Trying to understand 2010 Roth conversion/recharacterization

[Tramper Al]

Hi,

I am just trying to get my head around the 2010 Roth conversion/recharacterization, and particularly the risk/return profile. The conversion decision (present tax rate, expected tax rate) is separate, and I am trying to focus here on something different.

For example, say I have a tIRA of $125,000 which I value after-tax (assume a 20% rate) at $100,000. And say that this particular account is 100% stock index fund and my AA% in equities is right on target.

Then say I make a Roth conversion in 2010. Between conversion and the recharacterization deadline, the stocks in the account may go up or down, who knows. I continue to count my equities exposure in terms of my old 80% share of the IRA, for now.

If the stocks go way down as in 2008, I do recharacterize, and end up with the same stock losses on an after-tax computed basis that I would have had if I had never converted. In other words, the IRS shares my loss as they always have in my tIRA. So I had the same old $100,000 (and declining) at risk during that time of the bear.

If the stocks go way up as in 2009 and I don't recharacterize, I take ownership of that 20% in stocks that has now gone up substantially in value. In other words, the gain is all mine. So I had returns on $125,000 (and rising) during the time of the bull.

So my question is probably obvious. Didn't the IRS let me carry that 20% excess stock position risk free? Heads I win, tails I never made the "bet"?

Note that I could also look at it in terms of pure risk reduction. In effect, I could reduce my stock allocation by that $25,000 during this time frame, yet enjoy the same old level of upside exposure.

Just starting to think about this, so forgive me if I have overlooked something obvious. Thanks in advance . . .

[Tramper Al]

Seriously no comment? Free lunch, reward without risk, so please shoot it down. It's not as if I am suggesting sneaking food into a ballgame.

[Doc]

Hi,

I am just trying to get my head around the 2010 Roth conversion/recharacterization, and particularly the risk/return profile. The conversion decision (present tax rate, expected tax rate) is separate, and I am trying to focus here on something different.

For example, say I have a tIRA of $125,000 which I value after-tax (assume a 20% rate) at $100,000. And say that this particular account is 100% stock index fund and my AA% in equities is right on target.

Then say I make a Roth conversion in 2010. Between conversion and the recharacterization deadline, the stocks in the account may go up or down, who knows. I continue to count my equities exposure in terms of my old 80% share of the IRA, for now.

If the stocks go way down as in 2008, I do recharacterize, and end up with the same stock losses on an after-tax computed basis that I would have had if I had never converted. In other words, the IRS shares my loss as they always have in my tIRA. So I had the same old $100,000 (and declining) at risk during that time of the bear.

If the stocks go way up as in 2009 and I don't recharacterize, I take ownership of that 20% in stocks that has now gone up substantially in value. In other words, the gain is all mine. So I had returns on $125,000 (and rising) during the time of the bull.

So my question is probably obvious. Didn't the IRS let me carry that 20% excess stock position risk free? Heads I win, tails I never made the "bet"?

Note that I could also look at it in terms of pure risk reduction. In effect, I could reduce my stock allocation by that $25,000 during this time frame, yet enjoy the same old level of upside exposure.

Just starting to think about this, so forgive me if I have overlooked something obvious. Thanks in advance . . .

I don't think you are looking at this in the right way. You are not sharing anything with Uncle. Uncle owns $25k of your TIRA and you own $100k. If the market goes up or down your $100k goes up or down and Uncles $25k goes up or down by the same percent. Basically you are just acting as manager for Uncle's $$$ but you don't charge any management fee.

I think the way to play the recharacterization is to hope the market goes up and you have to pay more taxes. The more you pay the more money you move from taxable to tax advantaged and therefore the more tax free income you have in the future. (I am assuming you pay the taxes with outside funds.)

I know this sounds ass backwards but think it through. Remember there is no difference between your share of the TIRA and the ROTH except the possible difference in tax rate. You said that at the beginning of your post and then you ignored the consequences.

[Wagnerjb]

Al: I read your post twice, and came away confused both times. I believe the confusion arises from the fact that you describe the balances in two different manners. One place, you describe the balances on an absolute basis, then you switch to refer to the balances on an after-tax basis (the valuation of what is "yours"). I think if you stuck to one basis of presenting the figures, it would be easier to follow. I don't think of my assets in "net" terms, so I am having a hard time following you. Maybe others are too.

I know you are a regular contributor here, so I would certainly be willing to try harder to help you....if you would consider framing the issue in a different manner.

Best wishes.

[DSInvestor]

I would keep it simple and not bring in the tax adjust asset allocation of 20% for your Tax deferred accounts. Focus on the tax cost of the conversion and the market timing for the conversion.

If you have an IRA that is worth 125K on JAN 02, 2010 and you convert the entire balance on JAN 02, 2010, 125K of IRA distribution income will be added to your 2010 income and you'd pay taxes at marginal tax rates. If you're still working in 2010 and are in a 33-35% tax bracket, that's the rate that applies for the ROTH conversion. The income from the ROTH conversion could bump you into a higher bracket.

There's a market timing issue to ROTH-conversions. Assuming you have will pay for conversion tax from an outside source, your ROTH-IRA will have 125K on JAN 03, 2010. If the holdings in the ROTH-IRA rise in value after the conversion, you're in great shape. If the holdings in the ROTH-IRA fall to say 60K, you would have converted at the wrong time. If you could somehow undo the conversion for 125K and redo it (reconvert) at 60K, you'd save a bunch of conversion tax. I believe this is the conversion, recharacterization and reconversion scenario.

See IRS pub 590. See page 30 in the PDF linked below:
http://www.irs.gov/pub/irs-pdf/p590.pdf

Reconversions
You cannot convert and reconvert an amount during the
same tax year or, if later, during the 30-day period follow-
ing a recharacterization. If you reconvert during either of
these periods, it will be a failed conversion.
Example. If you convert an amount from a traditional
IRA to a Roth IRA and then transfer that amount back to a
traditional IRA in a recharacterization in the same year,
you may not reconvert that amount from the traditional IRA
to a Roth IRA before:
• The beginning of the year following the year in which
the amount was converted to a Roth IRA or, if later,
• The end of the 30-day period beginning on the day
on which you transfer the amount from the Roth IRA
back to a traditional IRA in a recharacterization.

