When stocks are down, many articles come out trying to calm people down. They typically say “don’t sell” or “do nothing.” Here’s a recent one from James Osborne at Bason Asset:
Here’s what you should probably do when stocks go down: nothing. Boring advice, I know. But usually you should do nothing. Sometimes there’s an opportunity to take some tax losses. Sometimes it will warrant rebalancing (though rarely upon a 10% correction, depending on your rebalancing rules). Most of the time you’re going to do nothing. We’re not good at doing nothing (more on that later) but give it a try. Go outside or read a good book and tell yourself “This Is What Stocks Do,” and do nothing.
Link: Four (Or Is It Six?) Years In The Making
In addition to doing nothing, it mentioned two actions you can take sometimes: tax loss harvesting and rebalancing. Let’s take a look at each one. Are they better than literally doing nothing? If you have a psychological budget to do only one of the two, which one should you do?
Tax Loss Harvesting
Tax loss harvesting refers to realizing some losses for tax purpose. You sell low, but you put the proceeds into a similar fund. The tax loss you realize can be used to offset your realized capital gains, if any. If your realized losses are greater than your realized gains, you can offset up to $3,000 of your income. If you still have more losses, the unused losses can be banked and carried over to next year. They don’t expire until you die.
Because you are not taxed in tax advantaged accounts such as 401k or IRA, tax loss harvesting only applies if you have a regular taxable account. Because only recent purchases before the market drop have losses, your selling and generating losses is limited to those recent purchases.
I have shares I bought in 2011 in my taxable account. Even though they lost value in the recent downturn, they are still worth 50% more than my cost. Chances are slim that I will ever be able to do tax loss harvesting on those shares.
Most people have more money in tax advantaged accounts than in taxable accounts. If you have been investing for some time, your recent purchases make up only a small percentage of your account. Suppose 30% of your portfolio is in taxable accounts. Suppose your recent purchases make up 10% of that. After a 10% market correction, you have 30% * 10% * 10% = 0.3% of your portfolio in losses. Say harvesting that 0.3% in losses will save you 15% of the losses in taxes. Performing tax loss harvesting will then save you 0.3% * 15% = 0.045% of your portfolio in taxes. In finance lingo, you will have a one-time tax saving worth 4.5 basis points of your portfolio.
Over time, your recent purchases will make up less and less of your total portfolio. As a result, the benefits from tax loss harvesting as a percentage of your portfolio will become smaller and smaller.
Rebalancing refers to shifting your investments to realign your ratio of stocks and bonds to your original asset allocation. After a drop in the stock market, it usually means selling bonds and buying stocks.
Rebalancing can be performed in both tax advantaged accounts and taxable accounts. It’s actually better done in tax advantaged accounts because there are no tax consequences there.
Let’s also run a back-of-the-envelope example.
Suppose you start with $100k, $70k in stocks and $30k in bonds. After stocks drop 10%, suppose bonds are flat, you now have $63k in stocks and still $30k in bonds. If you rebalance back to 70% in stocks, you will sell $30k – $93k * 30% = $2,100 from bonds and buy $2,100 in stocks. After rebalancing, now you have $65,100 in stocks and $27,900 in bonds, still $93k total.
When stocks eventually recover to the same level as before, suppose bonds are still flat, you will have 65,100 / 0.9 = $72,333 in stocks and $27,900 in bonds, for a total of $100,233. If you literally did nothing the entire time, you will be back at $100k. The extra $233 if you rebalanced came from buying $2,100 in stocks when stocks were 10% lower.
The benefits from rebalancing is 0.23%, or 23 basis points, of your portfolio. It’s not much but it’s several times larger than the tax saving from tax loss harvesting. Because rebalancing isn’t limited to taxable accounts or recent purchases, its benefits do not decay over time.
If stocks continue dropping beyond the 10% correction, rebalancing at a later time may produce a higher gain when stocks recover. Still, as long as stocks beat bonds on its way to recovery, rebalancing when stocks are lower will always be better than literally doing nothing.
