What’s wrong with Target Date Funds?

November 9, 2020

Amazingly, after 4+ years of blogging and 200 posts, I haven’t written anything about Target Date Funds (TDFs). For some folks, they are certainly a neat tool. Your fund provider automatically allocates your regular retirement contributions to a portfolio that they deem appropriate for your age and/or the number of years you’re away from your retirement date. It’s a hands-off approach for people who don’t want to think about their asset allocation and simply outsource that task to a fund manager.

But I think not all is well in the TDF world. People planning for FIRE should stay away from TDFs. But even for traditional retirees, there are some unpleasant features. Let’s take a look…

What’s right with TDFs?

First, in all fairness, I want to point out some of the features I really appreciate about target-date funds. TDFs are an easy, hands-off life-cycle asset allocation tool for people who, without any other guidance, might just “invest” their 401(k) savings in the money market option at a 0.1% annual yield. Or even worse, without the TDF option in their 401(k) plan, people might capitulate altogether and not participate at all in the retirement plan for fear of “doing something wrong.” 

I also like the whole TDF philosophy because it has made it a lot more palatable for a lot of employers to “auto-enroll” employees in their 401(k) and direct those contributions. Research has shown (see here and here) that participation rates increase if new employees are enrolled by default rather than employees having to spend the effort to sign up. It’s almost hard to believe that a lack of participation is due to laziness, essentially. And more participation in companies with auto-enrollment is due to people being too lazy to opt-out. But that’s what research appears to indicate. How crazy is that? In any case, the auto-enrollment is a lot easier to sell if employers can default employees into a generally accepted and approved investment allocation. It’s a CYA measure! You lower the lawsuit risk. If Joe Shmoe finds out that his 401(k) account went down from $10,000 to $7,000, then the company simply tells him, “Hey, don’t sue us! This is what Fidelity recommended for your age group!”

Why is there a glidepath toward retirement anyway?

Glidepaths toward retirement normally involve starting with a high equity share and low bond share for young investors and shifting to a portfolio with lower risk and thus lower equity share over time as they approach retirement. See the glidepath parameters for the Vanguard TDF family with retirement dates 2015 through 2065. Notice that they still keep the 2015 fund around, so this for folks who retired 5 years ago!

Vanguard TDF Weights Table
Glidepath asset allocation as of 9/30/2020. Source: vanguard.com

If we now assume that the glidepath parameters as a function of years to retirement stay the same over time (i.e., the 2065 fund in 2065 will look like the 2020 fund in 2020), then we can plot the projected path of the asset allocation over the life-cycle of a typical employee, see below:

Vanguard TDF Weights Chart
Glidepath over the life-cycle. Source: vanguard.com as of 9/30/2020

There are (at least) three explanations for this glidepath shape: 

  1. People become more risk-averse as they age: If I look at my driving style 25 years ago vs. today I definitely notice that I’m a lot more cautious now. If older folks are more risk-averse they should have lower equity shares. Plain and simple!
  2. People want to take less risk as their assets grow: Becoming more cautious may not even be due to age alone. If you look at two investors, one with a $10,000 retirement account and one with a $1,000,000 portfolio, who do you think would be more willing to allocate 100% into equities where you’d occasionally observe a 30% or even a 50% drop? Well, measured in percentages alone both investors face the exact same risk. But a 50% drop in a $10,000 portfolio seems a lot easier to swallow than a 50% drop in a $1m portfolio. Since older investors likely have a larger net worth you can apply the “if you won the game already, stop gambling” logic and push for a slowly declining equity weight as investors age. Thus, a lower equity share might be justifiable even if young and old folks don’t differ substantially in their risk parameters. 
  3. A changing “implicit” asset allocation over the life-cycle: Related, but not necessarily identical to the points above, one of the reasons why younger investors should take more risks with their portfolio (i.e., a higher equity allocation) is that they have more “runway” ahead of them in the form of future cumulative earnings potential. A stock market drop is much easier to swallow if you have 30 more years of future contributions than for an investor who is 3 years away from retirement. In other words, one could view the cumulative sum of future contributions as an “implicit bond portfolio.” And the equity glidepath in the financial portfolio is there to counterbalance the shrinking implicit bond portfolio as the investor ages.

So far, so good. Let’s take a look at eight reasons why I still dislike TDFs:

Dislike #1: An additional layer of fees

Most of us in the FIRE community are extremely cost-conscious. Almost maniacally cost-conscious. We don’t like actively-managed funds with high expense ratios and prefer low-cost passive index funds instead. So, if you ever even consider a target-date fund make sure that the underlying components of it are all passive funds. That would be the case with the Vanguard TDFs. Fidelity apparently has two TDF fund families one with expensive managed funds and one with passive funds only. For example, for the 2030 TDF, they have the FXIFX with a 0.12% annual expense ratio, but also FFFEX with a 0.68% expense ratio.

But even with Vanguard, the TDFs have an additional layer of fees. For example, the Vanguard expense ratios for the 2015 to 2065 TDFs range from 0.13% to 0.15%. If you had built the same asset allocation with the underlying funds yourself you’d pay only 0.07%. You could even move your money to Fidelity and achieve the same result with an even lower expense ratio. All the way down to 0.003% for the 2045 to 2065 funds, thanks to the zero-expense-ratio funds for both domestic and international equities!

Vanguard TDF Expense Ratios
Expense Ratios of Vanguard TDFs vs. reconstructing them from scratch with Vanguard (VTSAX, VTIAX, VBTLX, VTABX, VTAPX) and Fidelity funds (FZROX, FZILX, FXNAX, FBIIX, FIPDX). Source: vanguard.com, fidelity.com

Of course, 0.10% to 0.15% may not seem like a lot of money, but most (all?) of us in the FIRE community strive to accumulate seven-figure retirement accounts, and in a million-dollar account, that would be $1,000-$1,500 a year. Why would you want to pay that kind of money every year to someone just shifting around a few fund allocation weights every once in a while? And again, this is not active management, not stock picking, not market timing, nothing that requires an army of finance, economics, and physics PhDs to run. It’s merely rebalancing a portfolio of about 5 mutual funds back to the target weights. Every DIY investor, every Bogelhead, every FIRE disciple should be able to do this!

Dislike #2: Potential Tax Inefficiencies

Most of the time, TDFs are held in tax-advantaged accounts. In fact, as I stated above, you’ll most often find them as the default allocation in your 401(k) plan. If you hold your TDF in your 401(k), there’s no tax inefficiency. Not so in a taxable account, though. Because of the frequent rebalancing in the TDFs due to 1) rebalancing back to the target weights and 2) changing the target weights over time, you’ll generate taxable events. Not only that, most of the taxable events will be capital gains:

  1. Rebalancing back to the target weights means that you sell winners and buy losers.
  2. Lowering the equity weight over time means that you will likely sell equities who have a higher potential for generating capital gains than bonds.

Dislike #3: How do you handle your non-retirement accounts?

For those of you who objected that item #2 is no big deal because it only impacts taxable accounts and you will hold your TDFs in a tax-free or at least tax-deferred retirement account, there’s still a headache I like to point out to you. Let’s assume for a second that you’ve convinced yourself that the Vanguard TDF asset allocation path is indeed your favorite and accepted asset allocation over the life-cycle. I don’t think it should be (more on that below) but let’s ignore the pesky objections from cranky old Big Ern for now. Most of us, especially in the FIRE community, will also have regular taxable non-retirement brokerage accounts. If our retirement account is invested 100% in your TDF, then by simple arithmetic, your taxable account has to match the TDF allocation as well in order for your overall allocation to match your target allocation. So, you do introduce the tax efficiency headaches into your taxable account after all. 

