Author Topic: What to Invest in for Tax Loss Harvesting?  (Read 2466 times)

bb11

  • Pencil Stache
  • ****
  • Posts: 503
What to Invest in for Tax Loss Harvesting?
« on: August 17, 2018, 10:24:09 AM »
My taxable income is in the marginal 24% federal bracket + 10% state/local (NYC), so I'd really like to use TLH to offset capital gains and hopefully the $3k allowed in regular income. But how are people doing it, what do you invest in?

I've just been using a 3-fund portfolio, and only 2 funds in my taxable account (US Total Market Equity, International Total Market Equity). The problem with this is that there are rarely an opportunity to harvest losses in these funds. If I literally sold my entire International taxable holding I could harvest about $800 in these losses, but that's probably not ideal.

My thought has been I could:

Use a small (10% or less) portion of my portfolio to invest in individual stocks. I would not necessarily even need to evaluate them (if the Efficient Market Hypothesis is true my evaluations would be a waste of time). I would just hold a stock in each sector for example. Given there is much more variance in an individual stock I should be able to harvest considerably more losses. Then any stock I sold I would just use the proceeds to buy another company in the same industry, keeping my diversification without triggering a wash sale.

Example:

Market return: 10%
Current capital gains/dividends income taxed at 15%: $2k
Current marginal income tax rate: 34%
Current tax incurred on $2k income + $3k in allowed harvesting potential: $1,320

Using individual stocks (the six stocks below average a 10% return):
 
Stock A return: 40%
Stock B return: 25%
Stock C return: 10%
Stock D return: 5%
Stock E return: - 5%
Stock F return: - 15%

I sell Stock E and stock F and invest in competitors in their industry, getting the $1,320 in reduced taxes. However, in order to cancel out my $2k in capital income and $3k in regular income I'd need to have $45k in stocks E and F alone, which is the biggest problem I see with this method. My current portfolio is ~$300k so I'd want to keep the individual stock investment to $30k or less. However, as my portfolio grows this becomes more and more doable, and my capital income will keep rising so the benefits grow as well.

How are others TLH'ing? I see nothing when I search for ways to do it other than basic info. It seems difficult to do with just index funds. I could see slicing and dicing to hold a bunch of different index fund sectors, though I'm not sure if they would have enough volatility. Perhaps I am letting the tax tail wag the dog, and adding too much risk for the benefit.

Systems101

  • Stubble
  • **
  • Posts: 218
Re: What to Invest in for Tax Loss Harvesting?
« Reply #1 on: August 17, 2018, 11:19:48 AM »
First, I think you're missing a bit of the point.  If you are capturing losses and swapping between stocks, then on average, you actually HAVE lost money.  That's really backwards from what you are trying to achieve.  You'd be better off just giving $3K a year to a charity if you want to reduce your taxable income.  It will take less time and produce better outcomes.

In my ETFs, I am rarely harvesting.  I don't see this as a problem.  I avoid individual stocks at this point - the statistics are very much against using them.

I think it's a matter of scale and timing.  Do you have $100Ks of assets?  That's where you will have enough to see and capture losses without losing money to transaction risk.  Also, are you purchasing every week or month so you have ongoing purchases?  Also, how long have you been investing?  The market for the last few years has been harder to capture losses.  Then again, a situation like 2008 can provide you years of losses to work off...

The last time loss harvesting was useful in my mind was early 2016 (the flat-ish market in late 2014 through 2015 produced lots of purchases in the same price range, well above the early 2016 low).  The drop in 2018 was off a swift rise, so there was very little purchased above where the market fell... so part of the lesson is that harvesting is a rarer event.  My international fund also has a bit of a loss, but also in the hundreds; not anywhere near enough to worry about capturing.

