As I have said several times in this blog, developing a reasonable spending budget would be a lot easier if we just knew when we were going to die. In fact, one of my long-time actuary buddies, Bruce O suggested that I would probably get more readers if I simply renamed this site, “How long can I afford to live in retirement?”
So, when I saw the title, “When and how you will most likely die” recently in Business Tech, I knew that this article was going to be must-reading for this mortality nerd. The article points its readers to an interesting interactive tool developed by Nathan Yau of Flowingdata. Mr. Yau has mined data maintained by the Centers for Disease and Prevention containing information for people who died in the U.S. between 1999 and 2014 to develop probabilities of death by age and associated probabilities regarding reported cause of death. Mr. Yau further segregates the data into eight reported categories (four each for males and females): White, Asian, Black and Native.
In terms of developing a spending budget, I’m considerably more interested in age at death than I am in the reported cause of death. More specifically, I was interested in probabilities of survival from a given retirement age (once I discovered that Mr. Yau didn’t really have a magical crystal ball). I played with Mr. Yau’s tool to determine probabilities of survival to various ages for males and females currently age 65. I wasn’t sure why, but Mr. Yau’s tool gave me slightly different probabilities each time I ran the tool. It also takes a little dexterity to pause the aging process at the right moment. Here is a summary of the results:
 |
| (click to enlarge) |
The final column labeled “SoA” shows comparable probabilities of survival from the Society of Actuaries’ 2012 Individual Annuity Mortality table with a 1% per year mortality improvement projection. This table is available in the “Other Calculators/Tools” section of this website. I had to estimate probabilities of survival until age 70 for this column. In general, the SoA table shows much higher probabilities of survival to various ages than the information gleaned from Mr. Yau’s tool. In my opinion, the primary reasons for this are: 1) the SoA table is based on mortality of individuals who buy annuities where Mr. Yau’s experience is based on all individuals in the U.S., 2) the SoA table includes mortality improvement projections while Mr. Yau’s is based on experience from 1999-2014, and 3) as Mr. Yau’ admits, his data has some deficiencies.
Mr. Yau’s tool shows that age, gender and race are generally factors in one’s mortality/longevity. Of course, there are many other factors. For example, studies have shown factors influencing longevity can include, wealth, general health, smoking, diet, weight, exercise, participating in community activities, marital status, taking part in dangerous hobbies, geographical location, etc. I wouldn’t be surprised if some researcher has even found a positive correlation between longevity and developing a spending budget for retirement. Suffice to say that there are a lot of variables at play and caution should be used selecting your period of retirement for budget setting purposes.
When developing a spending budget, I recommend that retirees and their financial advisors plan on living until age 95 or the retirees’ life expectancy if greater. Yes, this is a conservative recommendation. I don’t recommend assuming probabilities of death at every future age even though we all have non-zero probabilities of dying every year in the future. The fact of the matter is that either 100% of you will be alive in a given year or 100% of you will be dead. To assume that 2% of you will die next year would just make the calculations more complicated and it wouldn’t add much value to the spending budget you are trying to determine. Unless you are quite old, I also don’t recommend that you use your life expectancy when determining your spending budget. The reason for this is that as you age, your expected age at death increases. Thus, your expected period of retirement will not decrease by one year for each year you age and your spending budget will therefore decrease in the future in each year when your expected age at death increases by that one year, all things being equal. This decrease is illustrated in my post of December 3, 2014. By assuming death at age 95, the retiree avoids these budget decreases until approximately age 89, when budget decreases may be more acceptable to the retiree. You can assume a shorter period of retirement, but this will lengthen the potential period when your survival will result in decreases in your spending budget, all things being equal.
Well, enough time spent on this morbid subject. I leave you today with a Mr. Spock’s emoji wishing you long life and prosperity in your retirement.
This is a follow-up to my post of November 5, 2015 that focused on the two problems with any proposed reforms to Social Security that simply reduce the system’s 75-year actuarial deficit to zero.
As Yogi Berra said, “It’s tough to make predictions, especially about the future.” If we knew for sure what the future held, it would probably make it easier for us to plan for it. But if we did, we wouldn’t need people like actuaries who make a living by making educated assumptions about the future.
Last month, the Congressional Budget Office (CBO) released, Social Security Policy Options 2015 which examined the impact of 36 specific policy options on their estimate of the 2015 75-year actuarial deficit. But, instead of using the assumptions for future system experience adopted by the Social Security Trustees and the Social Security Administration actuaries, those rascals at CBO tweaked a few of the Trustee’s assumptions. The effect of this tweaking of was to increase the estimated 2015 75-year actuarial deficit from 2.7% of taxable payroll to 4.4%. Wow, over a 60% increase in the estimated deficit from tweaking a few assumptions!