You will also likely run afoul of some of the tax arbitrage analysis I performed in Part 35 of the Safe Withdrawal Series. There, I made the point that it’s highly unlikely that you want to keep the same asset allocation in both taxable and retirement accounts. Let’s look at the following example. Imagine an investor who has $100,000 in a taxable account and $100,000 in a retirement account and wants to implement an 80/20 overall allocation at the current spot on your glidepath. If bond yields are high and you face a high marginal tax rate on ordinary income (bond interest = ordinary taxable income) you will likely want to hold the entire bond portion of $40,000 in the retirement account. So you would end up with a 100/0 portfolio in the taxable account and a 60/40 in the retirement account. The classical recommendation of holding tax-inefficient assets in the retirement account.   (side note: we’d also have to factor in the different after-tax values of $1 invested in taxable vs. tax-deferred vs. tax-free accounts, but let’s abstract from that for this simple example!)

Conversely, if bond yields are low and/or you face a low marginal tax rate you might benefit from the White Coat Investor recommendation of holding the low-yielding asset in the taxable account, which means you have a 60/40 allocation in the taxable account and 100/0 in your retirement account.

Another headache would be the REITs allocation in some of the TDF families. REITs are extremely tax-inefficient because dividends are considered non-qualifying on your U.S. tax return and thus taxed at a higher rate. So, if you prefer a TDF allocation with REITs in your overall allocation, you’d be smart to keep all of your REITS in the retirement accounts and reshuffle the rest to fit the overall asset allocation.

Consequently, even if you really prefer the TDF asset allocation on an aggregate level you’ll likely leave money on the table through tax inefficiencies unless you get your hands dirty and maintain and rebalance the whole shebang yourself not just within but also across different accounts and especially account types; taxable vs. tax-deferred vs. tax-free. So for example, if your overall allocation drifted too far away from the target because equities performed strongly, then you don’t liquidate equities in your taxable account, but rather do the rebalance entirely in the tax-deferred account. So you’d have to overcompensate in the retirement account because you’re stuck with the overweight in the taxable account. Which defeats the whole purpose of simplicity in TDFs, doesn’t it?

Dislike #4: Young investors should have 100% equities

I’ve looked at various different TDF families from different providers and one feature sticks out: They never start at 100% equities. The highest equity shares are just about 90% at the large TDF providers: Vanguard, Fidelity, and T. Rowe Price. Even lower in some other fund families. I would normally recommend that investors who are just starting out on their path to retirement savings should just keep it simple and put 100% into equity index funds. Nobody can tell me that a 25-year old making $50,000 can’t take a 100% equity allocation in their $10,000 retirement account. This person has another 40 years of work-life and contributions, another 5 or so recession/expansion economic cycles, and probably more than 5 bull/bear market cycles ahead of him/herself. My personal opinion is that a 10% bond allocation is 99% fig leaf and CYA on behalf of the plan administrator and 1% diversification benefit. 

For full disclosure: Some experts find that “Target-Date Funds Are Too Risky for Savers” (Yahoo! Finance) but that’s not based on any actual analysis, just on “feelings”. To make this less about feelings and more about analysis, let’s take a look at how “Optimal” glidepaths would look like pre-retirement. I put the “Optimal” in quotation marks because nothing is really optimal, it’s all subject to the objective function and the assumptions you make along the way and subject to another really big bummer, see item #7 below! In the chart below, I contrast the Vanguard TDF to my own optimization calculations. I don’t want to get too much into the weeds of my simulations and optimizing, but for the math wonks:

  • It’s a Monte-Carlo simulation
  • There is a concave (risk-averse) utility function over the final payoff. Constant Relative Risk Aversion (CRRA) with a gamma parameter at 3.0.
  • I don’t search over the entire universe of possible glidepath shapes (a gazillion different combinations!) but restrict myself to a start point, an endpoint, and several “kink points” along the way so I’m able to simulate flat spots and concave or convex shapes.

