Help me understand MMM's case for Tax Loss Harvesting

[mattpew]

More specifically, the following quote from [his Betterment experiment post](https://www.mrmoneymustache.com/betterment-vs-vanguard/): "a $79k deduction saves me over $25,000 in income taxes right now"

Losses can only be deducted from ordinary income up to $3k/year.  Does the average investor have any way to make immediate use of the other $76k in carryover loss?  If not, his subsequent analysis to justify Betterment's fees is just wrong: "...which I can use to buy still more investments. If this $25,000 goes on to earn a conservative 7% ($1750/year), it is more than paying for the fees Betterment charges me". 

His logic seems based on the assumption that the loss is immediately deductible, but I'm struggling to understand how that would be the case (he doesn't make any mention of realized investment gain that needs to be offset).  With a significant enough portfolio value (and investing at regular intervals) it's trivial to find at least $3,000 in harvestable losses per year that will provide an immediate benefit.  Anything additionally harvested will only help in the case a scenario is encountered later in life where you are forced into realizing a large gain.

Is there something I'm missing here?

[FLBiker]

PTF, mostly -- your understanding jibes with my own.  Many years ago, I remember claiming more than a $3000 loss, and I got my hand slapped by the IRS and had to carry the balance forward.  I don't really mess around with TLH myself.  We moved from the US to Canada, and our taxes are complicated enough. :)

[wageslave23]

Paying for it probably doesn't make sense. It's easy to do yourself though and it's a nice $3000 per year deduction.

[mattpew]

Hmm - I don't know what PTF means in this context, Google pulls up lots of definitions.

[secondcor521]

Hmm - I don't know what PTF means in this context, Google pulls up lots of definitions.

PTF means posting to follow.  If a person posts in a thread, then that thread will show up if they click on the "Show new replies to your posts" link above.

OP, I think there's two reasonable interpretations of the blog post:

1.  MMM is being overly optimistic and very hand-wavy and a bit ignorant about how federal income taxes work, by ignoring the $3K per year against ordinary income tax limitation.  (You're right about that part, BTW (by the way).)

2.  MMM is being moderately hand-wavy by assuming that he'll have ~$80K of realized capital gains that he can use to offset the loss.  He's still doing optimistic math anyway, because his Betterment losses are erasing capital gains taxes, not ordinary income taxes.  Thus the applicable tax rate on most of that is probably more in the low- to mid-twenty percent - only the last $3K is saving his 40%-ish marginal rate, if he ends up with a small net loss.

ETA (edited to add):

3.  In point #2 above I was assuming LTCG, because that's all I ever do.  If he manages to erase ~$80K in STCG with the Betterment losses, then his math would be correct.  IIRC (if I recall correctly) STCG are taxed at ordinary income rates.

[Telecaster]

More specifically, the following quote from [his Betterment experiment post](https://www.mrmoneymustache.com/betterment-vs-vanguard/): "a $79k deduction saves me over $25,000 in income taxes right now"

Losses can only be deducted from ordinary income up to $3k/year.  Does the average investor have any way to make immediate use of the other $76k in carryover loss?  If not, his subsequent analysis to justify Betterment's fees is just wrong: "...which I can use to buy still more investments. If this $25,000 goes on to earn a conservative 7% ($1750/year), it is more than paying for the fees Betterment charges me". 

His logic seems based on the assumption that the loss is immediately deductible, but I'm struggling to understand how that would be the case (he doesn't make any mention of realized investment gain that needs to be offset).  With a significant enough portfolio value (and investing at regular intervals) it's trivial to find at least $3,000 in harvestable losses per year that will provide an immediate benefit.  Anything additionally harvested will only help in the case a scenario is encountered later in life where you are forced into realizing a large gain.

Is there something I'm missing here?

Nope.  You are correct.  And MMM overlooked a couple other things too.   One of the benefits of tax loss harvesting is that you get to deduct from top marginal ordinary income rate.  That steps up your basis for higher future capital gains rates, but income is typically taxed at high rates than capital gains, so that's probably a good trade.  Plus, bird in the hand.

Now, let's say you have a $500,000K portfolio.  And let's say you tax loss harvest $3000.  If you are in the 33% tax bracket, that saves you $990.  That's probably worth doing.  But Betterment charges a 0.25 fee.   So it will cost you $1,250 to save $990, and because you stepped up the basis it is even more expensive than that.     Worse, if you have a $500K portfolio it is dead easy to come up with $3K of losses on your own.   

[MustacheAndaHalf]

I'm generally opposed to paying a fee for tax loss harvesting.  Over time, the market generally goes up, which makes it less and less likely the earliest contributions get tax loss harvesting.  So you pay a fee on all assets, but only the newest contributions get the benefit.

That said, and setting aside the tax issue raised above, I was impressed with Mr Mustache's methods.  In Portfolio Visualizer I entered a starting balance of 100,000 and 1,000/month contributions and ran that from Oct 2014 - Dec 2021.  Those results matched the VTI and VXUS displayed in his graph, which means he remembered to include the 1,000/month contributions when comparing to the U.S. market and international.  So I'm glad he paid attention to that detail.

