RILA buffers and floors: how an index loss lands on your account

Two registered index-linked annuities can both promise "protection from the first 10% of losses" and treat you very differently when the index falls. One absorbs that first 10% for you. The other lets you take it and then catches everything below. The labels are buffer and floor, and the choice between them decides whether a modest decline costs you nothing or something, whether a severe decline costs you a lot or a limited amount, and which of the two the insurer charges you for in the form of lower upside.

This comparison uses a round illustrative loss and a $100,000 account to show how each design works in dollars. The numbers are invented for clarity, not real products or forecasts, and they ignore fees and taxes. The overview of the product is in the guide to Registered index-linked annuities (RILAs): buffers, caps and risk; here the focus is the mechanics of the downside.

The two designs in one paragraph each

A buffer is a first-loss shield. The insurer absorbs the first slice of an index decline, up to the buffer percentage, and you bear the losses below that. With a 10% buffer, a 10% decline costs you nothing, and a 25% decline costs you 15%.

A floor is a last-loss limit. You absorb the first slice of any decline, down to the floor percentage, and the insurer absorbs everything beyond it. With a floor of −10%, a 25% decline costs you 10%, and a 4% decline costs you 4%.

The same words, "10% protection," therefore describe opposite arrangements. In the buffer, the 10% is a deductible the insurer pays. In the floor, the 10% is the most you can lose.

Who absorbs which slice

Take an index that drops 35% over the term, with a 10% buffer in one contract and a −10% floor in the other, starting from $100,000.

In a severe decline the floor wins by a wide margin. But severe declines are not the only case. Here is the same $100,000 across several outcomes, with the loss shown as a change in account value.

Two features of the table are worth noticing. The two designs cross at a 20% decline, because a buffer of 10 and a floor of 10 give the same result when 10 points are absorbed by each side (that is, when the index has fallen by twice the protection level). Above that point the floor is better for the owner. Below it, the buffer is better. In general, for a buffer and a floor of the same percentage, the crossover is at double the percentage.

The second feature is that the buffer fully covers modest declines and the floor never covers any decline completely. A buffer owner can go through a mild bear market without a scratch; a floor owner would have lost money in that same market.

Why the price is not the same

Neither design is free. The insurer hedges the risk it takes on with options, and the cost of that hedge comes out of what it can offer on the upside, through the cap or participation rate (see cap rate and participation rate). The two designs cost it different amounts.

A buffer covers the first slice of loss, the slice that occurs most often, because moderate declines are far more common than severe ones. A floor covers the tail: it pays nothing on small declines and everything beyond its level, however deep. Because the two cover different parts of the distribution of outcomes, neither is automatically cheaper, and the price an insurer pays depends on market conditions and its own pricing. So the useful comparison is not which design sounds safer but which pairing of protection and cap you are actually offered. You can only compare them by running the numbers on actual terms.

Buffers are the more common design in the market, partly because buyers find them easy to understand ("my first 10% is covered").

The upside side of the ledger

Protection is half of the contract. The other half is how much of a gain you get, usually described by a cap (the most you can be credited) or a participation rate. A hypothetical comparison in an up market, with a six-year term: Contract A has a 10% buffer and a 40% cumulative cap; Contract B has a −10% floor and a 25% cumulative cap. If the index rises 60% over the term, Contract A credits 40% ($140,000) and Contract B credits 25% ($125,000). If the index rises 20%, both credit the full 20%. If it falls 35%, Contract A ends at $75,000 and Contract B at $90,000.

The pattern is the same trade-off that runs through all of this: whichever design gives you more protection against deep losses tends to pay less in good years. These are invented caps; real ones differ and change at each new term.

Interim values: what a mid-term exit looks like

Both designs share a feature that catches many owners out. The buffer or floor applies to the index change measured at the end of the term. If you leave before the term ends, through a withdrawal, or a surrender, most contracts calculate an interim value by a formula in the contract. The formula typically adjusts the account for index performance to date, the time remaining and interest rates, and it may not credit the full buffer or floor.

A hypothetical example. Say you hold a six-year buffered term, and at the end of year three the index is down 12%. If the term ended today, a 10% buffer would turn this into a 2% loss. But an interim value might show a larger loss, say 8%, because the formula reflects the current value of the protection and the growth potential that are still outstanding, which is not the same as the value at the end of the term. The numbers vary by contract and by market conditions. A surrender charge may apply as well. This is the reason that "protection" is better described as "protection if you hold to the end of the term."

What to ask for: the interim-value formula in writing, and a worked example for a withdrawal in year two of a six-year term with a falling index.

Term length and reset: why they change the outcome

Terms are commonly one, three, five or six years. They interact with the buffer and the floor in ways that are easy to overlook.

The protection is per term. A one-year 10% buffer resets every year. Say the index falls 15% in year one, then rises in year two. The annual-term owner takes a 5% loss in year one (15 minus the 10-point buffer), which stays in the account at $95,000, and begins year two from that lower base with a fresh buffer. A multi-year term measures only the net change at the end, so a deep fall that recovers by the end can leave the buffered owner with no loss at all.

