Structured Products and Autocallables: How They Are Built, Priced, Hedged, and Where They Fail

A structured product is a note issued by a bank whose payoff is built from a bond and one or more options on a stock, index, or basket. An autocallable is the most common modern form: it pays a coupon while the underlying stays above a barrier, repays early (“calls itself”) if the underlying is at or above a set level on an observation date, and exposes the investor to the full loss of the underlying if it ends below a lower barrier. The investor is, in effect, selling a put option to the bank and receiving the premium as coupon.

In short: almost every retail structured product is a combination of three parts: a zero-coupon or funding leg from the issuer, an option the investor buys (upside participation) or sells (downside risk), and a fee. Reverse convertibles and autocallables have the investor short a put, often a barrier put that only activates after a large fall. Issuers hedge the resulting option exposure in the underlying market, and that hedge changes sharply near barrier levels. US regulators (FINRA Regulatory Notices 09-73, 10-09, 12-03, and 22-08; a 2011 SEC staff report) have repeatedly flagged suitability and disclosure problems, and the 2024 losses on Hang Seng China Enterprises Index linked products in Korea, documented by the Financial Supervisory Service, showed what happens when a widely sold barrier is breached.

This article describes how professional issuers build and hedge these products. It assumes you know what calls and puts are (see put options explained) and how barrier and digital options work (see exotic options explained). It describes mechanics, not recommendations.

The misconception: “the coupon is interest”

A 7% or 8% coupon on a note from a large bank looks like a bond yield. It is not. A bond pays interest because the issuer borrows your money. The extra coupon on an autocallable is mostly option premium: payment for taking on the risk that the underlying falls through a barrier. That is why the coupon rises when volatility rises, why it is higher on a single stock than on an index, and why “worst-of” notes on several indexes pay more than notes on one. The premium is compensation for a real, if infrequent, loss.

The product family

Principal-protected notes

A principal-protected note combines a zero-coupon bond (which grows to the face value at maturity) with a call option on the underlying. FINRA Regulatory Notice 09-73 (December 17, 2009) reminded firms that “these products are not risk-free, and their terms and structures can be complex,” and that “most importantly, the principal guarantee is subject to the credit-worthiness of the guarantor.” The protection is a promise by the issuer, not a deposit guarantee.

Reverse convertibles

FINRA Regulatory Notice 10-09 (February 12, 2010) described the structure directly: “the investor in the reverse convertible is selling the issuer a put option on the reference asset,” in return for a higher coupon. Many have a knock-in or barrier level: the put only becomes active if the reference asset falls below it. FINRA called reverse convertibles “complex investments that may be difficult for investors to understand” and suggested firms consider limiting sales to accounts approved for options trading.

Autocallables

An autocallable adds an early-redemption feature to the reverse convertible. On each observation date, if the underlying is at or above the call level, the note repays principal plus coupon and ends. If it is never called, the final payoff depends on whether the barrier was breached. Variants include “step-down” call levels (common in Korea), memory coupons that catch up missed payments, and worst-of notes linked to the lowest performer of several indexes.

The regulatory view

FINRA Regulatory Notice 12-03 (January 17, 2012), on heightened supervision of complex products, listed structured notes with embedded derivatives, “worst-of” notes, reverse convertibles, range accrual notes, and products with conditional or partial principal protection, and said complexity “presents an additional risk to retail investors.” Regulatory Notice 22-08 (March 8, 2022) returned to the subject, naming newer products such as reverse convertibles, volatility-linked products, and defined outcome ETFs with buffers and caps among complex products, and requesting comment on account approval, knowledge checks, and other guardrails.

The SEC’s Office of Compliance Inspections and Examinations reported in July 2011 (“Staff Summary Report on Issues Identified in Examinations of Certain Structured Securities Products Sold to Retail Investors”) that US sales of structured securities rose from about $32 billion in 2004 to more than $100 billion in 2007, that the Lehman Brothers bankruptcy produced defaults on structured products it had issued, and that retail sales were estimated at $34 billion in 2009 and $45 billion in 2010. Examinations of eleven broker-dealers found unsuitable recommendations, secondary-market trades with retail investors at disadvantageous prices, and supervision that often failed to treat these products as a distinct class.

A real term sheet

Pricing supplements filed with the SEC under Rule 424(b)(2) show the mechanics in plain view. One example (described for its structure only): JPMorgan Chase Financial Company LLC, Auto Callable Contingent Interest Notes Linked to the S&P 500 Index due June 11, 2029, pricing supplement filed June 10, 2025, guaranteed by JPMorgan Chase & Co.

