February 05, 2007	Monday	Muharram 16, 1428

THE planned switch at stock exchanges from broker-level to client-level netting has been delayed by one month-- from February 1, 2007 to March 1, 2007. Client-level netting is the last and the most important element of new netting methodology introduced by SECP in 2006 to improve management of settlement risk.

It is safe to say that it would remain a major issue in coming months, if not all of 2007. Here we discuss what is meant by switching to client-level netting, its justification and issues involved in its implementation.

Broker-level netting means offsetting security-wise unsettled sales and purchases across different clients of a broker to reduce a broker’s cumulative unsettled trades called “exposure.” For instance, if a broker buys 500 shares of PTCL for one client and later sells the same for another client, stock exchanges offsets his purchase and sale, reducing his exposure to nil.

Client-level netting means that unsettled purchases of a client would only be offset against his own sales. That is, in our example, exposure of broker would not be nil but equivalent to value of 1000 PTCL shares (500 buy + 500 sell). Client-level netting also means that trading done by a broker on his own account (proprietary trading) would not be netted with that of his clients.

Netting is important because (i) margin deposits that exchange collects from a broker to manage risk of broker default are a function of his exposure and (ii) capital adequacy of a broker (or his maximum permissible exposure as a multiple of his net capital balance) depends on how exposure is calculated. Lesser netting means higher exposure which in turns means higher margins and lower effective capital adequacy. Both of these would reduce ability of stock brokers to carry out speculative trading, from which they earn most of their commission and trading profits.

Brokers are supposed to trade as per their financial capacity and keep adequate margins on both proprietary trading as well as trading for clients. Not every one complies and it is the mechanics of broker-level netting that allow large exposures on thin margins, which in volatile times create threats of settlement crisis.

Core idea underlying client-level netting is to pre-empt under-margining by forcing every one trading in market to deposit applicable margin to the exchange. Case for client-level netting is supported by (i) investigation reports in May 2000 crisis and March 2005 crisis and (ii) anecdotal evidence in June 2006 crisis, that identify excessive netting as a cause of these crises. The first directive by SECP on client-level netting came almost two years ago on March 4, 2005. Later, after extensive consultation, another directive was given in September 2006 setting February 1, 2007 as its implementation date. With the problematic ‘badla’ financing setting a new record-high level every other day, regulator is aware that it has to act before any shock creates a new crisis.

According to brokers, client-level netting could reduce turnover at a time when it has already dropped substantially. Low turnover, they argue, affects liquidity and price discovery. The argument is not very convincing. It is common knowledge that most of market activity is leveraged speculation rather than investment. High levels of leveraged speculation lead to price bubbles (not price discovery) and then sudden disappearance of buying interest (not liquidity) as experienced in March 2005. However, brokers have rightly pointed out that institutional clients, such as mutual funds, due to external or internal rules, do not deposit any margin to anyone. Since only large brokers have the resources to submit margins for institutional clients, margin regime has a bias against less-resourceful brokers, which would be magnified in client-level netting. KSE is working on providing a new facility to allow institutions to deposit margins directly to the exchange, however, it is the now proverbial “capacity constraint” that is said to be hindering progress.

Institutional investors, which are not submitting margins to the exchanges, appear open to the idea of a new system of direct margining to the exchange. Individual investors trading small number of shares would remain unaffected because they are already depositing adequate margins to their brokers. Affected by this reform would mainly be those brokers who exploit broker-level netting for leveraged speculation and its beneficiary would be market at large because it would help reduce chances of more settlement crises.

There are a number of issues involved in Implementation of client-level netting. First, an unhealthy precedent has already been set in implementation of new netting regime. Just days before implementation of its first phase in November 2006, new netting regime drew strong protests from brokers who got its implementation stretched to October 2007. They also secured a number of other concessions, such as revision of initial margin rates for a number of securities, increase in broker-wise quantitative limits on margin securities, sharing of interest on cash margins in futures etc. Given this precedent, few are willing to believe that client-level netting can be implemented without fireworks and re-negotiations.

Second, UIN system, which is the foundation for client-level netting, may not be fully ready for the purpose. To hasten its implementation in 2006, UIN system was launched without any central registration of investors or verification of information being entered. Instead of a standardised key, different types of number formats were allowed and investors were also not given any control over the use of their UIN by a broker. Add to it that six months have passed but exchange regulations governing use of UIN are yet to be formed and no audit has been initiated to verify investor information. All this has caused quality and reliability problems in UIN data base which, reportedly, is being cleaned for client-level netting.

Third, some basic questions about client-level netting that have not been answered by exchanges and SECP. Consider the following. By how much exactly would it increase margins and shrink capital adequacy? Would it encourage off-system trading? Since NCCPL would continue to use multilateral netting in settlement, when would margins taken by exchange against in-house broker-level settlement be released? Have exchanges procured the required IT hardware? Would adequate time be available for testing software modifications and mock trading? Would incremental margins be taken immediately or phased in? How would it be ensured that one-client’s securities are not pledged against another client’s exposure? How would it cover inter-exchange trading where UIN system is not fully applicable? How would exchanges detect if a broker bypasses pre-trade UIN verification by using same client-code for different clients?

Fourth, there is an important legal issue in margining that has not been resolved for far too long. Simply put, do stock exchanges have the legal right to sell margin securities of a client if his broker defaults? It is possible that securities could be of clients who didn’t even have any exposure with their broker. Where such securities are pledged from his sub-account, investor would have irrefutable evidence in the court of law that it was his securities that were liquidated by the exchange for no fault of his. Given that investor protection is a fundamental responsibility of exchanges and SECP, this issue must be addressed while implementing client-level netting.

In sum, switching from broker-level to client-level netting is the right thing to do but given its strong impact and the many unanswered questions surrounding it, its implementation is bound to be challenging. Let’s hope that in the interest of investing public, SECP and stock exchanges would be able to live up to this challenge.