Do VA ‘Roll-Ups’ Have the Desired Effect?

Deferral bonuses, a key feature of many VAs, may not prevent contract owners from taking withdrawals, according to Milliman’s latest survey of issuers of VA guaranteed living benefits. Photo courtesy of

When insurers price their variable annuities and annuity riders, they have to anticipate more than movements in the stock market or decisions at the Federal Reserve. They also have to anticipate how contract owners will use those riders.

In the case of guaranteed lifetime withdrawal benefit (GLWB) riders, profitability may depend on how many people buy the riders as well as when they start taking income. Insurers often offer deferral bonuses or “roll-ups” to discourage early withdrawals, but it’s unclear—except in retrospect—whether those bonuses have the desired effect.

Milliman, the global actuarial firm, periodically publishes a survey, called the Variable Annuity Guaranteed Living Benefits Study, to reveal some of that behavior. In the most recent survey, based on sales in the first half of 2010, 18 insurers responded to Milliman’s questions. Milliman provided an executive summary of the study to RIJ.

One result surprised Sue Saip and Noel Abkemeier, the Milliman consultants who co-authored the study: a higher-than-expected proportion of contract owners seemed to be taking out money in the first year of the contract. The deferral bonus wasn’t having the expected impact.

“If you look company by company it doesn’t seem to have had a significant impact on withdrawals,” Saip told RIJ.  “That was our take.” The highest first-year rate was 40%, reported by a major VA seller whose product included a roll-up. “I would think that they did not expect withdrawals at that level,” she said.

Among the 18 companies in the survey, “about 15% of GLWB purchasers are pulling the trigger and exercising their GLWB in the first year,” said Abkemeier. “That’s not particularly different from the last few years. When the survey first came out, I was surprised that the withdrawal rate was as high as it was.

“My expectation was that if people wanted income right away they’d buy a SPIA as opposed to this, and that the percentage of VA owners who took withdrawals right away would be 5% or lower,” he added.

In other words, some contract owners seemed to be leaving money on the table by not taking advantage of the roll-ups. That doesn’t seem optimal or rational. On the other hand, they may simply have been reacting to the fact that their contracts were “in the money.” That is, the guaranteed value exceeded the account value—so they were more likely to take withdrawals.

“In-the-moneyness does significantly influence withdrawal benefit exercise rates,” Saip said. Among those whose contracts were at least 50% in the money (where the guaranteed base was 50% or more greater than the account value), a high percentage of people—56.5% on average across 18 companies (with a range of 7.5% to 98.9%)—started exercising their income benefit.  The older they were, the more likely they were to exercise withdrawals.

But did the contract owners who took withdrawals from depressed accounts immediately invest their withdrawals in the depressed stock market, as some advisors were recommending? If they did, there’s no clear evidence of it. They may simply have needed the income. But no one knows whether people truly understand the features of the variable annuities they buy.

The study confirmed that most people are buying VAs with lifetime income riders, but in some cases the riders are automatically included in the annuity. Among companies that offered an optional GLWB on some products, on average, 90.4% (median: 95.1%) of variable annuity purchases (by dollar value) in the first half of 2010 were of products that offered a guaranteed lifetime withdrawal benefit as an option. But where a GLWB rider wasn’t automatically bundled in, the GLWB was purchased in 59.8% of sales on average (median: 69.2%).  

Another finding was that more new money has been coming into the VA industry in recent years, as opposed to money simply moving from one insurer to another. That was partly to be expected, since new contracts tend to be less generous than old ones, and because many existing contracts may have been in-the-money at the time the survey was conducted.

“The number of external exchanges has fallen as the benefits have gained value,” Saip said. We haven’t seen as many customers rolling over. We had only eight companies reporting data on that, but of those, some had as much as 80% of sales being new money coming in. And that included some pretty major players.”

The product de-risking process continues to go on, with companies raising fees in subtle ways. Besides moving toward less frequent step-ups in the value of income base subsequent to an increase in the account value, Saip noted that some companies are reserving the right to raise a rider fee at any time, not just when the client elects to take a step-up. 

“Companies are moving to a little more frequency, where they can change the expense ratio on the anniversary or change it at anytime. That gives companies a little more control, which means they may not have to set the maximum as high. If you can only change the fee when there’s a step-up in the benefit base, you tend to set a higher maximum,” she said.

The Milliman survey also asked companies about their hedging habits. “Our participants nearly all said that they hedge 100% of delta and 100% percent of rho. They don’t all hedge vega and gamma. Four companies said they were hedging gamma, with an average of 75% of gamma hedged. Of those who hedged vega, 62.5% was hedged on average.  By far, companies are hedging 100% of the Greeks that they hedge,” Saip said.

[Delta is the rate of change of the option value relative to changes in the value of the underlying asset. Gamma is the rate of change of delta relative to changes in the value of underlying asset. Vega measures sensitivity to volatility, rho measures sensitivity to the interest rate.]

According to Milliman, companies that participate in the survey use it mainly to benchmark their own results against industry averages. Reading the results like tea leaves to determine the motives of investors is more difficult. That information remains largely the object of speculation. The issuers have a better idea of why advisors sell variable annuities than they do about why individuals buy them.

“My view is that people buy the variable annuity because they’re sold it,” Abkemeier told RIJ. “Aside from that, their understanding may be that the product provides a floor of investment protection that’s denominated in lifetime income benefits. They may also understand that if their account goes up a lot, their income may also go up. But I think they’re primarily buying an accumulation product and see the income guarantee as a secondary consideration.

“I think that investors want to keep their options open,” he added. “They’ve heard about the downsides of the income annuity. There’s an element of optimism where they think, ‘My variable annuity will grow at a nice attractive rate. I’d like to keep my money growing.’

“In the backs of their minds, they may also think, ‘If the account value goes way down, I’ll still have this guarantee that will let me take out more money than the contract has value,” he said. “But they don’t necessarily figure out all the arithmetic, which would tend to show that they may win in the end, but they’ll win not by hitting a home run but on a walk with the bases loaded in the bottom of the ninth inning.” 

© 2011 RIJ Publishing LLC. All rights reserved. Photo by