So, how do “optimal” glidepaths look like? My initial “Optimal” GP kept equities at 100% until 20 years before retirement and then shifts down to 60%. But notice that the entire red line is always above the Vanguard GP. Maybe I didn’t set the risk aversion parameter high enough and that explains the 100% initial allocation. So, then I jacked up the risk-aversion to a pretty unrealistically high level and get the yellow line. And I’d indeed push the final equity share to 40%, well below the Vanguard 50%. In fact, this very risk-averse investor would spend about 27 years before retirement with a lower equity share than under the Vanguard assumptions. But the initial equity allocation still stays at 100%! Even the craziest risk-averse investors would still want to start with 100%. And the rationale, again, are the two points I made in the intro. 100% equities might imply a high risk in percentage terms. But in dollar terms, it’s still relatively low-risk, considering the big chunk of future contributions, which almost works like an implicit non-shortable bond allocation when starting out.  (and yes, yes, yes, I’ve done a lot of sensitivity analysis and the 100% starting point is indeed quite robust. I am not going to get into the details here but it’s something I worked on 4 years ago while still advising clients on this issue)

Vanguard TDF Weights vs MC Sims
Vanguard TDF vs. optimized GPs in a Monte-Carlo simulation. It’s very hard to push the initial weight below 100%.

So, my suspicion is that a lot of the industry TDF shapes are based on two assumptions baked into their “optimization”:

  1. A more than average risk-averse investor,
  2. An external constraint on the initial equity weight at about 90% (all the way down to 80% for some other TDF providers) that’s based not on hard science but mostly on legal and behavioral considerations.

Neither of the two applied to me. Nor should they apply to the FIRE-community. TDFs are calibrated for and, pardon the pun, “targeted” at a different audience!

Dislike #5: TDFs reduce the equity allocation too early

Another pet peeve I have with TDFs is that the equity share starts moving down too early. That’s the case both in the actual industry TDFs and also my Monte-Carlo simulations and optimization exercises. Imagine you just found the FIRE community and you want to employ a TDF on your path to early retirement in 10 years. Your Vanguard 2030 fund would currently allocate only 67% to equities and the remainder to bonds with absolutely pathetic yields currently below the inflation rate. And then your TDF would slowly shift to only about 50% stocks along the way.

I don’t think this makes a lot of sense for us in the FIRE community. I agree that you might want to hedge the risk of a market downturn if you are a traditional retiree at age 50 or 55 with a sizable nest egg and a savings rate of “only” 10-15% where you have less Dollar-Cost-Averaging opportunities to make up for major losses. But if you start out with $0 and a 50-60% savings rate you should have a much more aggressive equity share!

One workaround for this problem would be for FIRE savers to pick a TDF with a retirement date probably around 10-15 years after your planned retirement. For example, it’s 2020 now. A new FIRE enthusiast who likes to retire in the year 2030 would pick a Vanguard TDF with a retirement date of 2040 or 2045. Now you’d start out with 80-90% equities and would end up at around 67-75% equity share at your retirement date. The 67% to 75% equity share at the retirement date is not that far off from what I’d normally recommend for retirees both in the FIRE and traditional retirement scene. Yeah, the initial equity allocation is “only” 80-90%. I’d prefer 100% but if you must use the TDF, then so be it. 

Dislike #6: TDFs are potentially a bad strategy for people with a late start

Even for traditional retirees, TDFs may be overly simplistic. What if you get a late start to retirement savings? Imagine you’re a 50-year-old starting with zero retirement savings. It could be because you’ve neglected your retirement savings so far or you had some other life-changing event (business failure, divorce, medical bills, etc.) that reset your net worth and makes you start from scratch again. You face the same problem as the FIRE investor above. Starting with such a meek asset allocation will never get you to a comfortable retirement. Drastic circumstances call for drastic measures; you might have to aim for a higher equity share. Remember, the 75% equity share for the 2035 TDF is calibrated to a retirement saver who already has a sizable nest egg. It may make sense for that retiree to take some chips off the table but not for someone just starting out!

Dislike #7: TDFs violate an important mathematical principle!

For those of you with an economics or engineering or math background, you’ll agree that some of my objections to TDFs above can be succinctly summarized into: 

Target Date Funds violate Bellman’s Principle of Optimality

That principle is, let me quote:

“An optimal policy has the property that whatever the initial state and initial decision are, the remaining decisions must constitute an optimal policy with regard to the state resulting from the first decision.” Bellman, R.E. (1957), Dynamic Programming, Dover.

