1) Historical Background
Time was not long ago, paying tax on U.S. – source dividends was optional for offshore investment funds. It was easy for funds to avoid withholding tax; instead of purchasing shares, a fund would enter into a long position in a swap on a name or basket of names. Pursuant to the terms of the swap, the fund would receive periodic “reset” payments equal to appreciation (if any) on the underlying shares, as well as dividend-equivalent payments whenever the issuer paid dividends. In return, the fund paid its counterparty periodic payments equal to depreciation (if any) on the shares and a financing fee equal to the product of the value of the shares referenced by the swap and a funding rate. The result was that the fund could gain economic exposure to the underlying shares without being subject to United States withholding tax on dividend-equivalent payments.
Support for this position came from a treasury regulation promulgated in 1991, which specifies that the source of payments made on notional principal contracts is determined with reference to the residence of the recipient.  Therefore, under this rule, to the extent that an offshore fund was treated as a U.S. tax nonresident, payments received by the fund under an equity swap were treated as foreign-source income. Because the United States federal withholding tax regime only applies to United States-source payments, such as dividends paid on shares issued by United States – resident corporations, this exempted dividend-equivalent payments on equity swaps from the general 30% federal withholding tax on certain U.S.-source income.
At first blush, this rule appears to burden similarly-situated taxpayers differently. For example, assume that Fund A holds 100 shares of XYZ stock, and Fund B holds a long position in an equity swap that references 100 shares of XYZ stock. XYZ stock is trading at $100 per share. Fund A borrows $10,000 from Bank to fund its purchase of the shares, and pays Bank interest equal to $10,000 multiplied by 3-month LIBOR plus a spread of 100 bps. Pursuant to the terms of the swap, Fund B will receive payments equal to any appreciation on a notional position of 100 shares of YX stock and any dividends paid on the same notional position. In return, Fund B will pay Counterparty amounts equal to the depreciation on the same notional position, and a funding leg equal to 3-month LIBOR multiplied by a notional principal amount of $10,000. Counterparty hedges by buying 100 shares of XYZ stock (which it can fund at, say, 3-month LIBOR plus 35 bps).
Diagram 1: Fund A Borrows $10,000 to Purchase Shares.
Click here to view diagram.
Diagram 2: Fund B “Receives Performance” on an Equity Swap.
Click here to view diagram.
Now, assume that XYZ pays a dividend of $1 per share. Under prior law, after withholding, Fund A would receive $70, while Fund B would receive $100. Since Fund A and Fund B both have “delta one” exposure to 100 shares of XYZ stock, how could this possibly be the right result? 
Until 2010, the answer was, “It depends”. In many cases, the economic position of a taxpayer who holds a long position in an equity swap in fact differs significantly from that of a taxpayer who holds the cash underlier; however, in many other cases, an equity swap is, in substance, merely a custodial arrangement in drag. Attention to the facts of each case is needed to distinguish the two. Under common law rules in effect prior to 2010, dividend equivalent payments on equity swaps that were true derivatives were not subject to withholding tax. However, dividend equivalent payments on equity swaps that were in substance custodial arrangements dressed up like derivatives were treated as actual dividends, and, as such, were subject to withholding tax.
In cases involving a “true” derivative, a fund will generally enter into a long position on liquid, publicly-trades shares with a counterparty who is not required to hold the referenced shares (the “underlying shares”, or the “underlier”) under the terms of the swap. In these cases, the counterparty might purchase the underlier in order to hedge its short position in the swap, but it also might hedge by offsetting its position in the swap against other customer-facing positions, or it might hedge all or a portion of its position with other instruments. The swap’s strike price generally reflects the price of the underlier at some time during the day on which the swap was entered into, but the strike price generally does not reference the price at which the hedge was purchased by the counterparty. In cases such as these, the better view is that, although the long party has “delta one” exposure to the underlier, the long party should not be treated as the owner thereof. There are several reasons for this. First, under general common-law tax principles, ownership of liquid, fungible property is determined primarily by reference to possession and control: the party that has the power to dispose of, say, 100 shares of XYZ stock, should be treated as the owner thereof, regardless of whether another party has economic exposure to the shares.  Therefore, because the counterparty, rather than the fund, has the right to purchase, hold or dispose of the shares held as a hedge, the better view is that the counterparty, rather than the fund, owns these shares. Second, there are important legal and regulatory differences between the position of an owner and that of a long party to an equity swap. An owner of shares in a brokerage account is subject to “Regulation T” margin requirements, while the holder of a long position in an equity swap may qualify for thinner margin. In the event of a counterparty default, a fund that holds a long position in an equity swap is merely an unsecured creditor of the counterparty, while a fund that holds actual shares in a brokerage account may have a security interest in the shares or other property held by the broker in the event of a broker default. In cases such as these, the better view is that a party holding a long position in an equity swap ought not be treated as situated similarly to a party holding a long cash position in the asset underlying the swap.
