Harvesting capital losses is a popular way to generate tax savings. It’s so popular, in fact, that, not long after the U.S. income tax was created a century ago, the strategy was rapidly adopted as a tax savings technique and became a perceived tax abuse.

This, in turn, led to the establishment of one of the first pieces of legislation to close a tax loophole: the wash sale rule.

The challenge of the wash sale rule is that the requirement not to own a “substantially identical” stock or bond within the 61-day wash sale period has become outdated given the rise of pooled investment vehicles such as mutual funds — and especially given the explosion of index ETFs. Now taxpayers face oddly disparate treatment: It’s not permitted to harvest the losses on a group of stocks that are down and replace them with the same stocks, but it’s OK to harvest the loss on a fund holding those stocks and replace it with another fund that owns the same overlapping securities.

The wash sale rule was intended to require investors to endure some tracking error risk with the replacement security owned during the wash sale period. So doing loss harvesting among ETFs that are designed to track the same index — and have a 0.99+ correlation and no risk of performance difference — almost certainly violates the law’s intent.

It remains to be seen whether the IRS will become more aggressive in pursuing the issue, or whether lawmakers will attack this loophole once again. But the fact that there has been no action so far to limit the abuses does not necessarily protect taxpayers who act in clear violation of the rule’s intent. Advisors should at least be cautious in considering how far they push the limits with tax-loss harvesting of mutual funds and ETFs.


The concept of the wash sale rule is relatively straightforward: Its purpose is to prevent someone from selling an investment for a tax loss and quickly buying it back again, without substantively changing the investor’s economic position.

To the extent the investor has purchased a stock 30 or fewer days before or after a sale that recognized a loss, the purchase within that 61-day period would cause the otherwise deductible loss to be a wash. And when the wash sale rule is triggered, any losses from the original sale are no longer deductible. Instead, the disallowed loss is added onto the cost basis of the replacement security. In addition, the holding period of the new investment (to determine eligibility for long-term capital gains treatment) will include the holding period of the original investment sold at a loss.

Here’s an example: You bought 1,000 shares of XYZ stock for $80 per share in February 2013. On June 3 of this year, you sell the stock for $62, recognizing an $18-per-share (or $18,000) loss. Later that week, the stock rebounds, and you buy it back at $67 on June 9.

Because the purchase occurred within 30 days of the sale, the original $18-per-share loss is no longer deductible. The newly purchased stock now has a cost basis of $67 (purchase price) + $18 (wash sale loss) = $85.

Later on, if you sell the stock for $67, you will again recognize an $18-per-share loss. Alternatively, if you hold the investment until it tops $85, the first $18 of gain will not be taxable, effectively having been offset by the wash sale loss. And any further capital gains on those shares will immediately be eligible for long-term rates, as the 28-month holding period from the original stock is automatically tacked on to the replacement purchase.


A key aspect of the wash sale rule is that it is triggered only when the investor buys back a “substantially identical” investment to the one that was sold for a loss (and/or a contract or option to buy the investment again).

And in the context of individual stocks and bonds, the definition of a substantially identical security is fairly well established. Stocks of different corporations are not substantially identical — so while selling Ford and buying back Ford stock (or Ford call options) would trigger the rule, selling Ford and buying GM, or selling Dell and buying Hewlett-Packard (same industry but clearly a different company) would be acceptable.
Nor are bonds substantially identical if they have different issuers — or, of course, if their underlying provisions differ (e.g., different interest rates or maturity dates).

Note that the rule refers to “substantially,” not precisely, identical investments. IRS Publication 550 notes that different corporations may be substantially identical if they are predecessor and successor corporations in a reorganization, for instance. Preferred and common stock can still be substantially identical if the former is convertible into the latter and they have the same voting rights and dividend restrictions (and trade at prices similar to the conversion ratio).

When it comes to mutual funds, though, the situation is murkier — and there has still been remarkably little clarification on the issue. Former IRS Publication 564 acknowledged that “ordinarily, shares issued by one mutual fund are not considered to be substantially identical to shares issued by another mutual fund.” But there was no clarification as to under what circumstances two mutual funds could be substantially identical.
The root of the challenge is that, ultimately, a mutual fund is a pooled investment vehicle for the underlying stocks, bonds or other securities — the mutual fund only declines (or increases, for that matter) in value because of changes in its underlying securities.

This means executing a tax-loss harvesting sale of a mutual fund and switching to another fund isn’t actually about harvesting the loss of the mutual fund itself, but instead the losses on the underlying securities — even though a replacement fund could be holding the same losing securities.

For instance, consider two of the largest actively managed mutual funds: American Funds’ Growth Fund of America (AGTHX) and Fidelity Contrafund (FCNTX). Most observers would easily consider the two mutual funds substantively different. Yet a recent look at the underlying holdings — see the chart below — revealed that, among the top 25 holdings alone, almost a third of the securities overlap, accounting for nearly 12% of the funds’ assets.

And the sheer amount of overlap leads to remarkably similar investment performance. The bottom portion of the chart shows the returns for the two funds over the past 10 years. While cumulatively the funds have exhibited a significant compound return difference over time, the correlation of their daily price changes is a whopping 0.93.

