How to create a no-limit Synthetic Roth IRA in a taxable account

Tired of contributing a paltry $5,500 per year ($11,000 for couples) to your Roth? If you like to contribute more than that, why not find a way to generate returns in a taxable account that mimic those of a Roth IRA? Impossible, you say? Under very specific conditions it is possible to generate after-tax returns in a taxable account that replicate those of a Roth IRA. We call it the Synthetic Roth IRA.

Disclaimers:

  • In no way do we recommend foregoing a Roth IRA altogether. If you like your Roth, keep it and keep contributing. Our idea is for folks who already max out their Roth IRA (direct or back-door) every year and like to get additional exposure to Roth returns, but can’t due to annual contribution limits
  • Our approach involves the use of leverage and derivatives (index futures) and it’s not for everyone. Moreover, this only works for pretty large size accounts: $70K+
  • Our general, more detailed disclaimers all apply as well!

Why a Roth IRA?

The appeal of a Roth IRA is obvious: money grows tax-free and withdrawals are tax-free as well. If you can come up with the after-tax money to fund the Roth and like the idea of tax-free growth and withdrawals it’s a great investment vehicle. There is also the added flexibility that, in contrast to a regular IRA or 401(k), Roth IRAs have no minimum required distribution requirements (RMD) so you can allow your tax-free assets to grow for the longest period in your portfolio.

The disadvantages of the Roth are mostly the obstacles imposed by the tax-man:

  • Income limits for direct Roth IRA contributions. Most high-income households are ineligible. One way to circumvent the income limits is the “back-door Roth.” Make a contribution to a traditional (after-tax) IRA, then convert the same thing to a Roth the same day.
  • Annual contribution limits of only $5,500 per spouse, which increases to $6,500 post age 50. This applies to both the direct Roth and the back-door Roth.
  • While we can access the principal of a Roth at any time, capital gains are only accessible after age 59.5 to avoid the penalty on early withdrawals.
  • If rolling over an existing IRA into a Roth, the previously tax-deferred portion will become taxable income in that tax year
  • If the value of the Roth goes down, there is no way to write off a capital loss. In other words, there is no tax loss harvesting in a Roth IRA.

Quantify the advantage of the Roth IRA over a taxable account

In the chart below we plot the after-tax portfolio value of an initial $1,000 as a function of the capital market return r. Clearly, for the Roth IRA, the final value is 1000(1+r), while for the taxable accounts it is 1000(1+r(1-MarginalTaxRate)). In fact, you could do even slightly better in the taxable account if taking into account tax-loss harvesting, which we ignore for now.

Roth IRA calc chart01
Final after-tax portfolio value of Roth IRA and taxable accounts as a function of capital market return.

There are several implications from this analysis:

  1. The Roth IRA is only advantageous over a taxable account if you actually make money. If you lose you’re better off with a taxable account, and that advantage gets even bigger if you can use tax loss harvesting.
  2. The after-tax return distribution of a taxable account is exactly identical to the Roth return multiplied (scaled down) by a factor of (1-MarginalTaxRate). Equivalently, a Roth IRA is statistically identical to a levered taxable portfolio, scaled up by a factor of 1/(1-MarginalTaxRate)
  3. If we could find a cheap (or even no-cost) way to exactly lever up the taxable portfolio by that factor we could exactly replicate the Roth IRA performance, after-tax

How to gain leverage

There isn’t an easy way for most retail investors to achieve the leverage necessary. Hence, the advantage of the Roth IRA is due (in part) to the average retail investor’s lack of access to easy and inexpensive leverage. Here are two suggestions, but they are not exactly attractive and workable:

  1. Buying equities on margin: That could get expensive. Most brokers have pretty hefty margin interest rates: Check here for a comparison. We’re talking about 5% margin interest for account sizes around $100,000-200,000. So, if you like to lever up your account by a factor of 1.25 (corresponding to a marginal tax rate of 20%) we’re looking at a performance drag of about 1.25% p.a., which seems quite high. Interactive Brokers seems to be an outlier with only 1.36%, so for that 25% extra leverage, you lose around 0.34% p.a.
  2. Buying levered equity ETFs. To gain 1.25x leverage, one could buy 75% (unleveraged) equity funds plus another 25% in a 2x levered ETF. Fees for those funds are high, around 0.8-1.0% p.a., so for the extra 25% of leverage, we’re looking at 0.2-0.25% drag on performance. For some people, this might be already close enough to the Roth.

Still, there are several problems and issues with these two approaches. For buy and hold investors in the taxable account: what’s the marginal tax rate? If you intend to hold on to your equity funds in the taxable account to defer the capital gains as long possible, you will have to gauge not only what’s the marginal tax at a date decades in the future but also how to distribute that marginal tax throughout all the years of the holding period. Leveraged ETFs tend to have a return drag in sideways-moving markets.

So, while the leverage idea through these two routes seems interesting as a theoretical exercise, the implementation in practice is just such a can of worms, it may not be so attractive after all. For us, if those two routes were the only ones to implement this, we’d give this synthetic Roth a pass. But: if someone out there has experience on how to implement this efficiently with portfolio margin and/or levered ETFs, please let us know and leave a comment below!

Let’s not waste any more time and instead look at the actual implementation we use:

Low-cost leverage through equity futures contracts

Futures contracts are a very easy and inexpensive way of generating leverage. Here are the requirements to make this work:

  • An account with a futures trading platform. We use Interactive Brokers (IB), which seems to have very low trading commissions. For the record, we don’t make any referral fee if you sign up with IB, so our recommendation is unbiased.
  • Some basic trading experience and clearance from your broker to trade index futures. See our intro to derivatives.
  • A relatively clear picture of your likely marginal tax rates, both federal and state (and local if applicable), for the current calendar year. Your marginal tax rate for index futures and their derivatives (IRS Section 1256 contracts) is 60% long-term capital gains and 40% short-term gains, irrespective of the holding period.
  • A large enough account to make this worthwhile: the absolute minimum would be the value of one single futures contract times one minus your marginal tax rate on Section 1256 contracts. Example: marginal tax rate 30%, ES future quote 2,090, multiplier 50, then the principal necessary would be 50x2090x(1-0.3) equal to about $73,000.

Synthetic Roth IRA mechanics

  1. Divide your desired account balance by one minus your marginal tax rate. Example: you have $150,000 ready to invest and your marginal tax rate is 28% (combined state, local and federal tax for Section 1256 contracts). $150,000/0.72=$208,333. This is the equity futures exposure you target. That is pretty close to two ES contracts, so go long two contracts.
  2. Keep an amount equal to the maintenance margin of the long futures contracts (plus maybe a small cushion) as cash in the account. Currently, the maintenance margin is $4,200 for the ES future.
  3. Invest the rest of the cash in a low-risk interest-bearing fund. This could be a money market fund or something with a little bit more duration and/or credit risk to juice up returns a little bit more. In other words, why not do better than the 0.3% in a money market account: You could get around 0.9% yield in a floating rate ETF (no interest rate risk, only credit risk) or even more than 1% with short-duration (1-3 year maturity) corporate bonds.
  4. ES futures contracts expire on the third Friday of March, June, September and December each year. If you are currently holding the June ES Future you have until June 17, 9:30 am EDT to sell the contract and buy a later-dated contract such as the September or December contract. This process is also called “rolling a futures contract” and needs to be repeated every quarter, ideally a few days or weeks before the contract expiration. It should take only a few minutes.
  5. Sit back, relax and watch your synthetic Roth exactly replicate the after-tax return in a Roth IRA.

Numerical examples

Assume you’re in the 28% federal bracket for short-term gains, 15% for long-term gains and 5% bracket for state taxes. The marginal tax for the futures returns is 0.6×0.15+0.4×0.28+0.05=25.2%. Interest income is taxed at 0.28+0.05=33%. The first example looks at what if you invest the available margin cash exactly at the risk-free rate in a money market account. I pick the initial wealth such that the leveraged futures position comes out as exactly the value of an ES futures contract at a price of 2,100, close to where it was around writing this WO=$2,100x50x(1-0.252)=$78,540.

Roth IRA calc TablePart01
Example 1: Synthetic Roth IRA Parameter Assumptions
Roth IRA calc ResultsPart01
Example 1: Roth IRA vs. Synthetic Roth Returns

The synthetic Roth yields an after-tax return exactly 0.14% percentage points below the Roth IRA, irrespective of the market return. This return drag is due to two effects

  1. Not all the margin cash yields interest, because you need the $5,250 sitting around in cash, which doesn’t pay you anything in the IB account.
  2. The interest income on the funds actually invested is subject to income tax.

We can alleviate, maybe even reverse this drag by investing in slightly higher yielding assets (subject to adding a little bit of return variation vs. the Roth). For example, one could invest in a floating rate ETF from iShares (ticker: FLOT) for a little bit of extra yield. Current yield around 0.90%.

Roth IRA calc TablePart02
Example 2: Synthetic Roth IRA Assumptions: Higher yield investment

With the additional after-tax interest income, we’re able to beat the Roth IRA return by 0.18%:

Roth IRA calc ResultsPart02
Example 2: Roth IRA vs. Synthetic Roth Returns

Update (June 2017):

We implement this idea with options and/or on futures options (see more details in our Put Writing Post). It’s a slightly different strategy, not using the S&P 500 as the target, but the leveraging of the put options follows the same principle: scale up the leverage to overcome the tax drag from having to tax the option profits every year.

Also, in light of higher and rising interest rates, we need to jack up the yield we generate from investing the margin cash. We currently hold a Muni Bond fund and Preferred Shares. The yield on both easily overcomes the drag from taxes and the margin cushion. But we also introduce a certain degree of duration risk.

Advantages of the Roth replication

  • There is no upper limit on the account size. Except for how much money you feel comfortable putting into this. S&P500 futures are very liquid, with a daily trade volume of over 1 million contracts on most days (worth around $100 billion).

    AutobahnNoSpeedLimit
    A sign you love to see on the German Autobahn: No (speed) Limit!
  • Liquidity of assets outside the inflexible Roth IRA. No need to wait until age 59.5 and no need to wait 5 years to withdraw rollovers.
  • Futures contracts have a zero expense ratio. But there is a cost of trading and rolling contracts, which should amount to about 0.01 to 0.02%, still below the expense ratio of even the cheapest S&P 500 index fund. If you really want to skimp on the transaction costs you could also try to roll only twice a year, i.e., June and December, cutting that roll cost in half.
  • It’s easy to do other fun financial rocket science, such as covered call selling on your futures holdings, downside protection through puts or, if feeling really lucky, naked put writing (shorting) which is something we do very effectively. Another idea: mix in some 10Y Treasury Futures for some small extra yield and great diversification with the equity futures.
  • Keep your margin cash in a higher-yielding bond fund to further boost returns, maybe even a tax-free Muni bond fund to shield your interest income from the taxman. If you are willing to take on a little bit of credit and/or duration risk you could not only avoid any return drag but even beat the return of a Roth IRA invested in an S&P500 index fund. See example 2 above.
  • International index futures exist as well. Nikkei futures trade on the CME as well, and they even have a currency hedged version so you would take on only the pure equity return risk, while you are protected from currency fluctuations (which can be substantial between JPY and USD). I checked on iShares and they charge close to 0.50% expense ratio for those single country ETFs. With Futures contracts you can do the same for a fraction of the cost.
  • Your futures contracts don’t have to be rebalanced continuously the way the levered ETF do, which saves on transaction costs.

Disadvantages of the synthetic Roth through equity futures

  • Futures contracts require quarterly “rolling”, that is, selling the contract approaching expiration and buying a new contract with a later expiration date. It’s imperative that futures be held in an account where you can do this rolling in a cost-effective way. Some providers have very high fees that can make the whole exercise moot, but Interactive Brokers seems pretty reasonable.
  • Large minimum investments: the smallest denomination of one single S&P500 index futures contract is the CBOE e-mini contract at 50 times index value, currently over $100,000, as the levered-up amount. Probably around $70,000-$80,000 for most investors.
  • Depending on your tax situation and the level of risk-free interest rates there may be some performance drag relative to the actual Roth IRA. One can juice up the yield by holding higher yielding funds and mitigate or even reverse this drag into an advantage over the simple index fund in the Roth IRA, as mentioned above.
  • Limited selection of investments: S&P500 futures are very liquid, Dow Jones, Nasdaq 100 and a few international indexes as well, but anything exotic, like REITs or small cap index futures, doesn’t trade very much. Personally, we focus on the S&P500 and its options because there is a lot of liquidity and trading activity. Most other equity indexes are highly correlated and if we have demand for them, we just use our other accounts, IRAs, 401(k), etc. to hold them.
  • We need to fill out an additional tax form: IRS form 6781: Gains and Losses From Section 1256 Contracts and Straddles. It’s relatively easy: our brokerage company sends us a Form 1099, and there is no need to itemize the futures transactions. Rather, we simply enter one single number, the net profit or loss from Section 1256 contracts. That number will be split into 60% long-term gains and 40% short-term gains, to be added to our 1040 Schedule D (capital gains).
  • The potential loss of means-tested government benefits (e.g. Obamacare/ACA), college financial aid for your kids etc. This is a serious issue if it affects you. In our case, we plan to have taxable income in retirement too high for ACA subsidies. We have no idea yet what are the typical upper-income limits we want to avoid to maintain eligibility for financial aid for Little Miss ERN when she heads off to college. Looking for some advice from the blog community!
  • An actual Roth IRA is protected from the greedy fingers of ambulance-chasing trial lawyers. If someone sues you for damages, a privately-owned brokerage account for your replication portfolio is fair game in frivolous lawsuits. An LLC or a family limited partnership as a wrapper around your at-risk assets would offer some protection against greedy lawyers.
  • Marginal taxes aren’t really constant. When you increase/decrease your income you may go from one tax bracket to another. That means you will have to keep track of your year-to-date income (all income including your realized futures gains and losses so far) and potentially adjust the notional exposure if your tax bracket changes. For us personally, we are so far away from those kink points that we don’t have a problem. Our marginal tax rate is pretty much the same no matter what the market throws at us.
  • The dreaded $3,000 maximum of capital gains per year you can use to offset ordinary income. But the solution is that you can carry forward unused short-term losses to offset future gains. Even better: for Section 1256 contracts you can carry losses backward (!!!) to offset up to three years worth of prior gains.

Conclusion

Under the most basic assumptions, you’d currently experience a 0.14% p.a. return drag behind the actual Roth IRA. That’s still unpleasant, but some people are willing to sacrifice much more return for much less tax-arbitrage. For example, those who have maxed out all their tax-friendly accounts and go with “Deferred Variable Annuities” pay about 0.25% p.a. in fees even at the low-cost providers, like Fidelity. And with those annuities, you would only defer not eliminate income taxes. Before putting any serious money into a variable annuity, you’d definitely want to consider the synthetic Roth. What’s more, in the synthetic Roth you are able to juice up the interest income by simply investing in something with a slightly higher yield than the risk-free rate. Then you can even come out ahead of the Roth invested in an index mutual fund.

Has anybody else thought about the Roth IRA as a leveraged taxable account? We’re interested in hearing your suggestions from folks who have maxed out all traditional routes of tax deferrals and seek more tax-arbitrage!

88 thoughts on “How to create a no-limit Synthetic Roth IRA in a taxable account

  1. You should probably add a caveat that this method is only for people over the 15% tax bracket. Otherwise, the dividend and long term capital gains are taxed at 0%. Just like a roth

    1. Hi, thanks for stopping by and thanks for your comment!
      People in that tax bracket might still benefit from this if:
      1: they live in a state with high marginal state income taxes.
      2: they like the idea of splitting up the equity return into (Equity-RiskFree) and RiskFree, and then replace the RiskFree yield with a slightly higher yielding instrument, say a Floating Rate bond index or short-term credit fund. That can easily add a few dozen basis points over the equity return
      3: they like other fun stuff, like covered call writing or cash secured put writing with very low fees

      And also, for people in that tax bracket, it’s not so much that our procedure doesn’t work or doesn’t apply. It’s mostly that the Roth IRA is less attractive for them, because you might as well buy an equity fund in a taxable account.

  2. First time I hear about this. Are you sure this is legal?
    What do you do when you reach retirement and you have a close to zero % marginal tax rate?

    1. Everything legal. But please check the disclaimers and always check and make sure this fits into your personal financial situation.
      If you reach retirement you should sit on a very high tax-basis chunk of money because over the years your money grew and and you paid taxes out of your gross capital gains. In retirement you simply don’t use leverage any more because 1/(1-T_Fut)=1.0000.
      Or alternatively, you transfer this to a brokerage account at that tax-basis and invest in a mutual fund (no leverage). Then realize long-term capital gains (also at zero marginal tax) and live happily ever after. 🙂

      1. Ern, why does T_Fut = 0 in retirement? You still may have income from investment gains?

  3. I’ve also pondered the similarity between the leveraged taxable account and the Roth account, and I have a few considerations to add.