A hypothetical dip-and-recover path shows both sides. Take a three-year term with a 10% buffer where the index is down 25% at the end of year one and down 4% at the end of year three. The end-of-term change of −4% is inside the buffer, so the account ends at $100,000, as if nothing happened. Now take the same index path under three one-year terms with the same buffer and an assumed 10% cap. Year one: −25% becomes −15%, leaving $85,000. The recovery from 75 to 96 is a rise of 28%, which we split into +14% and +12.3% over years two and three; each is capped at 10%, so $85,000 grows by 1.10 × 1.10 = 1.21 to $102,850. In this invented case the annual terms end slightly ahead, because the caps are generous and the dip was front-loaded. Change the path or the caps and the order of results can flip. The lesson is not that one length is better; it is that you cannot evaluate a term length without a path and the caps that go with it.

Locking in matters differently. In an annual design, each year's credit is locked at the anniversary. In a multi-year design, nothing is locked until the term ends. A long term gives the buffer more time to absorb a dip that recovers, and gives you less chance to bank a gain.

Renewal. When a term ends, you choose a new one, with the buffer and cap that the insurer then offers. The buffer percentage and the cap are set at the start of each term. They can be different from the ones you bought.

Exits at term end. Many contracts waive the interim-value adjustment at the end of a term, which is the natural time to leave without it. The schedule of surrender charges is separate and may continue; see annuity fees, explained for how the layers add up.

Reading the outcome table in a prospectus

Every RILA prospectus is expected to explain how the crediting works, and many include a table or chart showing outcomes for a range of index changes. Use it the way this post uses its own tables.

1. Find the loss column at the end of the term. It should show a zero loss inside a buffer, and the loss beyond it, or the capped loss under a floor.
2. Find the gain column. Check where the cap starts to bind. Note the index change at which the credit stops rising.
3. Look for the interim row. If the document shows what happens on an early exit, read the formula underneath. If it does not, ask where the formula appears.
4. Compare the break-even. Ask yourself at what index change you would be better off with the other design. For equal percentages, a buffer and a floor cross at twice the protection level. For different percentages and caps, run the calculation on the two sets of terms.
5. Add the fees. The base contract may have no explicit annual fee, but riders and some options do. annuity fees, explained lists the usual layers.

For the arithmetic of a capped credit in a related product, the worksheet in work out an indexed annuity credit by hand with your own contract terms shows a step-by-step method that carries over to RILA caps and participation rates.

A decision list

These "if... then..." statements are a starting point, not advice, and each assumes you have already decided that a RILA belongs in your plan at all.

• If you could not tolerate a loss of more than about a tenth of the account, and the available floor is near that level, a floor design matches your limit, at the price of lower upside.
• If you mainly want to get through ordinary down markets unscathed and accept the tail risk, a buffer design matches that.
• If you may need to leave before the term ends, treat the interim value as the real protection level, not the stated buffer or floor.
• If you want to see every gain locked at least yearly, favor one-year terms and accept that each year's cap and protection reset.
• If you want the greatest chance that a dip recovers before you are measured, longer terms give it more time, with less chance to lock gains.
• If none of these fits and the principal guarantee is the real goal, look at fixed products instead of a registered one.

What neither design changes

• You can lose principal. Both designs limit losses; neither prevents them. Unlike a fixed indexed annuity, which has a zero floor (see fixed indexed annuities: caps, participation rates and trade-offs), a RILA is a security with risk of loss, regulated by the SEC and sold by FINRA-registered sellers.
• Guaranty protection is limited. State guaranty associations back the insurer's contractual obligations up to their limits, but they do not make up market losses that the contract allows.
• Dividends are generally excluded. If the index is a price index, the credit leaves out dividends. Confirm in the contract.
• Taxes still apply. Gains are taxed as ordinary income when withdrawn, with a 10% additional tax on gains before age 59½ in most cases, as for other deferred annuities.
• Caps change. The terms of each new period are set at that time.

Which suits whom

A buffer tends to suit someone who:

• Wants moderate declines to cost nothing and can accept the risk of a deep loss.
• Values simplicity ("the first 10% is covered").
• Expects to hold to the end of each term.
• Is comfortable with the cap that accompanies the buffer.

A floor tends to suit someone who:

• Is most concerned about a severe decline and wants a known maximum loss.
• Can accept small losses in ordinary down years.
• Understands that the floor usually comes with a lower cap or participation rate.

Neither suits someone who may need the money before the term ends, who cannot afford any realistic loss, or who is looking for a guarantee of principal. A fixed product is the right category for that. A saver comfortable with low-cost index funds also has a reasonable alternative that has no caps or surrender charges, though also no protection.

Questions to ask before choosing

1. For each option, what are the protection type and level, the term, and the cap or participation rate, and how long is each locked in?
2. What is the interim-value formula, and what would I receive if I left in year two?
3. Is the index price-only or total return? Standard or proprietary?
4. What are the surrender charges, and do they continue after the term ends?
5. What can change at renewal, and how will I be told?
6. Where in the prospectus can I find a table of outcomes for different index changes?

The prospectus and the summary prospectus are where the real terms live, and the rules governing sellers are described in best interest rules for annuity sales. A fee-only planner can compare the options against your own time horizon and risk, and the state insurance department and securities regulators take complaints about sales.

Key takeaways

• A buffer absorbs the first slice of a loss for you; a floor limits your total loss and covers everything beyond it, so the same percentage means opposite things.
• With equal percentages, the buffer is better when the decline is smaller than twice the protection level, and the floor is better beyond it.
• Protection usually applies at the end of the term; leaving early triggers an interim value that can be well below the buffered or floored result.
• Term length and reset rules change which dips are absorbed and which gains are locked, and the answer depends on the market path and the caps.
• Both designs can lose principal, so compare caps, terms, exit values and prospectus tables on real terms rather than the protection label alone.