Wednesday, September 30, 2020

Alarm Raised on Stock Loans for "Withholding Tax Schemes”

Findings Point to a New Role for Emerging Fintech

Author: David Schwartz J.D. CPA

European commissioners are reviewing a study from their securities and market authority (ESMA) that includes a recommendation for new laws to combat unfair trading practices and an extended remit for National Competent Authorities (NCAs) to conduct snap audits of securities loans and transactors. Loans deemed to be suspicious would prompt an inquiry to determine penalties for unfair strategies and inappropriate beneficiaries. However, useful audit results may be doubtful based on our preliminary review that uncovered shortcomings in the proposed SFTR surveillance datasets, as well as possible flaws in the study’s basic methodology. 


The ESMA report describes a series of unfair trades involving securities loans, called "WHT Schemes" in reference to the withholding tax on dividend payments. The first named among the disputed WHT Schemes is dividend arbitrage, wherein an investor in a high tax regime lends an income-producing security to a borrower in a lower tax regime. [1] The ESMA report lays out a variety of similar WHT schemes, which are labelled Cum/Ex and Cum/Cum. The report's conclusion is that, while most of these trades are not necessarily illegal, they have to be stopped because they’re “unfair.”  The loans are considered unfair because, by their nature, the loans create positions for investors which avoid and reduce tax revenues for certain EU member states. However, “At this stage no criminal court judgment has been made available to ESMA, declaring the above practice [“cum-ex” trades] as fraudulent.” 


Datasets and Limitations


ESMA’s report bases its recommendations on a study performed with the help of unidentified “commercial data providers.” It was understood that, in both the trading and lending datasets, proxy activity and recalls might distort the spikes around concurrent dividend record dates in the securities lending analysis. Unstated was that even more recalls might have been generated by institutional lenders bringing back securities on loan because they are not permitted to accept manufactured dividends nor cash-in-lieu from borrowers. Despite these shortcomings in the datasets, significant weight is being given by ESMA to the consistency of increases in stock lending volumes near dividend dates. Given that most WHT schemes involve high trading levels, per ESMA, the analysts concluded that evidence of the disputed trades, notwithstanding their unavoidable data limitations, could be found in the volume spikes on record dates for dividend-paying stocks. Therefore, while awaiting new legislation, NCAs can design their snap audits to be selective, targeted at high-volume transactors in likely issues.


Methodological Background


ESMA’s methodology is reminiscent of the 2006 study (the "Christofferson Study") that led to an SEC investigation of “empty voting.” In that matter, a team of academics had misconstrued the record date spikes as proof positive of manipulation by hedge funds of the proxy voting process. Their theory assumed (incorrectly) that all trades were new loans. In reality, many were either substitutions or returns/recalls of existing loans. A subsequent and similarly misinformed article by law professors Henry Hu and Bernard Black indicted securities lending as a vehicle for widespread abuse of the corporate governance process. In response, a multi-year RMA/SIFMA study of 800 million transactions was conducted by CSFME and presented to the SEC in 2010. Based on the rebuttal, the SEC concluded that there was no manipulation. Quite the opposite, the spike was caused by institutions who were recalling their securities loans to vote and collect dividends to which they were entitled.[2] The benign causes of the dividend-related spikes in the stock loan dataset were again confirmed in 2011 by an independent academic study.[3]


Predictable Analytic Constraints


The ESMA report does not explain how their analysts calibrated the volume statistics, so there is room to question the findings at this point. Moreover, even if dividend spikes can be isolated and investigated, the results may duplicate those of earlier audits in the UK that concluded most suspicious loans were originated by institutional investors who were legitimate dividend recipients -- as expected, since it is the borrowers who are usually the initiators of disputed WHT loans. Surveying or auditing lenders alone is not enough: unless the entire fungible transaction is mapped, it will not be possible to uncover and prove the strategy being used in the trade. For that reason, a more realistic goal for ESMA may be simply eliminating the permissible trades from further scrutiny. However, even that may be beyond reach of the SFTR datasets: the original field layouts for the transmission tables did not include codes to distinguish transactions that are generated by recalls by lenders and returns from borrowers. Therefore, NCA auditors will not be able to distinguish new loans from the terminations of existing loans. That will be a severely limiting constraint on ESMA's surveillance efficiency, especially at a time when volatility linked to recalls seems poised to increase, perhaps significantly.


A recent digital stock dividend for generated very high volumes around record date. The operational uncertainties of digital stock dividends on stock loans was too much for intermediaries to handle as hedge funds closed out their positions with prime brokers, and more than half the surveyed lending agents stopped making new loans and simultaneously recalled existing loans.[4] Overstock's digital stock dividend passed through its own alternative trading site (tZero) into investors' retail brokerage accounts in much the same way that ESMA hopes to trace the dividend payout directly to the beneficial owner. According to the RMA, the solution may lie in the use of distributed ledger technologies: 


This case reveals the existing infrastructure around ownership is outdated, creating risk for both brokers and shareholders alike. A distributed ledger provides an effective real-time track record of ownership, making it easy to attribute relevant entitlements such as dividend payments. ... The industry must come together, considering that digital assets may be the new normal, and decide how best to prepare.[5]


Market regulators, like ESMA, and issuers, like Overstock, are waiting to see how quickly the securities finance industry will figure out how to manage the entitlements for corporate actions in a token-friendly market infrastructure.

by Ed Blount and David Schwartz



1. The two share the tax savings and the borrower returns the shares.

2. CSFME, and RMA. “Borrowed Proxy Abuse: Real or Not?”, 2010,  Accessed 30 September 2020 

3. Moser, Shane M., et al. “Securities Lending Around Proxies: Is the Increase in Lending Abuse or a Result of Dividends?” The Journal of Financial Research, vol. XXXVI, no. 1, Spring 2013, pp. 1-17.  

4. “RMA’s Financial Technology & Automation Committee found that 54% of agent lenders who owned shares decided to recall outstanding loans and restrict lending,” said Fran Garritt, RMA’s Director of Credit Risk, Global Markets Risk, and Securities Lending.

5. Risk Management Association, "The New Normal: Digital Asset Corporate Actions," September 2020, page 7.

Categories: All, Commentary, Cross-Post