It’s commonplace to evaluate “gross” equity returns (how much we get through dividends and sale proceeds compared to how much we invest). It’s also important to be aware of how we’re doing “net” of taxes and make sensible decision regarding tax efficiency.
The Basics
Our starting point is that everything we earn in the stock market is going to be reduced by the taxes we pay on what we get. This refers to taxes on the income (dividends) we receive and the capital gains we get from when we sell stock for more than we paid to buy the shares.
In the real world, however, the basics are altered by many things, some of which we choose (the tax status of the account in which we hold the shares, whether we choose to sell or simply keep score of “paper” returns earned on stock we still hold, how long we hold before we sell) and other things the world foists on us (whether we make money in the shares or loose money, whether dividends are “qualified”, and of course, our personal tax profile including where we live and what kinds of exemptions and deductions we have).
For those who are not inclined to dive into details (on their own or with the assistance of personal tax advisers), here are some general principles:
- Tax efficient equity investing (investing in ways that minimize the the tax burden) requires lower turnover.
- The rate that applies to profits from sales of equities held longer than a year (a long-term capital gain) is lower than that on profits from short-term gains, which are taxed at “ordinary” rates. (In other words, there is no penalty for selling quickly; instead, you forfeit a special reward — a lower tax rate — you could have gotten by holding for more than a year).
- This is because you do not pay taxes on gains accumulated by stocks you still hold (“paper gains”). You’ll pay taxes someday in the future when you (or your estate if we look far enough ahead) sells, but even without going through the mathematics of “present value,” it’s easy to recognize that money with which I must part today today is more burdensome than money paid out in the uncertain future.
- Taxes need to be considered, not in terms of individual stocks, but in terms of the sum of all you buy and sell. That’s because gains that might otherwise be taxable for some stocks can be offset, at least partially, by losses you incur with other assets.
- Dividends are usually taxed at a the lower rate that also applies to long-term capital gains.
- This is a separate topic in and of itself, but for now, suffice it to say that the exception is for dividends the IRS deemed “unqualified.” Without going into all the gory details, dividends are generally unqualified if they come from REITs or MLPs, which do not pay taxes, Dividends that come from regular corporations, (shares of which are held for at least 60 days) which are paid on proceeds left over after payment of corporate income taxes, are “qualified” and, hence, taxed to the shareholder at a rate lower than what the shareholder usually pays on ordinary income.
Simple Implications
Dividends aside, it’s clear that tax efficiency (minimizing the tax impact of gains/losses) requires minimizing trading. When we trade less frequently, we can push gains from short term to long term and get a lower rate, or push gains years into the future and have the money that would have been paid to the IRS at the end of the year available for our use (consuming or investing) for a long time.
So if we compare an investing strategy that returns, say, 7% per year to 6% per year from the S&P 500 Spider ETF (SPY), chances are we’re losing money, even if the portfolio risk is low enough to cause Alpha to be solidly positive. If one’s personal marginal tax rate is 20%, then the strategy would be returning 5.6% per year, which is less than what SPY is delivering Given that SPY, as an ETF, is a buy-and-hold vehicle, it would not produce any gain/loss tax impact until it gets sold who knows how many years down the road.
Subtleties
What if the respective strategy and SPY returns are, in our view, a temporary phenomenon. Suppose we expect long-term annual returns of 14% and 9% from the strategy and SPY respectively. Now, the edge goes to the strategy, which is projected to return 11.2% after tax versus 9% for SPY (and for the latter, there would be a potentially big tax payment to be subtracted when we sell and book taxable gains some time down the road).
This raises two important-but-not-necessarily-easy-to-answer questions:
- How credible is the 11.2% return assumption for the strategy, and more to the point of tax planning, to what extent is the credibility of the assumption enhanced or diminished by the amount of trading necessary to implement it?
- What, exactly, is the marginal tax rate we should assume when we do the analysis?
Turnover/Trading and Strategy Effectiveness
It’s easy and commonplace to say that lower turnover means better tax efficiency.
On the other hand, higher returns often require higher turnover (more trading). This has nothing to do with churning or day trading. It has to do with selling when investment considerations dictate sale, however long or shot a time period that may take.
Regardless of one’s tax bracket, the worst thing one can do is convert a gain to a loss by holding on to a stock beyond the time when it should have been sold. We never want to be so tax crazed as to slash our income in order to keep money away from Uncle Sam (unless we’re trying to make a some sort of libertarian political point).
At Portfolio123, we invest based on objective data-driven rules. Simpler screen-based investing dictates that we sell whenever a stock no longer meets the criteria for inclusion in the screen. Other more sophisticated models specify data-driven sell rules that do not precisely match the criteria we used for purchase.