Fairmark.com's ROTH-IRA page has some information on conversions and re-characterizations:
http://www.fairmark.com/rothira/index.htm

Fairmark's "Roth IRA Reconversion to Reduce Taxes":
http://www.fairmark.com/rothira/recon.htm

[Doc]

There's a market timing issue to ROTH-conversions. Assuming you have will pay for conversion tax from an outside source, your ROTH-IRA will have 125K on JAN 03, 2010. If the holdings in the ROTH-IRA rise in value, you're in great shape. If the holdings in the ROTH-IRA fall to say 60K, you would have converted at the wrong time. If you could somehow undo the conversion for 125K and redo (reconvert) it at 60K, you'd save a bunch of conversion tax.

I think you have it backwards.

First "Assuming you have will pay for conversion tax from an outside source", is a good thing. If you pay the tax from an outside source you are essentially converting money in a taxable account to a tax advantaged account - in this case a ROTH.

Second "you'd save a bunch of conversion tax" is thinking in favor of big government not your own pocket book. People get themselves misled in this type of calculation because they "try to save taxes". That is the wrong metric. The correct metric is the "amount of money left after tax." Maybe that means you pay more tax and maybe less.

Which would you rather do: Have a $200k salary and pay taxes at 50% or a $100k salary and pay only 25%? If you choose the lower tax rate I have a job offer for you.

Despite the way the tax on a TIRA is calculated on your tax return and despite the popular term "tax deferred" the effective tax on a TIRA is the tax on your contribution only. The earnings on your share grow tax free just like a ROTH. And you share remains a constant percent regardless of any future growth assuming no change in tax rate now and then. Al specifically excluded changes in tax rate for this question is his first sentence.

In the case at hand, with no change in tax rate, you want to pay the most taxes you can as long as money comes form an outside source because this increases the size of the tax advantaged account. You are going to pay those taxes anyway. The fact that they may be greater in absolute terms is only that the government made money by letting you invest the governments money for a while. And just like when you let Vanguard invest your money for you, when you take it out you want both the original principle and any earnings that have accrued.

As far as whether you can convert, recharacterize and then convert again all within the rules I have no opinion.

[grabiner]

Hi,

I am just trying to get my head around the 2010 Roth conversion/recharacterization, and particularly the risk/return profile. The conversion decision (present tax rate, expected tax rate) is separate, and I am trying to focus here on something different.

For example, say I have a tIRA of $125,000 which I value after-tax (assume a 20% rate) at $100,000. And say that this particular account is 100% stock index fund and my AA% in equities is right on target.

Then say I make a Roth conversion in 2010. Between conversion and the recharacterization deadline, the stocks in the account may go up or down, who knows. I continue to count my equities exposure in terms of my old 80% share of the IRA, for now.

If the stocks go way down as in 2008, I do recharacterize, and end up with the same stock losses on an after-tax computed basis that I would have had if I had never converted. In other words, the IRS shares my loss as they always have in my tIRA. So I had the same old $100,000 (and declining) at risk during that time of the bear.

If the stocks go way up as in 2009 and I don't recharacterize, I take ownership of that 20% in stocks that has now gone up substantially in value. In other words, the gain is all mine. So I had returns on $125,000 (and rising) during the time of the bull.

So my question is probably obvious. Didn't the IRS let me carry that 20% excess stock position risk free? Heads I win, tails I never made the "bet"?

There is a free lunch, but it's not quite as large as your tax bracket. Here's an example with numbers worked out.

You have $10K in stock in your IRA, and $2500 in cash outside. You are in a 25% combined tax bracket and expect to retire in the same bracket.

If you don't convert, you have a $7500 tax-free investment, and a $2500 taxable investment which will lose about 30% to taxes, so the after-tax value of your investments is $9250.

If you convert and never look back, you use the $2500 in cash to pay the tax on the conversion, so the after-tax value of your investments is $10,000 as you now have $10,000 in a tax-free Roth. This is the $750 you have gained from conversion when you paid the tax with outside funds, independent of any recharacterization tricks.

If you convert and the stock market rises by 20%, the after-tax value of your investments is $12,000. If you convert and the stock market falls by 20%, the after-tax value of your investments is $8000.

If you convert in 2010 and the stock market falls by 20%, you can recharacterize and get back the $2500 in cash. You can then reconvert in 2011, now paying only $2000 in tax, and retaining $500 in cash which you can invest. The after-tax value of your investments is $8350; that is, the IRS took back 17.5% of your losses. (The 17.5% is the product of the 25% tax bracket and the 70% of your taxable account that you will be able to keep after tax.)

I am currently in the process of pulling off a similar trick with my 2008 conversion. I converted an IRA to a Roth in 2008, and when the market hadn't recovered to its pre-conversion levels in September 2009, I recharacterized. I will reconvert in 2010.

[DSInvestor]

Hi Doc, The only reason I assume that the taxes are paid from an outside source is to keep the value of ROTH-IRA and TradIRAs the same before and after the conversion. If you held 10,000 shares of VWO in TradIRA, you'd still have 10,000 shares of VWO after the conversion in the ROTH-IRA because no money from the IRA was used as tax withholding. You'd have to pay taxes from the taxable accounts.

Let's pretend that 2009 was the year that the income limits for conversion were removed.

10,000 shares of VWO was worth 248K on JAN 02, 2009
10,000 shares of VWO was worth 197K on MAR 09,2009.
10,000 shares of VWO is worth 401K today.

The best time to convert that TradIRA with 10,000 shares of VWO was on March 09, 2009. It would have resulted in the less tax than if the investor had converted on JAN 02, 2009.

IRS seems to have a rule to allow investors who made that early conversion at a higher value to do over.
http://www.fairmark.com/rothira/recon.htm

My mom is in the zero tax bracket and made a mistake filling in her ROTH conversion form several years ago. She didn't check the box to say do not withhold taxes on the ROTH conversion and Ameriprise withheld taxes which reduced the size of her ROTH-IRA account.

[Tramper Al]

OK, well thanks for the tries on this. My story must be very unclear! It's just a simple extension of the proposal I have often seen to recharacterize a Roth conversion that drops substantially before the deadline. The article linked in the other Roth conversion thread says as much. I was just trying to put it in terms of risk/return and the IRS as investment partner.

[Tramper Al]

I am not the only one who has thought of this, the opportunity in recharacterization, and I have not explained it well. Basically the idea is that you have the benefit of several months of performance after a Roth conversion before you have to decide if you want to own it for good. And how this can lead to various schemes.

This article is pretty clear.

http://badmoneyadvice.com/2009/06/the-r ... ategy.html

This one a little less so.

http://www.emarotta.com/article.php?ID=334

Early in year 1, do five Roth conversions of equal amounts into five separate accounts. You aren't going to keep them all, so you can convert five times as much as you want to end up keeping and actually paying tax on. Invest each Roth account in a different asset class (e.g., large-cap U.S. stock, small-cap U.S. stock, foreign stock, emerging markets and hard asset stocks).