Given the incredibly high correlation between these two funds, the grounds for calling them “not identical” may begin to break down. Not only does the substantive overlap in their securities lead to remarkably similar price returns — especially over a cycle as short as a 30-day wash sale period — but the reality is that big losses in just a few stocks could drive down the price of both mutual funds in a similar manner.

Thus, selling one and replacing with the other could mean replacing the exact stocks that were down with identical stocks held in another fund.

For instance, if a loss in Growth Fund of America was triggered by a big decline in tech sector stocks such as Amazon, Google and Microsoft — overlapping holdings in both — while the rest of the market was flat, and the investor sold for a loss and replaced it with Contrafund, the investor would replace the exact stocks that were down with identical replacement stocks and still claim the loss.

If the investor held the underlying stocks directly, harvesting those losses could be done on a stock-by-stock basis — but the replacement securities to avoid the wash sale would actually have to be different stocks, not the same stocks just held within a different pooled mutual fund wrapper.

Over the years, the IRS has not pursued wash sale abuses against mutual funds, perhaps because it just wasn’t very feasible to crack down on them, or perhaps because it just wasn’t perceived as that big of an abuse. After all, while large diversified mutual funds often have a big chunk of overlap, there are clearly at least some differences in the underlying stocks, as well as the manner by which the manager buys and sells those stocks over time. Given these differences, mutual funds have kept their “ordinarily not substantially identical” treatment.


With index ETFs, though, the situation gets murkier, as nominally different ETFs and providers may be mimicking the exact same underlying index and basket of stocks.

For instance, SPY (the State Street S&P 500 fund) and IVV (the iShares version tracking the same index) have different fund managers and expense ratios and may execute the underlying mechanics slightly differently. But because they are designed to track the exact same index, they own virtually identical stocks.

So does swapping these run afoul of the wash sale rule?

For active funds, this was never an issue; fund managers might hold similar views and similar stocks, but they weren’t by design holding the same stocks with the same target allocation. And in the past, the smaller size of index funds limited the impact. But indexing has now become huge; S&P 500 ETFs SPY, IVV and VOO (the Vanguard version) account for nearly $300 billion.

Swapping from SPY to IVV would almost certainly violate the intent of the wash sale rule. But what about swapping from the S&P 500 to the S&P 100? Yes, these are different index funds. But any losses driven by the top 100 stocks will by definition be felt in the replacement fund. Overall, almost two-third of the S&P 500 is made up of the stocks in the S&P 100, and the two exhibit a whopping 0.983 correlation on their daily price returns over the past decade.

In fact, even switching index providers doesn’t actually yield much of a difference. For instance, take the S&P 500 and the Russell 1000 (both large-cap U.S. stock indexes).

These may appear to be different indexes containing different companies. But the top holdings of the funds that mirror them are almost entirely identical; when I checked recently, the only difference in their top 25 was that Comcast and IBM alternated in the No. 24 and No. 25 spots. Over the past decade, the indexes’ daily price changes had a whopping .0.991 correlation.

In fact, because the performance of large-cap stocks is so dominating in cap-weighted index funds, IVV (tracking the S&P 500) actually has a 0.989 correlation to the Vanguard Total Stock Market fund (VTSMX) as well.

Unless the fund is extremely constrained — for example, by some non-traditional weighting approach, or is narrowly confined to a particular industry or sector — a huge portion of ETFs based on differently constructed indexes or from different providers still end up with extremely similar performance tracking, especially over shorter time periods, due to market-cap weightings.


Ultimately, the problem with the wash sale rule is that it simply was not intended for a world of pooled investment vehicles.

If two funds overlap by 70% in their underlying holdings, how do you determine whether they are substantially identical? Is there a threshold where the rules would apply? Does it matter whether the losses being harvested are attributable to the 70% of overlapping stocks or the other 30%, and is there an easy way to figure this out?

These issues crop up in regular practice for advisors. And there will no doubt be a large number of murky situations.

I suspect that until the IRS provides better guidance — or Congress rewrites the wash sale rule altogether — the easiest red flag to look for is simply the correlation between the original investment being loss-harvested and the replacement security. At correlations above 0.95, and especially at 0.99+, it’s difficult to argue that the securities tracking the same index are not substantially identical to each other in performance.

Of course, the common goal of finding good replacement securities when doing tax-loss harvesting is specifically to find those that have the lowest possible tracking error in anticipated performance relative to the original security. Yet the whole purpose of the wash sale rule is to ensure that tracking error is present. In other words, the wash sale rule forces a risk/return trade-off to be considered whenever you harvest a loss.

If you believe you’ve found a replacement investment that removes any danger of a material difference in performance during the wash sale period, you’ve probably just identified a substantially identical security that would violate the wash sale rule.

If the replacement security does present a less-than-near-perfect correlation to the one being harvested, it should pass muster with the wash sale rule. But then you’ll have to decide if the tracking error risk is worth the value of harvesting the loss in the first place.

Michael Kitces, CFP, a Financial Planning contributing writer, is a partner and director of research at Pinnacle Advisory Group in Columbia, Md., and publisher of the planning industry blog Nerd’s Eye View. Follow him on Twitter at @MichaelKitces.

Read more: