Cash in while LTCG rates are at historic lows?

The stock market is close enough to being efficient that one's

*opinions* of where a stock is headed should often be given less weight than tax effects, which are more knowable.

That's not quite correct, depending on how far back your way-back machine can look. From 1913 to 1921, capital gains were taxed at ordinary rates, initially up to a maximum rate of 7 percent. In 1921 the Revenue Act of 1921 was introduced, allowing a tax rate of 12.5 percent gain for assets held at least two years. From 1934 to 1941, taxpayers could exclude percentages of gains that varied with the holding period: 20, 40, 60, and 70 percent of gains were excluded on assets held 1, 2, 5, and 10 years, respectively.

You're right - I missed it, as you put it, maybe in part because Rich Carreiro from previous postings is a sharp guy, and selling to reinvest immediately doesn't make sense if my math is correct. E.g. I think the math is to use the same percentage gain, but in alternative a) reduce that percentage by 15% (an 8% estimate becomes 6.8%). It isn't necessary, if I'm right, to consider the cost basis - only the gain is relevant.

If on the other hand you are planning to re-allocate capital into a substantially different investment (based on fundamentals), then you might consider doing that a year early at 15% as opposed to waiting for a presumed 28%.

Then Elle (of course) can compute the dividends one foregoes by selling sooner.

You are ignoring the 0% cap gain rate. The only cost is the 2 commissions ($20?) plus the bid/ask spread (I was going to say 1/8, but bid/ask is in cents now, so this will change depending on the stock, let's assume 5-10 cents/share?)

A thousand share position sitting on a $5000 gain may cost $120 to churn. That's a 2.5% hit, and is likely on the high side. Of course this is made up. As long as this gain would be taxed in the owner's 10 or

15% bracket, for this year (through 2010) it's zero. Does this change your thoughts on this approach? JOE

Forgive me, but I still think you're missing the point. Rich is suggesting that an investor with an appreciated asset, who plans to continue holding that asset, could sell and rebuy that asset to lock in a capital gain taxed at 15% under the assumption that the capital gain tax rate will increase in the future. This can make sense depending on the difference between the two tax rates and the amount of gain you expect to have past the point where the 15% tax rate is no longer available.

-Will

william dot trice at ngc dot com

Will Trice -

If my numbers are correct THIS time, you caught me being sloppy and posting something mathematically incorrect - significantly incorrect - I tried a shortcut that didn't work. I still find it hard to believe that selling at 15% gives a significant advantage if tax rates subsequently increase. But I have the equations below so anyone can check them.

Two guys, Abe and Bill hold an indentical investment position with

100k long term capital gain. Abe sells in 2008 to pay a 15% tax rate and immediately reinvests the remaining 85k. Bill simply holds onto his existing position. The problem is to determine who ends up with more cash in his pocket at some future date when they both sell. The variables are the rate of return, the future tax rate, and the number of years. Rich Carreiro specified "substantial" gains so I think it's not unreasonable to assume a full 15% and full 28% effective tax rate. I also think an 8% rate of return is reasonable.

Y = number of years Ry = rate of return compounded over Y number of years (it should be a superscript, representing the rate of return to the power of Y) T = future tax rate P = present tax rate

Bill = 100,000 x Ry x (1-T) Abe = (100,000 x (1 - P) x Ry - 100,000 x (1 - P)) x (1 - T) + 85,000

In our example, P = 15%, T = 28%, R = 8%.

Y = 1 years (2009)

Bill = 100k x 1.08 x 72% = 77,760 Abe = (100,000 x 0.85 x 1.08 - 100,000 x 0.85) x 72% + 85,000 (91,800 - 85,000) x 72% + 85,000 = 89,896

In English, for Abe, first we reduce the capital invested by 15% tax paid, then apply the rate of return. To determine the taxable portion of that return, we subtract the established cost basis of 85k, then take out the 28% tax leaving 72% of the new capital gain. Then we add back the 85k cost basis to get final or total realized profit, so that we can compare that total to Bill.. Abe = (85k x Ry - 85k) x 72% +

85k.