    1) Aren’t you overselling your proposal a bit? When you lever using futures/options, you pay the baked-in implied financing rate. And since you don’t explicitly pay this interest (it is baked into the price, so you pay it indirectly), it is not tax deductible. Generally you’ll pay about the risk-free rate for your leverage. Meanwhile, the “leverage” you get from your Roth is free.

    2) Aren’t you underselling your proposal a bit? If you lose money on your futures, you get to offset ordinary income at your higher ordinary income marginal rate. If you make money on your futures, you get additional income that is taxed at a weighted average of your short-term and long term rates which is less than you your ordinary income marginal rate.

    Ultimately, the $3k limit makes this scheme uninteresting for me. As an investor who realizes very few gains, I already have no trouble realizing enough losses to offset my gains + 3k. Come the next crash, I expect I’ll harvest enough losses to last me decades.

    I think there is a more general, more versatile strategy. First, let me establish that I think that the upside from the strategy you describe comes from the offseting power of losses. You take on more risk by leveraging, but it’s subsidized risk because if you lose money you can reduce your tax bill.

    So here’s a more general strategy: first, calculate how much in losses you’d like to generate (i.e. your realized gains for that year + 3k – any losses you’ve carried over from the past). The general idea is to make appropriately sized bets that have a pre-tax expected value of 0, and a positive post-tax EV (since you have losses that will offset your gains). It is essential that winning the bet does not cause you to realize a gain. One option is to buy deep out of the money options (if we assume the market is efficient, then these options are a bet with a pre-tax EV of ~0).

    E.g. let’s say SPY is trading at $200 and you buy a call option with a strike price of $250. If you lose the bet (i.e. your option expires with SPY < $250) then you realize a loss. If you win the bet (let's say SPY reaches $300), then you exercise the option and you're left with SPY with a cost basis of $300 + $1. You've won the bet and had an unrealized gain. You repeatedly make these bets until your losses (from worthless expired options) reach the desired loss amount.

    1. Hi, thanks for stopping by and thanks so much for your very thoughtful comments.

      1: we are not overselling our approach. First, our approach is only recommended for folks who have maxed out their Roth. We never intended this as replacing your Roth. Second, the margin cash is invested in interest bearing assets, so you do not lose the entire safe interest rate. You lose a little bit due to investing only slightly less than 100% of the portfolio and because you pay ordinary income tax on the interest income. In the baseline this drag is 0.14% p.a. But if you are willing to take on a little bit of credit and/or interest rate risk you can even beat the Roth by 0.18% when you achieve a slightly higher yield with your margin cash than the risk free interest rate used in pricing the Futures contracts.

      2: tax loss harvesting can be performed, but the impact is small once you implement this on a large scale. That’s due to the $3,000 maximum write-off per year as you stated. But the value of that TLH option might be enough to overcome the 0.14% performance drag in the baseline. Good point.

      Your approach to the tax arbitrage is quite intriguing. Do you have a blog where you wrote about this in detail? Do you want to write a guest post? Or a joint post with me?

      1. No blog yet, but I’ve been meaning to start one for ages. In the meantime, a guest/joint post sounds good. I’d also be interesting in talking some of this over with you if you have the time. I’ve commented using my real e-mail address this time so feel free to get in touch.

        Also, you have a good point in 1) regarding the ability to beat the rf rate when investing your margin cash. Seems like there’s a widely applicable (near?)-free lunch here: rather than invest $x in y, invest $x/c with a leverage ratio of c in y with an implied financing rate ~ the rf rate, and park your $x*(c-1)/c cash in a savings account that earns a risky return but is rf to you because of the FDIC. With CDARS this could be viable not just for an individual investor but also for a fund.

        However, I’m moderately uncomfortable with the prospect of keeping lots of the margin cash outside of the brokerage. I’m concerned about a worst-case scenario where a skittish or incompetent (or whatever) brokerage unwinds your leveraged position before you have a chance to send in money from your external interest bearing account. Brokerages can change margin requirements just about anytime I believe (do you know anything about limitations on this? who regulates this?), and sudden crashes are possible. What would have happened to this strategy during the 2010 flash crash? What about during an even worse flash crash? Ideally the brokerage would pay you interest on your margin cash so you could just keep your cash in the brokerage, or you could have your brokerage account linked in some way to the interest-bearing account where you park your margin cash.

        1. I sent you an email directly.

          “However, I’m moderately uncomfortable with the prospect of keeping lots of the margin cash outside of the brokerage.”

          That’s why you keep all the margin cash in the brokerage account (I use IB, Interactive Brokers) where the derivatives are held. Instead of using a Money Market fund I use a Muni Bond Fund, maybe around 70-80% of the margin cash. Pretty decent interest and tax-free and even a little bit of negative correlation with equities (very helpful in July 2015 and Jan-Feb 2016). The other 20-30% are just sitting idle as cash. Now the nice thing about this is that IB uses that 20-30% of margin cash plus three-quarters of the muni bond as “marginable” funds. So I never run into trouble with margin calls because I have many times more than the maintenance or even initial margin that way. With that, I never ran into issues. Not even close, not even intra-day on Feb 11, 2016, or August 24-25, 2015. 🙂

          1. Have you read Nassim Taleb’s books? The selling puts makes me think of his work. My understanding is that these bets can go very wrong during large downturns because the price goes down and implied volatility goes up at the same time making the losses much bigger than normal with potential of completely wiping out your whole portfolio. We have had some very low volatility the last few years. How would it done during much bigger crashes? I think there was one around 1987 that was very quick. Or maybe a time during the housing bubble crash?

