Mind the tax when hedging TRS

Total return swaps (TRS) rarely appear in the pages of Cutting Edge. This is partly because they do not require complex pricing models. However, their valuation is still affected by a number of factors beyond the underlying dynamics – among them, collateral funding costs, counterparty risk, regulatory asymmetries and taxation.

This last variable has largely been overlooked in the literature and in practice, which might be a costly mistake, as it can often mean the difference between a profitable deal and a loss-making one.

“The impact of taxation effects varies greatly depending on several factors, such as a portfolio’s composition, direction of trade, tax regime and rates. We noticed in some circumstances it is comparable to that of the TRS spread itself,” says Stefano Scoleri, a quantitative analyst at Be Consulting in Milan.

Banks use different strategies to hedge TRS. For example, they can buy and hold the underlying share, or buy and lend it. These have different tax implications, which the bank should consider before selecting the optimal one.

In 2019, this problem landed on the desk of Andrea Pallavicini, head of equity, FX and commodity models at Banca IMI. The bank wanted his team to look at the TRS spreads and determine whether the choice of hedging strategy had any effect on profitability. The quantitative team was asked to take all relevant factors into account, and taxation in particular.

The results of the research project are detailed in a paper Pallavicini co-authored with Be Consulting’s Scoleri and Stefania Gabrielli. Taking the existing pricing framework for valuation adjustments (XVA), they developed a model that also considers taxation and is agile enough to give a trader a rapid view of the profitability of a trade.

“In essence, we simplify the way XVAs are incorporated in the pricing and include the effects of taxation in it as well. That provides the trader with a rule of thumb to tell whether a deal is worth entering or not,” says Pallavicini. 

In a TRS contract, the performance of an equity index is exchanged for a Libor-linked floating leg. These represent a sizeable and growing share of the market for equity forwards and swaps – $3.142 trillion of notional outstanding in the first half of 2019, according to the Bank for International Settlements.

In essence, we simplify the way XVAs are incorporated in the pricing and include the effects of taxation in it as well

Andrea Pallavicini, Banca IMI

TRSs are used as a funding tool by banks, and by buy-side firms – particularly hedge funds – to take exposure on single equities.

Economically, they are closely related to repos and are increasingly replacing them in the market. “You can think of repos and the total return swap as the front and the back end of the same product. On the front, you look at the repo as a loan product, a secured loan. On the back side there is the derivative form of the funding product, which is the total return swap. So repo and TRS serve the same purpose,” explains a repo market and funding expert at a global bank.

The model can also be used to gauge the profitability of other equity-linked products, such as repo, reverse repo, equity lending and analogous structures.

Because TRSs are often used in cross-border deals, the model is designed to be implemented in a multi-currency setting. While some jurisdictions, such as the US, do not tax financial transactions, at least 10 European countries do.  

“We incorporate all these effects into the discount curves, so that lengthy simulations are not necessary for the valuation,” explains Pallavicini.

The fact that the model is simple and readily informative makes it particularly attractive and easy to implement at desk level.

“This is where it is useful to a practitioner. Practitioners are not looking at differential equations, they want to understand how their TRS spread is going to change. So in terms of usability, it has high value,” says Amit Pandey, director of liquidity risk management at Bank of America in Charlotte.

The model sacrifices some accuracy for the sake of simplicity. In the paper, the authors assume TRS notionals remain constant, which would require the basket of the hedging assets to be rebalanced. The transaction cost associated with this is not considered in this model.

“This is the trade-off to get their simple formula and the speed of calculation they are seeking. And with the rebalancing of the hedges being infrequent, and the bid-ask-spread small, this is not going to have an impact on the validity of the model,” Pandey says.