Real World Data: Wealthfront Tax-Loss Harvesting

Editor’s Note: This blog post was updated on June 14, 2017. We’ve extended the analysis of the our clients’ actual average benefit from Wealthfront’s Tax-Loss Harvesting to include calendar years 2015 and 2016.

Nothing we write about at Wealthfront raises people’s interest like tax-loss harvesting. The spectrum of opinion is extreme; with some believing it’s a brilliant advantage for investors and others convinced it’s chicanery.

We launched our automated daily tax-loss harvesting service in 2012 and have been running the service ever since. As a result, we have been able to gather a significant amount of empirical evidence on the actual performance of our automated strategy.

We recently analyzed our empirical results over that time period and thought we would share that data to further understanding of this innovative and beneficial investment strategy.

A History of Tax-Loss Harvesting

The concept of tax-loss harvesting has been around for more than a century. For that entire time it has been one of the primary ways wealthy investors have minimized their tax bills.

The concept of tax-loss harvesting is simple. Every year, due to market volatility, some assets go up and some assets go down. If you sell the assets that are down, you can realize the tax loss and use that loss to lower your overall tax bill. Our tax system requires payment on an investment transaction only in the year it takes place, so there is a clear incentive to take losses quickly, while letting gains compound as long as possible.

Let me illustrate with an example. Imagine you purchased 10 shares of an ETF that tracks the S&P 500 at $100 per share and two months later it declines to $90 per share. If you sell that ETF, and replace it with an equal amount of an ETF that tracks a different, but highly correlated index (like the Russell 3000), you can recognize a loss of $100 without substantially changing the risk and reward characteristics of your portfolio. Those losses may be applied up to $3,000 of ordinary income every year, and beyond that to any capital gains you might have incurred elsewhere. These losses can also be banked indefinitely against future gains.

The Misperceptions that Surround Tax-Loss Harvesting

Given the long history surrounding the concept of tax-loss harvesting, I’m constantly amazed by people who claim it has no value. Their primary argument is tax-loss harvesting doesn’t avoid tax; it simply delays it. To play that out, in the scenario above, let’s imagine you start with $1,000 invested in the S&P 500 and it falls to $900. You sell, harvest $100 in losses, and reinvest your money in the Russell 3000. Over the next two years, that $900 investment appreciates back to $1,000. If you sell, you’d owe taxes on the $100 gain, offsetting the $100 loss you harvested earlier.

It’s true your gain and loss are equal, but they are taxed at very different rates. The average Wealthfront client has a combined marginal federal and state ordinary income tax rate of 42.7%. Therefore any losses applied against ordinary income or short-term capital gains have a value equal to the harvested loss times 42.7%. If the ETF bought at $900 was held for more than a year, then the resulting $100 gain will be taxed at the much lower long-term capital gains tax rate (the average Wealthfront client has a combined federal and state long-term capital gains rate of 24.7%). Even though the gain and loss were equal, the different tax rates would have created an $18 benefit (($100 * 42.7%)-($100 * 24.7%)). Tax-loss harvesting’s ability to arbitrage tax rates always leads to a benefit as long as the gain is long-term because long-term capital gains tax rates are always lower than ordinary income and short-term capital gains tax rates. That’s one of the reasons why Wealthfront discourages individuals from opening accounts unless they hope to keep them for several years.

Most people who believe tax-loss harvesting has no benefit also claim small investors have no gains against which they can apply their harvested losses. They forget that harvested losses can be applied against up to $3,000 of ordinary income each year. At a combined federal and state marginal tax rate of 42.7% that represents $1,281 in tax savings per year ($3,000 * 42.7%). Based on our estimated tax-loss harvesting annual after tax benefit of 1.55%, that suggests accounts below $82,000 will get the full benefit of tax-loss harvesting based on the ordinary income allowance alone ($1,281/1.55%).