Well, of course the CBO report got a bunch of folks upset. In her recent article, Alicia Munnell, the Director of the Center for Retirement Research at Boston College, calls the CBO estimate “astounding” and implies that the CBO report was issued for the purpose of setting “the stage for benefit reductions.” She is upset because she feels that “cutting benefits [now] would be a huge mistake.” She argues that the 2015 75-year deficit probably isn’t any higher than her Technical Panel’s “preferred alternative” estimate of 3.42%. Her article implies that the CBO report makes the system’s financial problems sound insoluble.
Whose assumptions are a better estimate of future system experience? The Trustees’? The CBO’s? Alicia’s Technical Panel’s? I have no idea. And frankly, no one does (now that Yogi has passed). As the ultimate sentence in Alicia’s article indicates, “Only time will tell which of us comes closer.” And that is the primary point of this post. Since we don’t know what the future holds, we need to solve our best estimate of the size of the problem in the near future. But equally important, we also need to use sound actuarial principles to ensure that the system automatically maintains its actuarial balance in the future when actual experience deviates from assumed experience (because trust me on this one, it will). The alternative is to simply assume that the “best estimate assumptions” will be accurate for the next 75 years and proclaim the program “fixed” as members of Congress did back in 1983. We just have to look at the mess the system is in now to see how the lack of automatic adjustments has worked for us.
Periodically, we read some article from a retirement expert touting their modification of the 4% Rule and why their modification will work better. As I have indicated many times in this website, the 4% Rule and its many variations (including Safe Withdrawal Rates, the RMD approach, the Guyton Decision Rules, etc.) have significant deficiencies when compared with using general actuarial principles to develop a spending budget. And while I have said it many times before, I believe it bears repeating. Since I recently included a very simple present value spreadsheet calculator in the section entitled, Articles and Spreadsheets (which is even more simple and basic than the Excluding Social Security V 3.1 spreadsheet), I will use that spreadsheet to calculate all the present values discussed below and to demonstrate one of the several problems with rule of thumb withdrawal strategies for retirees who have other fixed dollar sources of income such as pensions, immediate life annuities or deferred life annuities, etc.
As discussed previously, The Actuarial Approach for developing a spending budget in retirement involves matching household assets (current assets/investments plus the present values of other sources of income) with household liabilities (the present value of future expected expenses to be incurred by the retiree/retiree's family over the remaining lifetime plus the present value of assets remaining at death). This balancing (which normally takes place at least once a year) is referred to as an actuarial valuation and can be summarized in what is called an "actuarial balance sheet."
But enough background. Let's take a look at two simple examples of how using The Actuarial Approach blows away rule of thumb withdrawal strategies for retirees with fixed dollar sources of retirement income. We will be looking at two 65-year old male retirees, each with a Social Security benefit of $20,000 per year, each with $500,000 in accumulated savings and each with the same spending objective to cover essential expenses of $50,000 per year with any remaining assets to cover non-essential expenses. Both retirees will be assuming a 4.5% discount rate, 2.5% inflation and a thirty year retirement period. The only difference between the two retirees is that Retiree #1 has an immediate fixed dollar pension of $30,000 per annum and Retiree #2 has a fixed dollar deferred annuity to commence at age 80 of $30,000 per annum.
To be consistent with the assumptions inherent in the 4% Rule, both retirees assume that their essential and non-essential expenses will increase with inflation of 2.5% per annum. As discussed in previous posts, it may be reasonable when using the Actuarial Approach to assume different rates of future increases for different types of expenses. For simplicity purposes, we are only looking at essential and non-essential expenses. Certainly other categories of expenses may be included when using the Actuarial Approach.
Here is the actuarial balance sheet for Retiree #1
 |
| (click to enlarge) |
Under the assumptions discussed above, the present value of Retiree #1's essential expenses of $50,000 per year increasing by 2.5% per annum over a period of 30 years is $1,149,368. This amount is shown in the liability column. In order to cover these estimated essential expenses, Retiree #1 dedicates his Social Security benefit (with a present value of $459,747), his pension benefit (with a present value of $510,657) and $178,964 of his accumulated savings ($1,149,368 - $459,747 -$510,657). This leaves an asset of $321,036 in accumulated savings dedicated to non-essential expenses, which is also equal to the liability for such expenses.
Retiree #1's actuarial spending budget for the first year of his retirement is equal to the sum of his essential spending budget of $50,000 plus the annual amount payable over a period of 30-years increasing by 2.5% per annum with a present value of $321,036, or $13,966 in year 1 (determined using the present value calculator spreadsheet). So his total actuarial spending budget for the year would be $63,966 ($50,000 + $13,966).
If he spends all of his pension benefit and his Social Security benefit to satisfy his essential expenses, he will spend $0 from the accumulated savings he has allocated to cover his essential expenses. Thus, for this portion of his spending budget, his withdrawal rate from accumulated savings is 0%, not some rule of thumb withdrawal percentage. By comparison, the withdrawal percentage from the assets he dedicates to his non-essential expenses represents about 4.35% of such assets ($13,966/$321,036).