In other words, if you optimize the GP the way it’s currently done, setting equity/bond allocations merely as a function of the number of years to retirement, you will violate Bellman’s Principle. Take two investors, both 45 years before retirement, starting out with $0 initial assets and contributing $1,000 each month to their retirement plan TDF. They have an identical objective function over the final wealth. 10 years into their path, investor A has $100,000 and investor B has $300,000 in savings (because he/she started at a different time with better returns or got an inheritance). If we now re-optimize another glidepath for each investor with a 35-year horizon (again, the GP depending only on age), the same final value objective function but different initial net worth numbers ($100k vs. $300k) you can show that those glidepaths will have to differ because of the different initial wealth. At a minimum, one investor, probably both investors will now have reoptimized glidepaths looking different from the original glidepath for years 11 through 45. A direct violation of the Bellman Principle.

The mathematically correct way to construct a TDF would be to allow for a lot more “path-dependency” of glidepaths along the way, i.e., your asset allocation depends not just on your age but also on the asset level. This would be a stochastic dynamic optimization problem, which is actually not that complicated to solve. You would optimize this by “backward induction,” something that’s done routinely in economics and finance. But it might be a bridge too far for the typical 401(k) plan administrator! And since the lawyers will not understand it, we are stuck with the sub-optimal, one-size-fits-all TDFs.

Dislike #8: TDFs use the wrong post-retirement glidepath

At least I agree with the pre-retirement glidepath shape qualitatively, though not exactly quantitatively. The equity glidepath moves down as you age and we can argue and haggle how steep the slope should be. It’s a matter of a quantitative disagreement.

But that changes after retirement. Most TDFs shift further out of equities as evidenced in the Vanguard TDFs (49% equities at retirement start, 34% when 5 years into retirement). The same is true for most other fund families I’ve checked as well. 

But Michael Kitces pointed out that around your retirement date you should ideally have a “bond tent“: shift more into bonds before retirement and then shift out of bonds again in retirement. This is a viable solution to alleviating Sequence Risk. Not a perfect hedge – nothing ever is – but this glidepath out of bonds and back into equities certainly hedges a little bit of Sequence Risk. I confirmed that result in my simulations in Part 19 and Part 20 of the Safe Withdrawal Series. I got slightly different quantitative results because I rely on historical returns and Kitces/Pfau on Monte-Carlo simulations. But the same flavor prevails: early in retirement, you want to keep a lot of bonds around as a hedge against a bad bear market. But then shift back into stocks to make sure your assets last over the next 30 to 50 or even 60 years.

This one problem in the TDF design may not be the worst dealbreaker because there is an easy workaround. You most likely hold your TDF in a retirement account so if you’re happy with its allocation up until retirement but not during retirement, simply sell the the TDF at the end of your career and switch over to a mix of equity and bond index funds more in line with the Kitces/Pfau and ERN glidepath model. Or at least keep the allocation constant. And if your 401(k) plan doesn’t offer the inexpensive index funds, simply roll your account over to an account at Fidelity where you do have access to zero-expense-ratio funds. 

Conclusion

While some form of a glidepath shifting down the equity weight on the path to retirement may be desirable I don’t really care much for actual TDFs available today. Certainly, folks in the FIRE community should stay away. But there are even some headaches for traditional retirees. 

Also, this is not the last time I want to write about this topic. I’m particularly interested in the pre-retirement glidepath under different assumptions: Monte-Carlo simulations vs. historical returns. The impact of the initial portfolio value on the “optimal” glidepath. Different objective functions. How do glidepaths that do satisfy the Bellman Principle differ from the sub-optimal conventional glidepaths? I hope to address all that in one or even multiple future blog posts. If I find time in my busy retirement schedule to finally put that together! 🙂 Stay tuned!

Hope you enjoyed today’s post! Please leave your comments and suggestions below!

 

81 thoughts on “What’s wrong with Target Date Funds?

  1. All valid observations ERN; however TDFs are still the best choice for the average person who is not into personal finance like us.