This benefit is unassailable. It’s just math.  I think financial pundits dismiss it because they think the benefit of canceling out $3,000 in income is small potatoes. People are used to thinking about tax-loss harvesting through the lens of the very rich, but automated investment services have made it available to everyone, and it’s extraordinarily powerful for the middle class and upper middle class. For most people, saving $1,200 per year is a big deal, and tax-loss harvesting does that with extremely little risk. (By comparison, the annual management fee you pay on an $82,000 Wealthfront account that includes tax-loss harvesting comes to just $180, less than 1/6 that estimated after-tax value).

If you’re lucky enough to have a larger account, of course, additional losses can either offset gains or be banked indefinitely for the future.

The final benefit is the savings you create from tax-loss harvesting can be reinvested and compounded until you withdraw your money from your investment account. Compounded for 30 years at a 6% rate, the $1,281 per year that our prototypical Wealthfront client saves from the $3,000 income offset will grow to more than $7,000 – and that’s just the compounded value of savings from one year. Of course, you will owe taxes on the gains when you sell, but you’re still sitting pretty even after you pay the 24.7% long-term capital gains tax.

There are quite a few misunderstandings about how tax-loss harvesting works – if you are curious, you might enjoy my earlier piece on the topic: 10 Things You Probably Didn’t Know About Tax-Loss Harvesting (and should).

Wealthfront Empirical Data (as of mid-2015)

Historically, tax-loss harvesting was only done for the ultra-wealthy. It can be intimidating to implement yourself, so only those investors with sophisticated financial advisors had tax-loss harvesting at their disposal. Even for these investors, harvesting was typically done once a year at year end, because it’s typically far too complicated for the spreadsheet-based approach employed by the advisors to execute more frequently. Accounts with multiple deposits (typical of our clients) were not even a possibility.

Wealthfront has made tax-loss harvesting available to all accounts regardless of size, thanks to the power of automation. As such, we have long running data on the application of sophisticated tax-loss harvesting strategies to normal accounts.

When we launched our tax-loss harvesting service in October 2012, we claimed we could generate a 0.90% annual after tax benefit over the long-term (more than three to five years) for our clients. In 2014 we released a major enhancement to our tax-loss harvesting algorithm that improved its expected annual long-term after tax benefit to 1.55%. These claims were based on both back testing our algorithms, as well as through detailed Monte Carlo simulations of thousands of possible market scenarios (please read our tax-loss harvesting white paper for a detailed explanation of how we derived our estimates).

Now that we’ve been operating our tax-loss harvesting service for several years, we have enough data to analyze our clients’ actual average annual benefit. To simplify our analysis we present data on a calendar year basis:

As you can see we generated a tax alpha (the amount of annual after tax benefit divided by initial portfolio value) of 0.54% in our first full year, despite the S&P 500 being up almost 30%. We know the S&P 500 is not the appropriate benchmark for a diversified portfolio, but it helps provide context. We never expected to be able to generate this much benefit in a market that was up so much. Interestingly the tax alpha for a once a year, year end only tax-loss harvesting strategy would have been close to zero in 2013 because the vast majority of losses we harvested occurred in June and the portfolios recovered by year end.

The tax alpha generated between 2014 and 2016 not only met our estimate of 1.55% but actually exceeded it. In fact,  we do not expect our tax alpha to consistently hover around our annual expected benefit. We actually expect far more volatility in the annual benefit than we have observed thus far due to the inconsistent volatility of the financial markets. Also our tax-loss harvesting is implemented to ensure you will never realize short-term capital gains, and thus higher taxes, from our service. (Our white paper provides a more detailed explanation).

Please note: the tax alphas provided above represent just the tax benefit from the individual years, and do not benefit from the compounding that would result from reinvesting the savings.