Then again, I captured $10K+ in losses in 2016, so I have time left to work that off...

bb11

  • Pencil Stache
  • ****
  • Posts: 503
Re: What to Invest in for Tax Loss Harvesting?
« Reply #2 on: August 17, 2018, 11:54:13 AM »
Quote
First, I think you're missing a bit of the point.  If you are capturing losses and swapping between stocks, then on average, you actually HAVE lost money.  That's really backwards from what you are trying to achieve.  You'd be better off just giving $3K a year to a charity if you want to reduce your taxable income.  It will take less time and produce better outcomes.
Nope, I wouldn't have lost money. Check my example again. The example 6 stock basket I created achieves the +10% market return, but because I exchange the losers for comparable stocks I get subsidized on those "losses" via lower taxes. In other words, I achieve the same market return I get in index funds but get $1,320 in tax savings.

Quote
In my ETFs, I am rarely harvesting.  I don't see this as a problem.  I avoid individual stocks at this point - the statistics are very much against using them.
It's only a problem in that we're leaving money on the table by not harvesting. As for statistics, I agree that market timing and individual stock picking is problematic for most people. However, my method was more of a de-facto mini-index. I am not planning on using market timing, incurring fees, or MOST of the other problems with individual stocks. The one remaining one would be some concentration of risk, which I would plan to combat by making this a small percentage of my portfolio.

Quote
I think it's a matter of scale and timing.  Do you have $100Ks of assets?  That's where you will have enough to see and capture losses without losing money to transaction risk.  Also, are you purchasing every week or month so you have ongoing purchases?  Also, how long have you been investing?
Agreed on the scale part. I have about $300k and started in 2012, but again we're not talking about market timing or using a majority of my portfolio on this. I can perhaps see the logic that I should wait until my portfolio is even larger before taking advantage of TLH, but I am trying to think this through.

Systems101

  • Stubble
  • **
  • Posts: 218
Re: What to Invest in for Tax Loss Harvesting?
« Reply #3 on: August 17, 2018, 01:27:14 PM »
Check my example again.

I did, so I can be more specific in my criticism.

However, my method was more of a de-facto mini-index.

That's what you call it, but that isn't what it is.  You are massively underestimating the risk of concentration.  6 stocks isn't enough.  See:
https://www.investopedia.com/articles/stocks/11/illusion-of-diversification.asp

Even if you could do it with 6 stocks, your distribution of returns is incorrect.  How is the model of 6 stocks accounting for skew and kurtosis of the distribution of returns on individual stocks?  In less formal words, how are you accounting for the challenge highlighted here (and in the previous article):
https://www.economist.com/finance-and-economics/2018/06/23/most-stockmarket-returns-come-from-a-tiny-fraction-of-shares

The one remaining one would be some concentration of risk, which I would plan to combat by making this a small percentage of my portfolio.

Reducing the % of portfolio doesn't make it any less of a losing strategy in the long term.  Losing money to save on taxes is still a losing strategy.


bb11

  • Pencil Stache
  • ****
  • Posts: 503
Re: What to Invest in for Tax Loss Harvesting?
« Reply #4 on: August 17, 2018, 02:02:09 PM »
Quote
You are massively underestimating the risk of concentration.  6 stocks isn't enough.
To be clear, I was using the number six to provide a simple example of what I could do. There's no reason the number has to be just six. In fact, making smaller investments in more stocks would probably make this easier to pull off, not more difficult. With 30 $1k investments there would be plenty of variation in returns to TLH. The problem is fees, but I think the RobinHood platform would work for this. There are no fees for equity trades nor account minimums.

Quote
How is the model of 6 stocks accounting for skew and kurtosis of the distribution of returns on individual stocks?  In less formal words, how are you accounting for the challenge highlighted here (and in the previous article):
See above. 6 may not be enough, but even your linked article shows 15 stocks gets 75% of the benefit of total diversification. In my example Stock A accounts for nearly all of the stock market returns, but even excluding this stock the strategy still comes out to about breakeven because of the decrease in taxes.