            Also have you looked into tastyworks? They have very cheap prices for option writing.

            1. Selling puts would have worked extremely well during 2008/9. The drawdown of the put selling strategy was not as deep and the recovery was faster. While the equity market kept dropping, the implied vol and option premiums were already so rich that you hardly lost any money or made money already (because strikes were so far out of the money) during the second half of the crisis.

              tastyworks commissions are actually more expensive than IB. At least for ES options. Looks like $2.50 per contract plus exchange fees. IB charges $1.42/contract including exchange fees.

              1. Okay, so the worst case would be a quick downturn and then quick recovery, so you lose big on the sold put, and then the premium under performs the quick recovery. And Muni bonds interest would need to go up at the same time. Just thinking of worst case scenario, October 19, 1987, there was a 22% drop in one day. How would the strategy perform during something like this?

                Ah, because you are selling options on futures contracts. For stock options, it is only $1 to open and $0 to close a trade. I imagine you are just letting the trade expire each contract so you don’t need to worry about transaction costs on that side of the trade anyway though.

                What is the benefit of selling ES puts over selling SPY puts?

                1. Correct! You would get whipsawed in that scenario: quick unexpected drop and then a quick recovery. I answered a few questions about the Black Monday episode in the Options writing post part2.
                  Yes, for stock options tastytrade seems really competitive.
                  The benefit for me is that ES options are Section 1256 contracts on my tax return. Lower taxes! 🙂
                  But in general, you can trade ES, SPY and index options and they are almost interchangeable!
                  Cheers!

    2. ABABABABABABABABA: you mean cost basis of $250 + $1 if you exercise that (initially) DOTM option, right?

  4. Biggest drawback is only being able to invest in the index. You’re squabbling over basis points of tax benefits/costs when you can do much better than the S&P 500 . . .

    1. Thanks for stopping by. Remember what I’m doing here: I propose a way to generate the S&P500 returns after tax. But in a taxable account. I doubt you can achieve after-tax returns as high as the pre-tax returns of the index. You have tax inefficiency and trading costs through portfolio turnover. Heck, I even doubt that you’ll beat the market before taxes unless you show me some account statements. 😉
      But don’t get me wrong: I also would prefer to have more choices of investments. But it’s not in the cards. If I have the appetite for deviating from the index weights I can still do that in my real retirement accounts. But the contributions are all maxed out every year. In the Synthetic Roth, I am stuck with S&P500 futures and options on futures. I currently trade options and I do beat the index pretty consistently, see here:
      https://earlyretirementnow.com/2016/10/05/passive-income-through-option-writing-part-2/
      But that strategy is also quite a deviation from a straight index investment.

  5. I don’t know how old Little Miss ERN is, but I just calculated a series of BRK.B purchases compounded. $2000 invested in 1997 with a yearly $2000 investments till 2017 ($40K basis) yielded a near $5M account. Put it into a gift trust to minor account and never tell her about it. Maybe tell her when she is 50. BRK.B is invisible when it comes to taxes.

    I like this idea

  6. There were a few things I didn’t quite understand here, but the main one is the assumptions that losses in a roth are magnified and losses are decreased in a taxable. I assume your projection assumes offsetting capital gains but I wonder whether that is a fair assumption for most people who would implement this strategy.

    1. Well, the Roth losses are not magnified. They are merely not reduced by the ability to write off losses.
      My assumption in this exercise is that this is for an investor like me. This is not suitable for many other investors.

  7. I don’t know all the details of your investment style, but my superficial understanding is that you are overall an efficient market/boglehead buy-and-hold-index-funds-type investor. You know market correlations better than I, but they are overall highly correlated. I don’t know about you, but if I lose money in one place, I expect to be losing in other places. I certainly don’t expect to have offsetting capital gains. Perhaps you have a larger and wider-ranging portfolio than I. Likewise, you know that there’s only partial gain unless you were planning on selling the investment with capital gains in any case. Otherwise, there’s only step-up in basis which is worth something but can’t be considered fully gained.

    1. Granted! Equity markets are highly correlated, especially when the market goes down! But if you lose money on the futures transactions you have multiple options:
      a) you can write off up to $3,000 of capital losses against your ordinary income.
      b) you can carry forward your losses and reduce future capital gains
      c) Section 1256 can even be carried back (!!!) to offset past gains (see https://www.law.cornell.edu/uscode/text/26/1212) and get a refund for previously taxed gains.

  8. So getting closer to personally leveraging to invest (though I hesitate in the current bull market). The comparison to Roth IRA in particular doesn’t really concern me. I was considering trading e-minis in a taxable account to accomplish this as discussed here but wonder whether buying LEAPS may be a better idea. As far as I know, you can keep leaps long term and only pay long-term capital gains, so they may work better. I’m not super-knowledgeable with futures and options, so please share your thoughts on the comparison. Obviously other factors may work the other way (bid-ask spread on e-mini low, etc) but taxes seems to be a huge difference. Please share any thoughts on which of the two works better. Thank you!

  9. Hello!

    Correct me if I’m wrong but, compared to holding a levered portfolio of dividend-paying stocks or ETFs, by using futures contracts you also avoid the unfortunate situation in which your broker (e.g. IB) provides you with payments in lieu of dividends when they have lent your shares out for a period of time over which a dividend is paid. Those payments in lieu of dividends, of which are taxed as ordinary income. Please advise.

