The Perils of Tax Loss Harvesting

As we approach the end of the year, we would like to take the time to discuss a question we often get from clients: “why isn’t there ‘tax- loss harvesting’ in my portfolio?” For the unfamiliar, tax- loss harvesting is the practice of selling investments to deliberately realize a loss in order to offset both gains and income elsewhere. Ideally, the sold investment is replaced with a very similar one, maintaining the overall expected return of the portfolio.

The practice of tax- loss harvesting has grown in popularity alongside the growth in passive fund/ETF investing. This is in large part because it is very easy to replace a passive fund with a very similar fund, pocket the tax savings, and then move on with the IRS being none the wiser.

But what happens when the investments you are harvesting have no similar replacements? For example, a strategy such as our own- investing in the 25 or so highest quality companies available on the market at an attractive price. By definition, these companies are going to be unique, and selling them in the short-term, even if it saves us on our tax bill, runs the risk of losing out far more in long-term after-tax compounded returns.

For a real-world example, consider our holding of Equinix (EQIX), a very well defended company that is benefiting from the secular trend of increased data consumption. Last December, we could have harvested a loss here as the stock fell from a high of $450 to a low of $335. To keep the numbers simple, let’s assume we had invested $100k at the top, for a $26k loss. Assuming a tax rate of 20%, that would have saved us about $5k on our tax bill.

Flash forward to this year, and Equinix is one of our best performers, recently trading above $570. From the same starting position, and at a 20% tax rate, that works out to an after-tax return of about $21k.

So that’s a gain of $21k made not by harvesting losses, compared to just $5K saved in taxes by harvesting.

A reasonable follow-up question is: “but what if Equinix went down instead?” This certainly is a risk, but the point is that our buying or selling shouldn’t be predicated on tax efficiency, but on whether the investment case remains intact or not. Often, we are in fact looking to buy more of our poorly performing investments- if our investment thesis remains the same but the price is now lower, this is an even more attractive investment. This is the strategy that is most conducive to maximizing our client’s long-term after-tax returns, not tax- loss harvesting.

This same idea also works in reverse, sometimes clients don’t want us to sell stocks that have done very well, as they don’t want to realize the gain. But bear in mind, often we may want to sell a stock that has performed extremely well- maybe now the valuation is too high, or the thesis is changing, and we want to sell out. This will almost always result in a gain, and likely incur tax. But a 20% tax on the profit is likely to be far smaller than the negative performance that can result when the stock experiences a significant drawdown.

The ‘Wash Sale Rule’

A follow-up question people often ask is: “why can’t I sell the stock, take the loss, then just buy the same stock back again?” Well, because the IRS doesn’t want you escaping tax. If you sell a stock and buy it back again within 30 days, the loss will not be usable for tax purposes. This is the wash sale rule. The risk you take is that you sell the stock, you then have to wait for 30 days until you can buy it back again. Will it be 20% higher by then? Will you have missed out on that return? That is the danger.

Conclusion

The key takeaway here is that any decision made on one’s investments should always be made in order to generate the best possible overall return. Taxes are certainly a negative: a real cost that depresses returns. But making investment decisions in order to reduce taxes can be even more destructive. Investors are often far better off just paying the tax now, rather than restricting their investments and suffering poor returns in the future.