Quote
Reducing the % of portfolio doesn't make it any less of a losing strategy in the long term.  Losing money to save on taxes is still a losing strategy.
Can you explain how I would be losing money? I do not see that at all. The portfolio would gain 10% annually, I'd just hold on to the winners so they don't incur any tax liability. The losers would be swapped out for comparable stocks.

seattlecyclone

  • Walrus Stache
  • *******
  • Posts: 7254
  • Age: 39
  • Location: Seattle, WA
    • My blog
Re: What to Invest in for Tax Loss Harvesting?
« Reply #5 on: August 19, 2018, 08:17:54 PM »
I think your basic premise is solid: that by investing in a variety of individual stocks you will have more frequent opportunities to harvest losses than if you invest in a single index fund that holds all of those stocks (and more). This is exactly what at least one of the big robo-advisors does. Instead of investing in an S&P 500 fund, they'll buy you all 500 stocks individually and harvest losses.

Trying to do it yourself, a complicating factor might be balancing diversification against trading commissions. I agree with Systems101 that six stocks is likely too few to be reasonably well correlated with an index fund; if I were doing it I'd probably pick a few dozen. Unless you have a good source of free trades lined up, the cost to set this system up could eat up a good portion of the benefit from increased loss harvesting.

bb11

  • Pencil Stache
  • ****
  • Posts: 503
Re: What to Invest in for Tax Loss Harvesting?
« Reply #6 on: August 20, 2018, 01:50:55 PM »
I think your basic premise is solid: that by investing in a variety of individual stocks you will have more frequent opportunities to harvest losses than if you invest in a single index fund that holds all of those stocks (and more). This is exactly what at least one of the big robo-advisors does. Instead of investing in an S&P 500 fund, they'll buy you all 500 stocks individually and harvest losses.

Trying to do it yourself, a complicating factor might be balancing diversification against trading commissions. I agree with Systems101 that six stocks is likely too few to be reasonably well correlated with an index fund; if I were doing it I'd probably pick a few dozen. Unless you have a good source of free trades lined up, the cost to set this system up could eat up a good portion of the benefit from increased loss harvesting.
Yep, definitely. And I have not tried it, but Robinhood seems to be a good service to do so. They don't have any trading fees (other than the SEC tax on selling securities, but that works out to about one penny per $10k traded), so as far as I can tell there shouldn't be any issue in pulling this off. Both of you are right I should probably use 25-30 rather than 6 stocks.

Are there any other concerns/issues I should be aware of?

Systems101

  • Stubble
  • **
  • Posts: 218
Re: What to Invest in for Tax Loss Harvesting?
« Reply #7 on: August 20, 2018, 04:30:38 PM »
I think your basic premise is solid: that by investing in a variety of individual stocks you will have more frequent opportunities to harvest losses than if you invest in a single index fund that holds all of those stocks (and more).

The premise of individual stocks and being able to capture losses is solid, but the problem is the gap between theory and practice.  The two root cause problems are: Not using all 500 stocks and transaction fees.

This is exactly what at least one of the big robo-advisors does. Instead of investing in an S&P 500 fund, they'll buy you all 500 stocks individually and harvest losses.

Yes, there are a few ways to get this, mostly at scale (multiple millions of dollars).  The problem is that below that point, you get swamped by transaction fees. 

Trying to do it yourself, a complicating factor might be balancing diversification against trading commissions.... Unless you have a good source of free trades lined up, the cost to set this system up could eat up a good portion of the benefit from increased loss harvesting.

It will certainly either (1) eat MORE THAN the benefit or (2) have sufficient tracking error that you have exposed yourself to risk of capital loss. 

To the cost of a trade, please be aware, no trades are free.  You may pay no commission, but there are ways for Robinhood to be paid for the transaction anyway.  See:
https://www.reddit.com/r/RobinHood/comments/7sxxbj/robinhood_paid_by_routing_companies_which_enable/

The net result is that you pay a bit more for each stock you purchase and a bit less on each stock you sell.  This increased bid-ask spread will eat up just as much as the commission otherwise would have... There is no such thing as a free lunch.

Both of you are right I should probably use 25-30 rather than 6 stocks.

Actually - just so we are clear - my recommendation is "Don't track the S&P 500 with individual stocks it unless you can afford to buy all 500 stocks (in the right ratio), which is practical at about $20M in assets".  The $20M is calculated from looking at the stocks and where a $5 loss is into fractions of a percent of the amount being traded.  This is about where the investment level is needed to get 100 shares of the least-share quantity stock... from a quick glance I used stock #501, ticker BHF

Are there any other concerns/issues I should be aware of?

To be honest, I'm concerned you don't understand the issues that have already been raised.  Since I've already stated my concerns, let me try it a different direction.  If you were trying to convince me to do this here are the questions I would ask:

1) What will I lose to transaction costs (this is more than commission and SEC fees - see above for "It's not free")
2) You've stated 30 stocks has an R squared of 0.75.  Of the tracking error that occurs, what is the distribution of the tracking error?  Is the distribution random? Skewed?  What is the likelihood I will lose vs gain vs the index?
3) Wealthfront starts at $100K before it will buy individual stocks in this methodology, and even then is using an ETF for the smaller companies.  How much are you putting into this strategy, and if it's not well over $100K, why do you believe your strategy is superior to theirs (given they have more money and have done the research)
4) From the investopedia article I linked in an earlier post, it notes that "64% of stocks underperformed the Russell 3000" and "25% of stocks were responsible for all of the market's gains".  You state you are not stock picking, but how are you going to ensure the 30 stocks you pick will select a sufficient number of the winners in order to track the index in any meaningful way?  (Said another way - if you are randomly selecting stocks, then what are the implications of the skew and kurtosis of stock returns to the implication of an incorrect random selection?)
5) How will you ensure your selection doesn't include Bear Stearns or Lehman or Enron or GM or another company that goes bankrupt?  What is the impact to a concentrated portfolio of 30 stocks if one goes bankrupt?  How long will it take to recover? (Can it ever recover?)
6) What is the differential IRR you expect to produce?  See: https://research.wealthfront.com/whitepapers/stock-level-tax-loss-harvesting/

bb11

  • Pencil Stache
  • ****
  • Posts: 503
Re: What to Invest in for Tax Loss Harvesting?
« Reply #8 on: August 20, 2018, 09:21:30 PM »
Quote
To the cost of a trade, please be aware, no trades are free.  You may pay no commission, but there are ways for Robinhood to be paid for the transaction anyway.  See:
https://www.reddit.com/r/RobinHood/comments/7sxxbj/robinhood_paid_by_routing_companies_which_enable/
That still makes the trade free to me.

Quote
The net result is that you pay a bit more for each stock you purchase and a bit less on each stock you sell.  This increased bid-ask spread will eat up just as much as the commission otherwise would have... There is no such thing as a free lunch.
Source? The link you gave makes the bolded seem highly unlikely.

And yes, a "free lunch" can occur. Facebook is free, because they make money on the ads. Fidelity just started offering free index funds, because they have many ancillary services. Loss leaders are huge in retail, etc. I disagree with the premise that the commission is just hidden based on the evidence you gave.

Quote
1) What will I lose to transaction costs (this is more than commission and SEC fees - see above for "It's not free")
From my calculations, about $1-2 per year.

Quote
2) You've stated 30 stocks has an R squared of 0.75.  Of the tracking error that occurs, what is the distribution of the tracking error?  Is the distribution random? Skewed?  What is the likelihood I will lose vs gain vs the index?
I said 15 stocks has an R squared of .75, according to your link. 30 has an R2 of 0.86. The distribution is random for larger US equities, which is what these stocks would be replacing.

Quote
3) Wealthfront starts at $100K before it will buy individual stocks in this methodology, and even then is using an ETF for the smaller companies.  How much are you putting into this strategy, and if it's not well over $100K, why do you believe your strategy is superior to theirs (given they have more money and have done the research)
I'm not claiming it's superior, I'm just avoiding the fees.

