Wash sale rules are pretty straight forward when you're trading stocks directly. However, how are these rules applied to ETFs?

Right now I own shares of CMF and want to do some tax-harvesting on some of my short-term losses for 2016. I want to reinvest the cash in some other security for the 30 days that I can't buy CMF back, and want it to be a bonds ETF.

I'm considering AGG as a candidate, but I'm not sure if it will be considered "substantially identical" to CMF by the IRS. I think it wouldn't, since the only overlap between CMF and AGG is bonds from California. How can I know for sure?

How about NYF? I'm pretty certain they wouldn't be considered "substantially identical" since they do not overlap at all.

How can I know objectively if two ETFs are "substantially identical"?

2 Answers 2


It sounds like this is an entirely unsettled question, unfortunately. In the examples you provide, I think it is safe to say that none of those are 'substantially identical'; a small overlap or no overlap certainly should not be considered such by a reasonable interpretation of the rule.

This article on Kitces goes into some detail on the topic. A few specifics.

First, Former publication 564 explains:

Ordinarily, shares issued by one mutual fund are not considered to be substantially identical to shares issued by another mutual fund.

Of course, what "ordinarily" means is unspecified (and this is no longer a current publication, so, who knows).

The Kitces article goes on to explain that the IRS hasn't really gone after wash sales for mutual funds:

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 the rules might allow you to loss-harvest a particular stock you couldn’t have otherwise, it also limits you from harvesting ANY losses if the overall fund is up in the aggregate, since losses on individual stocks can’t pass through to the mutual fund shareholders.

But then goes to explain about ETFs being very different: sell SPY, buy IVV or VTI, and you're basically buying/selling the identical thing (99% or so correlation in stocks owned).

The recommendation by the article is to look at the correlation in owned stocks, and stay away from things over 95%; that seems reasonable in my book as well.

Ultimately, there will no doubt be a large number of “grey” and murky situations, but I suspect that until the IRS provides better guidance (or Congress rewrites/updates the wash sale rules altogether!), in the near term the easiest “red flag” warning is simply to look at 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 are not ”substantially identical” to each other in performance.

Basically - use common sense, and don't do anything you think would be hard to defend in an audit, but otherwise you should be okay.


Nobody knows for sure what "substantially identical" means because the IRS hasn't officially defined it. Until they do so, it would come down to the decision of an auditor or a tax court. The rule of thumb that I have always heard is if the funds track the same index, they are probably substantially identical. I think most people wouldn't consider any pair of AGG, CMF, and NYF to be substantially identical, so you should be safe with your tax-loss harvesting strategy.

You must log in to answer this question.