Contrast this with cases where a party holding a long position in an equity swap is in economic substance the tax owner of the underlier thereof. These include cases where the underlier is so illiquid that the short party is economically compelled to hold it as a hedge, and cases where the long party transfers the underlier to the short party to initiate the swap, or where the short party transfers the underlier to the long party at the end of the swap. The rights of the long party to a transaction of this type are more closely akin to that of an owner of the underlier than those of a party who merely places a bet as to the performance thereof. In the former case, the taxpayer’s “papering” of the arrangement as an equity swap, rather than as a custodial arrangement, could be seen as an attempt to circumvent applicable withholding tax rules and margin regulation.
The IRS was aware of the potential for abuse issue for a long time, and chose not to impose withholding tax on dividend-equivalent payments on equity swaps that qualified as true derivatives. The history of the current regulations indicates that this was a conscious policy choice. Equity swaps have existed since the mid-to-late 1980s. The IRS first issued guidance regarding the source of swap income in 1987.  This was soon followed by proposed regulations,  which were finalized in 1991.  Shortly thereafter, in 1992, the IRS published proposed regulations regarding the treatment of substitute payments on securities loans. These rules were finalized in 1997.  The history of these two sets of regulations is illuminating in two ways. First, the 1991 equity swap regulations only applied to notional principal contracts. They did not preempt the pre-regulation common law rules regarding tax ownership described above. As such, the 1991 regulations only served to characterize dividend equivalent payments on equity swaps to the extent that the applicable swap was a notional principal contract in the first place . To the extent that a transaction constituted a disguised custodial arrangement under common law rules, the 1991 regulation did not serve to characterize payments made thereon; under the 1991 regulations, payments dressed up as dividend equivalent payments that were, in fact, dividends, were treated as dividends.
Second, when the IRS issued regulations regarding the treatment of dividend-and interest equivalent payments on securities lending transactions, it chose to use a rule that was very different from that which it chose to apply to similar payments on notional principal contracts. By way of background – a securities lending transaction is a transaction pursuant to which a taxpayer (the securities lender) loans bonds or shares of stock to a another taxpayer (the securities borrower), in order for the securities borrower to have the use of the loaned assets. In most cases, securities borrowers enter into these transactions in order to sell the borrowed assets “short”. Although the securities borrower is not obligated to return the same bonds or share certificates that were borrowed, he is generally required to return fungible securities in order to terminate the securities lending transaction. So long as a securities lending transaction is outstanding, the securities borrower is obligated to make payments (dividend- or interest equivalent payments) to the securities lender equal to any dividends or interest paid on the loaned securities.
Diagram 3: Loan of XYZ Stock.
Click here to view diagram.
Although a securities lender continues to have full economic exposure to the loaned assets while the securities loan is outstanding, courts and the Service have long held that a securities lender disposes of beneficial ownership in securities when the securities are transferred to a securities borrower. This is because, as discussed above, in the case of liquid, fungible property, tax ownership is generally held to rest with the party with possession and control thereof. Since a securities lender parts with possession and control of the loaned assets when they are transferred to the securities borrower, a securities lender is not treated as the tax owner thereof so long as the securities loan is outstanding. Instead, the securities lender’s ownership rights in the loaned assets are replaced with contractual rights against the securities borrower, and dividend- or interest equivalent payments are fees paid by the securities borrower to the securities lender, rather than dividends or interest. This distinction is relevant in many contexts, not least in determining the source of these payments. The source of dividend and interest payments is generally determined by reference to the residence of the payor. Dividends and interest payments made by U.S. residents are, generally, U.S. source, and dividend and interest payments made by U.S. non-residents are, generally, foreign-source. However- assume a case in which a U.S.-resident securities lender loans shares issued by a U.S. nonresident corporation to a U.S. resident securities borrower. The foreign issuer pays a divided, and the U.S-resident. securities borrower makes a dividend-equivalent payment to the U.S.-resident securities lender. What is the source of this payment?
Diagram 4: Loan of German Stock by U.S. Securities Lender to U.S. Securities Borrower.
Click here to view diagram.
The dividend equivalent payment made by the securities borrower to the securities lender is a fee paid pursuant to a contract between two domestic residents, but it is calculated with reference to a dividend paid by a foreign resident. In the 1997 regulations, the IRS held that the source and withholding tax character of these payments is determined with reference to the actual dividend or interest payments that they replicate. By contrast, under the 1991 notional principal contract regulations, the source of dividend-equivalent payments on equity swaps is determined with reference to the residence of the payee, independently of the source of the underlying payment.
