Independent
Thinks s/he gets paid by the post
- Joined
- Oct 28, 2006
- Messages
- 4,629
What number do I put into FireCalc? Should I defer Social Security?
Life expectancy, or life span, or 10% survival ages, often comes up in our discussions. When people quote numbers, I always wonder what mortality table they used to get those numbers. After all, mortality data comes from groups of people. If I’m going to use that data, I should probably have some notion that I’m kind of like that group.
But, that’s theoretical. I don’t know if this is a big deal in practice. Are the differences between tables big enough to matter?
So, I thought I’d run some numbers for a few tables and see what I got for a 62 year-old male.
The first row says that if we start with a large group of 62 year-olds, slightly more than 50% of them will be alive 20 years later, but slightly less than 50% will make it 21 years.
Similarly, slightly more than 10% will survive for 31 years.
The next two numbers are just the ages that correspond to 20 and 31 years.
I’m showing six tables. There’s a range of 7 years in the 50% column and 5 years in the 10% column. I’m not sure if that’s “significant”, but those are the numbers.
The first table, SS 2010, is probably the most quoted. It’s the top response if you Google “mortality table social security”. This table represents the experience in 2010 of the population covered by Social Security. I don’t think it’s the best choice for posters here. This table includes people who have stage 4 cancer, or are living in nursing homes, or even living in hospices. I’m claiming their mortality rates aren’t particularly relevant to people deciding whether to put 30 or 35 into FireCalc.
RP-2000 is the experience of people with private pensions in 2000. I included it because Vanguard has a nice calculator that allows the user to pick two different ages for last-survivor numbers. That calculator happens to use the RP-2000 table, so I thought I’d include it for people who might be using that calculator.
https://personal.vanguard.com/us/insights/retirement/plan-for-a-long-retirement-tool
RP-2014 is again the experience of people who are collecting on private pensions, but it’s somewhat more recent. I included it because I wanted an example of some other grouping rule. In this case, it’s “collar”. No, the word “blue” does not refer to these guys, https://www.youtube.com/bluemangroup , it means retirees who had blue collar jobs when they were working. The next row is people who had white collar jobs. As expected, the white collar group lives a little longer.
IAM 2012 is Individual Annuitants Mortality. This includes people who bought private SPIAs, or took life settlement options on life insurance policies, or life settlements in lawsuits. I’d expect these people to have the lowest mortality among the choices here. Most of them self-select. If you’re in any sort of poor health, you probably won’t buy a private SPIA.
Some of these tables say “Period”, others say “Cohort”. If I construct life expectancy for a 62 year-old from a 2012 period table, I’ll use the 2012 experience for 62 year-olds, and attach it to the 2012 experience for 63 year-olds, and then the 2012 experience for 64 year-olds, etc.
For a cohort table, I’ll use the 2012 experience for the 62 year-olds, but I’ll use projected 2013 rates for 63 year-olds, projected 2014 rates for 64 year-olds, etc.
Since I believe that the trend in slowly dropping mortality rates is likely to continue, I would use a cohort table instead of a period table for my own planning. In fact, since we were both in fine health when I retired, I would have used the IAM cohort table (if it had been available).
Of course, only a minority of us are single males. We’ve got some single females, and a lot of couples. More numbers follow.
This table is the last-to-die for a couple, both age 62.
Of course, all this comes with the usual caveats:
I'm been known to make mistakes before, there could be some here.
I checked a couple numbers against published sources, but most of this is strictly my calculations.
The cohort tables are particularly tricky to work with.
etc.
Life expectancy, or life span, or 10% survival ages, often comes up in our discussions. When people quote numbers, I always wonder what mortality table they used to get those numbers. After all, mortality data comes from groups of people. If I’m going to use that data, I should probably have some notion that I’m kind of like that group.
But, that’s theoretical. I don’t know if this is a big deal in practice. Are the differences between tables big enough to matter?
So, I thought I’d run some numbers for a few tables and see what I got for a 62 year-old male.