Term	Value
Initial value (S&P 500 close on June 6, 2025)	6,000.36
Interest barrier and trigger value	70% of initial, 4,200.252
Contingent coupon	7.30% a year, paid 1.825% a quarter if the index closes at or above the barrier on that review date
Automatic call	From the fourth quarterly review date, if the index closes at or above its initial value
At maturity, if not called	Full principal plus the final coupon if the index is at or above the trigger; otherwise principal falls one for one with the index
Estimated value at pricing	$977.20 per $1,000 note

Three features generalize to almost every product of this type:

The loss is not cushioned once triggered. Above 70% you are repaid in full; below it you lose the full decline from 100%. A fall to 69% costs 31%, not 1%. The supplement states that below the trigger “you will lose more than 30.00% of your principal amount at maturity and could lose all of your principal amount.”
The estimated value is below the issue price. The supplement explains that it is the sum of a debt component valued at the issuer’s internal funding rate and the value of the embedded derivatives, and that the issue price also includes the projected profit from hedging and the estimated hedging cost. The $22.80 gap on a $1,000 note is the issuer’s margin and costs, before any distribution fee. (This issue was sold to fee-based advisory accounts with no selling commission.)
Credit risk is the issuer’s. Every payment depends on JPMorgan Financial and the guarantor being able to pay.

Worked example 1: pricing a step-down autocallable

To see where the coupon comes from, we priced a hypothetical three-year step-down note of the Korean kind on a single index, by Monte Carlo simulation (60,000 paths, daily steps, 3% interest rate, 3% dividend yield, no issuer credit spread):

Six semiannual observations with call levels of 90%, 90%, 85%, 85%, 80%, and 75% of the starting level
If called, the investor receives 100 plus 5% a year of coupon for the time elapsed
Knock-in barrier at 50%, observed on every daily close
At maturity, if never called: 115 (principal plus three years of coupon) if the index is at or above 75%, or if it never closed below 50%; otherwise the index level (a loss one for one with the index)

Index volatility assumed	Fair value of a 100 note with a 5% coupon
20%	99.65
25%	97.11
30%	94.48

At 25% volatility, the coupon that makes the note worth exactly 100 is 8.92% a year. Selling it at a 5% coupon leaves the issuer about 2.9 points of value per 100 to cover hedging, distribution, and profit. Two facts follow. The note loses value as volatility rises, so the investor is short volatility and the issuer is long it. And the coupon investors see is set in large part by implied volatility on the underlying, which is why coupons on these products jump after a sell-off. Options usually trade above the volatility that later materializes (see the variance risk premium), which is the economic case for selling them and also the reason the occasional loss is large.

How issuers hedge

The issuing bank, having effectively bought a barrier put (and sold some digital coupons and early-call features) from its clients, runs the position on its derivatives book:

Delta hedge. It holds the note’s delta in index futures, the index ETF, or the stocks. Because the investor’s payoff behaves like a long position in the index near the barrier, the issuer’s hedge is a long position that it adjusts as the index moves.
Vega. Being long the embedded options, the issuer is long volatility. Banks that issue large volumes of these notes often sell volatility elsewhere (listed options, variance swaps) to offset it. The BIS argued in its March 2024 Quarterly Review that dealers who buy options from structured-product investors hedge by “buying when the index goes down and selling when it goes up,” which dampens price moves.
Dividends and correlation. The issuer’s hedge leaves it exposed to dividend changes (the note ignores dividends) and, for worst-of notes, to correlation between the indexes.
Barrier risk. Near the knock-in level, the hedge stops being smooth.

Worked example 2: the hedge near the barrier

Using the same simulated note, halfway through its life (1.5 years elapsed, no call yet, 25% volatility), here is the note’s delta at different index levels, and the long index exposure the issuer holds per 100 of face value to hedge it:

Index level (start = 100)	Knocked in?	Note delta	Issuer’s hedge (index exposure per 100 face)
100	No	0.12	12
80	No	0.65	52
65	No	1.66	108
55	No	2.35	129
52	No	2.49	129
49	Yes	1.16	57

As the index falls toward the barrier, the issuer’s long hedge grows to more than the face value of the note, because the investor’s loss jumps once the barrier is touched. The moment the index closes below 50, that cliff is gone and the hedge falls back to roughly one times the note’s value. The issuer must sell about 72 of index exposure per 100 of face value into a falling market. Multiply that across a few trillion won of notes with similar barriers and you have hedging flow that amplifies the move. The BIS made the same point in a box on leveraged products and Korean chip-stock volatility in its September 2026 Quarterly Review, noting that hedging autocallables “can require large, destabilising trades near the barriers.” Dampening when the index is far from the barrier, amplifying near it: the same product does both. More on the general mechanism in dealer gamma and market impact.