On the flip side it’s important to understand that tax alpha defines the maximum possible benefit from tax-loss harvesting, not the actual realized benefit. Your realized benefit will depend on your tax rate, when you withdraw your money and depending on your account size, the amount of gains you have available against which you can apply your harvested losses. As mentioned, this should be no issue for most investors; accounts with less than $82,000 should, on average, be able to use all their losses against just the $3,000 ordinary income and everyone will ultimately generate gains when they liquidate their portfolios, even if they do it gradually.

You Will Likely Realize The Vast Majority Of Tax Alpha

In our post New Research on the Efficacy of Tax-Loss Harvesting, we examined the likely percentage of tax alpha our average client should expect to realize under different scenarios assuming they had enough ordinary income and gains against which they can apply their harvested losses. As you can see from the table below, you can expect to derive 70% to 100% of the hypothetical tax alpha depending on how long you invest and the rate at which you withdraw.


We think a realization of more than 85% of theoretical tax alpha is most likely for how our clients are likely to withdraw their money. That’s because Wealthfront’s tax-efficient withdrawals always prioritize selling shares with the highest cost basis first, so withdrawing a small amount each year upon retirement will lead to a number not terribly dissimilar from the no liquidation column. Very few of our clients withdraw all their money at once, as the most common reasons for withdrawal are either the down payment on a house, or for their children’s education. However we display the full liquidation column because we know many people want to know what that scenario would look like.

The bottom line is you get overwhelming benefit no matter what your withdrawal or investment horizon is (as long as it’s long-term) because of the power of tax rate arbitrage and compounding.

Our Real-World Results Line Up With Theoretical Expectations

It’s hugely gratifying to see our real-world results line up with theoretical expectations. Tax-loss harvesting has always been a huge tool for the wealthiest investors, and we think our automation enhances that result. But more importantly, it puts the value of tax-loss harvesting within reach of the majority of investors for the first time ever. And the data shows it’s producing real results.


Nothing in this blog should be construed as tax advice, a solicitation or offer, or recommendation, to buy or sell any security. Financial advisory services are only provided to investors who become Wealthfront Inc. clients pursuant to a written agreement, which investors are urged to read carefully, that is available at All securities involve risk and may result in some loss.  For more information please visit or see our Full Disclosure.  While the data Wealthfront uses from third parties is believed to be reliable, Wealthfront does not guarantee the accuracy of the information.

This article is not intended as tax advice, and Wealthfront does not represent in any manner that the tax consequences described here will be obtained or that Wealthfront’s investment strategy will result in any particular tax consequence. The tax consequences of this strategy and other Wealthfront strategies are complex and uncertain and may be subject to challenged by the IRS.

For our Monte Carlo analysis we simulated the future returns of two portfolios: the Wealthfront diversified risk level 7 portfolio with daily tax-loss harvesting and the Wealthfront diversified risk level 7 portfolio with no tax-loss harvesting. We evaluated the performance of our daily tax-loss harvesting algorithms across three investment periods (10 years, 20 years and 30 years) and three different withdrawal rates (none, 50% at the end of the investment period and full withdrawal at the end of the investment period). In our Monte Carlo simulation, we first sample asset class returns from the asset class return joint distribution model and transform them into ETF prices. For the daily tax-loss harvesting portfolio, any tax savings generated by tax-loss harvesting are reinvested into the portfolio at the beginning of the next tax year. For both portfolios, we assume an initial deposit of $100,000 and follow on deposits of $10,000 at the beginning of every quarter.

We assumed a married couple filing jointly in California, with a combined federal and state short-term capital gain tax rate of 42.7%, a combined federal and state long-term capital gain tax rate of 24.7%, and a portfolio risk score of 7. Dividends, earnings and interest were not considered.  Review further details in our Tax-Loss Harvesting Whitepaper.

Hypothetical performance is not an indicator of future actual results.  The results do not represent returns that any client actually attained. Hypothetical results are calculated by the retroactive application of a model constructed on the basis of historical data and based on assumptions integral to the model which may or may not be testable and are subject to losses.

