Investment outperformance, part 2: how you invest

You've determined your risk tolerance and risk capacity, and you've used them to create an efficient asset allocation of uncorrelated asset classes that balances risk and reward. You've taken factors of increased returns into account, adjusting your allocation and/or mutual fund choice accordingly. So far, so good. What next?

Everything we've talked about so far has been somewhat abstract, portfolios drifting in the ideal theoretical space inhabited by physics homework involving frictionless surfaces and spherical cows. The real world, though, has friction and cows that are less than geometrically perfect; it has markets that go up and down, taxes, and 401(k)'s with funds that don't quite match what you were hoping to invest in. And therein lie several more opportunities to optimize your investments.

Opportunistic rebalancing

If you set up a portfolio that's 60% stocks and 40% bonds and then just let it run, chances are pretty high that after a few decades it won't be 60/40 anymore. Odds are it'll be tilted towards stocks, given their penchant for outperforming bonds -- more risk will have snuck into your portfolio, like a thief in the middle of the bull market. Hence the need for rebalancing: a periodic "reset" of your portfolio to bring it back in line with your chosen asset allocation.

You could rebalance your portfolio once a year, and that would be fine -- but is it optimal? In 2008, researcher Gobind Daryani proposed that "opportunistic rebalancing" could produce higher returns, in addition to controlling risk, and this method is now widely accepted as optimal in the investing research community. So: what does opportunistic rebalancing look like?

• For each asset class, set a target error percentage. Daryani's research indicates 20% as optimal.

• Check your asset allocation frequently. According to the paper, biweekly works well, and weekly or daily seem to be even better.

• When an asset class exceeds the error threshold, rebalance. For example, if your large-cap asset allocation target is set at 20% of your portfolio and your error threshold is 20%, rebalance if it exceeds 24% or goes lower than 16%.

Why does this work? For one thing, it's making use of the advantage of uncorrelated asset classes: as they swing-out-of-phase with each other, you're able to "buy low and sell high". For another, by setting a somewhat-loose error threshold of 20% and by only rebalancing assets that are beyond that threshold, you're letting the momentum premium do its thing, allowing outperforming asset classes to continue to outperform for a bit before selling them off.

But don't take my word for it -- read the paper!

Tax-loss harvesting

If all of your assets are in 401(k)'s, 403(b)'s, or IRA's, you can skip this section, as it only applies to taxable investments. Otherwise, read on!

The idea behind tax-loss harvesting is relatively simple, though it has a few wrinkles that you need to be aware of. The theory is this: if a taxable investment dips below the price you bought it for, you can sell it at a loss for a deduction on this year's taxes. By "harvesting" these taxable losses, you're able to increase your gains by lowering your overall tax bill. There are some caveats, though!

You can't just turn around and buy back the same investment (which would be ideal). If you sell an investment at a loss and then buy it back within 31 days, that's called a "wash sale", and the tax loss is ignored by the IRS. You have to either hold the proceeds of the sale in cash for 31 days (not ideal!), or buy an investment that isn't "substantially identical". What exactly is a "substantially identical" investment? Well, the IRS hasn't come out with recent guidance to make it crystal clear, but one thing most experts agree on is that different ETF's that track the same index are, in fact, substantially identical. Beyond that, you'll have to work with your CPA to figure out something they could defend to an auditor.

You can only deduct a certain amount from this year's taxes. While investment losses can be used to offset an unlimited amount of taxes on investment gains, they can only give you up to a $3,000 deduction from taxes on other income (e.g. your paycheck). That doesn't mean the excess is lost, though! It's just carried forward to next year, and the one after that, etc., offsetting investment and other gains until it's used up.

It's a tax deferral, not a tax avoidance. When you sell that investment at a loss and buy a nonidentical one, you're effectively lowering your cost basis (the price you bought the investment for). That means that when you sell the investment down the road, you'll be taxed on an additional gain equal to the loss you harvested. For example: let's say you buy an investment at $20. If you sell it in 20 years at $40, you're taxed on $20 of gain. But if it dipped down to $10 in the meantime and you harvested those losses, the price you bought that investment at would go down from $20 to $10, which means selling at $40 would get you taxed on $30 of gain.

So with all those caveats, is it still worthwhile? Yes! The $3,000 limit only applies to non-investment income, and the old accountant saying of "a tax deferred is a tax avoided" holds up to empirical research in this case.

Efficient asset location

No, that wasn't a typo: I'm speaking here not of asset allocation, but asset location. You've chosen which asset classes you want to hold and in which amounts, but which accounts do they go in? For example, let's say you've decided on a 75/25 stock/bond allocation. Do you mirror that allocation across all accounts, holding 75% stocks in your 401(k), and 75% in your IRA, and 75% in your brokerage, and so on? That's certainly fine, but again, it's not optimal. There are two items to look at when it comes to optimization here: available implementations and tax efficiency.

With regards to available implementations, sometimes an account just doesn't have a good implementation of what you're looking for. Maybe your 401(k) doesn't have access to real estate investment trusts (not uncommon), or maybe its implementation is super expensive (extremely not-uncommon, though getting better). In this example, you could allocate what would otherwise go into real estate into something similar -- for example, mid-cap stocks -- or you could cut that allocation out of the 401(k) entirely and overweight it in another account that does have access to REIT's.

Tax efficiency can be even more beneficial to take into account. The aforementioned REIT's are an excellent way to invest in real estate...but they're very tax-inefficient, as they must receive 75% or more of their income from rent and pass 90% of their taxable income onto shareholders, which means a big chunk of their returns are taxed as ordinary income. Contrast this with domestic US stocks, whose returns are often primarily in the form of capital gains, and whose dividends are often "qualified", and thus are taxed at the capital gains rate as well. Therefore, to optimize on this front means to put the tax-inefficient classes in your tax-advantaged accounts, and to leave the tax-efficient ones in your taxable accounts.

Before you go off to optimize for tax efficiency, there's another issue to take into account: tax efficiency matters in direct proportion to the expected returns. In other words: sure, short-term bond dividends are taxed at ordinary income, but the expected long-term growth is tiny compared to, say, that of emerging markets stocks! When determining what goes where, make sure you take this into account as well.

Do or delegate?

If all this sounds like a fair amount of work: it is, especially if you don't have the tools to automate it all. The good news is that any investment manager worth their salt will do this for you (but many aren't, so you should definitely ask). Of course, they're going to charge you for it, so the question then becomes: is it worth it? Some things to take into consideration. (Full disclosure: Seaborn does investment management.)

I'm not interested in doing this myself. If the idea of implementing daily opportunistic rebalancing, tax-loss harvesting, and efficient asset location is so mind-numbing it makes your teeth hurt, then it's a no-brainer: find a good financial advisor who will! Otherwise, you're just lighting money on fire -- money that could be compounding over time.

I'm interested in doing this myself, but I can't find the time. See above. Like I just said, this is money that could be compounding over time -- don't put it off until some later date, when you magically find more free time!

I'm interested in doing this myself, and I've got nothing better to do. Go at it! While good investment managers are worth much more than their fees, they're not free -- if you're looking for a way to make money long term, there are worse things than being your own investment manager. Just make sure to do your research; a few tenths of a percent in optimization can easily be worth tens of thousands of dollars over the long run!

(And yes, Seaborn is all about this kind of optimization. Here's the tech we use to make it happen.)

Britton is an engineer-turned-financial-planner in Austin, Texas. As such, he shies away from suits and commissions, and instead tends towards blue jeans, data-driven analysis, and a fee-only approach to financial planning.