This distinction is illustrative. Securities loans and equity swaps share many similarities. Most importantly, the “long” party on an equity swap and a securities lender both have “delta one” exposure to the underlying security. Nevertheless, the fact that the IRS consciously chose to treat the character and source of dividend equivalent payments on securities loan transactions in one way, while treating dividend equivalent payments on equity swaps quite differently, indicates that the government knew of these similarities, and concluded that differences predominated. This is because a securities lender by necessity owns the loaned asset prior to the inception of a securities loan, and receives it back at the end of the term. The term of the securities loan is merely an interval that punctuates the term of the securities lender’s ownership of the loaned assets. By contrast, a long party under an equity swap that is a true derivative may never own the underlier; the swap is merely a bet on the performance thereof. In cases such as these, the better view was that the relationship between the long party on an equity swap and the underlier was so attenuated that dividend-equivalent amounts should not be treated as dividends.
The foregoing was the state of affairs until 2010. By the late 2000s, only the most ill-advised offshore funds purchased dividend-paying shares issued by United States corporations outright. Instead, funds seeking economic exposure to dividend-paying U.S. equities did so by entering into equity swaps thereon. This would likely have continued, had market practice not gravitated toward equity swaps better characterized as custodial arrangements than true derivatives. For example, by the mid to late 2000s, it would not be uncommon for a fund manager to purchase shares in a dividend-paying stock and hold them until a day or so before the ex dividend date. At that point, the fund would sell the shares to a U.S. broker-dealer counterparty, and immediately enter into an equity swap thereon, with a strike price equal to the price at which the shares were sold from the fund to the counterparty. Shortly after the ex dividend date, the counterparty would sell the shares back to the fund and terminate the swap at the price at which the shares were sold. When the counterparty received the dividend on the underlying stock, it would make a dividend-equivalent payment to the fund without reduction for withholding tax. The net result was that the fund retained all of the economic risk associated with ownership of the shares, the counterparty assumed none of the economic risk associated therewith, and dividends were paid without reduction for withholding tax.
Congress held a series of hearings in 2009 to address perceived abuse inherent in transactions of this type. Spurred to action by the legislative branch, the IRS published guidance issued to certain government personnel in charge of auditing potentially abusive equity swaps in early 2010. Later that year, Congress passed Section 871(m) of the Internal Revenue Code, which characterizes dividend equivalent payments made on certain equity swaps and stock loan transactions as dividends for United States withholding tax purposes. The IRS published proposed regulations under Section 871(m) in January of 2012 (the “2012 proposed regulations”). In response to widespread opposition from the financial services industry to the 2012 proposed regulations, which many banks viewed as unworkable, the IRS withdrew the 2012 proposed regulations and issued new proposed regulations (the “2013 proposed regulations”) in December, 2013. The 2013 proposed regulations were finalized with certain changes on September 17, 2015. New temporary regulations regarding certain transactions not covered by the 2013 proposed regulations were issued on the same date.
The rules contained within the four corners of Section 871(m) (the “statutory rule”) and the rules contained in the current final and temporary regulations are described below. As will become evident, the IRS significantly broadened the scope of the statutory rule when it passed the current final and temporary regulations.
2) The Statutory Rule
The rule contained in Section 871(m) is narrow, and is consistent with the common law rules distinguishing a true derivative from a custodial arrangement discussed above. Under Section 871(m), a dividend equivalent payment on, inter alia , a “specified notional principal contract” is treated as a dividend from sources within the United States. For these purposes, a “specified notional principal contract” is a notional principal contract with an underlier consisting of shares issued by a United States-resident corporation if any of the following is the case:
- In connection with entering into the contract, the long party transfers the underlying security to the short party (a “cross-in”);
- In connection with the termination of the contract, the short party transfers the underlying security to the long party (a “cross-out”);
- The underlying security is illiquid; or,
- In connection with entering into the contract, the underlying security is posted as collateral by the short party to the long party.
By drafting Section 871(m) in this manner, Congress effectively targeted the abuse addressed in the 2009 hearings. Payments on transactions where there is a transfer of the underlying asset to or from the counterparty in connection with the swap are subject to withholding, as are payments on transactions in which the counterparty is economically compelled to hold the underlying asset as a hedge, by virtue of its illiquidity.
3) Current Final and Temporary Regulations
In the author’s experience, traders’ initial reaction to the Section 871(m) statutory rule swaps was, “So we will do forwards, instead of swaps”. The current final and temporary regulations make this impossible. Very briefly – the new regulations include provisions similar to the statutory rules, and they also contain provisions which subject all “dividend equivalents” on “specified equity linked instruments” (“specified ELIs”) to withholding. Because the definition of a specified ELI is broader than the definition of a “specified notional principal contract” under the statutory rule, and because the definition of “dividend equivalent” under the regulations is broader than the definition of “dividend equivalent” under the statutory rule, the regulations significantly broaden the scope of Section 871(m) beyond that of the statutory rule. In so doing, they prevent certain types of transactions that the IRS perceives as abusive. However, as described in more detail below, they also haul in many dolphins with the tuna.
The meaning of the terms “ELI”, “specified ELI”, “dividend equivalent”(referred to herein as either a “dividend equivalent” or a “dividend equivalent payment”) and “payment”, and their application in the new rules, are discussed below.
- Equity Linked Notes
i) Delta. The final regulations define an ELI as any transaction, other than a securities lending transaction, a sale-repurchase transaction or a notional principal contract, that references the value of one or more shares of stock, dividends on which are subject to withholding when paid to a United States nonresident (“underlying securities”). Examples of ELIs include forward contracts, futures contracts, options, convertible debt, or any other contractual arrangement that references the value of one or more underlying securities.
Except in the case of “complex ELIs”, discussed below, a specified ELI is an ELI that has a delta of 0.8 or greater. “Delta” is a value used in derivative pricing that measures the rate of change in a derivative’s value relative to changes in the price of the underlying asset. 
Example 1. On 1/2/2016, a call option to purchase 100 shares of XYZ stock is issued. The option may be exercised at any time up to its maturity date on February 19, 2016. XYZ is currently trading at $100 per share. The strike price is $95 per share, and the initial purchaser of the option pays $700 for the option. The issuer of the option determines that, if the value of XYZ stock were to increase by $1, the value of the option would increase to $770. The option has a delta of 0.7. Because this is less than 0.8, the option is not a specified ELI. 
The 2013 proposed regulations defined any ELI with a delta of 0.7 as a specified ELI. This rule met widespread opposition by the financial services industry, because a threshold of 0.7 was seen as being over-inclusive. The preamble to the final regulations indicates that the delta threshold was increased to 0.8 because long exposure to a derivative with a delta of 0.8 or higher may be taken as a proxy for full exposure to fluctuations in the price of the underlying asset, while long exposure to a derivative with a lower delta may not.
Note that the delta measure used to determine whether an ELI is a specified ELI is the delta of the ELI as calculated by the relevant party at the time of the ELI’s issuance . This raises two areas for potential uncertainty. First, delta is a model-based value; the delta attributed to a given derivative may vary, depending on the model used and the inputs fed into the model. Under the final regulations, if one party to a potential 871(m) transaction, the broker dealer is required to calculate delta and to report it to the other participants in the transaction; if either both parties, or neither party, to an initial transaction is a broker dealer, the short party is required to calculate delta and to report it to the other participants. Because models and inputs may vary, this means that deltas used to determine whether a transaction is an 871(m) transaction may vary. Second, delta changes over time; therefor, the delta of an ELI at issuance may differ significantly from its delta later in its life. Under the 2013 proposed regulations, the delta of an ELI was measured at the time the ELI was acquired by a taxpayer. The IRS changed this to the current rule in the final regulations because the administrative burden of re-measuring delta each time an ELI changed hands was seen as outweighing any marginal increase in accuracy. While ease of administration is welcome, this rule may cause economically similar instruments to be treated differently in certain cases:
Example 2. An over-the-counter six-month call option with a strike price of $100 on XYZ stock is issued on 1/1/2016, when XYZ stock is trading at $80 per share. At the time of its issuance, it has a delta of 0.3. Taxpayer 1 purchases the option and holds it for three months.
On April 1, 2016, when XYZ stock is trading at $125 per share, the option’s delta has increased to 0.85. At this time, Taxpayer 1 sells the option to Taxpayer 2. Because the option had a delta of less than 0.8 at the time of its issuance, it is not a specified ELI in the hands of either Taxpayer 1 or Taxpayer 2.
Example 3. The facts are the same as in Example 3, except the call option is issued by, and traded on, the CBOE options exchange. On April 1, 2016, Taxpayer 1 terminates her position in the option, and Taxpayer 2 purchases an option listed on the CBOE with the same underlier, strike price and maturity. Pursuant to the rules of the exchange, Taxpayer 1 exits her position by entering into an offsetting short position in a call option that is fungible with her long position, and Taxpayer 2 enters into a long position with the exchange. Both Taxpayer 1 and Taxpayer 2 are in contractual privity with the CBOE, and the CBOE only. Because Taxpayer 2’s option was newly-issued by the CBOE on April 1, at which time it had a delta of 0.85, Taxpayer 2’s option will be a specified ELI.
ii) Combined Positions. One of the most difficult aspects of the delta test is its applicability to combined positions. It is common for traders to combine derivative positions to increase exposure to the underlier, to hedge risk, or to reduce the up-front cost of entering into a transaction. Delta is additive; combining positions may have the effect of either increasing or decreasing aggregate position delta:
Example 4. XYZ stock is trading at $99.80 per share. A taxpayer purchases a call option to buy 100 shares of XYZ stock for $100 in two months’ time, and sells a put option to sell 100 shares of XYZ stock with the same “strike price” and the same maturity date. Both the call and the put cost $5.00, resulting in a net cash flow on Day 1 of $0.
The call option has a delta of 0.5 (i.e., for each $1.00 increase in XYZ stock, the value of the call option is expected to increase $0.50). The put has a delta of -0.5 (i.e., for each $1.00 increase in XYZ stock, the value of the call option is expected to decrease $0.50). However, since the taxpayer is “short” the put, her position in the put has a delta of 0.5, and the delta of the combined position is 1.0.
Because the put and the call have the same underlier, strike price, and maturity date, it is economically certain that one or the other will be exercised. Because of this, the taxpayer will profit (or lose) an amount equal to the difference between the strike price and the fair market value of XYZ upon maturity. This reflects the delta of 1.0 of the combined position.
Example 5. XYZ stock is trading at $100 per share. A taxpayer purchases a call option to buy 100 shares of XYZ stock with a strike price of $92 that expires in two months’ time. In order to offset the cost of the first option, the taxpayer sells a call option on XYZ stock with a strike price of $105 and the same expiration date. The taxpayer pays $9 to purchase the $92 strike price call option, and receives $3 in exchange for selling the $105 strike price call option.
The $92 strike price call option has a delta of 0.85. The $105 call option has a delta of 0.35. However, since the taxpayer is “short” the $105 call, her position in the $105 call has a delta of -0.35. Therefore, aggregate position delta is 0.5.
The risk profile of the combined position is significantly different from that of the $92 strike price call option in isolation. If the taxpayer held only the $92 strike price, she would benefit on a dollar-for-dollar basis from every increase in the value of the option above $92. However, the pay-off from the combined position is capped at $105. In exchange for foregoing this up-side potential, the taxpayer’s cost of entering the position is reduced by the proceeds from the sale of the $105 strike call. The aggregate position delta of 0.5 reflects this.
In a perfect world, all positions with respect to the same underlier entered into in connection with each other would be tested for delta on an aggregate basis. However, the rub is how to determine whether one or more transactions were entered into “in connection with each other”. Both funds and broker-dealers hold large portfolios, at the entity level. Traders at one desk usually do not know what positions their colleagues at other desks are entering into; it is not uncommon for different desks to enter into multiple transactions with respect to the same underlier without knowing about each others’ activities. Therefore, in order to craft a perfect rule, it would be necessary to use a test that only combines deltas of positions that were entered into with the intent of modifying each other’s delta. The IRS struggled to find an accurate test in the 2013 proposed regulations, and in the preamble to the final regulations, the IRS said that it found it difficult to enunciate a rule on point that is neither over- nor under-inclusive.
Under the 2013 proposed regulations, all separate transactions that referenced identical underliers were combined, to the extent that they were entered into in connection with each other, and to the extent that position delta was increased by their being combined. In the preamble to the final regulations, the IRS noted that many commenters had suggested that this rule be amended to allow a reduction of delta in the case of transactions that reduce position delta, but said that they had declined to change the rule, because they could not think of an administrable test to net positive and negative deltas.
Under the final regulations, for purposes of determining whether a potential Section 871(m) transaction is a Section 871(m) transaction, two or more potential Section 871(m) transactions are treated as a single transaction with respect to an underlying security if the following are the case:
- A single person (or two or more persons who are related persons, within the meaning of Sections 267(b) or 707(b)) is the long party with respect to the underlying security ,
- The potential 871(m) transactions reference the same underlying security;
- The potential 871(m) transactions, when combined, replicate the economics of a transaction that would be a Section 871(m) transaction if the transactions had been entered into as a single transaction, and
- The potential Section 871(m) transactions are entered into in connection with each other (regardless of whether the transactions are entered into simultaneously or with the same counterparty).
The final regulations also reiterate the “heads we win, tails you lose” rule that transactions may not be treated as combined in order to reduce aggregate position delta used in determining whether a transaction is subject to Section 871(m).
As the IRS has noted, the weakness with this rule lies in the intent-based test in the fourth bullet point above. In order to determine whether two or more transactions are entered into “in connection with” each other, the final regulations include the following presumptions:
- A short party that is a broker may presume that positions held in separate accounts are not entered into in connection with each other, absent actual knowledge to the contrary;
- A short party that is a broker may presume that transactions entered into two or more business days apart are not entered into in connection with each other. For this purpose, every position entered into during a trading day is treated as entered into at 4:00 PM.