The five accounts will appreciate differently, but the entire portfolio will be fairly well balanced. Before April 15 of year 2, decide if you will be keeping only one account or more than one. If more than one has appreciated significantly, you may want to keep more than one account's conversion. Compute your tax liability for the year and pay the tax, but instead of filing your return, file an extension.

Before the October 15 extension deadline, decide which of the five accounts you are going to keep. By now, nearly a year and three quarters has elapsed. You can easily determine which account has appreciated the most. Keep that one and recharacterize the other four. Because you only have to pay taxes on the amount you originally converted, it's like betting on the horse race after the winner has already been determined. After recharacterizing the accounts, file your tax return before the October 15 extension.

If all of the accounts decrease in value, recharacterize them all and pay no tax. Financially you are none the worse for having filled out a folder of paperwork. If only one account appreciates significantly, you only keep one conversion. But you have increased the odds of your Roth account going up by five times.

[Chip]

Didn't the IRS let me carry that 20% excess stock position risk free? Heads I win, tails I never made the "bet"?

Yes. I think that's a very good way to look at it.

Another way to frame it is that your marginal tax rate gets "adjusted" by the appreciation between the date of the conversion and the date the conversion becomes irrevocable.

Example: You convert 10K, all taxable within the 15% bracket. Or $1500 of tax. At the time you decide to make that conversion permanent, it's worth 15K. Your "adjusted" marginal bracket is $1500/15K, or 10%, not the 15% you originally expected. So in effect that changes the inputs to the conversion decision (tax rate now, tax rate at withdrawal).

Right now I have four conversion accounts, all from the first couple of months of the year. They've appreciated 21%, 37%, 67% and 83%. I'll probably recharacterize the first two next year.

[Wagnerjb]

I am not the only one who has thought of this, the opportunity in recharacterization, and I have not explained it well. Basically the idea is that you have the benefit of several months of performance after a Roth conversion before you have to decide if you want to own it for good. And how this can lead to various schemes.

Al: I don't see any fundamental issues with the "schemes". However, people should examine their own situation to see if this strategy fits for them:

a) The guy with $20,000 in an IRA won't want to bother splitting his money up 5 ways. The payoff is probably too small to make the hassle worthwhile.

b) The guy with $1,000,000 in an IRA may not want to do it either. You see, he has $1 million in an IRA, invested in bonds, and $1 million in taxable, invested in stocks. This strategy would have him switch his IRA investments to stocks (and break them into 5 buckets). Logically, this would mean he would shift his taxable from stocks into bonds. Two problems. First, this is very tax inefficient...holding bonds in taxable. Second, he is probably triggering capital gains taxes when selling the stocks in taxable.

c) Be careful how you execute this strategy. The Bad Money article you linked has a dumb idea:

Ideally, you would want some collection of investments that behaved such that after a year or so, although you had about the same amount of money overall, one of those investments would likely have vastly outperformed the others. In the most ideal situation, it would be winner-take-all.

I’m having trouble coming up with investments that behave that way. My first impulse is to use a long position in a volatile stock in one account and a counterbalancing short position in the same stock in another, but as far as I know, no broker will allow an IRA to short stock.

Owning a long and a short position would - in a perfect world - result in your portfolio having ZERO gain after 18 months. How can this be better than just investing in stocks and having a 15% gain after 18 months? This scheme only saves taxes on the gains in the IRA, so I suspect the market gains would be superior to playing games but forfeiting market returns.


IMO, the bigger strategy revolves around tax rates. You can cherry pick your tax rate, which is even more valuable if you have variable income. For example, if a significant part of your pay is due to performance bonuses, or if you got laid off one year, or if you retire during this time.

While you can use both options simultaneously (cherry picking the accounts that rose, and cherry picking the better year to pay taxes), keep in mind that these may conflict or partially offset each other. You may find that one account rose substantially, but that your tax rate is particularly high as well.

One final note. You are correct that you can choose which of the 5 Roth accounts to recharacterize. But for the ones that you do not recharacterize, you cannot choose between them as far as which year to pay the conversion taxes. You can either pay the taxes in 2010 or pay half in 2011 and half in 2012. But the decision applies to all conversions made in 2010.

From the Marotta article:

As if tax matters couldn't get any more complex, Roth conversions during 2010 are taxable 50% in 2011 and the other half in 2012 unless the taxpayer elects to have them taxed completely in 2011. Generally, with rising tax rates, paying the tax in 2011 could be best, but individual situations may warrant spreading the tax over two years.

I think there is a typo in there. They mean to say "unless the taxpayer elects to have them taxed completely in 2010".


Best wishes.

[Tramper Al]

Al: I don't see any fundamental issues with the "schemes". However, people should examine their own situation to see if this strategy fits for them:

Thanks Andy for taking a look, and Chip for a little confirmation.

Personally, between my wife and I we have 4 separate IRA accounts at Fidelity at the moment, and those locations are devoted to 3 tax-inefficient asset classes, each quite volatile (e.g. CCFs). If we did convert all to Roth in Jan 2010 and kept them as 4 separate accounts, it would just be a matter of segregating (moving things around) so that each account held 1 class each. That might cost a few $8 commissions. The 4th account amounts to the conversion tax, the amount by which the Roth location value would be higher than the old tIRA location. So for us, we certainly would not sell taxable stocks and realize cap gains. Obviously, when the Roth conversion is final, that tax must be paid from taxable.

So, I'm saying that if we decided to Roth convert our IRAs, we already have segregated accounts in relevant asset classes to apply a Segregated Roth Recharacterization Put "scheme", for lack of a better name. I would not change my AA, preferred locations, or accelerate CG taxes for this approach. So if we already have volatile and potentially low correlated assets in those IRA accounts, why not.

I don't think the examples of 4-5 kinds of stocks in 4-5 accounts were all that instructive. I guess if you already had that allocation to stocks, you could pick cherry pick the 1 or 2 that did best. But I tend to think that stocks are pretty highly correlated to stocks, and you are likely to end up wanting to recharacterize all of the accounts or none of them. And we don't hold stocks in our IRAs typically. It's a little easier to see how a Permanent Portfolio proponent, for example, might expect a good bit of variation of short term returns in his 4 allocation accounts.

