So I responded to a mailbox flier that offered an interesting booklet: "Minimize Taxation of Your Social Security Retirement Benefit." I met with the insurance agent who sent the booklet. A week later he offered a $494,000 life insurance policy with five premiums of $50,000 that would go into an equity index life insurance policy. The money would be allowed to grow for an additional 10 years. In the 16th year (at age 80), I would be able to borrow $25,000 a year for the rest of my life, tax-free, even as the death benefit increased.

All this was based on an equity index return assumption that was about 8.6 percent.

If the policy performed at the guaranteed rate of 2 percent, however, the outcome was very different -- the policy would exhaust my $250,000 of premium payments in the 12th year. There would be no lifetime income of $25,000. The only way to benefit in any way would be to die within 12 years -- before the policy collapsed -- so my family could collect the death benefit.

High taxes sound pretty good compared to that.

If you examine recent equity index returns (the return on large common stocks excluding dividends), that 8.6 percent index return looks pretty reasonable. From 1991 through 2000, according to Ibbotson Associates, capital appreciation of the S&P 500 index was 12.2 percent a year, compounded. Even burdened with an annual cap or limited to a percentage of the gain -- as most equity index policies are -- 8.6 percent looks very likely.

Unfortunately, the '90s were an unusual period.

While equity appreciation (excluding dividends) was over 10 percent through the '80s and '90s, it was under 3 percent in the '60s and '70s. In the 78-year period from 1926 through 2003, according to Ibbotson Associates, index appreciation ran at a compound annual growth rate of 5.9 percent. About 1 percentage point of that came from rising price-to-earnings multiples.

Bottom line: With P/E ratios at relatively high levels today, future index appreciation is likely to be closer to 5 percent than 6 percent -- if the policy has no cap or participation limit. If earnings multiples decline, it could be closer to 4 percent.

That's better than the 2 percent guarantee -- but far below the 8.6 percent assumed and projected. Andrew uses a 7.75 percent assumption that is nearly as unrealistic as what I was presented.

Then there are the last two burdens: commissions and life insurance costs. These can be devastating. After five years and $250,000 of premiums, the cash value of the policy I was offered would have been $140,000 at the guaranteed 2 percent. That's a loss of $110,000. At projected returns, it would have been worth $220,000, a loss of $30,000.

No doubt some can do better, and the actual result would be somewhere in between. The only thing certain is that most people will be disappointed, perhaps disastrously, with the consequences of exchanging home equity or tax-deferred retirement accounts for life insurance.