  10. This is really interesting but I have some questions.

    1. Wouldn’t there be some exposure to increases in dividends that you’d be missing by following this strategy?

    2. This post seems to have little to do with a Roth IRA and mostly be describing a general way to leverage index returns with minimal drag from margin interest. Replicating a Roth IRA is just one example of how to use this. As such, isn’t choosing your marginal tax rate as the amount to leverage by fairly arbitrary? What’s to say this is the ideal amount of leverage to use? For me personally, I’d be interested in using this to implement a lifecycle investing strategy where I use more leverage early on in my career and dial it back to maintain steady exposure to equities as my savings grow. Am I missing something here?

    3. If it’s possible to get a reverse drag by investing your cash in something with higher interest. Wouldn’t this strategy used to mimic index returns with no leverage be strictly preferable to actually buying the underlying security?

    Thanks!

    1. Good questions!
      1: Dividends are obvisouly priced into the Futures price. Certainly even estimated increases in dividends, otherwise someone would have found a way to arbitrage this already. Now, unexpected increases in dividends are clearly not, but so are unexpected decreases. In the end, the non-arbitrage condition should ensure that the two balance each other out.
      2: Yes, it has nothing to do with a Roth. That’s the point of the post.
      The leverage ratio is chosen to exactly replicate the Roth. THere is no guarantee that this is the actual desirable ratio. I could envision you investors with very little capital leveraging a little bit more (to replicate a >100% equity share in a Roth IRA) and older investors with a lower equity ratio in their portfolio with a little bit less.

      3: With no leverage desired, simply buy the index mutual fund from Fidelity in your taxable account and forget about leverage.

  11. Great idea about impact of taxes on leverage. One issue with leverage is that in case of a large drop (e.g. 75% plus), there is a risk that your account gets wiped out, albeit the risk is small. So you try to minimize this by using less leverage or don’t invest all your portfolio. The refund from tax can get out of this dilemma, but only if the refund is available relatively quickly after the loss. For futures, can one file for a tax refund right after a loss has been realized (and assuming their are gains in the last 3 years) and get a refund quickly? Or do we have to wait for the year end before filing for a refund?

    1. Great question!
      The 75% drop will not happen over night (at least not in the S&P500 – different story for other assets and individual stocks).
      You’d have to adjust the number of contracts over time to recalibrate the proper leverage ratio. If you start with 20 ES contracts and the market keeps falling and takes your principal down, you’d reduce the # of contracts to 19, then 18, then 17, etc. until you reach the bottom. And on the way up you’d raise the leverage levels again.

      1. Thanks for the response. If a drop and loss happens in January. Can one quickly file for a 1256 tax refund and expect to get some money by say June? Or do we have to wait till end if year to collect net losses for the year before we can file for a refund?

        1. I’m not an accountant but I doubt you can do that. You’d first have to get the tax form from the broker (early the next year) and then file your taxes by 4/15.
          You have the option to offset any taxable losses against gains for the current year and up to $3K against ordinary income.
          But after a large January loss you could probably already change your withholdings and/or estimated tax payments for that year so you do get some of the money back already during the current tax year.

  12. That is a good point that after a large January loss, you could get a beneficial impact relatively quickly by changing your withholdings or estimated tax payments all the way to zero if needed. However, you cannot have your estimated payments go negative, i.e. get a refund before the end of the year if you had a large loss; or at least we don’t know how to do it.

  13. Complete newbie to the derivatives game; I’m all-indexing, all-the-time at the moment, haha. But would it make sense to, on top of this strategy, sell covered calls on the futures contract(s)? Seems to me that collecting those premiums could provide some extra return, especially if you go with near-to-expiration weekly options like you do with your put writing strategy.

    I love your site, and I’ve learned so much from your articles! Thank you, and keep up the good work!

    1. Great idea! You could definitely do that! Enjoy the extra income. Or, alternatively use the call-writing income to simultaneously buy a but to hedge the downside. Or at least the downside to the leveraged equity portion.
      Feel free to mix with other strategies! 🙂

  14. Please ignore. I’m not sure how to subscribe to the comments besides making a comment so that’s what I’m doing.

  15. Big Ern – there was a comment above regarding the use of futures vs LEAPS for gaining leverage in a taxable account regarding tax efficiency. I too am curious on this topic. Any wisdom here?

    Any other advantages/disadvantages?

  16. Have you looked at the effect that compounding your taxes yearly will have vs. buy and hold? If you replicate SPX returns by using derivatives, and realize all of your gains annually, the gains you are left with after y years are:
    (1+r(1-MarginalTaxRate))^y -1
    Whereas buy and hold leaves you with:
    ((1+r)^y-1)*(1-MarginalTaxRate)
    And this drag only gets worse as you apply more leverage.
    One advantage of using levered ETFs instead of derivatives would be that you don’t compound the taxes.

    1. There is no compounding of taxes. The leverage of 1/(1-tax) will exactly offset the marginal tax rate:

      (1+r(1-tax)/(1-tax))^t -1=(1+r)^t -1.

      Also, leveraged ETFs use derivatives as well, so you’d get the exact same result, but with the added expense ratio.

  17. Hi ERN. I’ve read this and your other posts several times over several months and there’s one thing I still can’t wrap my head around – why do you subtract the risk-free return rate from your pre-tax futures return (R_EQ-R_RF)?

    -With an IRA, 100% of available cash is invested at an equities risk level. There’s no portion of the IRA earning at a risk-free rate.
    -With the futures, a levered nominal value is also invested at the same level of equities risk, but this time with a bonus lump of cash set aside that you can put into MM/bonds/etc. Rather than subtracting a RF return, it adds to the return and diversity of the portfolio.

    So why exactly do you subtract the risk-free rate? What am I missing?

  18. After reading this post I started trying to understand the fundamentals of the technique(s) being discussed. Before this month I’ve never looked into derivatives in any detail, so my grasp is still pretty limited.

    I’ve gotten myself hopelessly confused trying to understand the fundamentals how one is able to use futures to track the S&P 500, mirroring what someone could achieve by holding the fund inside a Roth vehicle, minus a small drag.

    The bullet list from the original post discusses purchasing futures calls (“go long two contracts” in the example). Later in the 2017 update it is stated that “options on futures” were used, and a link is provided to a post that discusses writing put options, collecting premiums to generate a positive return.

    These seem like two totally separate techniques, what am I misunderstanding? Or is the idea simply that one uses some form of leverage to boost returns while keeping a large chunk of money in municipal bonds or some other income generating vehicle to offset costs and provide a slush fund to handle any change in margin requirements?

    If someone were to “go long” on an ES futures contract per the original example, how are you supposed to select a strike price? Would you be selecting a strike price as close to the current price as possible? When it comes time to roll the contract, is the idea that whatever you sell the futures contract for is going to reflect the current underlying and generate a gain or loss that mirrors the underlying?

    1. Futures on the S&P 500 index replicate the S&P 500 index, by definition.
      Futures don’t require choosing a strike price.

      The put writing strategy I mention does not replicate the S&P 500. But the idea of using leverage is the same as with the synthetic Roth. You simply choose the leverage to overcome the drag from taxation. I never intended to imply that I the put option strategy is identical to the S&P 500 index. I changed the wording in that section to make it not ambiguous.
      Thanks for pointing that out! 🙂

      1. Thanks so much for the reply and clarification!

        My confusion over a strike price was due to my misunderstanding that you were talking about buying the futures in the original article, not options on futures (my incorrect inference from the 2017 update prior to the clarification).

  19. Hi ERN,

    I just want to clarify something with you that might seem obvious for leveraging with futures vs a margin loan. If we assume similar margin rates across the two for a low-cost broker (like IB) is the main difference that for futures you don’t have to pay back the principal on the leverage?

    Is there a detailed assessment you’ve come across (or written about) clarifying these concepts?

    It sounds like the debt is being sold to someone else when you sell the future and realize the gain.

    1. The margin loan is usually more expensive. Even at IB it’s still ~1.75-2.0%, while the futures use the true non-arb implicit margin rate equal to the short-term cash rate, somewhere around 0.25%.

      There is no debt arrangement with futures. It’s a pure derivative transaction where the buyer and seller transfer money equal to the multiplier times the change in the price of the underlying every trading day (“Mark-to-Market”).

  20. Strategies like this and put selling require large sums dedicated to them. Is there an optimal way to transition into using them? The problem is not strictly account value, but rather an account which is already fully invested in stock indices. Should one eat the LTCG and swap in a day? Begin by building up the margin cash / bond portfolio with new investments? Or begin with a futures contract and minimum required margin cash? I’m young and loathe to greatly reduce my equity exposure for an extended period of time while building up funds for a new / additional strategy.

      1. I already have a 6 figure account, but it’s invested fully in stock index funds. For a synthetic Roth or the put options strategy, the margin cash should be held in safer assets like bonds. I was curious if there is any best way to switch to a margin based strategy without have to force capital gains while minimizing the time the returns would be dampened by accruing cash or bonds up to a sufficient quantity, i.e. how could a $200k portfolio fully invested in VTSAX efficiently transition to a synthetic Roth or put selling strategy? Is the only answer: build up bonds and cash until you have ~$100k of safer assets for margin?

        Appreciate your insight,
        Phil

        1. Interesting problem you have there. One could estimate the tax drag from this. Say 20%. Then hold a smaller size futures position to make up for that drag. Say, if you pay a tax rate of TF on the futures, then hold an additional 20%/(1-TF) on S&P500 futures on margin.

        2. Not sure if brokers treat VTSAX differently than the same etf (VT) for margin purposes but I just kept my VT but added micro-ES futures (5x) as needed. For instance assuming they treat VTSAX as marginable collateral for futures and you wanted to transition to ~1.2x leverage you could buy 2 MES futures and you’d have $200k+$38k = $238k/$200k or 1.19x leverage. For further additions to your portfolio you can add bond etfs and start adding the futures contracts in 1 lot of future chunks ($19k).

  21. Any reason this strategy wouldn’t work with the Micro E-Mini S&P 500 future contracts? Would seem like you could replicate with 1/10 the portfolio size, or get better alignment with the portfolio since Micro E-Mini is $5 vs. the Mini being $50. Is there something I am missing here?

      1. Appreciate the response! It looks like at least on IB the commissions are reduced for the Micro compared to the Mini, but certainly not 1/10. I think the big appeal to me is the ability to replicate pre-tax returns – saving that extra 15% (assuming all LTCG) should ideally offset most drag.

        You didn’t touch on roll cost when rolling over the contracts quarterly – is this generally nominal or could this also have potential drag? At the current price difference from March to June (selling at 3,742, buying at 3,732) it looks like it’d be a 1% annualized impact to returns.

        Again, appreciate the insights!

        1. The bid/ask spread looks excessive. Make sure you use the b/a spread during normal trading hours. It should be back at 0.25.
          At least for the e-mini contract (50x) I once computed that the annual commissions plus roll costs are low:

          Commissions: $2.50 one way, 4 contracts times 2 = $20
          B/A spreads: You can do the IB combo trades (sell near, buy next out) in one single transaction, with one single B/A spread (also in 0.25 steps). If the roll-cost is half the B/A spread ($0.125) that’s $6.25 per roll, so another $25 per year.

          $45 per year is a little bit above 0.01% for a contract worth $380,000. Cheaper than a Vanguard fund!

          1. Thanks for the additions detail. The B/A spread was around 10 points all day Friday too – could it be related to the longer time horizon on the June contract and pessimism around the S&P500?

            Or could it be that it’s still a week or two until everyone starts rolling contracts forward? I’ll see what happens over the next 14 days to see if it starts to contract before 3/19. If it doesn’t, this would be an added cost to the strategy, but still would outpace after-tax returns for about 1.3X leverage. Is that a fair way to think about it?

            Appreciate the dialogue!

            1. It’s still too early to roll to the June contract, one indicator of liquidity is to look at the volume traded, which on Friday was 2.2million March vs. 14k for June, it’s better to wait until this gets a little more even, which will happen closer to expiration. Also, not sure if you knew this but to roll a long March contract from March to June you can trade a futures spread contract (sell March & buy June) in one transaction which usually has a 0.25 B/A spread especially when it’s time to roll.

              1. Thanks for the response, was monitoring volume so figured it was too early to roll. Did not know about the futures spread contract so that’s super helpful. It still shows the 10 point differential but I’ll monitor to see how that changes over next two weeks.

                1. What time of day were you checking it? I just checked at ~8:45am eastern and the Mar/Jun future spread is 10.00/10.05 or a nickel wide!

                2. Maybe I’m looking at the wrong thing then? I was looking at the leg details which had the ~10 point differential. Now when I just check the ES calendar spread it’s b/a 10.50/9.75, so about 0.75 or about $37.50.

                  Is that the right place to be checking?

                  Again, appreciate all the help while I try to figure this out!

                3. Currently, the June MES contract has a spread of 0.50 with very low bid/ask sizes (1, sometimes 2 or 3). One might have to wait until the March contract goes closer to expiration. But even today, 0.50 spread is better then a 10.00 spread! 🙂

                4. The leg details are the underlying which have a difference of 10 points right now, last I checked the Mar/June spread for MES Bid/Ask is around -10.10/-10.05 and ES is similar, meaning if you have an March contract now, you’d collect 10.05 ticks if you paid the spread ask price (only giving up 0.05 bid/ask) to roll your March to June, the calendar spreads can trade inside the regular 0.25 tick increments and instead trade in 0.05 increments. After you execute the spread trade, they tell you what leg prices you bought/sold based on your spread price.
                  There are a number or reasons that different months trade at different prices, carry costs, dividends, etc.

                  See #5 regarding spread price increments in
                  https://www.cmegroup.com/trading/equity-index/eminifaq.html

                5. Maybe the confusion stems from mixing up b/a spread with the slope of the term structure. There will likely always be a large spread between the March and the June contract, due to the dividend yield vs. risk-free rate spread. Currently the March/June contracts are indeed around 10 points apart. But that’s natural. It’s a non-arbitrage condition.
                  The b/a spreads for the contracts are indeed lower during trading hours. Usually 0.25 to 0.50.

        2. Mystery solved. See my comment below from earlier today. the 10-point spread you mention is not the b/a spread. It’s the roll yield that compensates you for the missing dividend income.
          The b/a spreads for the individual contracts are 0.25 for the near contract and 0.50 for the June contract.

          1. I think I understand it now! The appropriate way to measure the roll cost is the difference in the b/a spread between the March and June contracts. I was missing that the 10-pont spread between March and June is for the roll yield and absent that would create arbitrage opportunities.

            Appreciate the patience!

  22. What combining the “no limit synthetic roth” idea with the put option writing strategy in the same brokerage account, but instead of borrowing from the broker at 1.5%+ for the equity position and holding bonds against the put selling position just borrow from “yourself” and not carry a margin loan?

    For example assuming your tax rate is 22% and you are targeting 1/(1-.22)=1.28 leverage ratio:
    Portfolio A:
    $632k in equity etfs @ 1.28 leverage ($139k of which is borrowed on margin from IB at 1.5%)
    $139k in bonds against your 3x leverage put selling strategy selling 1 SPX put

    Portfolio B:
    $632k in equity etfs (0 borrowed on margin)
    $0 in bonds against the 3x leverage put selling strategy selling 1 SPX put

    With portfolio B you likely will get better risk adjusted after tax returns by paying off the margin loan than your bond holdings in portfolio A and don’t have to pay taxes on taxable bonds or hold muni’s which tend to pay lower interest for comparable risk. You also don’t have to deal with the extra taxes of payment in lieu of dividends on your equities portion since you aren’t borrowing them on margin.
    I know you can create the equities portion of the portfolio with futures along with holding bonds against them but one advantage of setting it up similar to portfolio B is you can take advantage of being at a likely lower tax bracket in retirement when you realize your equity etf gains. Futures are realized each year and you could be paying 15-25% (even with the 60/40 tax treatment) on your gains while working whereas many retirees are in the 0% in long term capital gains tax bracket when they start selling their equities.

    1. Yes, good points. i never recommend using a leveraged portfolio with borrowing at the broker interest rate. At IB it’s “only” about 1.5% above market rate, but other brokers really take you to the cleaners.
      When I propse leverage as in the Synthetic Roth, I always advise to use futures or futures options of SPX options. They are S.1256 contracts, they have an “implicit” borrowing rate close to the 3m Tbill.

  23. Very true, another one benefit of doing it with the portfolio “B” method is you have better control of your AGI in retirement from year to year which can help with qualifying for certain programs such as ACA subsidies. In a strong equities year with 20%+ equity gains, one might have hundred’s of thousands of dollars in taxable income if they are holding futures and might miss out on lots of tax breaks that are available to those with more moderate taxable incomes.

Leave a Reply to Bennett Cancel reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.