The other thing that learning more about the recharacterization process does for me is that I get the impression that the decision for a Jan 2010 Roth conversion of separate accounts is not a permanent one, at least not until April or October of 2011. That somehow makes the decision a little easier to contemplate.

c) Be careful how you execute this strategy. The Bad Money article you linked has a dumb idea:

Ideally, you would want some collection of investments that behaved such that after a year or so, although you had about the same amount of money overall, one of those investments would likely have vastly outperformed the others. In the most ideal situation, it would be winner-take-all.

I’m having trouble coming up with investments that behave that way. My first impulse is to use a long position in a volatile stock in one account and a counterbalancing short position in the same stock in another, but as far as I know, no broker will allow an IRA to short stock.

Owning a long and a short position would - in a perfect world - result in your portfolio having ZERO gain after 18 months. How can this be better than just investing in stocks and having a 15% gain after 18 months? This scheme only saves taxes on the gains in the IRA, so I suspect the market gains would be superior to playing games but forfeiting market returns.

I'm not sure I agree that this part is so dumb, though. The most efficient result for 2 segregated IRA accounts would be for all the value of one to be shifted to the other, I think he's clearly right about that, as it would halve the conversion tax bill. Ideally this could be accomplished in a very short time frame, but I know what you are saying about opportunity cost. This would certainly be an example of an investor changing his AA (to allow some zero sum pair) as part of his tax scheme. I would not do that.

However, many people do have a cash allocation that is not specifically meant to be at the ready for emergencies at all times. I could see justifying a riskless pair allocation on that basis. But again, I am not inclined to take such an aggressive approach anyway.

IRS audit guy: "So, Mr. Tramper, when did you and your wife start investing in the 10X overnight all or nothing leveraged UP & DOWN ETFs, respectively? It seems like it was right after your Roth conversions?"

[Wagnerjb]

I'm not sure I agree that this part is so dumb, though. The most efficient result for 2 segregated IRA accounts would be for all the value of one to be shifted to the other, I think he's clearly right about that, as it would halve the conversion tax bill. Ideally this could be accomplished in a very short time frame, but I know what you are saying about opportunity cost. This would certainly be an example of an investor changing his AA (to allow some zero sum pair) as part of his tax scheme. I would not do that.

Let me quantify an example to show how this could be dumb. You and I both have $200 in our IRAs, and we are both in the 30% tax bracket.

You convert $100 and invest in stock A, and convert another $100 and invest in a short of stock A. The stock appreciates 25%. You pay the taxes on the $100 conversion at $30 and recharacterize the other account, which has dropped to $75. You convert this other account next year, paying $22 in tax. At the end, you have assets worth $200 and have paid $52 in tax. Congratulations, you paid less than the 30% tax rate on your conversions, thanks to the "schemes".

I convert $100 and invest in stock A, and convert another $100 and invest in stock B. They both appreciate 25% like the market did. I pay taxes on both conversions, and don't recharacterize either one. I pay $60 in tax. At the end, I have assets worth $250 and have paid $60 in tax.

I have $42 more than you as of today. That's huge when you consider that we both started with only $200. Yes, the fact that I assumed a 25% return serves to highlight the issue, but I suspect I still win with lower rates of return.

The "scheme" only pays off with very extreme (and unrealistic) assumptions....like one asset gaining 80% and the other falling 80%.

I don't see any problem with you - or anybody - separating your IRA into different buckets. But the idea of investing in two opposite assets (such that the sum equals zero return) won't work in the real world.

Best wishes.

[Tramper Al]

I'm not sure I agree that this part is so dumb, though. The most efficient result for 2 segregated IRA accounts would be for all the value of one to be shifted to the other, I think he's clearly right about that, as it would halve the conversion tax bill. Ideally this could be accomplished in a very short time frame, but I know what you are saying about opportunity cost. This would certainly be an example of an investor changing his AA (to allow some zero sum pair) as part of his tax scheme. I would not do that.

Let me quantify an example to show how this could be dumb. You and I both have $200 in our IRAs, and we are both in the 30% tax bracket.

You convert $100 and invest in stock A, and convert another $100 and invest in a short of stock A. The stock appreciates 25%. You pay the taxes on the $100 conversion at $30 and recharacterize the other account, which has dropped to $75. You convert this other account next year, paying $22 in tax. At the end, you have assets worth $200 and have paid $52 in tax. Congratulations, you paid less than the 30% tax rate on your conversions, thanks to the "schemes".

I convert $100 and invest in stock A, and convert another $100 and invest in stock B. They both appreciate 25% like the market did. I pay taxes on both conversions, and don't recharacterize either one. I pay $60 in tax. At the end, I have assets worth $250 and have paid $60 in tax.

I have $42 more than you as of today. That's huge when you consider that we both started with only $200. Yes, the fact that I assumed a 25% return serves to highlight the issue, but I suspect I still win with lower rates of return.

The "scheme" only pays off with very extreme (and unrealistic) assumptions....like one asset gaining 80% and the other falling 80%.

I don't see any problem with you - or anybody - separating your IRA into different buckets. But the idea of investing in two opposite assets (such that the sum equals zero return) won't work in the real world.

Best wishes.

Thanks, but really don't you think your example is more than a bit stacked?

I already said I agree that a zero sum pair would be an inappropriate departure from a stock allocation. I also used the example of a cash allocation and riskless pairs. As you continue to increase the stock returns assumption (I think you first assumed 15%, now 25%?). Are you not simply demonstrating that when markets go up a lot, they return more than zero? As in, "don't abandon your AA for a prolonged period of time for schemes"? Did I suggest anyone should? No. Would anyone be much swayed by a contrived example where stocks declined by 20%, and the zero sum investor mysteriously came out ahead? I would not think so.

The author of one of the linked articles described the riskless pairs thing fairly well, I think in one case using the absurd example of 16 accounts with double or nothing pairs whittled down to one. I agree that absent the investment vehicles, it's an abstract academic example. He also found the devil in the details, that the zero sum pair trading vehicles were a little hard to come by. Personally, I don't find indentifying a pair that may give an 80% gain / 80% loss to be all that unbelievable - except of course that those gain/loss figures you gave are not symmetrical! The couple of UP/DOWN ETF pairs (first there was oil, now Schiller RE) might be candidates for this. I don't trust them for symmetrical (and thus risk free) returns, however.

Maybe the more relevant exercise, though, if you want to consider it likely that stocks may appreciate 25% in 18 months, would be the one I alluded to in the first post. Roth convert a 100% stocks tIRA, the portion of which you own (1-tax)% is what you have allocated to stocks in your AA. Gamble with the (tax)% that belonged to the IRS in the tIRA. Wait 18+/- months into 2011, and now you have the option to buy out your government partner at the old Jan 2010 prices, or not. If stocks have gone way up, you can then commit to that higher %stocks and own the (in excess of AA) gains of the last 18 months. If stocks have tanked, you recharacterize and have the same old loss share as dictated by your old AA. It's as if you never made the bet. Closest thing to a time machine I've run across lately.