In total, withdrawals from accumulated savings for Retiree #1 are $13,966, or about 2.8% of his total accumulated savings. By comparison, it is not entirely clear how the 4% Rule would be used to determine separate withdrawals from Retiree #1's assets dedicated to fund essential and non-essential expenses, but in total, 4% of $500,000 would be $20,000. Therefore, in total, withdrawals would represent 4% of Retiree #1's total accumulated savings vs. 2.8% under the Actuarial Approach.
Retiree #2 has the same situation as Retiree #1, except instead of having a fixed dollar pension of $30,000 per annum, he has a fixed dollar deferred annuity benefit of $30,000 per year commencing at age 80.
Here is the actuarial balance sheet for Retiree #2
 |
| (click to enlarge) |
The present value of Retiree #2's expected future essential expenses is the same as Retiree #1's, $1,149,368. The present value of Retiree #2's Social Security benefit is also the same as Retiree #1's at $459,747. The present value of Retiree #2's deferred annuity, however, is only $203,814 leaving $485,807 from his accumulated savings ($1,149,368 - $459,747 - $203,814) to cover the present value of Retiree #2's expected future essential expenses. This then leaves only $14,193 ($500,000 - $485,807 to cover future expected non-essential expenses, which is also the liability for such expenses.
Retiree #2's actuarial spending budget for the first year of his retirement is equal to the sum of his essential spending of $50,000 plus the annual amount payable over a 30-year period and increasing by 2.5% per annum with a present value of $14,193, or $617 in year 1 (determined using the present value calculator spreadsheet). So his total actuarial spending budget for the year would be $50,617 ($50,000 + $617).
Retiree #2 expects to spend $30,000 ($50,000 - $20,000 from Social Security) from the accumulated savings he has allocated to essential expenses. This withdrawal represents about 6.2% of such assets. In total for both essential expenses and non-essential expenses, he plans to spend $30,667, or about 6.1% of his total accumulated savings. By comparison, the 4% Rule would give him spending for the first year of retirement of $40,000 ($20,000 from Social Security and $20,000 from accumulated savings).
We see from these two examples that simply adding 4% of a retiree's accumulated savings to other retirement income payable during the year to determine a spending budget may either overstate or understate an actuarially sound spending budget. And this is only one of the problems associated with the 4% Rule or any rule of thumb that focuses on "tapping your savings" rather than properly coordinating with other sources of retirement income to develop a reasonable spending budget.
If you are a financial advisor, does the advice you give your clients depend on how you are paid? If you are a retiree and work with a financial advisor, does his or her advice depend on how your advisor is compensated? In another fine blog post, Michael Kitces discusses the possible impact of financial advisor compensation on the partiality of advice given. He states, “the fact that financial planning remains rooted primarily in product sales [insurance products] and asset gathering [Assets under management or AUM] has influenced – perhaps more than most advisors realized – the way that our financial planning software is built.” In his post, Michael focuses on development of financial software, but his ultimate goal is to encourage advisors to provide impartial, “real” advice. Thank you for bringing up this unsavory subject, Michael.
Full Disclosure: As noted in my Biography, I’m a retired pension actuary with no expertise in financial planning or investments. I receive no direct or indirect compensation from visits to this website or from any activity associated with this blog. I will not try to sell you insurance products, try to manage your investments or give you advice regarding how to reduce your taxes. My mission is simply to encourage you to use basic actuarial principles to manage your spending in retirement. Sorry, I won’t be providing you with any sexy software for this purpose. Just the basics: 1) Using the Actuarial Approach to develop a reasonable spending budget (usually annually), 2) Monitoring actual spending vs. the budget developed using the Actuarial Approach and 3) making necessary adjustments to keep spending on track. To use an old football metaphor, I focus on the “blocking and tackling” of spending in retirement.
What follows are admittedly generalizations that don’t apply to all financial advisors, and particularly do not apply to financial advisors who charge an hourly rate for their advice.
Interestingly, Michael Kitces is a staunch proponent of the 4% Rule. This rule and other static (“safe) withdrawal rate approaches were generally developed by financial advisors who also encouraged their clients to invest significant portions of their retirement assets in equities. Many of these advisors down-played the benefits of insurance products as purchase of these products would reduce AUM, and therefore their potential compensation. Those who used these static approaches develop a spending budget by adding income expected to be received from other sources during the year to this static percentage withdrawn from accumulated savings (typically increased by inflation from year without regard to actual investment performance). Other than this addition, there was no coordination with other sources of income. So, the recommended amount to be withdrawn from accumulated savings during the year didn’t change if the client did or didn’t have a fixed dollar pension or immediate annuity or a deferred annuity.