Nothing holds you from doing both at the same time. Sell your shares that have a loss and buy a similar but different fund. Sell some from bonds and put the money into stocks. Either is better than literally doing nothing. You stack the benefits if you do both.
If psychologically you can only handle doing one of the two, then I would prioritize rebalancing over tax loss harvesting. Tax loss harvesting feels clever as you are beating the tax man, but rebalancing is much more powerful.
I created a one-question poll to see what you did in the recent downturn. Please take 10 seconds to answer it. You can also see what others did in that poll.
Say No To Management Fees
If you are paying an advisor a percentage of your assets, you are paying 5-10x too much. Learn how to find an independent advisor, pay for advice, and only the advice.
I think your assumptions on tax loss harvesting hold true only in certain circumstances. Especially with large drops eg 2008/09 potentially 12 years of losses could have been harvested in one year since the stock market fell below 1996-2008 prices. The magnitude of the drop also impacts the benefit of TLH. Investments in 2007 could have dropped by 50% giving you 5x the benefit over your 10% drop example.
It also depends on what investments you hold. The total international market is currently at a price lower than periods within 2011, 2013, 2014 and 2015. If you invested at a higher price in each of these years then potentially that could be 4 years worth of TLH in this year.
Though I doubt it is better than rebalancing it certainly can bring significant benefits.
Harry Sit says
2008-2009 already happened. You were able to harvest loss from the previous 12 years but you would have to go back maybe more than 30 years to find something of that magnitude. It likely won’t happen again in the next 12 years. When you average it out, on a per-year basis, the effect of tax loss harvesting is still very small. The conclusion still holds true. Rebalancing in 2008-2009 would have a much larger effect than tax loss harvesting.
2008 was a big opportunity to tax loss harvest and re balance. Most of the time, the advantages are small.
But , booking a large capital loss back in 2008, puts a smile on my face every April.
There is also a question of when to rebalance. Do you recommend rebalancing on a periodic schedule or is it better when certain tolerance bands are breached?
Harry Sit says
I use tolerance bands. I don’t know whether it’s better or not. People write papers and develop software trying to optimize it. You can optimize for the past but you can’t be sure it will be better in the future. Tolerance bands make more sense to me.
Excellent analysis. I’ve always thought that tax loss harvesting is largely a gimmick and that rebalancing is not going to make or break a portfolio, though better than doing nothing. Thanks for working through the numbers to confirm this.
It seems like taxes incurred on rebalancing within a taxable account could quickly eat away the small benefit. Using your example, it seems like it wouldn’t be difficult to incur a $233 tax bill on gains realized. Rebalancing may be best limited to a retirement account.
Also, in many cases, the benefit of tax loss harvesting is simply a deferral. The investment purchased to replace the harvested investment will have a commensurately lower basis and therefore a larger potential capital gain when sold. At the same tax rate, the only benefit is the time value of money on the up front tax savings. Of course, if the new investment is held until death, donated to charity, or taxed at a lower rate upon the eventual sale, then the analysis is different, but it’s difficult to be sure of any of these outcomes at the time the losses are harvested.
> It seems like taxes incurred on rebalancing within a taxable account could quickly eat away the small benefit.
Actually, in this scenario we assume that we sell bonds that stayed flat. It *is* difficult to incur $200 tax bill on $2K sale of bonds — they’d have to have gained 67% (assuming LTCG at 15%).
> At the same tax rate, the only benefit is the time value of money on the up front tax savings.
If one plays it right, one can deduct losses from ordinary income now, but pay only LTCG tax on the same amount later.
Can anyone elaborate on the FIFO (first in, first out) rules that apply to taxable accounts? If, for example, I have a mix of shares with the 2011 vintage still in the money, but losses on more recent purchases; am I able to switch to LIFO (last in, first out). I know for E-trade, I can specify my lots when I sell individual stocks; but how does Vanguard work?