Example 6. At 3:45 PM on Tuesday, October 6, 2015, a trader purchases a call option to buy 100 shares of XYZ stock with a strike price of 100 and a maturity date of November 20, 2015. The call option has a delta of 0.5. At 11:13 AM on Thursday, October 8, 2015, the trader sells a put option to sell 100 shares of XYZ stock with a strike price of $92.50 and the same maturity date as the call. The put has a delta of -0.35; since the trader is “short” the put, his position in the put has a delta of 0.35. The trader places both trades through the same account with the same broker. The call is deemed entered into at 4:00 PM on October 6, and the put is deemed entered into at 4:00 on October 8. Because the two transactions were entered into two days apart, the broker may presume that they are not entered into in connection with each other.
- The IRS will presume that a long party did not enter into two or more transactions in connection with each other if the positions are reflected in separate trading books. The IRS may rebut this presumption with facts and circumstances showing that transactions reflected on separate trading books were entered into in connection with each other or that separate trading books were used to avoid Section 871(m); and,
- The IRS will presume that a long party did not enter into two or more transactions in connection with each other if the transactions were entered into two or more business days apart. The IRS may rebut this presumption with facts and circumstances showing that the transactions were, in fact, entered into in connection with each other.
The foregoing presumptions only apply to broker short parties and the IRS. Long parties must treat transactions as combined transactions to the extent that they are entered into in connection with each other.
The final regulations stipulate that, in combining transactions, the rules must be applied in the manner that maximizes the number of Section 871(m) transactions.
Example 7. A taxpayer purchases two call contracts to buy 100 shares of XYZ stock at $100 per share, and sells one put contract to sell 100 shares of XYZ stock at the same strike price. The call contracts each have a delta of 0.5, and the taxpayer’s short position in the put has a delta of 0.5. The two calls are not combined because they do not provide the long party with economic exposure to depreciation in the underlying security. Instead, the put contract and one of the call contracts are combined, to produce a combined position with respect to 100 shares of XYZ stock, with a delta of 1.0.
iii) Indices. In drafting the 2013 proposed regulations and the final regulations, the IRS acknowledged that derivatives on broad-based indices whose composition is determined by rules determined ex ante or by third parties, are generally not instruments of abuse. Therefore, under the final regulations, derivatives on certain indices are not within the scope of Section 871(m). To the extent that an index constitutes a “qualified index”, dividend equivalent payments made with respect thereto are not subject to withholding tax under Section 871(m). A qualified index is an index that,
- References 25 or more component securities (whether or not the securities are underlying securities);
- Does not reference more than a de minimis amount of component securities; 
- References no component underlying security that represents no more than 15 percent of the weighting of the component securities in the index;
- References no five or fewer component underlying securities that together represent more than 40 percent of the weighting of the component securities in the index;
- Is modified or rebalanced only according to publicly stated, predefined criteria, which may require interpretation by the index provider or a board or committee responsible for maintaining the index;
- Did not provide an annual dividend yield in the immediately preceding calendar year that is greater than 1.5 times the annual dividend yield of the S&P 500 Index as reported for the immediately preceding calendar year; and,
- Is traded through futures or option contracts on a national securities exchange or a domestic board of trade, or a certain foreign exchanges or boards of trade, provided, in the latter case, that underlying securities, in the aggregate, make up less than 50% of the weighting of the component securities in the index.
In addition, an index 10 percent or less of the weighting of which is attributable to underlying securities, is also a qualified index.
Transactions that reference a security, such as an exchange-traded fund, that itself references a qualified index, are treated as referencing a qualified index.
iv) Complex ELIs. In contrast to the general rule, the delta test is not used in determining whether so-called “complex contracts” are within the scope of Section 871(m); instead, the “substantial equivalence test” is used.
(1) Complex Contracts. Under the final rules, a complex ELI is any ELI that consists of a “complex contract”. A complex contract is any contract that is not a “simple contract”. A simple contract is a notional principal contract or an ELI for which, with respect to each underlying security, all amounts to be paid or received are calculated by reference to a single, fixed number of shares of the underlying security, and the contract has a single exercise or maturity date with respect to all amounts other than upfront or periodic payments. For these purposes, a contract has a single exercise date even though it may be exercised at any time on or before the stated expiration date of the contract.
Example 8. Shares of XYZ stock are trading at $100 per share. On Day 1, the taxpayer purchases one contract of an ELI for $10,000. At maturity, the taxpayer is entitled to a return of the $10,000, plus 2X the amount of any increases in the value of 100 shares of XYZ stock up to a cap of $110, minusthe amount of any decrease in the value of 100 shares of XYZ stock below $95.
