. Thanking You. President ANMI