gmaynardkrebs wrote: ↑
Thu Mar 08, 2018 12:03 am
thx1138 wrote: ↑
Wed Mar 07, 2018 8:18 pm
gmaynardkrebs wrote: ↑
Wed Mar 07, 2018 5:50 pm
thx1138 wrote: ↑
Wed Mar 07, 2018 4:34 pm
<For example if you assume in retirement you will spend $10,000/year then a 15% savings rate in this case means you save $1,500 each year of your accumulation>
Thank you very much for taking the time to do this. Trying to understand what you mean by the above: is that the same as plugging in 15/25/35/50% of a $10K income (real dollars) over 40 years, in order to achieve $10K a year in retirement for 30 years? That is the way I visualize the problem. Perhaps it's the same thing, but I'm confused a bit. Thanks!
Yes that's the same thing, you got it!
So, if it really is the same thing, it seems to me that the reason TIPS fail in all of your scenarios (even your 50% scenario) is that you are significantly raising the standard of living in retirement, by failing to take into account that no more savings are being deducted in retirement. Just to take the simplest example, suppose you work for 30 years with a steady $10K (real) salary, and will then be retired for 30 years. If you put away 50% of your income every year for thirty years in TIPS, it seems to me that there is a 100% certainty that you will have the same disposable income in retirement as you did during your working years ($5,000/year). It cannot be otherwise, can it? So, if the Monte Carlo simulation is giving you an answer anything other than 100%, I would surmise that you may be addressing a different question, or perhaps that the 5-year note is not a good proxy for TIPS. Note also, that under the original 40/30 hypothetical, putting away 43% of your income every year gets you a disposable income of $5,700 over the 70 year period with 100% certainty. Now, I have not done the math on the following, but if one considers that 30 year TIPS have historically earned about 2% real interest vs 0% in my hypothetical, and that one's expenses often go down in retirement (the kids are gone, no more tuition etc), we probably get closer to saving 25%-30% of one's income during 40 working years to still have 100% certainty of maintaining one's former living standard for 30 years in retirement, at least in our hypothetical.
I now have this terrible feeling that what I've said is incredibly stupid; my math skills have severely eroded over the years (I was once pretty good.) So can you show me where I have erred, in this admittedly simple model?
Your math is spot on. The reason I wrote the savings rate as percentage of required spending in retirement is precisely because we expect retirement spending to go down and by how much is completely a personal choice. That's why I decoupled the two. For instance, for a frugal person with a high salary their retirement spending could be as little as 20% of their income while working. That's what all the early retirement folks are doing.
For more modest incomes options are more limited but the standard recommendation of 70% spending in retirement is probably not a bad guess so if we need to save 50% of our retirement spending that would be 35% of our income before retirement.
So the above were all just very simplified examples to show the scenarios in which "risk free" investing can actually result in lower chances of portfolio success where "success" is meeting a predicted spending requirement over a 30 year retirement with a 40 year savings period. And indeed, the more you save the more you can drive "success" to 100% using a "risk free" asset like TIPS.
There is a huge elephant in the room though. These simple "success" scenarios all only look at one half of the equation - the uncertainty of what your investments will return and the distribution of those returns. They all assume that you have 100% certainty of what retirement expenses will be. What if there is a 1% chance that your retirement expenses will be say double what you expected? And what if there is a 5% chance your expenses will be 50% more than you expected? If TIPS allow you to hit your "number" perfectly with 100% certainty but also make it 100% certain you won't have any *more* than that this creates a problem. There is uncertainty in your expenses too. With the above numbers the "risk free" use of TIPS actually fails at least 5% of the time because there was at least a 5% chance that your estimate of your expenses was wrong by 50% and there is a 0% chance that TIPS might have returned 50% more than expected.
Consider equities now, what if at a given savings rate my optimal equities included portfolio has a 99% chance of returning three times as much as the TIPS portfolio based on historical data? Let's be a bit pessimistic because of course the future isn't history, but still say that there is a 97% chance an equities included portfolio will support just two times the spending rate of a TIPS only portfolio and there is a 2% chance that in fact it will support less than the TIPS only portfolio (again, based on the simulations run before this is actually a significantly pessimistic estimate compared to past history).