Because the contract references either 100 or 200 shares of XYZ stock at maturity, depending on the value of the shares at that time, its pay-out is not calculated with reference to a single, fixed amount of XYZ shares. Therefore, it is a complex contract.
(2) Substantial Equivalence. Under the final regulations, a complex contract is subject to Section 871(m) if it is treated as such under the “substantial equivalence test”. A full description of the substantial equivalence test is outside the scope of this Alert. However, in general terms, a complex contract is treated as a Section 871(m) transaction with respect to an underlying security if, pursuant to a “stress test” in which the value of the underlier is assumed to both increase and decrease by one standard deviation, the change in value of the complex contract relative to the change in value of the initial hedge entered into by the issuer of the contract is equal to or greater than the change in value of a certain benchmark simple contract would be relative to the applicable hedge therefor.
The purpose of the substantial equivalence test is to treat contracts for which delta can not be calculated as within the scope of Section 871(m) if the value thereof closely approximates that of the value of an underlying security.
b) Payments. As discussed above, Section 871(m) treats dividend equivalent payments on specified ELIs as dividends for withholding tax purposes. This raises three questions, i.e. (i) What is a payment?, (ii) What portion of a payment is treated as a dividend equivalent?, and (iii) When is a payment made?
i) What is a “Payment”? in order to forestall potential abuse through the use of instruments that had the value of actual or estimated dividends “baked into” amounts payable at maturity, the final regulations include an expansive definition of “payment.”
Under the final regulations, a “payment” includes any amount that references an actual or estimated payment of dividends, whether the reference is explicit or implicit. A dividend payment is implicit if it is taken into account in computing one or more terms of a Section 871(m) transaction, including interest rate, notional amount, purchase price, premium, upfront payment, strike price, or any other amount paid or received pursuant to the Section 871(m) transaction.
Example 9. Shares of XYZ stock are trading at $100 per share. An offshore fund enters into a forward contract to purchase 100 shares of XYZ stock in one year. The price of the forward contract is determined by multiplying the number of shares referenced in the contract by the current price of the shares and the fund’s borrowing costs (currently 4.00%), and subtracting the value of any dividends expected to be paid during the term of the contract. It is expected that XYZ will make dividend payments of $2.5 per share over the life of the contract. The forward price is set at $10,150 (= $10,000, plus $400 interest charge, less $250 expected dividends). Under the final regulations, the $250 offset to the forward price attributable to expected dividends is a dividend-equivalent amount subject to withholding under Section 871(m).
This is the rule that thwarts the seemingly easy solution of replacing swaps with forwards.
It should be noted that dividend equivalent payments are calculated on a gross basis, rather than on a net basis. This is counter-intuitive, because cash payments on equity swaps and other ELIs are usually calculated on a net basis:
Example 10. An offshore fund enters into an equity swap with a bank on 100 shares of XYZ corp. Pursuant to the terms of the swap, the bank will pay the fund an amount equal to the increase in value of 100 shares of XYZ stock above a stated strike price, plus amounts equal to any quarterly dividends paid on 100 shares of XYZ stock. In exchange, the fund will pay the bank an amount equal to the decrease in value of 100 shares of XYZ stock below the strike price, plus a quarterly funding rate of 4% multiplied by the value of 100 shares of XYZ stock on Day 1.
Two months after the swap is entered into, XYZ declares and pays a dividend of $0.63 per share, and a financing leg payment of $100 from the fund comes due on the swap. In order to settle the offsetting liabilities, the fund makes a net payment of $37.00 to the bank.
Although no cash has been paid by the bank to the fund, the final regulations make clear that a gross dividend equivalent payment of $63 is deemed paid from the bank to the fund. Therefore, a withholding tax of $18.90 is due on the deemed $63 payment even though there is no cash from which to withhold this money . If we assume that an equity swap is the economic equivalent of a levered loan and that actual dividend payments are subject to gross withholding tax, this makes tax-policy sense. However, most banks do not have systems set up to account for and deduct withholding tax in the absence of cash payments – and cashless withholding causes cognitive dissonance for buy-side participants.
For a comparison of an equity swap with a leveraged purchase, See Diagram 1 and Diagram 2, above. The problem of cashless withholding is discussed in more detail below.
The final regulations also contain certain simplifying rules that clarify the way in which dividend equivalent payments on baskets, indices and exchange-traded funds may be calculated and withheld on.
ii) Amount of Dividend Equivalent. Once it has been established that there is a payment, it is necessary to determine what part of the payment should be treated as a dividend-equivalent amount. This calculation depends on whether the applicable instrument is a securities lending or sale-repurchase transaction, a simple ELI, or a complex ELI.
