High-return tax-efficient or low-return tax-inefficient assets in Roth IRA?

From Reddit Personal Finance we read: I see a lot of guides (e.g. Bogleheads) recommend putting tax-inefficient assets in tax-advantaged accounts, which makes sense. However I would think that due to the contribution limits on tax-advantaged accounts like a Roth IRA you would also do better putting an asset with a high rate of return in rather than one with a low rate of return. My reasoning is as follows:

If you put $100 of bonds yielding 3% in dividends into a Roth and $100 of 7% stocks into a taxable account taxable at a 25% capital gains tax rate, both set to reinvest earnings, and cash out after ten years, you end up with $34.99 in untaxed bond dividends and $101.38 in capital gains taxed at 25% (so $76.04). This results in a net return of $111.03.

On the other hand, if you put $100 of bonds yielding 3% in dividends distributed quarterly into a taxable account taxed at a 33% marginal tax rate and $100 of stocks yielding 7% into a Roth, both set to reinvest earnings, and cash out after ten years, you end up with (I think) $22.26 in bond dividends and $101.38 in untaxed capital gains. This results in a net return of $123.64.

Is there something off with my math? Am I misunderstanding or missing something to do with investments and taxes? Or is this the correct way to manage the portfolio with respect to taxes this combination of tax efficiency and yield? Is there anything else I should consider?

Since I'm not sure I calculated bond return in the second scenario correctly, here's my method: I had a gross return of 3% annually, but that was taxed at 33%, which lowered my net return to (3%)*(67%)=2.01%. I then calculated the continuously compounded return at 2.01%.

all 8 comments

sorted by: 

best

[–]kurds_way 2 points 17 hours ago

The Bogleheads advice is still usually correct, though less important than it once was due to low interest rates and tax law changes.

At a glance, what % of that 7% stock return is divs and what's cap gains in your model? And don't forget to adjust for risk - $1 in a pretax taxable account is less risky than $1 in a Roth.

• permalink

[–]Kibatsu[S] 1 point 16 hours ago

I was modeling the stock return as 100% capital gains to make the point about tax efficiency vs return rate while keeping it as simple as possible. I'm less interested in the precise numbers than knowing if high enough returns can trump tax-inefficiency for space in a tax-advantaged account, which it seems is possible. The difference between calculating expected returns and variance to choose allocations versus not worrying about it and following the Bogleheads advice, I suppose.

• permalink
• parent

[–]kurds_way 2 points 16 hours ago

I was modeling the stock return as 100% capital gains

Well there's your problem. The reason stocks in taxable works is that most of the stock gains grow tax deferred for decades, sometimes forever. You've just turned a tax efficient investment into a very tax inefficient one.

Also, note my point about risk adjusting (picture what you'd be left with if stocks dropped 99.99% at the end of the period with each alternative).

The difference between calculating expected returns and variance to choose allocations versus not worrying about it and following the Bogleheads advice, I suppose.

Better to prove stuff to yourself - sometimes Bogleheads gets it right, sometimes not so much ;)

• permalink
• parent

[–]Kibatsu[S] 1 point 16 hours ago

You've just turned a tax efficient investment into a very tax inefficient one.

Is this because the stocks are being sold after too short a period, then? Or because of something about capital gains vs income taxes I've missed?

• permalink
• parent

[–]kurds_way 2 points 16 hours ago

You're assuming every year you sell you entire stock holding, pay cap gains on it, then buy it back. A little spreadsheet work will show this costs you a lot more in taxes that modeling what actually happens, where you only pay taxes on the divs, and don't pay cap gains until you sell (or avoid cap gains completely).

• permalink
• parent

[–]Kibatsu[S] 1 point 8 hours ago

I don't think that's what I did, unless it was implicit somehow. I allowed the stock value to appreciate at 7% annually for ten years, and then took the profit at the end of the ten year period and took 25% off. The capital gains tax applies to the whole profit at the point of sale, right? Not just the profit in that year? ($100 * exp[0.07 * 10] - $100) * 0.75 = $76.03

• permalink
• parent

[–]kurds_way 1 point 4 hours ago

Oh, so you're assuming no div taxes paid each year. In that case your $100 grows to $134.39 (100*1.07^10), then you pay 25% taxes on the $34.39 gains, for an after tax value of $125.79.

• permalink
• parent

[–]Plum12345 1 point 4 hours ago

Its not that you made a mistake, but it all depends on the numbers you choose. Why did you pick 10 years? If you pick 30 then its a whole different story because of compounding.

Another thing to consider is the state you live in. I live in California. My state income tax is over 10%. Some general obligation bonds are both state and federal tax free and very safe.