Quote
4) From the investopedia article I linked in an earlier post, it notes that "64% of stocks underperformed the Russell 3000" and "25% of stocks were responsible for all of the market's gains".  You state you are not stock picking, but how are you going to ensure the 30 stocks you pick will select a sufficient number of the winners in order to track the index in any meaningful way?  (Said another way - if you are randomly selecting stocks, then what are the implications of the skew and kurtosis of stock returns to the implication of an incorrect random selection?)
I've already answered this. It's possible this part of my portfolio gets no returns (though very unlikely if you look at the numbers). That would still not be devastating given the tax benefits. It would be a loss compared to holding a US Equity index, but not by a great amount.

Quote
5) How will you ensure your selection doesn't include Bear Stearns or Lehman or Enron or GM or another company that goes bankrupt?  What is the impact to a concentrated portfolio of 30 stocks if one goes bankrupt?  How long will it take to recover? (Can it ever recover?)
You're making me think you're not understanding my strategy or I'm not explaining it well or something. First you thought I was talking about losing money, now you're talking about a devastating loss from one company. This entire individual stock part of my portfolio would be capped at say 10% of my total portfolio (could be capped at less if there was reason too). Each stock could be no more than one third of one percent of my total portfolio (1k out of 300k). The impact of picking a company that went bankrupt would be inconsequential. It's a tiny tiny slice of my overall equity investments.

Systems101

  • Stubble
  • **
  • Posts: 218
Re: What to Invest in for Tax Loss Harvesting?
« Reply #9 on: August 21, 2018, 12:12:34 PM »
Overall, I think you are being taken in by perceptions and surface analysis rather than actually doing deep, hard analysis.

Source? The link you gave makes the bolded seem highly unlikely.

The problem is that HFT makes it *appear* as if there is a narrow bid-ask spread.  However, since the spread is being narrowed solely by the HFT trades, the *effective* bid-ask spread that an actual consumer sees (especially one that doesn't do limit orders - and Robinhood doesn't) is wider than it would have been without HFT.

It's not only likely that this is occurring, it's real and extremely lucrative.  Go read "Flash Boys" for how HFT works and how it eats away at investor profits during a transaction.

The distribution is random for larger US equities, which is what these stocks would be replacing.

Source?  In reality, the distribution is not normal. The last part of one of the earlier links showed this.  The actual distribution is skewed with significant kurtosis.  As a result of assuming it's random, the rest of your analysis is flawed.  If the distribution was random and normal, then the answer I'd be giving you would be different. 

It would be a loss compared to holding a US Equity index, but not by a great amount.

I don't know how you are supporting "but by not a great amount".  Have you backtested that claim using actual random selection of stocks?

Also, if you're losing relative to being in an index, why do it?  You still haven't shown with any real data that overall you will gain more than you lose.  IMHO, the strategy is penny wise (to capture losses) and pound foolish (taking losses you don't need to take).

You're making me think you're not understanding my strategy or I'm not explaining it well or something.

Fair enough.  I'm concerned you have based a decision on a simplified model, and continue to make dangerous assumptions.  You should be searching for contrary evidence, but when I provide a view of what to look for in your logic, you are pushing back with assumptions and marketing content, not real data.

The Fidelity example you give is because they are making money on loaning stock (for example).  Other companies that are more transparent with fees (Vanguard) are giving those gains back to the holders of the mutual fund/ETF.  You think it's free, because it appears that way in absolute terms.  However, it's really lost opportunity... therefore, RELATIVE to the other options, it is very much not free... So Fidelity's marketing has been successful at deceiving you as to the best choice for your money.