That's actually the question I had up top for the thread. Didn't the IRS let me carry that (tax)% excess stock position risk free?

[Wagnerjb]

The author of one of the linked articles described the riskless pairs thing fairly well, I think in one case using the absurd example of 16 accounts with double or nothing pairs whittled down to one. I agree that absent the investment vehicles, it's an abstract academic example. He also found the devil in the details, that the zero sum pair trading vehicles were a little hard to come by. Personally, I don't find indentifying a pair that may give an 80% gain / 80% loss to be all that unbelievable - except of course that those gain/loss figures you gave are not symmetrical! The couple of UP/DOWN ETF pairs (first there was oil, now Schiller RE) might be candidates for this. I don't trust them for symmetrical (and thus risk free) returns, however.

Al: I didn't mean to suggest that your strategy was flawed, nor that you intended to change your AA. However, I was trying to highlight that the author's recommendation would likely fail in the real world....even if you do get symmetrical returns.

Here is what he says:

I’m having trouble coming up with investments that behave that way. My first impulse is to use a long position in a volatile stock in one account and a counterbalancing short position in the same stock in another, but as far as I know, no broker will allow an IRA to short stock.

I hope to dissuade other investors (not you) from trying to find apparently symmetrical investments as I am convinced that this strategy will fail - unless the results are extreme gains and losses. Moderate - and symmetrical - gains and losses will likely generate inferior results unless the overall stock market declined during that time....in which case you got lucky by investing in a zero return asset when the market dropped.

Best wishes.

[Wagnerjb]

Thanks, but really don't you think your example is more than a bit stacked?

The market assumption isn't a sensitive factor in this issue. The example I showed assumed a +25% and -25% riskless pair, plus a 25% market return. If you have a +25% and -25% riskless pair, the market only needs to increase by 4.1% or more, and I win.

If the riskless pair is +50% and -50%, the market only needs to generate a return of 7.6% or more, and I win.

It takes a VERY EXTREME set of riskless pair returns, plus a very low market return....to make the author's strategy work. IMO, that is pretty unlikely in the real world. (The casino example cited by the author has a +3600% and -100% set of returns).

Again....Al, this isn't aimed at your strategy, which makes sense to me.

Best wishes.

[Tramper Al]

Again....Al, this isn't aimed at your strategy, which makes sense to me.

Hey, thanks for the further explanation!

[xerty24]

Let me quantify an example to show how this could be dumb. You and I both have $200 in our IRAs, and we are both in the 30% tax bracket.

You convert $100 and invest in stock A, and convert another $100 and invest in a short of stock A. The stock appreciates 25%. You pay the taxes on the $100 conversion at $30 and recharacterize the other account, which has dropped to $75. You convert this other account next year, paying $22 in tax. At the end, you have assets worth $200 and have paid $52 in tax. Congratulations, you paid less than the 30% tax rate on your conversions, thanks to the "schemes".

I convert $100 and invest in stock A, and convert another $100 and invest in stock B. They both appreciate 25% like the market did. I pay taxes on both conversions, and don't recharacterize either one. I pay $60 in tax. At the end, I have assets worth $250 and have paid $60 in tax.

I have $42 more than you as of today. That's huge when you consider that we both started with only $200. Yes, the fact that I assumed a 25% return serves to highlight the issue, but I suspect I still win with lower rates of return.

I don't think this is a fair example - you are long the market $200, and Al is long/short net $0, so it should come as no surprise that you beat him when the market is up 25%. What about the 2008 scenario, -25% across the board?

Al's long/short approach makes the same amount of money, only it's the short account that wins and he recharacterizes the long one. Same $200 in assets and $52 in taxes after converting the losing account value next year. Aftertax assets = $148.

You convert $100 and invest in stock A, and convert another $100 and invest in stock B. They both fall 25% like the market did. You recharacterize both, and they are worth $75 each at the end of the year. At the start of the next year you convert them both, paying $45 in taxes. After tax assets = $105.

Perhaps it's more instructive to look at the flat market case. In that case, Al saves $8 in taxes ($60 vs $52), which is a 4% excess return. That's a pretty good return for something with no market risk, but maybe is arguably more like a fixed income investment (in tax savings) than a stock.

In both cases, a natural extension of this approach is to keep trying to convert losing accounts in later years, and, should they lose money, recharacterize and try again until you (or the market) gets lucky.

Remember that shorting in your IRA is very difficult to do; it's allowed by the IRS rules, but I don't know of any normal brokerages that will let you do this. You would typically have to go to a specialty (read expensive) IRA custodian. An alternative would be to invest in "short ETFs" or similar products, but those have their own host of issues and I can't really recommend them.

[Tramper Al]

Let me quantify an example to show how this could be dumb. You and I both have $200 in our IRAs, and we are both in the 30% tax bracket.

You convert $100 and invest in stock A, and convert another $100 and invest in a short of stock A. The stock appreciates 25%. You pay the taxes on the $100 conversion at $30 and recharacterize the other account, which has dropped to $75. You convert this other account next year, paying $22 in tax. At the end, you have assets worth $200 and have paid $52 in tax. Congratulations, you paid less than the 30% tax rate on your conversions, thanks to the "schemes".

I convert $100 and invest in stock A, and convert another $100 and invest in stock B. They both appreciate 25% like the market did. I pay taxes on both conversions, and don't recharacterize either one. I pay $60 in tax. At the end, I have assets worth $250 and have paid $60 in tax.

I have $42 more than you as of today. That's huge when you consider that we both started with only $200. Yes, the fact that I assumed a 25% return serves to highlight the issue, but I suspect I still win with lower rates of return.

I don't think this is a fair example - you are long the market $200, and Al is long/short net $0, so it should come as no surprise that you beat him when the market is up 25%. What about the 2008 scenario, -25% across the board?

Al's long/short approach makes the same amount of money, only it's the short account that wins and he recharacterizes the long one. Same $200 in assets and $52 in taxes after converting the losing account value next year. Aftertax assets = $148.

You convert $100 and invest in stock A, and convert another $100 and invest in stock B. They both fall 25% like the market did. You recharacterize both, and they are worth $75 each at the end of the year. At the start of the next year you convert them both, paying $45 in taxes. After tax assets = $105.

