Still Buffering: Options-Based Funds Gathering Traction and Critics

In this issue’s column, we look at a hotly contested derivative strategy known as the defined outcome fund, or more colloquially the buffer fund. Buffers typically comprise both downside protection and rolling contracts and have enjoyed mainstream retail investor adoption. Buffers are having a moment and are attracting many managers, including ARK fund manager Cathie Wood1 who has made submissions to the SEC for a downside-protected equivalent of her Innovation fund. Reuters recently covered the growth in the space, suggesting $70 billion in ETFs and mutual funds in almost 350a individual funds.

Buffers come in many flavours and serve different risk assets, including Bitcoin. As of August 2025, the leader in volume and net cumulative flows is an S&P 500 participation fund-of-funds with 10% downside protection by First Trust Vest (ticker and hereafter: BUFR).

BUFR’s approach is similar to the oft-quoted S&P 500 Cboe Zero-Cost Put Spread Collar Index (CLLZ). The Cboe index mirrors the performance of S&P 500 exposure hedged with a put spread funded by an OTM call.2 BUFR adapts the idea with a deep ITM call instead of buying the index (stock replacement) downside-protected with a 15% OTM call that funds a 10% OTM put spread. And it does this every month--BUFR is 12 funds wearing an ETF trenchcoat. In a ladder, the 12 ETFs held in roughly equal weighting represent each month of the year, rolling the average downside protection and cap. These monthly holdings are described as "series" or "vintages". For this reason, we take Reuters’ headline-grabbing funds tally with a pinch of salt. SEC N-Port reveals 276 ETFs with a defined outcome strategy. And bucketing them into these vintages, where each January–December (or variations therein) ETF is counted as a vintage of one fund, rather than as separate funds, reveals a more sensible figure than 350: 37. There are 37 vintage families of ETFs from First Trust Vest, AllianzIM, PGIM, Innovator, Pacer ETFs and iShares. Alliance Bernstein and Goldman Sachs have buffers with three-monthly vintages. Invesco has buffer funds that are not ETFs.

However, popularity is not performance. Buffers have come under sharp scrutiny by Cliff Asness and Daniel Villalonb at AQR. In their article “Rebuffed”, published in the Journal of Portfolio Management, the authors suggest that buffers underperform a comparable portfolio of stocks and cash, failing to beat benchmark cumulative returns and drawdown protection.

The underperformance of the category is a disputed claim.3 It somewhat defies expectations. How are buffers, prima facie a well-hedged alternative to stocks and cash, underperforming stocks and cash? Ask Jeff Chang, President of Vest – an early buffer innovator – and the answer will be that they do not underperform an appropriate benchmark. And AQR, Chang says, are wrongly comparing buffers to equities, and are including funds in their comparison set which carry out different strategies entirely.4 "Derivative Income funds, for example, use covered-call strategies for yield, not protection. Including these in the same sample distorts any conclusions. It’s incomplete, outdated [and] misleading." AQR and Vest have been going back and forth. Vest writes blog posts, and AQR retorts in podcast appearances. I, for one, enjoy the debate between fund providers with different points of view. We all benefit from the revelations and clarifications they publicly provoke each other into giving.