Even if it's 10% of your portfolio, if you lose a portion of it and have to backfill it from the other 90%, you are losing more money than you would ever gain in tax benefit.  That's what makes no sense to me... but it's your money... the tools are in this thread to educate yourself and to build backtests and more complex models.  It's up to you whether you want to do that.

bb11

  • Pencil Stache
  • ****
  • Posts: 503
Re: What to Invest in for Tax Loss Harvesting?
« Reply #10 on: August 21, 2018, 01:49:46 PM »
First of all, thanks for your feedback. I appreciate the discussion.

Quote
The problem is that HFT makes it *appear* as if there is a narrow bid-ask spread.  However, since the spread is being narrowed solely by the HFT trades, the *effective* bid-ask spread that an actual consumer sees (especially one that doesn't do limit orders - and Robinhood doesn't) is wider than it would have been without HFT.

It's not only likely that this is occurring, it's real and extremely lucrative.  Go read "Flash Boys" for how HFT works and how it eats away at investor profits during a transaction.
Do you know of a way to quantify the estimated cost here? I see what you're saying, but I need to know the actual estimated cost rather than just that the bid-ask spread might be larger.

Quote
(especially one that doesn't do limit orders - and Robinhood doesn't)
This is incorrect btw. They do support limit orders.

Quote
Source?  In reality, the distribution is not normal. The last part of one of the earlier links showed this.  The actual distribution is skewed with significant kurtosis.  As a result of assuming it's random, the rest of your analysis is flawed.  If the distribution was random and normal, then the answer I'd be giving you would be different.
You keep using these terms. Can you please explain why you believe the sample would be skewed with significant kurtosis? I see no reason to believe that would be true. I can't access the Economist article if it's the one you're referencing, as it's behind a paywall.

Quote
I don't know how you are supporting "but by not a great amount".  Have you backtested that claim using actual random selection of stocks?

Also, if you're losing relative to being in an index, why do it?  You still haven't shown with any real data that overall you will gain more than you lose.  IMHO, the strategy is penny wise (to capture losses) and pound foolish (taking losses you don't need to take).
This statement makes me believe you're still not understanding the strategy. Let me try explaining the strategy with a (possibly) clearer example:

Current allocation:
$300k in index funds
Market Return: 10%
2019 portfolio value: $330k

Updated allocation: $270k in index funds, $30k split between ~25-30 individual stocks
Portfolio return: 10%
Tax loss harvesting: Because there will be a much wider distribution of returns in the individual stocks compared to the index fund, I can sell the worst performing stocks (while still maintaining the 10% market return across the 25-30 stocks!) to reduce taxes
2019 portfolio value: $330k +$1k-1.5k in annual tax reduction

You seem to be thinking the money invested in individual stocks must be a net loss in order to tax loss harvest. That is not at all the case. In fact, the individual stocks could greatly exceed the market return (say reaching a 20% CAGR) and I could still sell the few losers at a loss to reduce taxes and benefit from the strategy. Of course this is unlikely, but the key is I am not counting on this. Simply having the stocks anywhere near the market return means the strategy will be profitable, because the tax benefit represents a 3.3-5% return on the $30k invested on it's own.
« Last Edit: August 21, 2018, 02:13:17 PM by bb11 »

Rocketman

  • Stubble
  • **
  • Posts: 118
Re: What to Invest in for Tax Loss Harvesting?
« Reply #11 on: August 21, 2018, 03:05:20 PM »
If you really want to tax loss harvest - signup with Personal Capital. The fees are higher than vanguard or fidelity, but for domestic stocks they buy you the shares of individual companies. Tax Loss harvesting is part of their plan.

bb11

  • Pencil Stache
  • ****
  • Posts: 503
Re: What to Invest in for Tax Loss Harvesting?
« Reply #12 on: August 24, 2018, 08:49:12 AM »
Thanks Rocketman. Anyone else? I was hoping to get some more feedback.