In both cases, a natural extension of this approach is to keep trying to convert losing accounts in later years, and, should they lose money, recharacterize and try again until you (or the market) gets lucky.

Remember that shorting in your IRA is very difficult to do; it's allowed by the IRS rules, but I don't know of any normal brokerages that will let you do this. You would typically have to go to a specialty (read expensive) IRA custodian. An alternative would be to invest in "short ETFs" or similar products, but those have their own host of issues and I can't really recommend them.

Thanks xerty, for going through that. In considering something like this, I do try to separate any "change of AA" effect, so at least it is not an apples to oranges thing.

If we do decide to Roth-convert in Jan 2010, I might have accidentally fallen into this "strategy" anyway, if the performance of accounts diverged substantially over the course of the year. Knowing the recharacterization segregation approach, I would now make an effort to put a different asset class in each account, but will still stick with long positions that are already in our AA. I think I might include REITs (located elsewhere currently) as one of the 4 corners.

In terms of zero sum pairs? Our IRAs at Fidelity do seem to allow the use of options, at least of the limited liability long puts and long calls variety. Maybe the right pair could lock in a major shift in value from one account to the other without too much trouble. I really don't know.

[xerty24]

Getting back to Al's question about AA issues related to a recharacterization, suppose you start with his AA:

tIRA $100 in stocks
(other stuff)

Once you convert, you've tentatively got an AA that looks like this:

Roth IRA $100 stocks
$20 tax liability (using 20% tax rate)

So if you do this in January (the best time), you're essentially borrowing $20 from the government (for free) and betting it on the same stocks. Come the end of the year (or more precisely the end of your last filing extension), you decide whether or not to keep the Roth or recharacterize. If you keep it (since stocks were up), your "margin loan" has paid off. If you recharacterize (since stocks were down), you no longer have your tax liability and hence have only lost pretax money.

Basically there are many ways to think about this, but the right to recharacterize based on market performance is a free option. Options are always worth something positive, and you can make them more valuable by extending their duration (1.5 years), increasing the volatility of what you own, etc.

[Tramper Al]

Basically there are many ways to think about this, but the right to recharacterize based on market performance is a free option. Options are always worth something positive, and you can make them more valuable by extending their duration (1.5 years), increasing the volatility of what you own, etc.

Nice, I was hoping someone would come around and offer some affirmation on this. I find that most threads that begin with something like "Am I getting potential reward without the usual risk?" don't end so well.

On a side note, I am now tracking in real time - on my spreadsheet - our projected Roth conversion tax cost, if we were to convert in January of 2010. It is essentially the marginal tax rate times the account balance less basis. And this tax cost has been going steadily down all week long! Excellent, right?

[sscritic]

I tried a little spreadsheet. Start with 100 each in two IRA's. Tentatively convert both to Roths. Investment A always returns more than investment B in year one, whether the return are 40% and 20%, 30% and -10%, or -10% and -30%. Assume zero return in year two for both.

There are four strategies that I can see over the two year period:
1) recharacterize both. Result: two IRAs.
2) recharacterize B, pay tax on A's conversion in year one. Convert B in year two and pay tax on that conversion in year two. Result: two Roths.
3) don't recharacterize, pay tax on both conversions in year one. Result two Roths.
4) recharacterize B, pay tax on A's conversion in year one. Don't convert B in year two. Result: one IRA and one Roth.

Assume 20% tax rate and $46 in taxable in a money market with 0% return.

Here are some results:

First Case: A up 40%, B down 10% (accounts go to 140 and 90).
1) result: $230 in IRAs and $46 in taxable. Govt share is $46. Net = $230.
2) result: $230 in Roths. Tax of $20 + $18 leaves $8 in taxable. Net = $238. ** BEST CHOICE
3) result as in 2, except tax is $40. Net = $236.
4) result $140 in Roth, $90 in IRA. Tax of $20; remaining taxable $26; Govt share $18. Net = $238. ** BEST CHOICE

Second Case: A up 40%, B up 20% (accounts go to 140 and 120)
1) result: $260 in IRAs and $46 in taxable. Govt share is $52. Net = $254.
2) result: $260 in Roths. Tax of $20 + $24 leaves $2 in taxable. Net = $262.
3) result as in 2, except tax is $40. Net = $266. ** BEST CHOICE
4) result $140 in Roth, $120 in IRA. Tax of $20; remaining taxable $26; Govt share $24. Net = $262.

Third Case: A down 10%, B down 20% (accounts go to 90 and 80).
1) result: $170 in IRAs and $46 in taxable. Govt share is $34. Net = $182. ** BEST CHOICE
2) result: $170 in Roths. Tax of $20 + $16 leaves $10 in taxable. Net = $180.
3) result as in 2, except tax is $40. Net = $176.
4) result $90 in Roth, $80 in IRA. Tax of $20; remaining taxable $26; Govt share $16. Net = $180.

Note that 2 and 4 always give the same result since the assumption of zero returns in year two means the govt share of the B IRA in year two is the same as the tax on the conversion.

The results are what we expect. If they both go up, convert both the first year. If one goes up and one down, convert just the one that goes up. If they both go down, don't convert either. The value of having choices (as compared to just leaving the money in the IRAs) is $8, $12, and $0 for the three example cases.

I left out the strategy of recharacterizing both in year one and converting in year two using the lower basis for the tax on the conversion. Assuming zero return in year two for both should mean that it gives the same result as leaving both in the IRAs (same reason 2 and 4 are equivalent, i.e., tax is exactly govt share). Of course we really need a multi-year strategy, but using this one year strategy repeatedly should be good enough since we can't predict returns two years out.

[Wagnerjb]

Perhaps it's more instructive to look at the flat market case. In that case, Al saves $8 in taxes ($60 vs $52), which is a 4% excess return. That's a pretty good return for something with no market risk, but maybe is arguably more like a fixed income investment (in tax savings) than a stock.

In the flat market case, we tie. Al doesn't save any taxes if both of his matched pairs have a 0% return.

I think we are all on the same page....that is, a "matched pair" strategy is a change of your Asset Allocation. The author seemed to suggest that a matched pair strategy was a clear winner, and I am pointing out that the AA change could cause the strategy to backfire. Just something to be aware of before diving in

Best wishes.

[Wagnerjb]

Basically there are many ways to think about this, but the right to recharacterize based on market performance is a free option. Options are always worth something positive, and you can make them more valuable by extending their duration (1.5 years), increasing the volatility of what you own, etc.

Yes, and don't forget that the option to choose your tax rate after the fact (by selecting the year to pay the taxes) is probably just as valuable to some people...