Male - Age 62 in 2015 | 50%. | 10%. | 50%. | 10%. |
SS 2010 Period | 20 | 31 | 82 | 93 |
RP-2000, Period | 21 | 31 | 83 | 93 |
RP-2014, Blue, Cohort | 24 | 36 | 86 | 98 |
RP-2014, White, Cohort.. | 27 | 37 | 89 | 99 |
IAM 2012, Period | 25 | 35 | 87 | 97 |
IAM 2012, Cohort | 27 | 36 | 89 | 98 |
The first row says that if we start with a large group of 62 year-olds, slightly more than 50% of them will be alive 20 years later, but slightly less than 50% will make it 21 years.
Similarly, slightly more than 10% will survive for 31 years.
The next two numbers are just the ages that correspond to 20 and 31 years.
I’m showing six tables. There’s a range of 7 years in the 50% column and 5 years in the 10% column. I’m not sure if that’s “significant”, but those are the numbers.
The first table, SS 2010, is probably the most quoted. It’s the top response if you Google “mortality table social security”. This table represents the experience in 2010 of the population covered by Social Security. I don’t think it’s the best choice for posters here. This table includes people who have stage 4 cancer, or are living in nursing homes, or even living in hospices. I’m claiming their mortality rates aren’t particularly relevant to people deciding whether to put 30 or 35 into FireCalc.
RP-2000 is the experience of people with private pensions in 2000. I included it because Vanguard has a nice calculator that allows the user to pick two different ages for last-survivor numbers. That calculator happens to use the RP-2000 table, so I thought I’d include it for people who might be using that calculator.
https://personal.vanguard.com/us/insights/retirement/plan-for-a-long-retirement-tool
RP-2014 is again the experience of people who are collecting on private pensions, but it’s somewhat more recent. I included it because I wanted an example of some other grouping rule. In this case, it’s “collar”. No, the word “blue” does not refer to these guys, https://www.youtube.com/bluemangroup , it means retirees who had blue collar jobs when they were working. The next row is people who had white collar jobs. As expected, the white collar group lives a little longer.
IAM 2012 is Individual Annuitants Mortality. This includes people who bought private SPIAs, or took life settlement options on life insurance policies, or life settlements in lawsuits. I’d expect these people to have the lowest mortality among the choices here. Most of them self-select. If you’re in any sort of poor health, you probably won’t buy a private SPIA.
Some of these tables say “Period”, others say “Cohort”. If I construct life expectancy for a 62 year-old from a 2012 period table, I’ll use the 2012 experience for 62 year-olds, and attach it to the 2012 experience for 63 year-olds, and then the 2012 experience for 64 year-olds, etc.
For a cohort table, I’ll use the 2012 experience for the 62 year-olds, but I’ll use projected 2013 rates for 63 year-olds, projected 2014 rates for 64 year-olds, etc.
Since I believe that the trend in slowly dropping mortality rates is likely to continue, I would use a cohort table instead of a period table for my own planning. In fact, since we were both in fine health when I retired, I would have used the IAM cohort table (if it had been available).
Of course, only a minority of us are single males. We’ve got some single females, and a lot of couples. More numbers follow.
Female - Age 62 in 2015 | 50%. | 10%. | 50%. | 10%. |
SS 2010 Period | 23 | 34 | 85 | 96 |
RP-2000, Period | 24 | 35 | 86 | 97 |
RP-2014, Blue, Cohort | 27 | 38 | 89 | 100 |
RP-2014, White, Cohort.. | 29 | 39 | 91 | 101 |
IAM 2012, Period | 27 | 37 | 89 | 99 |
IAM 2012, Cohort | 28 | 38 | 90 | 100 |
This table is the last-to-die for a couple, both age 62.
Joint - Both Age 62 in 2015 | 50%. | 10%. | 50%. | 10%. |
SS 2010 Period | 27 | 35 | 89 | 97 |
RP-2000, Period | 27 | 36 | 89 | 98 |
RP-2014, Blue, Cohort | 31 | 40 | 93 | 102 |
RP-2014, White, Cohort.. | 33 | 41 | 95 | 103 |
IAM 2012, Period | 30 | 38 | 92 | 100 |
IAM 2012, Cohort | 32 | 40 | 94 | 102 |
Of course, all this comes with the usual caveats:
I'm been known to make mistakes before, there could be some here.
I checked a couple numbers against published sources, but most of this is strictly my calculations.
The cohort tables are particularly tricky to work with.
etc.
Last edited: