What is the worth of net worth?

The Securities and Exchange Board of India (SEBI) announced today:

Presently, mutual funds are not allowed to appoint a custodian belonging to the same group, if the sponsor of the mutual fund or its associates hold 50 per cent or more of the voting rights of the share capital of such a custodian or where 50 per cent or more of the directors of the custodian represent the interests of the sponsor or its associates.

The Board has decided that the custodian in which the sponsor of a mutual fund or its associates are holding 50 percent or more of the voting rights of the share capital of the custodian, would be allowed to act as custodian subject to fulfilling the following conditions i.e. (a) the sponsor should have net worth of atleast Rs.20,000 crore at all points of time, …

To provide a perspective on this, the last reported net worth of Lehman was $19.283 billion which is about five times the Rs.20,000 crore stipulated in the above announcement. (The Lehman figure is from the quarterly 10-Q report filed by Lehman on July 10, 2008 about two months before it filed for bankruptcy.)

Even assuming that the reported net worth is reliable, what I fail to understand is the implicit assumption in the world of finance that wealthy people are somehow more honest than poor people. As far as I am aware, the evidence for this is zero. This widely prevalent view is simply the result of intellectual capture by the plutocracy.

Capital in finance has only function – to absorb losses. I would have understood if SEBI had proposed that a variety of sins of the custodian would be forgiven if it (the custodian and not its sponsor) had a ring fenced net worth of Rs.20,000 crore invested in high quality assets.

CBOE 2013 versus BSE 1993

Earlier this week, the US SEC imposed a fine on the Chicago Board Options Exchange (CBO) for conduct reminiscent of what used to happen in the Bombay Stock Exchange (BSE) two decades ago. In the early 1990s, the BSE board was dominated by broker members, and allegations of favouritism, conflict of interest and neglect of regulatory duties were very common. At that time, many of us believed that these were the kinds of problems that the US SEC had solved way back in the late 1930s under Chairman Douglas. India might have been six decades behind the US, but it is widely accepted that security market reforms in the 1990s solved this problem in India, though this solution might have created a different set of problems.

The SEC order reveals problems at the CBOE which are very similar to those that the BSE used to have in the early 1990s:

• The SEC order concerns the CBOE’s treatment of a member firm whose CEO was a member of CBOE’s Board of Directors and sat on CBOE’s Audit Committee (para 30 and 60). Incidentally, if you glossed over these two sentences while reading the entire 32 page order, you would completely miss this crucial fact. The SEC clearly did not want to highlight this issue at all and does not mention it in its press release.
• “Not only did CBOE fail to adequately detect violations and investigate and discipline one of its members, CBOE also took misguided and unprecedented steps to assist that same member which was under investigation by the Commission’s Enforcement Division staff and failed to provide information to Commission staff when requested.” (para 22)
• The CBOE helped the member firm to respond to a notice from the SEC by providing it with a summary of its (CBOE’s) investigative file despite knowing that CBOE investigations, and information obtained from other regulators during those investigations, were to be kept confidential. The member firm asked CBOE to provide “review, modification and insight” into its response to the SEC notice and the CBOE edited the response and emailed the “redlined” edits to the member firm. (para 34-35)
• When informed that the CBOE Department of Member Firm Regulation planned to issue a notice to this member firm for operating a non-registered dealer, CBOE’s former President and Chief Operating Officer asked that the notice not be issued until after an upcoming meeting with the member firm’s CEO (para 60).
• “CBOE made several financial accommodations to certain members [including the above member], and not to others, that were not authorized by existing rules. … The accommodations were made for business reasons and were authorized by senior CBOE business executives who lacked an understanding of CBOE’s legal obligations as a self-regulatory organization.” (para 65 and 66)
• “CBOE staff responsible for the Exchange’s Reg. SHO surveillance never received any formal training on Reg. SHO, were instructed to read the rules themselves, did not have a basic understanding of what a failure to deliver was, and were unaware of the relationship between failures to deliver and a clearing firm’s net short position at the Depository Trust and Clearing Corporation (‘DTCC’). … In fact, the investigator primarily responsible for monitoring the Reg. SHO surveillance from the third quarter 2009 to the second quarter 2010 had never even read the rule in its entirety, but only briefly perused it.” (para 14)

In financially regulation, no problems are permanently solved – potential problems just remain dormant ready to resurface under more favourable conditions.

The Baselization of CCPs

There was a time when central counter parties (CCPs) used robust and coherent risk measures. Way back in 1999, Artzner et al. could write that “We do not know of organized exchanges using value at risk as the basis of risk measurement for margin requirements” (Artzner, Delbaen, Eber and Heath (1999), “Coherent measures of risk”, Mathematical Finance, 9(3), 203-228, Remark 3.9 on page 217). During the global financial crisis, while Basel style risk management failed spectacularly, exchanges and their CCPs coped with the risks quite well. (I wrote about that here and here).