Best wishes.

[xerty24]

In the flat market case, we tie. Al doesn't save any taxes if both of his matched pairs have a 0% return.

True, although a more reasonable model might be that each stock in question gives the market return plus/minus some noise (idiosyncratic volatility). In practice, this means that there will still be expected value to the option in a flat overall market since you could still see some dispersion on the long/short position in the particular stock you picked.

Yes, and don't forget that the option to choose your tax rate after the fact (by selecting the year to pay the taxes) is probably just as valuable to some people...

Very true. This is probably one of the most valuable options, actually, especially since it can easily result in 5-10% savings by avoiding state and local taxes by moving away for a year (or forever!). Moving beyond just the "now vs retirement" tax rate comparison, there are also advantages for converting now... by expanding your total tax-sheltered account size, you gain the benefit of tax-free compounding on the value of the taxable funds used to pay for the conversion. Still, that one takes a long time to matter (much), and saving 5-10% off the bat by moving out of CA, NY, etc, can be a clear winner even if you have to wait a while for it.

[Wagnerjb]

In the flat market case, we tie. Al doesn't save any taxes if both of his matched pairs have a 0% return.

True, although a more reasonable model might be that each stock in question gives the market return plus/minus some noise (idiosyncratic volatility). In practice, this means that there will still be expected value to the option in a flat overall market since you could still see some dispersion on the long/short position in the particular stock you picked.

I don't think there is any value to the "matched pairs" option from an "expected value" standpoint. Remember, we are talking about the zero-sum matched pair option. From an expectation standpoint, you "expect" the stock market to generate a normal return...say 8%. My previous analysis showed that you would need the matched pairs to be greater than +50% and -50% in order for the tax savings to offset the loss of market returns. The expectation for stock volatility is much less than that...so the bottom line is that the "matched pair" option isn't like a classical option - where you can win, but not lose. You can lose by not being exposed to the market, and it is likely that you WILL lose.

I agree that - ex post - you might end up with a flat market and dispersion in your matched pairs...generating a benefit. But aren't you just as likely to end up with stock market gains and a matched pair of 0% returns....generating a loss?

Best wishes.

[Tramper Al]

But aren't you just as likely to end up with stock market gains and a matched pair of 0% returns....generating a loss?

I think you'd have to be profoundly unlucky or quite bad at selecting matched pairs in order to have each half remain more or less equal for 18 months! Just as likely? Come on.

Didn't an example in the linked article start with something like 16 accounts, paired to reduce to only one? I believe that would optimally reduce the conversion tax to 1/16th of the original. Extreme, perhaps, but I would probably put forth an example with a single pair and zero mismatch of returns only if I wanted to doom the "analysis" from the start.

I'm pretty sure there is general agreement that if you depart from an AA for a particular period of time, that there is no reason to expect that the returns of your new (non-AA) portfolio would match that of your AA portfolio. Personally, I think this rather goes without saying.

[Wagnerjb]

I think you'd have to be profoundly unlucky or quite bad at selecting matched pairs in order to have each half remain more or less equal for 18 months! Just as likely? Come on.

Al: I believe you will find that it is actually more likely that the stock will return zero and the market will be positive.

There are only two variables here - the return of the stock market, and the return of the individual stock you chose. If we expect the stock market return to be 8%, the most likely outcome is that both the stock market and the individual stock will return 8%. Sure, that particular outcome may only have a 3% chance of happening, but it is the single highest probability. But what about the many other possible outcomes?

The returns of the stock market (in any one given year) are distributed like a bell curve, with the top of the curve around the 8% mark. The curve is fairly high, with shorter tails....maybe -25% on the left and +40% on the right. The bell curve for the individual stock has a lower peak, with longer tails...maybe -75% and +150%. However - and this is the important point - the peak of the individual stock curve is around 0%. Studies have shown that the average individual stock underperforms the market, and that a few huge gainers like Microsoft or Dell contribute to the nice overall market return. This study explains this effect better:

http://www.efficientfrontier.com/ef/900/15st.htm

Anyway, I believe the data will show that a 0% return for an individual stock is more likely than a 0% return for the market.

Didn't an example in the linked article start with something like 16 accounts, paired to reduce to only one? I believe that would optimally reduce the conversion tax to 1/16th of the original. Extreme, perhaps, but I would probably put forth an example with a single pair and zero mismatch of returns only if I wanted to doom the "analysis" from the start.

Al - I agreed with you that the 16 account roulette example would work perfectly. And I agreed with you that not changing your AA is a wise strategy.

But I am simply trying to provide a "real world" example that demonstrates how the matched pair strategy would fail. I am happy to discuss why my examples may not be realistic, but I hope you don't feel that I am in any way disparaging your strategy or the value of the option to recharacterize your Roth.

Best wishes.

[Tramper Al]

I am happy to discuss why my examples may not be realistic, but I hope you don't feel that I am in any way disparaging your strategy or the value of the option to recharacterize your Roth.

No, it's no biggie. It's just been rather hard to get much commentary on the Roth-recharacterization-put option thing, and so the whole "don't change your AA!" sidebar is sort of a "no kidding" distraction for me. Potential reward without risk, that bit is what makes it interesting for me. But you guys discuss whatever facet seems important to you. I think I have enough of the basic facts to look ahead to Jan 2010 at least. I definitely don't want to be that guy who starts his thread with "this is only about X, so let's have no discussion of Y" and so on.

[Wagnerjb]

Potential reward without risk, that bit is what makes it interesting for me.

Tax management of a stock portfolio has similar characteristics, so this issue of an "option" isn't entirely unique to the Roth conversion....although it certainly is attractive for the Roth.

With individual stocks, you get to wait and see which ones gained and which ones lost money, just like the Roth conversion allows you to decide after you see which account gained and which lost. With stocks, you hold the ones that gained (paying no tax) and you sell the losers (getting valuable tax loss benefits). With the Roth, you keep the winners and rechacterize the losers.

You can do this with mutual funds, but it works better with stocks because you have a wider dispersion of returns. So, you get an effect somewhat similar to the matched pair strategy without sacrificing the market returns....some stocks will drop 20% and some will surge 75%. You get to choose which one will be taxed after you see the outcome.

Best wishes.

[xerty24]

Al: I believe you will find that it is actually more likely that the stock will return zero and the market will be positive.

There are only two variables here - the return of the stock market, and the return of the individual stock you chose. If we expect the stock market return to be 8%, the most likely outcome is that both the stock market and the individual stock will return 8%. Sure, that particular outcome may only have a 3% chance of happening, but it is the single highest probability. But what about the many other possible outcomes?