Harry Sit says
Vanguard explains it here: https://investor.vanguard.com/taxes/cost-basis/
You can change your cost basis method: https://investor.vanguard.com/taxes/cost-basis/methods
If they are stocks then in most cases yes you can sell specific lots, but Vanguard has a slightly different procedure to go through, but your best bet is to just call them and tell them what you want to do – they have always been more than helpful.
Still, harvesting a net $3000 in losses to offset ordinary income can save a reasonable $750 (25% tax bracket x $3000), feels good regardless of the size of the portfolio. The fallacy is equating relative impact versus absolute. The real risk in TLH is if cap gains rates rise substantially in the future, so when you sell your investments with a new, lower cost basis, you end up paying more than if the loss was unrealized and the basis remains higher.
Rebalancing benefits can be estimated by the formula:
RB = X1 x X2 x (1/2 Var1 + 1/2Var2 – Covar(1,2) ), where X1 = % stocks, X2 = % bonds.
Depending on your Var and Covar assumptions, the benefit can be about 30bps per year. In a world where we will be lucky to earn 4-6%, 30 extra bps can mean a 5-7.5% proportional boost in total returns.
Do both, TLH and RB.
I agree with you completely on your first point, but you need to apply that critical thinking into point #2 to understand that we can not predict the future returns when rebalancing stocks or bonds, so as long as you concede your error boundary is + or – 100% then I guess I’m good with it.
As Steve points out below rebalancing a stock / bond portfolio is more about managing risk, except for the one point he missed and that is in the withdrawal stage of retirement in which keeping a fixed bond allocation even when stocks go up can improve your sequence of return results when things turn the other way on you later. In other words it’s counter-intuitive that a 60/40 stock/bond portfolio would have a better survival rate than a 90/10 one if your only mindset is that stocks always outperform bonds.
I disagree with your logic on the value of rebalancing. It’s only half-true.
If you rebalance after stocks take a dive, then yes–you’ll increase the return of your portfolio assuming stocks outperform bonds.
But the far more common side of rebalancing has been to sell your overweighted stocks and buy bonds. And, given the history of the stock and bond markets, this actually TAKES AWAY from your annual return (given stocks have outperformed bonds over long periods of time).
I think rebalancing is about managing risk. Not increasing return.
Harry Sit says
The other half of rebalancing is addressed in Rebalancing: Sell High, Buy Higher. This article only addresses the situation when stocks are down.
Fair enough. I stand corrected.
Tax Loss Harvesting benefit is highly situation dependent. I found myself recently with the ability to TLH $120k loss because ALL my international lots were down. I was in this situation because the shares are all from the last couple years partly because of a windfall business sale and partly because of converting to a self managed portfolio.
It was intimidating to have my first real TLH be of such a magnitude but I think it was the right thing to do. I can offset my gains from rebalancing early this year as well as play forward the $3k/yr off ordinary income for quite some time. I won’t really know if this makes sense for 20 years when I see what tax bracket I’m in, but you can only act on what you know now and it seems to make sense.
Harry Sit says
I agree. After taking this big bath, you will have less motivation to harvest more next year.
If one cannot carry forward the losses (at state level for NJ) there is less incentive to TLH more than $3000 (anything more than that you cannot offset, gains or income)
in the above example one saves 15% on federal, but will loose say 5% on state again, thus saving ~10% only.
For the curious, here are the slices showing benefits of “ideal” rebalancing for different stocks/bonds allocations and stocks loss/recovery: https://docs.google.com/spreadsheets/d/1-ygKOdWwrfOqvQB0m8SeEJOTJH9tanzd0LaVZI8GmdQ/edit?usp=sharing. Bottom line: under all but the most extreme scenarios the benefits are surprisingly small.
Harry Sit says
Thank you. Love the polynomial! Good to see that my back-of-envelope 0.23% number matches the spreadsheet.
I think the TLH calculation will be very different for each individual. In my case 70% of my portfolio is in taxable accounts and my combined fed+state bracket is close to 50% tax (!!). So for me the TLH calculation becomes:
70%*10%*10%*50%=0.35%=35bps which is significant tax savings for my portfolio!