Wealthfront assumed we would have been able to purchase the securities recommended by the model and the markets were sufficiently liquid to permit all trading.  Hypothetical performance is developed with the benefit of hindsight and has inherent limitations.  Specifically, hypothetical results do not reflect actual trading or the effect of material economic and market factors on the decision-making process.  Actual performance may differ significantly from hypothetical performance.  There is a potential for loss as well as gain that is not reflected in the hypothetical information portrayed. Investors evaluating this information should carefully consider the processes, data, and assumptions used by Wealthfront in creating its historical simulations.

For the chart displaying the Actual average annual benefit of Tax-Loss Harvesting, the results shown are from a composite which consists of all risk level taxable client accounts, and managed in accordance with our tax loss harvesting and direct indexing investment methodologies from 2013 to 2016.

Past performance does not guarantee future results. Commissions are not considered since clients on the Wealthfront platform are not charged trading commissions. For all periods the performance information includes the reinvestment of dividends and interest.

We assumed a married couple filing jointly in California, with a combined federal and state short-term capital gain tax rate of 42.7%, a combined federal and state long-term capital gain tax rate of 24.7%.  

Some of Wealthfront’s investment strategies can lead to high levels of trading that could result in (a) bid-ask spread expense; (b) trade executions that may occur at prices beyond the bid ask spread (c) trading that may adversely move prices, such that subsequent transactions occur at worse prices; (d) trading that may disqualify some dividends from qualified dividend treatment; (e) unfulfilled orders or portfolio drift, in the event that markets are disorderly or trading halts altogether; and (f) unforeseen trading errors.

Both daily tax-loss harvesting and stock-level tax-loss harvesting described in this release may generate a higher number of trades due to attempts to capture losses. There is a chance that Wealthfront trading attributed to tax-loss harvesting may create capital gains and wash sales and could be subject to higher transaction costs and market impacts. In addition, tax-loss harvesting strategies may produce losses, which may not be offset by sufficient gains in the account and may be limited to a $3,000 deduction against income. The utilization of losses harvested through the strategy will depend upon the recognition of capital gains in the same or a future tax period, and in addition may be subject to limitations under applicable tax laws, e.g., if there are insufficient realized gains in the tax period, the use of harvested losses may be limited to a $3,000 deduction against ordinary income and distributions.

The effectiveness of the tax-loss harvesting to reduce the tax liability of the client will depend on the client’s entire tax and investment profile, including purchases and dispositions in a client’s (or client’s spouse’s) accounts outside of Wealthfront and type of investments (e.g., taxable or nontaxable) or holding period (e.g., short- term or long-term). The performance of the new securities purchased through the tax-loss harvesting service may be better or worse than the performance of the securities that are sold for tax-loss harvesting purposes.  

Wealthfront only monitors for tax-loss harvesting for Wealthfront accounts. Clients are responsible for monitoring their and their spouse’s accounts outside of Wealthfront to ensure that transactions in the same security or a substantially similar security do not create a “wash sale.” A wash sale is the sale at a loss and purchase of the same security or substantially similar security within 30 days of each other. If a wash sale transaction occurs, the IRS may disallow or defer the loss for current tax reporting purposes. A client may request spousal monitoring online or by calling Wealthfront at (844) 995-8437. If Wealthfront is monitoring multiple accounts to avoid the wash sale disallowance rule, the first taxable account to trade a security will block the other account(s) from trading in that same security for 30 days.

The S&P 500® (“Index”) is an index of 500 stocks seen as a leading indicator of U.S. equities and a reflection of the performance of the large cap universe, made up of companies selected by economists. The S&P 500 (“Index”) is a product of S&P Dow Jones Indices LLC and/ or its affiliates and has been licensed for use by Wealthfront. For more information on any of S&P Dow Jones Indices LLC’s indices please visit

Check out Wealthfront's services. We support taxable account, IRAs,
401(k) rollovers, and 529 college savings plans.

Ready to invest in your future?

Open an account