- Securities Lending and Sale-Repurchase Transactions. For a securities lending or sale-repurchase transaction, the amount of the dividend equivalent for each underlying security equals the amount of the actual per-share dividend paid on the underlying security multiplied by the number of shares of the underlying security. This rule makes sense, because the long party in both securities lending and sale-repurchase transactions has “delta one” exposure to the underlying asset.
- Simple Contracts. For a simple contract that is a Section 871(m) transaction, the amount of the dividend equivalent for each underlying security equals the per-share dividend amount with respect to the underlying security, multiplied by the number of shares of the underlying security, multiplied by the delta of the applicable Section 871(m) transaction.
- Complex Contracts. In the case of a complex contract, the amount of the dividend equivalent for each security equals the per-share dividend amount multiplied by the number of shares in the issuer’s initial hedge of the complex contract. 
iii) Timing of Payment. A dividend equivalent is considered to be made on the later of the following:
- the record date of the dividend, or the day prior to the ex-dividend date with respect to the dividend (whichever comes first), or
- the date when a payment occurs with respect to the Section 871(m) transaction.
For these purposes, a payment generally occurs with respect to a Section 871(m) transaction when money or other property is paid to or by the long party, or when the long party sells, exchanges, transfers, or otherwise disposes of the Section 871(m) transaction, including by settlement, offset, termination, expiration, lapse or maturity. However, when a long party pays a premium or other upfront payment to the short party at the time a Section 871(m) transaction is issued, the premium or other upfront payment is not treated as a payment for these purposes.
c) Effective Date. The final regulations and the temporary regulations are effective for all payments made after January 1, 2017 with respect to transactions issued on or after January 1, 2017, and for all payments made after January 1, 2018 with respect to any transaction issued on or after January 1, 2016, and before January 1, 2017. For practical purposes, this means that the statutory rule of Section 871(m) remains in effect for all transactions until January 1, 2017, and remains in force for transactions issued during 2016 until January 1, 2018.
d) Potential Issues. As with many things, the devil is in the details – and details will become apparent as market participants implement these rules. A very preliminary list of potential issues with the final regulations would include the following:
i. Cashless Withholding. The definition of “payment” in the final regulations requires withholding in certain cases in which there is no cash from which to withhold. This is due both to the fact that a dividend equivalent amount may include amounts other than actual dividend equivalent payments, and to the fact that dividend equivalent amounts are calculated on a gross basis, rather than on a net basis.
Example 11. The facts are the same as in Example 9, above. Because the forward price has been reduced by $250 of expected dividends, the value that the offshore fund is deemed to receive by virtue thereof is a dividend equivalent payment, subject to withholding under Section 871(m). Except – no payments are made by the bank to the fund, from which the bank can withhold. By contrast, the fund will pay the bank in order to buy the underlying shares.
Example 12. XYZ stock is trading at $100 per share. A taxpayer purchases a privately-issued call option on 100 shares of XYZ stock from the option writer with a strike price of $90 and an expiration date six months after issuance. The option has a delta of 0.85 at issue. The taxpayer pays $2,000 premium for the option. The amount of the premium includes $100 that is attributable to a dividend that is expected to be paid on XYZ stock prior to the option’s maturity date. The taxpayer holds the option to maturity. At maturity, XYZ stock is trading at $80 per share, and the option expires worthless.
In both of the foregoing examples, the party receiving cash is not the party against whom withholding is usually assessed; instead, the party receiving cash is the seller, or the short party. The issue is compounded in Example 12, where the date on which cash changes hands (i.e, the issue date) precedes the date on which the payment is deemed made under the final regulations (i.e., the expiration date) by six months. The solution to this conundrum will likely be for taxpayers to satisfy withholding obligations on cashless payments from a deposit or margin account set up by the deemed payee at or before the beginning of the trade. Setting up, ab initio , a system to handle this type of payment for large broker dealers and buy-side funds is not a trivial task. The IRS has advised taxpayers to ensure that computer systems are updated to handle this by the January 1, 2017 effective date.
ii. Combined Positions. The IRS has intimated that the rules regarding combined positions may be subject to further “tweaking”. Change would be welcome: as discussed above, the rules in their current form are both over-inclusive and susceptible to abuse. The failure to allow taxpayers to combine transactions that reduce delta means that components of many “plain vanilla” spread transactions entered into for non-tax reasons will be treated as Section 871(m) transactions, while the short-party presumption that positions held in separate accounts are not entered into in connection with each other is ripe for taxpayer abuse.
iii. Complex Contracts. The substantial equivalence test applicable to complex contracts is a masterwork of Rube Goldberg design. Although it may represent one of many methods that certain traders have of estimating risk of certain exotic options, it could well prove to not be administratively feasible.