Y = 11 years (2019)

Bill = 100,000 x 2.331639 x 72% = 167,878 Abe = (100,000 x 0.85 x 2.331639 - 100,000 x 0.85) x 72% + 85,000 (198,189 - 85,000) x 72% + 85,000 = 166,496

There must be a more elegant formula for this, but I haven't found it yet. I think the above works.

The higher the rate of return, the less time it takes to make up the presumed tax increase, and conversely, the higher the increase in the tax rate, the longer it takes. In the real world, even considering all the variables and assumptions (the tax rate might be decreased, or remain the same, for example), it does appear to make sense to at least consider taking profits at the 15% rate. Clearly if one thinks the market might drop, then realizing gains to establish a higher cost basis would result in a capital loss that could be used to offset other gains, and not simply a reduction of gains if the cost basis remains untouched.

Yes. Your numbers were right, but the search for the pretty equation is probably tough for anyone who hasn't touched algebra in 20+ years. And since a spreadsheet makes it easy to see, I think your point was well made. 8% - breakeven is 10-11 yrs. So this decision is a bet on a) cap gain rate going up (and the exact value of that new rate, one variable) b) market return c) time until stock needs to be sold for good

JOE

the higher the tax increase, the less time it takes to make up the tax increase, no?

In the real world, even considering all

Yep, you've got it right. In the terms you've defined above, the break-even point is when the following expression is true (neglecting trading costs):

Ry = [T(1-P)]/[P(1-T)]

So for your example, if you expect to get more than ~120% further gain before you ultimately sell, then stepping up the gain is bad. If you expect less, stepping up is good. At a constant annual 8% rate of return you'll reach break-even in a little over 10 years.

-Will

william dot trice at ngc dot com

You are *much* too kind. That was Joe that mentioned 20 years since he'd touched algebra - it's been over 30 for me (yikes!) so you should definitely check my results, I've been busted here too often to count.

I thought it more useful to not specify a gain mechanism, but rather a total gain. This way someone can play around with volatility or whatever and not be forced to use a potentially unrealistic constant gain function (although that's probably perfectly fine for a first approximation). If you really care about the derivation shoot me an email at my cleverly encrypted email address in the sig below.

A little over 4 years, but where are you going to get a consistent 20% appreciation?

This is a good point that I hadn't thought of.

-Will

william dot trice at ngc dot com

Will, It would be nice if there were a simple decision-making rule like that but unfortunately it's not even remotely linear. I use software for doing these projections because they're impossible to do any other way. The problem is that the rate isn't really 15%, especially when gains become significant. I didn't really get this until I started doing projections for clients and saw how often you hit strange (higher) marginal rates.

AMT can turn a marginal dollar of AGI into a 21-22% tax rate, by slapping AMT on other income. Whether this happens is unpredictable without knowing all the other specifics of the tax return but in CA it's common. And a bunch of deductions are pegged to AGI so you end up with odd marginal rates when you add a dollar of AGI (or AMTI) from capital gains. Some examples: medical expense deductions, 7.5% AGI floor; misc itemized deductions, 2% AGI floor; exemption & deduction phase-outs; Social Security taxation; student loan interest deduction; PMI deduction. It's uncommon to have all of them, but common to have some of them, come into play.

And you have state taxes to consider, in most states -- with their associated AGI effects. In CA it'd usually be 24.3% in the range where the 15% federal rate applies. That's a lot of "certain" tax to swallow for what is typically a "potential" tax benefit.

Point being I wouldn't rely on that equation to make this decision, you could end up paying substantially higher rates for a marginal $1k of capital gains - more than double, in certain scenarios. The simplest ground rule of tax planning has no equation associated with it: "delay taxes."

-Tad

Re Will's equation below

Tad Borek wrote: [snip]

[snip]

Forgive me, but while I agree about tax rates (I noticed the same phenomenon trying to run models for taxes and mortgage interest deductibility), as far as I can see, once you have comparative taxes rates, the equation itself works to provide a solution to the problem presented. I haven't worked through the math to the equation yet, but inputting numbers into it matches the spreadsheet results I have.