I think MMM's point is that Betterment charges a reasonable 0.25% for tax loss harvesting and other services, which can simplify investing for people.  Note MMM's post is titled "the Betterment Experiment" rather than anything specific to tax loss harvesting.  I would argue his main point is this:

Why do I recommend this?

In one word: Simplicity. OK, maybe we could add a second word to that: Efficiency.

After ten years of writing this blog and hearing from readers of all types, I find that the same question keeps coming up: “What is the single step I can do to get started in investing?”

(bold in original)

[Telecaster]

I'm not sure Betterment really adds any simplicity.   You can do automatic investments at Vanguard into what ever mutual funds you like.

Even though 0.25% in fees isn't a lot, over a typical investing horizon it could easily cost you the price of a nice car. 

[maizefolk]

2.  MMM is being moderately hand-wavy by assuming that he'll have ~$80K of realized capital gains that he can use to offset the loss.  He's still doing optimistic math anyway, because his Betterment losses are erasing capital gains taxes, not ordinary income taxes.  Thus the applicable tax rate on most of that is probably more in the low- to mid-twenty percent - only the last $3K is saving his 40%-ish marginal rate, if he ends up with a small net loss.

ETA (edited to add):

3.  In point #2 above I was assuming LTCG, because that's all I ever do.  If he manages to erase ~$80K in STCG with the Betterment losses, then his math would be correct.  IIRC (if I recall correctly) STCG are taxed at ordinary income rates.

Remember state income taxes as well as federal though. Colorado taxes capital gains as defined on your federal return at 4.63%. Between that and the relatively big firehose of cash the blog represents it seems plausible MMM could be paying in high 20s rather than mid-to-low 20s.

That's still lower than the 31.6% that $79k of losses equating to $25k of savings represents, but it's substantially get closer.

[MustacheAndaHalf]

I'm not sure Betterment really adds any simplicity.   You can do automatic investments at Vanguard into what ever mutual funds you like.

Even though 0.25% in fees isn't a lot, over a typical investing horizon it could easily cost you the price of a nice car.

You cannot do automatic investments if you don't know what to buy.  You can read MMM's article to see how he viewed it, but my interpretation is that many people don't know what to buy at all, and Betterment could provide a single step solution for them.

Most people invest more over time, and markets have varied returns.  But if you compare 10%/year versus 9.75%/year over 30 years, you get 17.5x versus 16.3x.  Actual returns will be far more varied than that, making comparison difficult at best.  For me, managing my own investments is better.

For someone sitting on cash, and too scared to invest, Betterment is a much better "one step" solution.

[secondcor521]

2.  MMM is being moderately hand-wavy by assuming that he'll have ~$80K of realized capital gains that he can use to offset the loss.  He's still doing optimistic math anyway, because his Betterment losses are erasing capital gains taxes, not ordinary income taxes.  Thus the applicable tax rate on most of that is probably more in the low- to mid-twenty percent - only the last $3K is saving his 40%-ish marginal rate, if he ends up with a small net loss.

ETA (edited to add):

3.  In point #2 above I was assuming LTCG, because that's all I ever do.  If he manages to erase ~$80K in STCG with the Betterment losses, then his math would be correct.  IIRC (if I recall correctly) STCG are taxed at ordinary income rates.

Remember state income taxes as well as federal though. Colorado taxes capital gains as defined on your federal return at 4.63%. Between that and the relatively big firehose of cash the blog represents it seems plausible MMM could be paying in high 20s rather than mid-to-low 20s.

That's still lower than the 31.6% that $79k of losses equating to $25k of savings represents, but it's substantially get closer.

Sure.  I was just trying to get close.  MMM's blog article has a number of minor inconsistencies in it, so I figured if he wasn't being precise I didn't need to be either.

As an aside, even with state taxes, "high 20s" is still pretty far from the "marginal tax rate of 40%" mentioned early on in the blog post.  Eh, whatever.  I know MMM is a fire hose of optimism guy with sometimes a lack of attention to detail.  That's OK with me.

[mattpew]

Note MMM's post is titled "the Betterment Experiment" rather than anything specific to tax loss harvesting.  I would argue his main point is this:

I understand there's a broader theme to the article, this thread was just to drill into how he came up with the claim I quoted in the OP.

a number of minor inconsistencies... sometimes a lack of attention to detail.  That's OK with me.

Eeeh, this seems like a pretty generous dismissal in this case.

It's a fundamental misrepresentation of how our tax code works.  I trust Pete enough that this isn't being done on purpose, but would still like for him to correct it given that this misunderstanding is used to justify Betterment's cost, which he has a referral link on his site for.

[secondcor521]

a number of minor inconsistencies... sometimes a lack of attention to detail.  That's OK with me.

Eeeh, this seems like a pretty generous dismissal in this case.

It's a fundamental misrepresentation of how our tax code works.  I trust Pete enough that this isn't being done on purpose, but would still like for him to correct it given that this misunderstanding is used to justify Betterment's cost, which he has a referral link on his site for.

I accept it not because it's right, but because (a) I understand how the tax code actually works and I've decided not to use Betterment, so it doesn't affect me personally, and (b) I know I won't change Pete's behavior or the article, and I've accepted the fact that I can't personally correct everything wrong out there on the Internet.

https://xkcd.com/386/