PizzaSteve

  • Pencil Stache
  • ****
  • Posts: 501
Re: What to Invest in for Tax Loss Harvesting?
« Reply #13 on: August 24, 2018, 09:09:39 AM »
Tax loss harvesting is over emphasized as a strategy by robo advisory services that generally are looking for reasons someone would want to use them, when actually a portfolio of low cost index funds is such an efficient investment tool, these robo strategies add very little value, and may actually increase fees that drag returns.

Dont fall for the buzz words!

A good, low turnover index fund is already more or less doing this, using hundreds of stocks (plus leveraging ultra low transaction costs).  You will never match their efficiency.  These funds are buying thousands of stocks in millions of lots.   Some will fail to gain in price, but these losses are already automatically calculated in your Net Asset Value, reducing your net gains if you sell.  So your index fund is effectively harvesting losses already, over the whole market.   When you hold the fund, some capital gains are realized when fund holders net distribute funds or corporate actions force realization of a gain (eg buyouts and mergers force a fund to realize deferred gains and pass them onto fund holders as a taxable distribution), but this happens with individual stocks as well.  A forced capital gain is a common issue with individual stock and bond holders.

Sure, if you have the bad luck to start investing in a down year for the market, you might benefit from a few harvests of these losses, but that is likely for only one or two transactions.  Eventually you will have accumulated gains and few losses, unless you take the risky path of concentrating on a few stocks.

Remember, the goal of investing is not minimizing taxes, it is building wealth.  When someone is focusing you on minimizing taxes it is a distraction.  They are looking to sell you something complex, like an annuity or trading system to generate fees.

Having to pay taxes is a good thing.  It means you are making money.  Yes, we will all have huge capital gains to pay some day when we fire.  The only way to avoid this is to fail with your investment plan, or to advocate an overturning of the tax code.  Trying to be clever to avoid taxes will often increase chances of a FIRE fail and in my humble opinion should only be simple strategies, such as selling gains when a favorable tax bracket is available.

Believe me, I have harvested plenty of losses over my 40 years of investing.  I wish I had bought efficient funds with those dollars and had long term gains to realize, and taxes to pay, instead.
« Last Edit: August 24, 2018, 09:30:10 AM by PizzaSteve »

Systems101

  • Stubble
  • **
  • Posts: 218
Re: What to Invest in for Tax Loss Harvesting?
« Reply #14 on: August 24, 2018, 06:34:38 PM »
Do you know of a way to quantify the estimated cost here? I see what you're saying, but I need to know the actual estimated cost rather than just that the bid-ask spread might be larger.

I don't know of a way to estimate it given how private HFT folks are about their profits; but if you assume it's on the order of $0.01 per share per trade it will give you the potential risk.

You keep using these terms. Can you please explain why you believe the sample would be skewed with significant kurtosis? I see no reason to believe that would be true. I can't access the Economist article if it's the one you're referencing, as it's behind a paywall.

Sorry about the paywall.  I found it through Google and it didn't paywall me at that moment.

I'm going to focus on skew because the impact is better studied than the Kurtosis effects (or at least my Google-fu can find an article that talks about skew :) )

To be clear, I'm saying the sample would be skewed because if you randomly selected stocks, it is just randomly sampling from the underlying data (which is skewed).  So if you did a histogram of the return of stocks, it would not be a normal distribution.  It would be skewed.

This matters, and will cause you - statistically - to underperform the overall market.  I finally found a good article with both a histogram and a good explanation of the issue of what skew will do to your returns.  There are further links at the end of this that are interesting:
https://www.wallstreetphysician.com/market-skew-another-cause-investor-underperformance/

This statement makes me believe you're still not understanding the strategy. Let me try explaining the strategy with a (possibly) clearer example:

Current allocation:
$300k in index funds
Market Return: 10%
2019 portfolio value: $330k

Updated allocation: $270k in index funds, $30k split between ~25-30 individual stocks
Portfolio return: 10%
Tax loss harvesting: ...

You seem to be thinking the money invested in individual stocks must be a net loss in order to tax loss harvest.