One critique the providers would probably accept is that buffers can be hamstrung by a mandated lack of flexibility, the presence of skewness and put-call parity. The rigidity can be explored by comparing BUFR with CLLZ. The Cboe index is priced to enable the put spread collar to always be zero cost. Reviewing SEC-reported holdings of BUFR vintages reveals there are typically only four option contracts (plus cash)5: two puts and two calls. Disregarding the deep call (the S&P 500 exposure), there is one call and two puts: a sold OTM call and a bought put spread. This is similar to CLLZ. But there has never been an instance (as far as SEC records show) of BUFR making the same decision to sell a weighted average of calls. One could argue this isn’t needed as FLEX options are written at the fund’s behest and can offer precise strike prices, a point noted in a recent Bloomberg articlec. However, 12 times out of 12 over the past year, each vintage’s put spread was costlier than the premium collected on the OTM call.6 Maybe this decision to not make a weighted average of calls is because the weighted average can only correspond to premium cost, not to moneyness. It would do BUFR no good to have a neatly zero-cost put spread collar that either did not offer a cap at 15% or downside protection up to 10%. So, the fund’s mandate places handcuffs on what CLLZ does more dynamically. So, if the put spread is not free, how is it financed? Mostly through the OTM call, with a little help from the deep call. Skew giveth and it taketh away. The OTM call is cheaper than the put spread, but the paucity of premium on the collar is compensated for by the embedded leverage of the ITM call. The S&P 500 exposure in BUFR vintages is rarely exactly 100%. In the last 12 months it has bounced around from  98.64-107.39%, as of August 20, 2025. In short, BUFR doesn’t need $100 to give $100 of participation, that extra dollar or so covers the outstanding cost of the put spread not paid for by the OTM call. Then, on performance itself, perhaps Sharpe is not the only measure to consider. Our analysis of buffers bucketed into vintages shows a more nuanced trade-off between Sharpe and Sortino ratios. On average, Sharpe differences versus benchmarks remain slightly negative (mean -0.004, median -0.055), possibly due to the capped upside participation. Sortino differences are positive (mean +0.08) with a negative median (−0.044). At the extremes, the AllianzIM Buffer10 Allocation ETF underperforms the set (Sharpe diff −0.32, Sortino diff −0.48). In contrast, the Innovator Growth 100 Power Buffer ETF stands out with the strongest improvements (Sharpe diff +0.51, Sortino diff +1.89). Timing is also germane to these metrics, which are based on all available data and the funds launched on different dates in different market environments. Overall, these findings reinforce the context dependence of buffers: while they often lag simple benchmarks in terms of Sharpe efficiency, Sortino ratios demonstrate their potential appeal to risk-averse. The category is not a monolith, and the products cannot easily be considered as a group—design choices determine how much upside is sacrificed and how much downside is protected. Haphazard category or index-level grouping is the crime here, not any party’s analysis.

Going forward, the combustible effects of mandate rigidity, skew and the ever-present absence of equity put-call parity will continue to drive dynamic vertical spread design in funds like BUFR. On "Liberation Day" policy uncertainty directly translated into high sustained volatility.7 BUFR’s April vintage, rolled after Liberation Day, carried both its most expensive put spread and its most leveraged exposure to the S&P 500. These dynamics prompt a question: downside protection is coveted, but what, if any, are the hidden costs? How to price a 10% downside buffer? We could implement advanced models to show the fair price of the put spread. However, it is not always necessary. The price of the put spread $P(K_1)-P(K_2), 0≤K_2≤K_1$ is obviously bounded:

$0≤P(K_1,T)-P(K_2,T)≤e^{-rT} (K_1-K_2)$

where are the strike, is the time to maturity and is the risk-free interest rate. Adding more, but not excessive conditions, i.e., the underlying follows some Itô-processes with locally bounded drift coefficient, and diffusion coefficient is locally Lipschitz-continuous, we may have a more precise upper bound based on P. Krühner and S. Xu’s resultd. On the other hand, we may measure the put spread using the OTM call. While there is not a great deal of literature to address this specific question, we are trying to identify "stickiness", or to understand the bounds of relative cost along the moneyness and the interplay between puts and calls. An intuitive observation comes from the Black-Scholes model. Taking the put spread BUFR uses (long ATM put, short 10% OTM put), once the risk-free interest rate, time to maturity, underlying’s price and volatility are given, the strike of the OTM call, whose premium offsets the put spread, is implied. Since the OTM put is massively underpriced by BS, the implied strike curve given below is nothing but a lower bound of the actual strike.

Typically, SPY’s implied volatility is around 12%-18%, which is a rather flat area, as we can see from the curve. This might reveal some sort of "stickiness" in the strike relationship.

However, performance is not popularity. AQR’s research notwithstanding, buffers are a growing product category that tangibly seems to offer retail investors peace of mind. A comparison of net flows around times of distress – and we’ve had a lot of market distress lately – shows SPY panic selling is significantly higher than that of SPY when housed in a buffer.8

Buffer investors hold when SPY investors sell. And if peace of mind is driven by belief in one’s buffer, and that belief leads to ongoing participation in risk markets, some mild "underperformance" is perhaps a price worth paying.