negative profit generated in the recent subperiod is worth keeping an eye on not be-coming a new normal for the cointegration method.
Rasmus Bruun Jørgensen, AEF Discussion
managers for a fee (Pedersen, 2015). Despite the lack of willingness, it should be feasi-ble to set up an automated procedure for conducting pairs trading as it is a rule-based algorithmic trading strategy.
However, the pairs trading strategy would practically require a significant amount of capital in order to execute the strategy. A practitioner must be able to trade an equal amount in 40 ETFs at any given time while ensuring the requirements of dollar-neu-trality. With share prices up to USD 300 this would most likely require large invest-ment to be made. For private investors with the access to the required capital there are no further complications in this regard, but for others - most likely the majority of private investors - this might prove to be a restriction for the practical feasibility of performing the strategy. For large institutional investors or hedge funds that most likely already have implemented numerous similar trading strategies, this should not prove to be a restriction of applying the model but rather a new perspective on how a pairs trading can be modified (Pedersen, 2015).
The assumption that regardless of the price development of the traded ETFs, we would always have the necessary capital to uphold collateral in our short-position might also prove to be another important consideration in regard for whom the trading strategy is useful. Evidently in the quant crisis from 2007, hedge funds and other arbitrage traders also had to exit their positions due to funding issues related to their short-positions (Pedersen, 2015). It is thus evident that institutional investors are also ex-posed to the risk of not being able to provide collateral when prices of securities in the long-short positions continue to diverge. For private investors the margin require-ments would potentially be problematic well before the requirerequire-ments would be prob-lematic for institutional investors. Nevertheless, the assumption of these matters are important to consider for both institutional and private investors when implementing the trading strategy.
The biggest difference in practical implementation is the transaction costs. The ex-pense ratios of the ETFs are the same regardless of the pairs trader. In regard to the commission costs, the institutional investors are exposed to lower commission than the commission paid by private investors. However, Do and Faff (2012) showcased that the gap between commissions paid by institutions and private investors has narrowed. The authors more specifically show that the gap in commission between all trades and
institutional trades was 16 bps in 1963 while this dropped to 2 bps in 2000. Their anal-ysis does not go any farther, but indicates that the spreads between the institutional and general commission costs are narrowing. The commission thus makes the pairs trading strategy more favourable for institutional investors relative to private inves-tors due to different transaction costs paid. With this, the findings of this paper should be assessed with caution for the private investor as it relies on the commission costs associated with an institutional investor. Regarding the short-sell costs, we assume that the investor is able to receive the LIBOR overnight rate as a rebate rate, which is an interbank rate that does not reflect the borrowing rates to private investors. Typi-cally, brokerage firms charge further fees including fees for applying a margin account (Fidelity-b, n.d.). Hence, there might be a risk that the short-selling costs are higher for private investors compared to what is outlined in this paper. On the matter of short-selling, there might also be some practical obstacles for private investors as not all brokerage firms permit short-selling in the sense outlined in this paper. When it comes to the bid-ask spread this cost must be paid by both institutional investors and private investors. Throughout the paper it has been emphasized that the bid-ask spread is a significant portion of the transaction costs. A lesser exposure to the cost of the bid-ask spread would further enhance the profitability of the pairs trading strategy. Here larger institutional investors have an edge relative to the private investor. Larger banks or similar institutions typically have their own credit line of transactions that can provide a buyer or seller for some of the ETFs (Smith, 2010; Hartmann, Huang, and Schoenmaker, 2018). As such, the pairs traders in large banks are not exposed towards the same bid-ask costs as some of the costs are then paid in-house. Further, the timing of when to enter a position with regard to the bid-ask spread might be more suitable for institutional investors. As such, the application of this strategy intraday or with an actual trader allocating the necessary time to monitor the strategy will fur-ther improve the results. As we have learned in chapter 5 and 6, exploitation of arbi-trage opportunities is a timing game. The market movement throughout the trading day could enable the realisation of profit, unable to be exploited by only relying on close prices as these only reflect the price at one point in time. Having this in mind, it makes sense that trying to compete in cases where anomalies are obvious to all can be difficult on a daily basis only. The full effect of the divergence will most likely have diminished
Rasmus Bruun Jørgensen, AEF Further research
to a lesser degree at the end of the day, or only being apparent because it is not a profitable anomaly to trade on anymore. Thus, there may be an advantage for institu-tional investors who to a larger extent are able to monitor the pairs trading strategy intraday and thus harvest the profitability arising within a short window of time which a private investor only to a limited extent is able to.
We can thus conclude that the outlined pairs trading strategy of this paper is practi-cally implementable for both private and institutional investors and hedge funds. How-ever, the above considerations showcase that the implied usefulness of the trading strategy increases with the economies of scale, as transaction costs can be lowered by large institutions compared to private investors. For the private investors there might be several obstacles in setting up such a pairs trading strategy that would otherwise not be problematic for institutional investors. As such, we believe the most useful ap-plication of this trading strategy would be for hedge funds or other institutional inves-tors that already have experience in trading such arbitrage strategies. However, more sophisticated private investors that have the required capital will also be relevant em-ployers of our proposed trading strategy.