On the other hand, those in the insurance industry tended to tout the risk pooling and guarantee benefits of buying insurance (generally lifetime income products and long-term care insurance) and begrudgingly acknowledged that there might be benefits of investing in stocks, bonds and other investments (but they took great care to emphasize the non-guaranteed nature of such investments). Since they generally looked down on such investments, they were somewhat of an afterthought when it came time to develop a client’s spending budget. Thus, those employed by the insurance industry tended to point to the various static approaches advocated by the financial advisors (if they addressed the issue at all) as possible approaches to be considered for those retirees who had some money left after buying insurance.
More and more academics and retirement experts are discovering that the optimal investment of retirement assets may involve some combination of pooled risk lifetime insurance products and investments. The Actuarial Approach advocated in this website develops an actuarial spending budget that coordinates the two types of assets to meet retiree objectives by matching a retiree’s total assets with her total liabilities. It doesn’t just “tap” the retiree’s savings by using a static withdrawal percentage. It also tells the retiree whether his or her spending is on track from year to year and reflects changes in investments and actual spending. The 5-year projection tab in the “Excluding Social Security v 3.1 spreadsheet also allows retirees and their financial advisors to plan future actions should actual future investment returns and actual spending force the retiree off the track.
The Actuarial Approach is not designed to help you choose investments or avoid taxes. But it does “tick a lot of the boxes” advocated in Michael’s Manifesto. While it is not as comprehensive as Michael would like, it does provide “real” impartial advice for managing spending in retirement.
As discussed in my previous post, the basic principle underlying the Actuarial Approach is that to be considered in actuarial balance, a retiree's assets (current assets plus the present value of future expected benefit payments or payments from other sources of income) must be equal to the retiree's liabilities (the present value of future expected expenses/amounts left to heirs). Our last post illustrated this concept with Richard’s Actuarial Balance Sheet. Several of my readers have asked me, however, how the present values shown in Richard’s actuarial balance sheet were calculated. While all the values can be obtained by using the “Excluding Social Security 3.1” spreadsheet, use of this spreadsheet to calculate such values may not be all that intuitive. Therefore, I have added a new Excel spreadsheet to help readers called “Simple Present Value Calculator.” It is located in the “Articles and Spreadsheet” section of this website.
The Simple Present Value Calculator can be used to calculate 1) the present value of a single payment (or estimated expense) or 2) a stream of annual payments/expenses. If the spreadsheet is used to calculate the present value of a stream of payments, the stream must either be the same amount per year or increase at a specific rate per year. The spreadsheet assumes beginning of year payments.
For example, Richard’s Social Security benefit was assumed to be $20,340 per annum. We assumed that this benefit would commence immediately (T=0), increase by 2.5% per year and would be payable for 27 years, based on an assumption of death at age 95. We also assumed a 4.5% discount rate. Inputting these items in the spreadsheet, we develop the present value of his Social Security benefit under these assumptions equal to $432,037 (one of Richard’s assets).
If you want to play with this new spreadsheet, see if you can duplicate the present value of Richard’s life annuity benefit of $15,000 per year (with no future increases) payable for 27 years or the present value of Richard’s essential non-health related expenses of $50,000 per year, increasing by 2.5% each year and assumed in this example to be payable for 25 years.
You can also use this simple spreadsheet for other (perhaps even more important) purposes. For example, if you win the Lottery, you can use it to determine the effective interest rate used to discount future lottery payments for the lump sum option. Since the Powerball Lottery was big news this week, this was an issue discussed in the New York Times article, Dear Powerball Winner Take Our Advice and Take the Annuity. If we take the author at his word and the choice is either $50 million (pre-tax) annual payment over 30 years (a total of $1.5 billion) or $930 million in a lump sum (also pre-tax), then the discount rate for this transaction is approximately 3.7% per annum. The author must have done his calculations on a post-tax basis (or simply made an error) to come up with an effective rate of 2.843%.
No sooner had I finished my January 9th post on long-term care (LTC) expenses when I received an email from Derek Castle indicating (as nicely as possible) that he thought I was perhaps being too conservative in determining the impact on a retiree's current spending budget resulting from establishing a separate reserve for future LTC expenses. While Mr. Castle did not disagree with the assumptions I recommended for estimating future LTC costs and the resulting reserve for such costs, he argued (and I agree) that it is likely that other expenses will be reduced if and when a retiree enters a LTC facility. This post will discuss a possible approach for determining a retiree's current spending budget to reflect the reduction in certain expenses when a retiree expects to enter a LTC facility in the future and has established a separate reserve to fund future LTC costs. To do this, I will revise the budget for Richard Retiree, whose 2016 budget was discussed in my post of December 26, 2015.
In our December 26 post, Richard Retiree assumed that he would live another 27 years and die at age 95. He also developed a reserve for future LTC costs by assuming that he would spend three years in an assisted living facility and one year in a nursing home and that those costs would be covered by the equity in his home. In addition to these expenses, Richard developed his budget for 2016 assuming that all types of expenses (other than his unexpected expenses) would be incurred until his assumed death in 27 years at age 95. However, as Mr. Castle pointed out, it is likely that some of the expenses that Richard plans to incur prior to entering LTC will be reduced after he enters LTC (which he expects to be in 23 years). So how should we adjust Richard's previously determined budget for this apparent overlap period?