No, I am saying your claim (I put it in bold above) of getting 10% portfolio return in the sampling / subset of stocks is a false premise that colors the rest of your analysis.  IF you can generate that, the rest holds true.  The problems are:
  • You have fees (even if hidden)
  • You have sampling problems due to skew [but your model assumes you will get the average return]
  • Your model only considers one time period and you haven't extended the model to consider the implications of additional time periods

Again on skew:

Consider this: “When stated in terms of lifetime dollar wealth creation, the best-performing 4% of listed companies explain the net gain for the entire US stock market since 1926, as other stocks collectively matched Treasury bills” From: https://www.sciencedirect.com/science/article/pii/S0304405X18301521

If you sample the market with even 30 stocks, there is a decent chance you select NONE of those best performing stocks.  If you don't get enough of them, you could underperform even US Treasury Bills, much less the stock market.

This again comes back to how you are interpreting the statistics.  The subset of stocks might get you 75% or whatever correlation.  But if the remaining 25% is all downside, you will lag the market more than you will ever make in tax loss harvesting.

On time periods:

Eventually you will have accumulated gains and few losses, unless you take the risky path of concentrating on a few stocks.

^^This.  I was so focused on the skew issue I missed this bigger problem.  Hat tip to @PizzaSteve for catching it.

In your model, you have another problem: Your model considers one time period.  If you have stocks that return what you proposed earlier, then you can sell one at a loss at the end of year one, but then what?
  • Will you redistribute value out of the large winner?  That produces capital gains that eat up the loss.
  • Will you hold the winners?  If you hold, then in the following year, will the stock that loses money be one where it just goes back to where you purchased it, so there is no loss to capture?
  • Will you buy more of all the smaller winners and replace the loser with lots of a new stock?  Do you have the cash to do that?  How many years can you keep up the total growth without having to sell a winner?

AccidentalMiser

  • Pencil Stache
  • ****
  • Posts: 704
  • Age: 56
  • Location: SE Tenn
Re: What to Invest in for Tax Loss Harvesting?
« Reply #15 on: August 24, 2018, 06:42:16 PM »
I think this would make an interesting research project but I wouldn't trade this way.  Too much risk of things going awry due to fees and inconsistent execution in practice.

PizzaSteve

  • Pencil Stache
  • ****
  • Posts: 501
Re: What to Invest in for Tax Loss Harvesting?
« Reply #16 on: August 26, 2018, 12:45:12 AM »
One added point.  Studies show that most index gains are from a few huge winners.  Gains are not evenly distributed.
« Last Edit: August 26, 2018, 12:49:37 AM by PizzaSteve »

MustacheAndaHalf

  • Walrus Stache
  • *******
  • Posts: 6633
Re: What to Invest in for Tax Loss Harvesting?
« Reply #17 on: August 27, 2018, 12:06:21 AM »
There's a flaw with tax loss harvesting that is only apparent years later: stocks go up more than down, so eventually tax loss harvesting doesn't work on old investments - only recent ones.

For example, if you bought the S&P 500 a decade ago with $1000, you'd have $2750 now.  To tax loss harvest that investment, the market would need to lose more than -63%.  The 2008 crisis didn't even do that - only the great depression did.  So over time, tax loss harvesting becomes less and less likely.

If you sign in to Personal Capital another service that charges an additional fee, you might get some tax loss harvesting that first year or two.  But over time, you'll still pay the added fee but won't get a benefit.

Also, the biggest chances to tax loss harvest are probably events you'll see on the news.  Most brokerages offer tracking your sales in "lots", where you can decide which sale paired up with which buy.

One

  • Stubble
  • **
  • Posts: 247
Re: What to Invest in for Tax Loss Harvesting?
« Reply #18 on: August 27, 2018, 01:18:26 PM »
The advantage to index is you don't have to sell.  With the individual stocks you will want to occasionally sell the winners which will incur taxes.  I think holding some individual stocks is a smart move but doing it to save on taxes probably won't work out.