I think the point of the equation is that you have to have accurate numbers to put into it, E.g. if you estimate a current tax rate of

22%, then estimate a future tax rate of perhaps 40% (both after the adjustments for "Area 51 Tax Phenomena" you mentioned), and plug the numbers into the equation, you have a valid comparison.

Tad, We may have to disagree on this. Someone attempting this strategy should have a good handle on current real marginal rates for capital gains including various AGI floors, etc. Given this, this person can make a guess at future marginal rates that will probably be just as good as your fancy sofware projections. After all, the software will have to make assumptions as well. What's worse is that the assumptions may not be explicitly stated limiting the ability of the user to play with the assumptions to test different scenarios.

Of course. And state and other taxes typically makes this strategy less attractive since it is dependent on the ratio of the two tax rates.

Again, we'll have to disagree. This is the same question that comes up here with regards to Roth vs. deductible IRA/401(k) contributions. Sometimes the Roth looks like a better choice, but that fails your "delay taxes" rule.

This equation is perfectly valid, but requires valid inputs. Garbage in, garbage out. But I think it is unlikely that if you make the decision to step up your basis, you'll find yourself in an unexpected tax position in the future that would have cut your taxes in half on the gains to the point where the decision was made as you suggest above. It could happen, though, especially if the "future" is a long ways ahead. But then, with a good length of time, the math works against you anyway. Typically good solutions will require selling relatively soon when tax rates are somewhat more predictable anyway. Besides, the decision to forgo this strategy may hurt you as well. Each person that considers this as a viable strategy must assess the probability and magnitude of a wrong decision.

And of course, you could get more gain than you expected, but most won't cry about that!

-Will

william dot trice at ngc dot com

I don't know anyone that has a handle on those! I rely on commercial software, it's too convoluted. BTW it's primarily for current-year estimates - "what will the tax be if I sell this?" You can enter assumptions for things like the AMT exemption, which has been amended every year recently. You could project further but I don't find that useful, the code changes too often.

A package like Turbotax could be a reasonable alternative, keeping in mind year-to-year changes in tax laws.

The main point I want to make is that the tax rates aren't constant so a linear formula will break down. It's correct for the incremental dollar in gains at either end, but maybe at $1k in gains, maybe at $100k, it will break down and go through several different rates. You could go from a 0% to low-30% tax rates, which is significant.

Here is a concrete example. Last year a retiree could donate their MRD to charity, and gains in the 15% and below brackets were taxed at 5%. So you might have retiree with a $750,000 IRA and substantial investments, realizing $85k in gains and paying only 5% tax - perhaps only on a portion of those gains. So the effective rate might have been in the low-single digits. I certainly would not have foreseen that ten years ago for someone with seven-figure net worth.

Similar issues come up with Roth conversions, but that's another topic!

Obama has proposed eliminating taxes for retirees with under $50k in income. Not making a political point, it's just an example of a possible change to the tax code that would have a very large effect on these long-term projections. Suddenly the end-rate for capital gains would be

0% for a lot of people. Meanwhile you'd paid out 15% in taxes that didn't continue compounding for maybe 10, 15 years or more.

Changes can work against you too, but you'd have to have a strong opinion about where they're headed to accept a "certain" tax. Where is that point? Back to Rich's question. For me, it would certainly be tested if the preferable LTCG rate was to be eliminated, as it was in the Reagan era tax code.

-Tad

I know mine, and I bet you know yours, too.

You're absolutely correct, but the formula does not break down in my view. Instead it is a point solution. To handle cases like you're talking about you figure your gains in lots. This is not difficult.

But think back 10 years ago to 1998, capital tax rates were heading

*down*. Not a time when you would employ this strategy with any confidence.

True, and he's proposed raising everyone else's LTCG tax rates to 28%, so for those not eliglble for the 0% tax bracket and planning to sell soon they may take an extra 13% in the shorts. Where is that point?

Not true, when you're looking out this far you will likely not get a favorable solution from the equation I presented.

But isn't this the very scenario we're talking about?

-Will

william dot trice at ngc dot com