    TDFs eliminate the analysis paralysis wrt which funds to select and how to figure out asset allocation. Many of my coworkers would either not opt-in to the 401K or leave the money in cash instruments because it was all too complicated for them.

    As I always told them, don’t let perfect be the enemy of good enough.

  2. Another great post Ern. Many thanks. I’ve often wondered about the shape of TDFs and hadn’t considered the political reasons. Learnt something new (again) today!

  3. Excellent post.

    Particularly appreciated the graph entitled, “Vanguard TDF vs. optimized GPs in a Monte-Carlo simulation” and the concept of the Bond Tent.

    I chose to use a VG TDF for more than 10yrs but, being generally aware of the conservative allocation, simply selected a TDF 15yrs beyond my planned ‘retirement’.

    Although we live by FIRE principles, I continue to love my job and have no desire to retire any time soon.

    Cheers from Seattle,

    Paul

  4. Thanks so much Karsten for once again a great in depth blog. I really enjoyed your thorough analysis of the downside of automated glidepath adjustion for FIRE adepts. I myself am struggling with this as well. Started out in 2018 with zero net worth. Building up my wealth in approx. 12 yrs to my estimated FIRE date in 2030. Now, of course, 100% equities and looking when to start reducing equity exposure in favor of bonds. Btw, age 40 now, age 50 at projected FIRE. SS and pension kick in at age 70 covering (nearly) all of required expenses. Any advice on this and/or more reading material?. Kind regards from your former neighbour country The Netherlands!

    1. Oh, wow, if you have to bridge only 20y between FIRE and then fully-government-funded retirement, you can certainly get away with a higher SWR. Check out my SWR Series Part 28 to check how much of a difference that additional income after 20 years would make!
      Greetings back to The Netherlands! We enjoyed our stay there in 2018 a lot!

      1. Thanks for your reply. For sure I will look into your SWR series part 28! Do you also have any (reading) tips specific for the glidepath question? Both when to start reducing equity exposure and when to ramp it up again post retirement.

        I’m thinking about keeping it (near to) 100% equities until FIRE date because in extreme market conditions I have the flexibility to keep working a few more years if necessary. And reducing it at FIRE date or few months beforehand for some period to some extend (40% bonds, 5 y period? just a thought) . And then raise the equity percentage a couple of years into FIRE. But I’m in doubt how to address this isssue, concerning the fact that I’ll probably only need to draw from my portfolio for some 20y. Can you help me out some more? And/or maybe one of the forum readers as well?

        1. “I’m thinking about keeping it (near to) 100% equities until FIRE date because in extreme market conditions I have the flexibility”

          There is your answer! I normally recommend keeping close to 100% when you’re flexible about the exact date!

          1. Yeah, that indeed is my plan pre FIRE. Thanks a lot Karsten, for your confirmation of my thoughts on the asset allocation whilst building wealth.

            Any thoughts on a sane asset allocation trajectory after FIRE date? Especially since I only really need to draw for 20-ish yrs. E.g. 70(stocks)/30(bonds and maybe small part cash) for the first X years, then ramping it up to for example 80/20? Or should you suggest a larger or smaller fixed income percentage in the first X years? And what would X be ideally?

            Hope I can pick your brain once more for the right side of the bond tent 🙂 Thank you for being so helpful so far anyway!

    2. I am intrigued what you believe your expenses will be at age 70 vs today – at age 40, if I have understood you correctly?
      Also, what confidence do you give to this forecast?

        1. ERN,
          Thanks for the response.
          Assuming you adjust for some inflation forecast, I agree your approach, on average, will probably provide a good (or possibly even slightly conservative) estimate.

          I was really aiming the question at Rolf – my apologies for not making that clear.
          I just wondered if he had enough of a spending history to be able to assess his current expenditure as he suggests that he started his journey only a couple of years ago. In our case, at least, it took a few years to get a good handle on spending and it took even longer for any pattern to emerge – ultimately I found it best to integrate/smooth consumption over several years to even out the inevitable year-to-year variability. We then did experience an [occasionally written about] expenditure drop on pulling the plug. However, I cannot yet say if that will be permanent or is just an aberration – possibly due to my paying a lot of attention to spending in the early years post retirement.