We can argue about which expenses might be reduced when Richard enters into LTC, and we can also argue about how much those expenses might be reduced. Fortunately, I am going to let Richard and his financial advisor make this decision. Richard decides that even though he plans to spend a total of four years in LTC facilities, he will not see a reduction in his essential health-related expenses when he enters the facility and while his essential non-health related and non-essential expenses may be reduced, he expects that they will not be totally eliminated (at least not until perhaps his final year). He decides that instead of assuming a 23-year period for incurring these expenses (i.e., the number of years before he plans to enter the LTC facility) or a 27-year period (as he did for the December 26 post), he will split the difference and assume that the reduction in his essential non-health related expenses and non-essential expenses will be equivalent to assuming that they cease completely 25 years from now.
The basic principle underlying the Actuarial Approach is that to be considered in actuarial balance, a retiree's assets (current assets plus the present value of future expected benefit payments or payments from other sources of income) must be equal to the retiree's liabilities (the present value of future expected expenses/amounts left to heirs). The chart below shows Richard's revised Actuarial Balance Sheet. Some explanation of the numbers in this chart follows:
 |
| (click to enlarge) |
Assets
Richard's accumulated savings of $1,095,193 (which includes $180,000 of home equity assets) is allocated to the following reserve accounts as follows:
- Reserve for future unexpected expenses: $102,000
- Reserve for future LTC expenses: $180,000
- Reserve for future Essential Health Expenses: $135,000
- Reserve for Essential Non-Health Expenses (for next 25 years): $361,330
- Reserve for death benefit expenses (assumed to be payable in 27 years): $3,047
- Reserve for Non-Essential Expenses: $313,816
The present value of Richard's Social Security benefit of $432,037 is the present value at a 4.5% discount rate of a stream of annual payments (assumed to be payable at the beginning of each year) starting at $20,340 and increasing by2.5% per annum for the next 26 years.
The present value of Richard's immediate life annuity payments of $242,199 is the present value at a 4.5% discount rate of a stream of annual payments of $15,000 per year for 27 years (also assumed to be payable at the beginning of each year).
The reserve for future LTC expenses of $180,000 is developed by assuming that Richard's home equity is $180,000 and will increase by 4.5% per year, or alternatively that it is more than $180,000 but will increase by less than 4.5% per year but will still cover Richard's expected LTC costs.
The reserve for Essential Non-health Expenses was derived by using the "Excluding Social Security V 3.1 spreadsheet and solving for the accumulated savings that, together with 25 years of expected Social Security and annuity payments would provide for expected essential non-health expenses of $50,000 per year increasing at 2.5% per annum.
Liabilities
The present value of essential non-health related expenses of $1,000,943 is equal to the present value of a stream of payments of $50,000 per year increasing by 2.5% per annum for 25 years.
The present value of the remaining expenses (except for non-essential expenses) are equal to the asset reserve allocated to that item above.
The present value of the non-essential expenses of $348,439 is equal to the reserve for non-essential expenses of $313,816 shown in the assets above plus the present value of the final two years worth of Social Security and annuity payments not used to support essential non-health expenses.
Revised Results
By assuming a 25-year funding period rather than a 27, the present value of Richard's essential non-health expenses (plus death benefit expenses) have been reduced by about $61,100 (there is some rounding involved). In addition to increasing his non-essential spending assets by about $61,100, Richard will be spreading his non-essential expenses over a 25 year period rather than a 27-year period, so his non-essential spending budget for 2016 will increase from $17,793 to $22,486, and his total spending budget will increase from $72,793 to $77,486, as follows:
- Essential Health Budget: $5,000
- Essential Non-Health Budget: $50,000
- Non-Essential Budget: $22,486
- Total 2016 Spending Budget: $77,486
For retirees who want to reflect the cost of LTC in their budget, but who do not want to get into this much detail in their calculations, the bottom line spending budget result for Richard is equivalent to using the original methodology (with a 27-year budget period for all expenses) but reflecting only about 58% of the expected present value of LTC cost. Amazingly, this was pretty close to the estimate made by Mr. Castle, whom I thank again for raising this issue with me.
Always happy to learn from others. Keep those good suggestions for improvements in the Actuarial Approach for spending budgets coming!
Like all other future expenses a retiree expects to fund with her accumulated retirement savings, a retiree (with possible help from her financial advisor) should plan on the possibility of incurring costs associated with long-term care. Under the actuarial approach advocated in this website, this is accomplished by including the present value of future expected long-term care expenses in a retiree's liabilities (and the present value of any long-term care insurance as an asset) when matching the retiree's assets and liabilities. In our post of December 21, 2015, we indicated that for someone currently living in California, that present value might be in the neighborhood of $200,000, based on an assumption of three year's stay at an assisted living facility, one year's stay at a nursing home facility and future increases in estimated current costs (which clearly will vary from facility to facility) of 4.5% per annum. This is only a rough ballpark estimate, and I encourage you or your financial advisor to investigate costs in your area and make your own reasonable assumptions. However, if you do set up a reserve for future long-term care expenses, you should understand that this money will not also be available for you to "tap" for other expenses.