For example, we may buy a stock when its Value rank rises above 90, but not sell until its Value rank falls to 50. We might also have a rule that prohibits sale of stock held less than a year. Either or both of these Sell rules would make for a much lower-turnover and more tax efficient portfolio than would be a stricter rule that forces us to sell when a stocks’ rank falls below 90.
But do the tax-targeted rules make for a better strategy? That’s not an easy question to answer. If our research and testing suggests that we should favor only stocks with high Value ranks, the lower-turnover strategy will cause us to be reducing the probability of investment success.
Here’s an example of how low turnover can cause investment problems. It involves an equity income model I developed in 2016 with a goal of minimizing turnover. At first glance, it seemed to be a success. Annual turnover was in the 10%-15% range and an approximately ten-plus year test that began 1/2/06 showed a full-period Alpha of 4.3% per year, and a last-three-year Alpha of 7.2%, and pretty good results each year. When I pushed the start date of the test out a year, to 1/2/07, things sill looked good; full-period annual alpha of 5.1% and last-three-years annual alpha of 4.5%.
Nice. Or so it seems until we look more closely. I wound up abandoning the model because as it turned out the low turnover (e.g. high tax efficiency) was leaving too much to dumb luck.
Table 1 focuses on backtested performance for calendar 2017.
Table 1
| 2017 % Return | |
| Test starts . . . | |
| 1/2/06 | 19.25 |
| 1/2/07 | 18.17 |
| 1/2/16 | 10.03 |
| 1/2/17 | 13.89 |
Because trading was so limited, the model’s Buy criteria, which looked pretty good at the start of the initial test period in 1/2/06, did not get enough opportunity to be matched up with the portfolio’s holdings later in time. It didn’t make much of a difference a year later: The world on 1/2/07 was similar enough to the world on 1/2/06 that I was able to successfully launch a portfolio using criteria that were then a year old.
But by 2016, a lot of change had taken place. The criteria used to successfully launch a portfolio ten years earlier were by now, no longer acceptable. It was a bit better a year later, but it’s clear that if I were to apply the 1/2/06 criteria to the launch of a new portfolio today, whether I’d succeed or not would depend on luck, not the model.
The only way to preserve the integrity of a model and lend credibility to the research and testing that supported its creation, is to keep the here-and-now portfolio as close as possible in time to the model’s buying criteria. That means portfolio turnover needs to be dictated as much as feasible by the data, and that means selling and replacing stocks when the data profile becomes too far removed from the original buy-criteria.
OK. Back to our hypothetical: Maybe it’s too extreme to use a rank of 50 as our sell signal. Perhaps we can change the rule to sell when the value rank score drops below 80. That would give us some leeway to hold onto stocks without risking our value portfolio become top heavy with overvalued positions. On the other hand, we would be doing more trading, realizing more gains, and seeing more of them turn out to be short term.
This is a delicate balance that lacks a precise answer.
We suggest avoiding steps designed primarily to suppress turnover. This doesn’t necessarily mean trading quickly. On a pure investment basis, many stories take time to develop and become manifest in stock prices, especially when working with fundamental criteria. It does mean letting investment considerations play out and accepting turnover for whatever it turns out to be. That’s the bad news. The good news is that there are many kinds of successful investment models, so you should have ample opportunity to focus on your own tax profile and choose among those strategies that work best for you.
Your Tax Profile/Marginal Tax Rate
The key here is, simply, to get it right.
The bad news is that this requires work (possibly a lot of work), if not by you then by your tax advisor. The good news is that it’s likely to be much lower than you might initially think.
There’s a lot of rhetoric out there about individual tax rates in and around the mid 30% range. But don’t use numbers like that in your analysis. Read the text, all of it. The rates writers and talking heads discuss are applied to “taxable” income, not you gross income (salary, etc.). All the blogs and columns I’ve seen do say taxable income, but few if any discuss the often vast differences between gross and taxable income. At the very least, everybody has an exemption for one’s self (and often more depending on other factors such as family composition). And at the very least, everybody can at least claim an IRS gimme, the standard deduction. But many taxpayers can “itemize” deductions and run the number far higher. So if your salary is $100,000 per year, the probability that your taxable income is $100,000 is zero.
And suppose you read something about the top individual tax rate for 2018 being 37% (which is true). Assume your salary is $100,000 per year and that after exemptions and deductions, your taxable income is $85,000. Don’t assume your federal taxes will be 37% of $85,000.
If ever there was a time to turn off the TV and its 2-minute talking head segments and to stretch beyond word-count limited blogs and dig into details, this is it. 37% is the top tax rate on a graduated scale. It applies only to “taxable” income above $500,000.