The returns of the stock market (in any one given year) are distributed like a bell curve, with the top of the curve around the 8% mark. The curve is fairly high, with shorter tails....maybe -25% on the left and +40% on the right. The bell curve for the individual stock has a lower peak, with longer tails...maybe -75% and +150%. However - and this is the important point - the peak of the individual stock curve is around 0%.

Even if we decide to model the market with an average annual return of 8% and our favorite stock with an annual return of 0%, the long/short strategy could still be better on average if the individual stock volatility was high enough. To make money on the market, you need a return of 0% or better, which corresponds to the right side of the bell curve or about a 60% chance. To make money after taxes on the recharacterization, you need the stock return to be any value not ~0%, essentially the whole distribution except the area exactly around 0% (you just recharacterize a different account for the left vs the right tail). Even though we're granting that 0% is the most likely value (the mean), there's still only a 2-3% chance of getting within [-1%,1%] assuming just the market volatility (which is certainly an underestimate).

Thus you're looking at a probability of after-tax profit of ~97% with the long/short approach, while only a probability of ~60% from the market. In addition, the long/short approach can't lose money (worst case is 0%), while the market could easily (40% chance) lose money, quite possibly a large amount if you're unlucky. This is part why I referred to the long/short approach as more like fixed income for AA purposes. It's certainly a lot less risky.

Of course if you wanted to be more careful in your analysis to figure out which approach had the higher expected return rather than probability of profit, you should really keep track of how much you make in each scenario. This involves counting the positive tail events more and the negative tail events less in the case of betting on the market, while counting both the positive and negative tail events more and the central events less in the case of the long/short position. While you were at it, you probably would want to throw in the correlation of your stock with the market too. I leave that exercise to someone else .

[Chip]

Al,

I for one do wish we could get away from the AA shift discussion since it was never advocated in the first place.

Here's my real world example of the value of the recharacterization option. If you recall from my earlier message, I made four conversions earlier in the year, each to a different Roth. Each conversion was 100% equities and there was no change in asset allocation.

If I finalized my conversions today (by recharacterizing everything except what I wanted to keep) here is the tax cost (assume a 15% bracket) to keep 10K converted:

Roth #1: $879
Roth #2: $812
Roth #3: $1146
Roth #4: $1217

If I had made those same conversions into a single Roth, the tax cost today of keeping 10K converted would be $977 (the weighted average of the four individual conversions).

It's not true in my case, but if each individual conversion was an amount that was more than I actually wanted to finalize, the value today of the option for this situation is $165 per 10K converted (977-812).

Frankly it's worth a small amount of my time to make a potential 1-2% since there is no downside risk.

[Wagnerjb]

Thus you're looking at a probability of after-tax profit of ~97% with the long/short approach, while only a probability of ~60% from the market. In addition, the long/short approach can't lose money (worst case is 0%), while the market could easily (40% chance) lose money, quite possibly a large amount if you're unlucky.

I agree with your analysis Xerty, but I don't think you have followed through to the conclusion. We both agree that you have a 97% chance of saving taxes using the matched pair. And we both agree that you have a 60% chance of having a positive market return. But I have demonstrated that a +25% and -25% matched pair does not 'win" unless the market returns under 4%. Even a +50% and -50% matched pair does not beat a "normal" market return of 8%.

Taking your analysis a bit farther, I would agree that 40% of the time, the matched pair clearly wins. That is, if the market declines the matched pair wins...not only do you get tax savings, but you avoid the market decline. But I suspect most of the 60% of the time when the market gains you will see the matched pair lose. For it to win, you need a high volatility AND and very low market return. It is certainly possible for some of the observations.

I also agree that taking an expected value is the best way of looking at this, but I was hoping you were going to volunteer for that..

If you and I can go around and around on whether the matched pair strategy has a positive expected value, maybe we can both agree that it clearly isn't a "no brainer" can't we?

Best wishes.

[xerty24]

Even a +50% and -50% matched pair does not beat a "normal" market return of 8%.

Taking your analysis a bit farther, I would agree that 40% of the time, the matched pair clearly wins. That is, if the market declines the matched pair wins...not only do you get tax savings, but you avoid the market decline.

I also agree that taking an expected value is the best way of looking at this, but I was hoping you were going to volunteer for that..

I compared an uncorrelated individual stock payout of 0% +- 35% with the long term market 8% +- 25%, and assumed a 20% tax bracket. The return to the long/short approach is only 2.7% +- 2.0%, remembering that it's about half what you expect since you're allocating one unit of capital to both the long and short sides (while the other person is betting both of these on the market). The Sharpe ratio is about 1.4 for this, meaning it's a very reliable but low return "investment", while the market has a Sharpe of 0.3 being much more volatile and with a much higher return under these assumptions. It is worth noting that the Sharpe ratio for the long/short approach is the same regardless of the individual stock vol you assume, and is essentially the same regardless of the mean you might assign to the individual stock.

The main issue here is that the returns of the low-risk long/short approach is related to both the tax rate and individual stock volatility. For example,

20% tax rate, 35% stock vol = 2.7% average return +- 2.0%
20% tax rate, 50% stock vol = 3.9% +-2.8%
50% tax rate, 50% stock vol = 9.7% +-7.1%

So you can see that if you're in a very high tax rate, this is more attractive (in principle). At relatively low tax rates, it's basically just a complicated low-return fixed income alternative (although most bonds still don't give a Sharpe over 1 and you'd need to go out to 5 year Treasuries to get paid ~3% these days).

If you and I can go around and around on whether the matched pair strategy has a positive expected value, maybe we can both agree that it clearly isn't a "no brainer" can't we?

I don't think there should be any question about whether the matched pairs is a positive expected value - you don't lose money and on average you make some. I agree it's not a "no brainer" that this is better than your other alternatives, like investing in equity index funds. However, that's a question of risky tolerance more than anything else, since this is much less risky than the market and essentially your only costs are opportunities lost.

[Wagnerjb]

I don't think there should be any question about whether the matched pairs is a positive expected value - you don't lose money and on average you make some. I agree it's not a "no brainer" that this is better than your other alternatives, like investing in equity index funds.

Xerty - thanks for the good analysis. This particular example is kinda like the saying about the tax tail wagging the (investment) dog. How is it different from the guy in the 10% tax bracket investing in Muni bonds? Sure, he can demonstrate that he is paying less income tax. But if his alternative is Taxable bonds, he has lost the total (after-tax) return game.

Best wishes.