            1. Yup – that is essentially it.
              FYI, over the last decade our (inflation corrected) min to max ratio on an annual basis is around 1:2, whereas on a monthly basis (not necessarily in the same year though) it is around ten times that.

      1. I live quite a lean and frugal life. Not wanting or needing much ‘stuff’ to be happy. And planning to have only a small if any mortgage left before FIRE. Now with young kids having some expenses that will be replaced by other categories when I’m much older. In the country I live, The Netherlands, we have a health insurance system – on which parts of Obama Care were based – that maximizes our health expenses through some sort of guaranteed maximum deductible. Of course no one knows what the future will look like and what rules will apply by then. But in my planning I use my current expenses as a rough estimate for future years. Using YNAB as a budgeting tool by the way helps me a lot in having full insight in my spending behaviour.

        1. OK.
          I also do not live in the US.
          AFAICT, all households have, to some extent, unique spending patterns through time – but it sounds to me like you have this aspect on your radar.
          Is it not the NL that wants/needs to make cuts to its state pension pay-outs?

          1. No, it’s not state pensions in NL that are to be reformed. Employer pensions are though, a new pension system is planned for 2026. In this system there won’t be any more defined benefits, but more or less personal pension buckets. Some risks will be shared among participants by pooling the capital. I’m very interested in how this will work out. Time will tell.

          2. FWIW, I’ve done some work on the importance of non-flat spending patters in retirement. Not published (yet) but the gyst is:

            seasonal patterns (intra-year fluctuations): irrelevant
            Y/Y fluctuations, so generally flat path but with +/-20% or even +/-50% fluctuations: also largely irrelevant
            Some very large expenses many years down the road (nursing home): a moderate impact on your SWR. Surprisingly low.
            A path slowly drifting up by: a large impact on the SWR.

            1. ERN,
              Thanks for the heads up.

              Are you familiar with the works of Banerjee (2012)
              (see https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2007190)
              and Blanchett (2013)
              (see https://www.morningstar.com/content/dam/marketing/shared/research/foundational/677785-EstimatingTrueCostRetirement.pdf)
              in this area?

              Both of which use US data.
              There are other studies around – not all of which are US based – but IMO these are the two best known ones for the US.

              1. Both papers are interesting. In the Blanchett paper, looking at figure 6 it seems that the largest downdraft in expenses is in the group low assets, high initial spending. They are the people who HAVE to reduce spending.
                I’m still going to plan for a flat path, especially between now and age 65 or 70. Whether my expenses go down after that, doesn’t make a huge difference for the current SWR. It’s too far in the future to matter much.

                1. I understand your conclusion [essentially flat] especially for the earlier years. Hence my comment above that flat/slightly rising will “probably provide a good (or possibly even slightly conservative) estimate.”

                  Will your (still to be published) work on “non-flat spending patters in retirement” consider spending reductions with age too? I suspect this scenario might throw up some interesting findings – even if/when coupled with a significant late spending ramp-up (aka care home),

                  Lastly, I agree both papers I linked to focus on “conventional” retirement ages/dates and thus are of limited use, at least initially, to early retirees. This point did occur to me after I posted the two links, and some more relevant info is available at:
                  https://ilcuk.org.uk/wp-content/uploads/2018/10/Understanding-Retirement-Journeys.pdf
                  IMO this is an excellent paper that hails from the UK, and whilst primarily focussed around conventional retirements (by age) does explore overall spending patterns – primarily to compare and contrast with elder spending – but, the younger spending data is presented. This paper, amongst many other things, also attempts to explore why spending declines with age – ie is it because it has too or does something else drive this behaviour.

                  Personally, I recognise the annual through-life spending shape described in this paper. However, I suspect it is highly idiosyncratic to each household and is probably based around life events rather than specific ages. Albeit there is some correlation of such life events with age – at least to the extent that some events generally happen after others and thus when you are older.