The "Excluding Social Security V 3.1" spreadsheet in this website can be used to estimate a current appropriate reserve (present value) for long-term care expenses by increasing current estimated long-term care costs with expected future increases in such costs, entering that amount as the desired amount of savings remaining at death (including possible reserve for long-term care), and solving for the accumulated savings that will leave that amount at death (with little or no withdrawals prior to the end of the input period).
This is a trial and error process.
One of my actuary friends indicated that I punted somewhat on this issue in my post of December 26, 2015 when I had my hypothetical retiree, Richard Retiree simply assume that the equity in his home would cover his future long-term care expenses. If Richard's home equity were significantly more or less than the estimated present value of his future long-term care cost and he plans on spending his home equity during his retirement, he could include the current value of his home equity as an asset in the spreadsheet and the future expected cost of long-term care as a liability as discussed in the paragraph above.
Readers may be interested in several recent articles on this subject from Dr. Wade Pfau:
http://retirementresearcher.com/costs-and-incidence-of-long-term-care/
http://retirementresearcher.com/the-continuum-of-long-term-care/
Some of the key takeaways from Wade's articles include:
- Long-term care is a generally a larger need (and therefore a higher expected cost) for women than men.
- "The odds for needing long-term care are higher for individuals with greater longevity in their family history and those with a family medical history including dementia, Alzheimer’s disease, and neurological disorders."
- "Costs for long-term care vary by geographic region, type of facility and services used, and reasons for care."
- "Generally, the cost of long-term care has risen faster than overall consumer price inflation."
Bottom Line: As Dr. Pfau points out, "Retirement planning is complicated." It involves much more than simply withdrawing X% of your accumulated savings each year and hoping that you won't run out of money. To do it right, you need to periodically compare your assets (your current investments plus the present value of future payments you expect to receive from various sources) with your liabilities (the present value of your future expenses). I understand that you may not want your spending budget to change significantly from year to year (as might be the case if you strictly balanced your assets and liabilities each year), but at a minimum, you need to periodically check to see that your spending does not significantly deviate from the budget value determined using sound actuarial principles.
The purpose of this quick post is to encourage you to track and save a comparison of your actual spending vs. your spending budget every year in your spending budget file. If your spending budget significantly deviates from the budget determined using the Actuarial Approach, you may also want to track that difference. While the Actuarial Approach automatically adjusts your future budgets for actual investment experience and actual spending (if used without smoothing), it can be helpful for future budget setting purposes to have a sense of how well you have done historically when it comes to following your spending budget. For example, if your history shows that you have constantly overspent your budget, you may be understating your spending needs when developing your budget. Or, if your history shows that your overspending relative to your budget is increasing from year to year, this can be a signal of financial problems on the horizon. On the other hand, a history of ever widening under-spending relative to budget provides evidence that you may be too conservative with your spending.
Do you need to track your spending exactly? While it may be helpful to do so, particularly if you develop your spending budget as the sum of several different categories, it probably isn't necessary. If you know all of the items in the equation below other than your spending, you can solve for the amount you spent during the year.
End of Year Assets = Beginning of year assets + investment income + income from other sources (such as Social Security, pensions, annuities, etc.) - amount spent
Regarding the other important assumptions used in developing your spending budget, you can also track how well you did on your investments vs. your assumed rate of return and actual inflation vs. assumed inflation. Of course, every year that you and your significant other (if you have one) survive to work on a new year's spending budget should be considered a good year.
Happy budgeting and wishing you all the best in the upcoming year.
It’s that time of year again to sit down to determine your spending budget for the upcoming year. See our post of December 21, 2015 for a discussion of our recommended assumptions for 2016 spending budgets. It just takes a little data gathering and number crunching, but you will find that the time it takes will be well worth your while. Yes, you may have to miss an entire half of one of the upcoming important football games.
The rest of this post will illustrate the Actuarial Approach for Richard Retiree, the hypothetical retiree we have visited each year around this time for the past three years. The last time we visited Richard was in our post of January 1, 2015 when he developed his 2015 spending budget. To refresh your memory, Richard retired on December 31, 2012 at age 65. After a good investment year in 2014, Richard decided to take $100,000 of his January 1, 2015 accumulated savings and segregate these assets in a "rainy day fund" for unexpected future expenses. This left him with $821,853 in investments, his annual Social Security benefit of $20,340 and an annual life annuity of $15,000 from which to develop his spending budget for 2015. He also owns his home, but he does not consider his home equity to be part of his assets for budget setting purposes, as he plans to use the proceeds from the sale of his home to cover future long-term care expenses. He has no heirs so, he plans on having $10,000 or more available upon his death to cover death expenses.
