Harvesting carry in equities by improving dividend tax efficiency

Higher dividends do not mean higher returns

Unlike in fixed income, we rarely talk about carry strategies in equities. In essence, a carry strategy buys high-yielding assets and sells low-yielding ones with the objective of benefitting from the yield differential. Dividends are effectively the income stream or yield from a stock. However, building a carry strategy based on dividends can be problematic: they account for only a small portion of total stock returns, and tilting a strategy towards high-dividend stocks can introduce a value bias into a portfolio, which may then be the main driver of long-term performance. 

In our view, taxes mean that dividends are not the most efficient way to capture equity returns. For example, US withholding taxes on dividends amount to 30%. In Switzerland, meanwhile, they are as high as 35%. Comparing gross and net performance of the MSCI World Index  shows that dividend taxes impose a performance drag of at least 0.5% every year. To reduce taxes, investors could move away from high dividend stocks in favour of low dividend ones. But this permanent tilt introduces a negative value bias, which may be detrimental for portfolio performance over the long term.

Read more: Keeping equity portfolios aligned to decarbonisation as US climate policy reverses

Tax optimisation can be implemented in a smarter way by a systematic portfolio rebalancing that temporarily reduces exposure to stocks that are about to pay dividends. This approach does not create a permanent negative value bias; however, it does result in higher transaction costs due to additional turnover. Nevertheless, since benchmark returns are typically calculated net of dividend taxes, such a strategy can potentially boost excess performance. In this article, we study the benefits of dividend tax optimisation using such a systematic strategy.   

True carry strategy in equities

The success of dividend tax optimisation depends on the ability to forecast upcoming dividends. While some companies declare dividends well in advance, others do it just before distributing them. For our study, we use a simple forecast based purely on past dividend payments. Specifically, at the end of each month, we forecast companies’ dividends in the next month, based on the actual dividends distributed during the same month a year ago.

Using the MSCI World Index as a starting point, we adjust stock weights at the end of each month, with the objective of reducing the next month’s dividend taxes by a certain percentage. The new portfolio is built by minimising the tracking error relative to the index, while keeping both region and sector weights unchanged.

Figure 1 shows the attribution of excess returns in our portfolio. These returns are boosted by the realised tax savings while additional transaction costs have only a slightly negative effect¹. The net excess return is further influenced by the gross excess return, which is the unpredictable part of this strategy. Interestingly, it contributed positively over the last decade. In attempting to understand this, we studied the gross excess return through the lens of equity factors. We found, as expected, that the strategy exhibits small negative value bias and an additional positive tilt toward higher beta stocks. However, total factor exposures explained only 5% of the gross excess return and 8% of its volatility.

The figure suggests that the ex-post benefits from tax reduction outweigh extra transaction costs even using a naïve dividend forecast. We see this as a true carry strategy where  the yield is equal to the magnitude of dividend taxes. Like a typical carry trade, our strategy rebalances the portfolio toward stocks with potentially higher (or less negative) yield.

FIG 1. Net excess return of optimised portfolio²

Enhancing TargetNetZero portfolios

According to Figure 1, more ambitious tax reduction results in higher returns. To determine the optimal target, however, we must consider the additional risk generated by the strategy.

In our article on optimal portfolio alignment, we emphasised the importance of incorporating uncorrelated sources of alpha in the construction of TargetNetZero portfolios. Prudent risk-budgeting allows us to potentially enhance portfolio returns without any material impact on the tracking error. We consider dividend tax optimisation as a source of possible alpha since the benefits outweigh the costs.

Figure 2 illustrates the impact of dividend optimisation on the tracking error of the TargetNetZero Global strategy. We note that the curve connecting the dots is almost vertical at small levels of tax reduction, meaning that the portfolio return can be improved without any material impact on tracking error. We call this the “free lunch zone”. Outside this zone, higher target levels involve a trade-off between risk and return, which may imply different optimal levels, depending on the risk tolerance of investors.  

FIG 2. Impact of tax optimisation on active risk and return of TargetNetZero portfolio³

In our TargetNetZero strategies, we implement a tax reduction target of 20%. According to our studies and past patterns based on naïve dividend forecasting, we estimate an additional potential return of 7bps, net of transaction costs⁴. In practice, we incorporate analysts’ estimates, which allow us to forecast dividends with much greater accuracy. This results in an expected potential return enhancement of 10bps per year⁵.

Adding alpha through carry

The study shows that our strategy of minimising dividend taxes enables us to boost potential returns for clients invested in our TargetNetZero equities strategies.

This carry strategy illustrates the potential to add alpha net of transaction costs, without raising risk levels. Over the long term, we believe this approach could materially enhance performance.