Here is the actual set of rates, based on IRS Bulletin 2018-10 (3/5/18):
Table 2 – Single Taxpayer
| If Taxable Income is . . . | Then tax is . . . |
| up to $9,525 | 10% of taxable income |
| above $9,525 and not above $38,700 | $952.50 + 12% of excess above $9,525 |
| above $38,700 and not above $82,500 | $4,453.50 + 22% of excess above $38,700 |
| above $82,500 and not above $157,500 | $14,089.50 + 24% of excess above $82,500 |
| above $157,500 and not above $200,000 | $32,089.50 + 32% of excess above $157,500 |
| above $200,000 and not above $500,000 | $45,689.50 + 35% of excess above $200,000 |
| above $500,000 | $150,689.50 + 37% of excess above $500,000 |
Even if your taxable income is $5000,000, you’re still not paying 37%. The tax rate is graduated: You pay 10% on your first $9,525, 12% on your next batch, 22% on the next batch and so on with the 37% rate applied only to income above $500,000. If taxable income is exactly $500,000, your actual tax is $150,689.50 or 30.1% of the total. If your gross income is $500,000 and exemptions and deductions take taxable income down to, say, $350,000, your tax is $98,189.50, which is 28.1% of your taxable income, and 19.6% of your salary.
Going back to the original hypothetical, suppose your salary is $100,000 and your taxable income is $85,000. Now your tax is $14,689.50, which is 17.3% of taxable income and 14.7% of gross income.
Even when you add more to cover state and local taxes, I really hope you weren’t analyzing after-tax portfolio returns using an assumed 35% rate or anything like that! If you did, delete and start again, after studying your past tax returns and coming up with a reasonable estimate of your likely taxable income.
And by the way, the foregoing assumed you are single and filing taxes on that basis. If you are married and filing a joint return, or if you are a surviving spouse, Table 3 shows your tax rates for the 2018 tax year:
Table 3 – Married Filing Jointly and Surviving Spouses
| If Taxable Income is . . . | Then tax is . . . |
| up to $19,500 | 10% of taxable income |
| above $19,500 and not above $77,400 | $1,905 + 12% of excess above $19,500 |
| above $77,400 and not above $165,000 | $8,907 + 22% of excess above $77,400 |
| above $165,000 and not above $315,000 | $28,179 + 24% of excess above $165,000 |
| above $315,000 and not above $400,000 | $64,179 + 32% of excess above $315,000 |
| above $400,000 and not above $600,000 | $91,379 + 35% of excess above $400,000 |
| above $600,000 | $161,379 + 37% of excess above $600,000 |
So now, instead of considering a single with $100,000 in salary and $85,000 in taxable income, let’s assume a couple each of which has that same and combines for a household gross income of $200,000 and $170,000 in taxable income. Now, the tax is $29,379, which is also 17.3% of taxable income 14.7% of gross income. If one $100,00 wage earner supports both spouses and has the $85,000 in taxable income, the tax would be $10,579, which is 12.4% of taxable income and 10.6% of gross income.
Bear in mind, too, that figuring out capital gains is not so simple either. Long term losses are subtracted from long-term gains. Short-term losses are separately subtracted from short-term gains. Up to $3,000 in “net” short term losses can be subtracted from “net” long-term gains, and leftover short-term losses can be “carried over”and used to offset gains in future years. And we’re not doing this stock by stock, or even portfolio strategy by portfolio strategy. We’re stirring the pot with all capital gains and losses one has and doing one grand set of computations.
How You Can Approach Portfolio Taxation
As you can see, this is a very complex topic and defies simple answers. This is not a back-of-the-envelope thing. It’s about details and fine print. The best approximate answers that can be given are as follows:
- If you absolutely positively unquestionably must add a simple tax assumption to your analysis of portfolio returns without going into detail, assume all returns are short term and taxed at a 20% rate. If that doesn’t feel reasonable, good: It means you have some sense of your own personal details, so as long as it’s backed by some awareness, I’m comfortable telling you to swap in your own number.
- Do not artificially suppress trading you would otherwise do in order to reduce taxable gains unless you’re following some sort of rational program of tax harvesting (a topic for another day).
- Try, to the extent possible, to have your strategic (i.e. data driven) strategies confined to tax-advantaged accounts such as IRAs, in which case you’d be able to put all of this on the back burner.
- If your personal numbers are such that you really do need to minimize taxable gains, I suggest you consider equity ETFs in lieu of equity strategies designed for low turnover that put realized returns at risk. (There are plenty of style-based ETFs you can consider in lieu of generic choices such as SPY.)
- Avoid open-end mutual funds, which are horrifyingly tax-inefficient. (You may wind up paying taxes on gains the fund earned but which you never benefited from; don’t ask, it’s pretty bad stuff.)
Acti-quant 
2 thoughts