Last year, he determined his essential expenses to be $55,000 and he decided that he would invest 33% of his assets in equities and 67% in fixed income securities. His total spending budget for 2015 was $66,884 ($31,544 from accumulated savings + $15,000 from his annuity + $20,340 from Social Security).
He earned a 2% rate of return on his rainy day fund and didn't use that fund for any emergencies during the year, so as of the end of 2015, he has $102,000 in the rainy day fund. His total investment return for 2015 on his non-rainy day investment funds was $15,000. Since his equity investments were down mid-way through 2015, Richard decided that he would try to limit his spending somewhat during the year and he ended up spending just $59,000 (including taxes). Thus, at the end of 2015, his non-rainy day accumulated savings are $813,193 ($821,853 (beginning of year assets) + $20,340 (Social Security benefit) + $15,000 (annuity benefit) + $15,000 (investment return) - $59,000 (total amount spent).
For 2016, Richard (who is now age 68) has decided that he will break up his total spending budget into several different component categories and dedicate assets to each separate category:
Long-term care: Richard has looked into the cost of assisted living and nursing home care in his area and has determined that the cost of three years of assisted living and one year of nursing home care would be about $180,000. He assumes that selling his home would cover this cost, even if long-term care costs increase at a faster rate in the future than the projected proceeds from selling his home, but he will continue to monitor the reasonableness of this assumption in the future.
Unexpected Expenses: Richard has $102,000 in this account and does not consider the money in this account for determining his 2016 spending budget.
Essential Health-Related Expenses: Richard's current health related costs are about $5,000 per annum. He goes to the "Excluding Social Security V 3.1" spreadsheet in this website and solves for how much of his accumulated savings he would need to provide a stream of payments starting at $5,000 per year and increasing by 4.5% per year over a 27 year period (95 minus his current age of 68) assuming he earns 4.5% per annum on the assets. He determines that this amount is $135,000.
Essential Non-health Related Expenses: Richard has determined that his annual essential non-health related expenses are about $50,000. Since his Social Security benefit for 2016 remains at $20,340, he will need $29,660 ($50,000 - $29,660) from his annuity and withdrawals from his accumulated savings. He again "backs into" how much accumulated savings plus his $15,000 annual annuity benefit will provide a stream of payments starting at $29,660, increasing by 2.5% per annum and leaving a fund of $10,000 at the end of 27 years. He determines this amount to be about $390,900.
Non-Essential Expenses: Richard has $813,193 in assets not allocated to his rainy day fund or his long-term care fund. He subtracts the amounts dedicated to his essential expenses discussed above to develop a total of $287,293 ($813,193 - $135,000 -$390,900). This is the amount he will dedicate to future non-essential expenses. He develops the portion of his spending budget attributable to non-essential expenses by inputting this amount in the spreadsheet as accumulated savings, 4.5% investment return, 0% future desired increases and a period of 27 years. This gives him a non-essential expense budget of $17,793. Richard knows that assuming 0% future increases for this item means that if all assumptions are realized in the future, his non-essential expense budget will not keep up with inflation. Richard also knows that he could be even more aggressive with this "front-loading" by assuming a payment period equal to his expected life expectancy (about 20 years under the Society of Actuaries' 2012 Individual Mortality Table with mortality projection) rather than a 27-year period, but he is happy with this result.
Total Spending Budget for 2016: Richard's total spending budget for 2016 is 72,793 ($5,000 for essential health related expenses + $50,000 for essential non-health related expenses + $17,793 for non-essential expenses). This budget is higher than the spending budget for 2015 of $66,884 even though Richard did not enjoy a particularly good year in 2015 investment-wise. This difference is primarily due to Richard's decision to somewhat front-load his non-essential expenses (in terms of real dollars). He decides that 2016 will be a year of more travel while he still able to do so. Richard notes what the expected amount of assets will be in each dedicated fund at the end of 2016 so that he can monitor his progress during 2016 and make spending adjustments if necessary (or possible).
Richard decides that he will talk to his investment advisor about how best to invest the various funds dedicated to his expense components. He feels that funds dedicated to essential expenses should be much more conservatively invested than his funds dedicated to non-essential expenses, and he may consider buying additional immediate annuities during 2016 if purchase rates become more favorable. He also decides that he feels much more comfortable with his decision to break up his expenses into separate categories, make different assumptions about how these expenses may increase in the future and treat them separately for investment purposes. And he feels infinitely more comfortable using this approach rather than using a "safe" withdrawal rate that doesn't consider his situation or goals.