Case study: the 2024 Hang Seng China Enterprises Index ELS losses

In Korea, equity-linked securities (ELS) are the dominant retail structured product. Many were linked to the Hang Seng China Enterprises Index (HSCEI, called the H-index in Korea), usually alongside other indexes. The Financial Supervisory Service published its preliminary inspection results and dispute-mediation standards on March 11, 2024 (press release “홍콩 H지수 ELS 검사결과(잠정) 및 분쟁조정기준(안)”, with an attached report and Q&A). Its figures:

Outstanding H-index ELS (including trust and fund wrappers) at the end of December 2023: 18.8 trillion won in 396,000 accounts. Banks sold 15.4 trillion won (81.9%), securities firms 3.4 trillion won. Individuals held 17.3 trillion won.
Investors: 84,000 accounts (21.5% of individual accounts) belonged to investors aged 65 or over. Bank sales were 90.6% offline, in branches; securities firm sales were 87.3% online.
Maturities: 15.1 trillion won (80.5%) were due to mature in 2024, concentrated in the first half.
Losses: of 2.2 trillion won maturing in January and February 2024, losses were 1.2 trillion won, a cumulative loss rate of 53.5%. Assuming the index stayed at its end-February level of 5,678, the FSS projected a further 4.6 trillion won of losses, about 5.8 trillion won in total.

The FSS inspected eleven sellers (five banks: Kookmin, Shinhan, Hana, NongHyup, and SC; six securities firms) between January 8 and March 8, 2024. It reported that sellers raised sales targets during a period of rising index volatility, designed branch performance indicators that rewarded ELS sales (one bank counted the product’s coupon as branch performance as long as no knock-in had occurred, regardless of how far the index had fallen), and in one case loosened an internal limit from 50% to 80% of the quarterly target and then added an exception limit of 2.5 trillion won. It found sales systems that allowed sales to investors whose stated loss tolerance was under 20% or who wanted short-term investments, omitted or distorted loss scenarios, and branch staff completing risk questionnaires on investors’ behalf.

The dispute-mediation formula set compensation as a base rate of 20% to 40% (depending on which sales principles were breached), plus 3% to 10% for seller-level control failures, plus up to 45% and minus up to 45% for investor-specific factors (age, purpose, prior ELS experience, amount), plus or minus 10% for other factors, giving a possible range of 0% to 100% of the loss. The FSS said it expected most cases to fall between 20% and 60%. In May 2024 the dispute mediation committee applied the standards to five representative bank cases, with compensation ratios reported at 30% to 65%. As of May 2026, the Financial Services Commission had not finalized sanctions against the sellers and had asked the FSS to supplement its findings.

Worked example 3: one note on the actual index path

Take a hypothetical H-index ELS with the terms from Worked Example 1 (call levels 90/90/85/85/80/75, knock-in at 50%, six-month observations), struck on February 17, 2021, the HSCEI’s highest close of that period. Using the index’s actual daily closing levels:

Observation	Date	HSCEI close	Percent of start	Call level	Result
Start	Feb 17, 2021	12,228.63	100%
1	Aug 17, 2021	9,057.88	74.1%	90%	Not called
2	Feb 17, 2022	8,711.40	71.2%	90%	Not called
3	Aug 17, 2022	6,755.31	55.2%	85%	Not called
4	Feb 17, 2023	6,991.27	57.2%	85%	Not called
5	Aug 17, 2023	6,292.33	51.5%	80%	Not called
Maturity	Feb 16, 2024	5,558.86	45.5%	75%	Loss

The index first closed below the 50% knock-in level (6,114) on September 28, 2022, at 5,958.62, and its lowest close was 4,938.56 on October 31, 2022, 40.4% of the starting level. At maturity the investor receives 45.5% of principal: 4,545,775 won on a 10 million won investment, a 54.5% loss, close to the 53.5% average loss rate the FSS reported for products maturing in early 2024. The note paid no coupons, because the coupon in this structure is paid only when the note is called or matures above the barrier.

What can go wrong

Barrier cliffs. Losses are not proportional to the shortfall below the barrier. They are measured from the starting level.
Path dependence. Daily barrier observation means one bad close during the life of the note can matter even if the index recovers.
Worst-of exposure. A note on three indexes pays according to the worst one. The coupon is higher precisely because the chance that at least one breaches is higher.
Issuer credit. FINRA’s 2009 notice and the SEC’s 2011 report both note that protection and coupons depend on the issuer, and the 2008 Lehman Brothers bankruptcy produced defaults on structured products it had issued.
Liquidity and value. The estimated value disclosed at issue is below the price paid, and any secondary-market bid comes from the issuer or its affiliate.
Sales incentives. The FSS findings and the SEC’s 2011 examinations describe the same pattern in two countries a decade apart: sales targets that rose as risk rose, and suitability checks that did not stop unsuitable sales.
Concentrated hedging. When many notes share the same underlying and similar barriers, issuers’ hedges become a single large flow near those levels.

The short version

Structured products repackage options inside a bank note. Principal-protected notes give the investor a call; reverse convertibles and autocallables have the investor sell a put, often a barrier put, in exchange for a coupon that is mostly option premium. Issuers hedge by holding the note’s delta and managing the volatility they have bought, and that hedge behaves very differently near the barrier, where it can force selling into a decline. US regulators have flagged these products as complex for more than a decade, and the Korean H-index losses of 2024, about 5.8 trillion won projected by the FSS, show the payoff working exactly as written on a market that fell more than half.