If we assume we know our expenses perfectly then it appears that TIPS is the winner. It has a 100% chance of meeting our spending requirements while the equities included has 2% chance of failing. But again, there is uncertainty in the expenses too. Small probabilities of *large* errors in that estimate (see later for examples). So using our example above, assume a 5% chance our expenses are 50% above what we expected. Now things look different. Now the TIPS portfolio fails 5% of the time. The equities portfolio has a 97% chance of meeting this unexpected extra spending which might occur in 5% of the cases. Combining both the return and expense uncertainties the equities portfolio actually has a slightly lower chance
of failing than the TIPS option. (EDIT: For clarity the approximate combined probability for the equities included profile is 2% failure to meet the expected expenses plus a 3% chance of failing to meet the 5% chance that expenses are 50% more than expected which is an additional 0.03*0.05 = 0.15% chance of failure for a total failure rate of about 2.15% for an equities included portfolio compared to a 5% failure rate for just TIPS. Depending on interpretation that difference might be "slight" as in 95% success vs. 97.85% success or rather extreme as in more than twice as likely to fail!).
The main issue that I think you are missing though is that by going TIPS only from the start the *cost* for a retirement success of greater than 99% or any other high probability number is more than three times as high for TIPS compared to "taking risk" with equities. That means for a given savings rate, whatever savings rate you choose, the equities person is going to have a 99% chance of having *at least* three times as much money to work with over their retirement as the person who stuck to the "risk free" all TIPS portfolio from the start. Three times as much is *the minimum* the equities person will have with a 99% probability based on historical data. They will likely have even more than that. Indeed they are taking a very small risk they might end up with less than the TIPS person. Based on historical data that risk is 0% which is of course silly, but the risk is still quite small. And they expect with very high (but not 100%) confidence that they will have *at least* three times the assets of the TIPS saver.
And this is where the whole "TIPS are risk free" thing completely falls apart. They are "risk free" only in the sense that you can perfectly predict your retirement expenses and all of your other life events from the age of say 22 when you start working and saving. Even if we say we can't predict capitalism and equity markets that factor of three is *huge*. Having three times as much money to spend over retirement gives you far, far, far more flexibility to deal with *spending* uncertainty that is very real.
The presumption that retirement spending is always lower is not actually true. Talk to anyone that has a spouse with Alzheimers. Costs of more the $4K/month for a period of 6-10 years is not unheard of. Because of our broken LTC system most retirees have been facing steadily increasing LTC insurance premiums leaving them the option of either significantly increasing their retirement expenses to cover the premiums or accept the risk of all of the retirement money meant for the surviving spouse to have to be paid for the end of life care of the deceased spouse. Not a pleasant choice.
Consider as well other late in life "unexpected" expenses. You mentioned kids being out of the house lower expenses. When do kids leave the house? Around mid-twenties typically. When do most debilitating mental health issues arise? Also mid-twenties. Hmmm... looks like there is a finite risk, one of around 1% or so (see https://www.nami.org/learn-more/mental- ... he-numbers
), that if you have kids when you expected your expenses to drop they will actually shoot through the roof. You always have the option of leaving your child's adult mental health care to the state of course - in most US locations that means prison these days. So most folks are going to have to pay for that themselves if they don't feel criminal incarceration of their mentally ill loved one is what they are looking for. The "risk free" TIPS saver is at this point up a creek with no options. They had a 100% "risk free" guarantee of meeting their expected spending requirements. Unfortunately they didn't have a 100% risk free estimate of those spending requirements. The "risky" equities holder, the one that now has a 99% chance of having three times the assets to work with actually has a lot of options now.
So we are *always* taking risks. We are trying to keep those risks as small as possible. The key point of "need to take risk" is illustrated above. When you add up all the uncertainties one of the biggest factors in reducing the risk of failure is higher returns over that long investment horizon. TIPS from birth gives a false sense of certainty. It minimizes only one aspect of volatility, the volatility of your returns. While it guarantees those returns to a high degree it also guarantees those returns will be very low compared to the alternatives with very high probability. Once we factor in the real uncertainties of expenses over that 70 year period we find that the "riskier" returns of equities actually result in a lower overall probability of failure when we properly account for the uncertainties in both *returns* and *expenses*.