This post updates my 2015 spending budget recommendations of February 14, 2015 and December 3, 2014. For 2016, Instead of recommending one set of assumptions for your entire spending budget, I will recommend different sets of assumptions for different component categories of expenses, as discussed in several of my recent posts (see the post of June 7, 2015 for an example). If you separate your budget into these different categories, your overall spending budget for 2016 will be the sum of the various component categories. You (and/or your financial advisor) may wish to consider segregating the assets dedicated to each expense category and adopting different investment strategies for these various dedicated funds. Each year you will compare assets in each dedicated fund with the liabilities for future expected expenses and make necessary budget adjustments and/or transfers between dedicated funds if appropriate. As a general rule, benefits payable from Social Security, pension plans and annuities are used to satisfy essential expenses first.
Essential non-Health Related Expenses
In brief, I recommend the same assumptions for this category of expenses as I recommended for 2015 in my post of February 14, 2015. As for prior years, I have recommended an investment return assumption that is close to the interest rates “baked into” immediate life insurance annuity quotes. Current immediate annuity purchase rates for a 65-year old male with a life expectancy of 22.9 years (based on the Society of Actuaries’ 2012 Individual Annuitant Mortality Table with mortality improvement) are consistent with an interest rate just a little bit lower than 4.5%. Therefore, I believe that the recommended investment return assumption for 2015 of 4.5% per annum is reasonable at this time. If the Fed continues to increase interest rates or immediate annuity rates change for other reasons, we may need to revisit this assumption during the year (as was the case in 2015). I continue to believe that a 2% spread between the investment return assumption and the inflation/desired increase assumption is reasonable and that 95 minus the retiree’s age, or life expectancy if greater, should be used for the expected payment period. See my post of December 3, 2014 for a graph that illustrates why I recommend using this expected payment period rather than life expectancy for this expense category.
If you use the recommended assumptions for this category, your effective goal for this budget component is to have relatively constant real dollar future essential non-health related spending at least until your late 80s.
Essential Health Related Expenses
Historically, health related expenses have increased about 1.5% to 2% faster per year than non-health related expenses. Given Medicare’s financial situation, there also appears to be trend toward passing these higher costs disproportionately to higher income individuals. If you consider yourself to be a relatively “higher income” retiree, I recommend that you use the same assumptions as above for this category, except with 4.5% annual desired increases (as opposed to 2.5%) for this category of expenses. If you do, then the fund you need to have to support these expenses this year will simply be your current expected payout period multiplied by your current essential health related expenses (adjusted if you considered some of your current expenses to be unusual and non-recurring in nature). So, for example, if you are currently age 65 (with a 30-year expected payout period) and $7,000 in current health related expenses, such as for premiums, co-pays, deductibles and non-insured prescription costs), then you should have current assets dedicated to future essential health related expenses of about $210,000 (30 X $7,000).
Long-Term Care Expenses
Since not everyone will require long-term care or they will pay for it with home equity that they don’t consider assets available for other expenses, this can be a difficult category to budget for. Data from the National Care Planning Council suggests that for those individuals who require long-term care, the average length of stay in an assisted living facility is about 2.5 to 3 years with many leaving the facility to go to a nursing home or another facility that provides more intensive (and expensive) care. The council also indicates that about 75% of the cost is covered by the family, not insurance. I recommend that you investigate current assisted living facility and nursing home costs in your area and plan on about 3 years at the assisted living facility and one year at a nursing home. For example in California average assisted living facility costs are about $45,000 per year and nursing home costs are almost double that amount, so budgeting about $200,000 for this expense would not be unreasonable if you plan on enjoying long-term care in California. I would also assume that these costs would increase by 4.5% per year (inflation plus 2%) so that you would need a dedicated fund of about $200,000 today to cover this future expense. If you do plan to use some or all of your home equity to pay for this expense, you should include an estimate of the home equity you plan to use in your current assets, when comparing your assets with your liabilities.
Other Unexpected Expenses
When establishing your spending budget, you will probably want to set aside a reasonable amount of assets dedicated to meeting unexpected future expenses (or a “rainy day” fund not dedicated to any of the other expenses discussed in this post).
Non-Essential Expenses and Bequests
If you have assets remaining after dedicating them to fund the other expense buckets above, you can use these assets to fund non-essential expenses such as travel, entertaining, dining out, gift giving, etc. Your bequest motive can also be included in this category or in the Essential non-health Expense category. Since some experts indicate that non-discretionary expenses are likely to decrease in real dollar terms as we age, it may be reasonable to assume less than 2.5% annual increases in future non-essential expenses or a shorter expected life period than recommended for essential expenses. If you do use these less conservative assumptions for determining how much you can withdraw from your non-essential asset account, you should realize that unless investment returns exceed the 4.5% assumption, your withdrawals from this account will likely decrease on a real basis over time. Note that amounts entered as bequests in our spreadsheets are nominal amounts and not real dollars, so if you want to leave real dollar amounts, you will have to increase the nominal amounts with assumed inflation.
We will illustrate operation of the 2016 assumptions in January when we revisit Richard Retiree and help him develop his 2016 budget and see how it compares with the budgets we calculated for him for 2015 and 2014.