But things are changing as CCPs gear up to clear OTC derivatives. The robust risk management of CCPs is not percolating to the OTC world; instead, the model-risk infested risk measures of the OTC dealers are spreading to the CCPs. The OTC Space has a nice discussion of how the systems that CCP use to margin OTC derivatives are different from the systems that they use for exchange traded derivatives. No, the CCPs are sticking to expected shortfall and are not jumping into value at risk. But their systems are becoming more model dependent, more dynamic (and therefore procyclical) and more sensitive to recent market conditions. These are the characteristics of Basel (even with Basel’s proposed shift to expected shortfall), and these characteristics are gradually spreading to the CCP world.

I am not convinced that this is going to end well, but then CCPs are also rapidly becoming CDO-like (see my post here) and therefore their failure in some segments might not matter anymore.

The NASDAQ Facebook Fiasco and Open Sourcing Exchange Software

Last week, the US SEC issued an order imposing a $10 million fine on NASDAQ for the software errors that caused a series of problems during the Facebook IPO on May 18, 2012. I think the SEC has failed in its responsibilities because this order does nothing whatsoever to solve the problems that it has identified. The order reveals the complete cognitive capture of the SEC and other securities regulators worldwide by the exchanges that they regulate.

The entire litany of errors during the Facebook IPO demonstrates that critical financial market infrastructures like exchanges and depositories should be forced to publish the source code of the systems through which their rules and bylaws are implemented. Of course, the exchanges will complain about the dilution of their “intellectual property”. But the courts have whittled down the “intellectual property” embedded in standard-essential patents and this principle applies with even greater force to software which implements rules and bylaws that are effectively subordinate legislation. Financial regulators have simply fallen behind the times in this respect.

What is the point of an elaborate process of filing and approval for rule changes, if there is no equivalent process for the actual software that implements the rule? The SEC order shows several instances where the lack of disclosure or approval processes for software changes made a complete mockery of the disclosure or approval processes for the rules and regulations themselves:

• “While NASDAQ’s rules had previously provided for a randomization interval at the conclusion of the DOP, in 2007, NASDAQ filed a rule change removing the randomization period from its rule. … However, the randomization function had never been removed from NASDAQ’s systems, and therefore the IPO Cross Application for Facebook – and for all other companies that had an IPO on NASDAQ since August 31, 2007 – was run after a randomized period of delay in contravention of NASDAQ’s rules.” (para 16).
• While installing an upgrade to the NASDAQ trading systems, an “employee misinterpreted the instructions associated with the upgrade and assumed that the SHO Through application was not needed and could be removed from the system. As a result, the employee removed the SHO Through application.” A second employee was responsible for checking the work of the first employee “also misinterpreted the upgrade instructions to mean that the SHO Through application could be removed” Personnel running the daily configuration test for the exchange’s trading systems“ received a system alert based on the fact that the SHO Through application was no longer part of the system. … they also thought the SHO Through application could be removed.” The error was detected only several days later in response to an enquiry from a trading member.(para 52-54)
• With inadequate understanding of the software bug that was causing problems in the Facebook IPO, the exchange implemented a hasty software change to bypass a validation check with full knowledge that this would cause “the exchange itself to take the opposite side of the mismatch ” caused by the removal of the validation check. However, “NASDAQ did not have a rule that allowed NASDAQ … to assume an error position in any listed security.” (para 24 and 28)

The Facebook fiasco was itself the result of an infinite loop in the software. This infinite loop would almost certainly have been detected if the source code had been publicly released and discussed with the same attention to detail that characterizes rule changes.

The lack of well defined processes for software testing is revealed in this tidbit: “Given the heightened anticipation for the Facebook IPO, NASDAQ took steps during the week prior to the IPO to test its systems in both live trading and test environments. Among other things, NASDAQ conducted intraday test crosses in NASDAQ’s live trading environment, which allowed member firms to place dummy orders in a test security (symbol ZWZZT) during a specified quoting period. NASDAQ limited the total number of orders that could be received in the test security to 40,000 orders. On May 18, 2012, NASDAQ members entered over 496,000 orders into the Facebook IPO cross.” It should be obvious that the one thing that could have been anticipated prior to the Facebook IPO was the vastly greater volumes than in small time IPOs. Doing a test that excluded this predictable issue is laughable. Proper rules would have required the postponement of the IPO when the volume exceeded the tested capacity of the system.

It is my considered view that the SEC and other securities regulators worldwide are complicit in the fraud that exchanges perpetrate on investors in their greed to protect the alleged “intellectual property” embedded in their software. I have been writing about this for a dozen years now: (1, 2, 3, and 4). So the chances of anything changing any time soon are pretty remote.