                  Very recently Blanchett co-authored a paper (see: https://cdn.ymaws.com/dciia.org/resource/resmgr/resource_library/DCIIA-RRC_RightSizing_091020.pdf)
                  where they use an approximation of fundedness immediately prior to retirement to further explore why spending declines with age.
                  Banerjee also touched on this question in 2018 at: https://www.ebri.org/docs/default-source/ebri-issue-brief/ebri_ib_447_assetpreservation-3apr18.pdf?sfvrsn=3d35342f_4

                  I trust you might find these of interest too?

  5. As always, another great article. I always learn something from your posts, and for each new thing I learn there are probably a dozen nuggets buried that I won’t understand until I come back again with some more knowledge and experience. 🙂
    I use Vanguard TDF’s in my 401k and HSA because it’s an easy way for me to get some basic index funds without having to sort through the available options and deciding what mix of other funds gives me somewhat of a total index. As someone else mentioned, this is great for people that aren’t super into the nitty gritty of personal finance. I use a TDF that’s much further past when I’ll actually retire because I wanted a higher percentage of equities. I then adjust my overall allocation when I buy/sell in my other brokerage accounts.

    1. Great idea! Glad we’re on the same page with the TDF with a later retirement date. But again, if you have the bandwidth, check if it’s worth to just replicate that TDF with the other, cheaper index funds if you have those in the plan! 🙂

    1. Nice. I agree with some of the items.
      I disagree with items 4 and 5, i.e., the whole small-cap and value stock mumbo-jumbo again. I doubt that these two styles will reliably outperform the broad index. But I guess you knew that. 🙂

  6. You say “My initial “Optimal” GP kept equities at 100% until 20 years before retirement and then shifts down to 50%” but plot is to 60% ?

  7. I’m a 28 year old making 76k a year and I am 48% equity and can barely sleep at night. Now you tell me I should be 100% equity and not sleep ever again? Something is wrong with me or you. I just suffer and sweat so much to earn my money that I can’t afford losing anything….

      1. I’m earning 1% garanteed and not risking losing 30 or 50%.what’s the point?
        This is why I don’t buy this whole fire thing. It’s so much easier just keep working and not just gamble with your sweat earned money. But I get if you make 6 figures or inherit a big chunk than you could risk more.

    1. Try to think about each dollar in the stock market as dollars you won’t touch for 10+ years. Take a peak at the worst 10, 20, and 30 year returns for equities and they don’t seem so scary. If that’s not enough, just throw it in a target date fund, don’t look at it and come back at retirement.

    2. I’ll reply to this in case you’re open to change your views on risk and this should have the necessary information for you to make a good decision. But you have to condition your emotions to not view this the same way as gambling at the blackjack table. While there’s some uncertainty in the total returns, it’s not gambling as long as you’re broadly invested in that asset class. Which you would be.

      With your proposed 48/52 split you’re at higher risk of outliving your retirement should you not want to live in abject poverty after working. This is called “Longevity Risk” and it’s a bigger threat than temporary drawdowns are for the 100/0 split. This risk is even more pronounced for FIRE goals. 100/0 is the historically optimal allocation for minimizing longevity risk because of the long-term return even though there’s more uncertainty around the average return. The only time people lose money during recessions is when they sell out of equities at unfavorable prices because of the same fear that you’re currently experiencing. Justin, it’s just a loss on paper. With 100/0, by the time you’ve encountered meaningful losses ON PAPER your total return will likely have already beat the 48/52. Go to a website like portfoliovisualizer.com and play around with the different allocations so you can see the big picture of what these gains/losses can look like when they’re graphed over multiple decades. You will find that 100% equities blows the other allocations out of the water. It’s not even a contest.

      If you want to do something that’s easier to look at but gets the majority of the returns of 100% equities, go to that site and look into 60% US Total Stock Market and 40% Long-Term US Treasuries. It’s important that the 40% is NOT in broad bonds or short-term fixed income or you WILL see a substantial reduction in total return.

      I gave you what you needed. Do what you want. Peace out.

  8. I work at a large 401k recordkeeper. While TDFs do lead to auto-enrollment and higher long-term retirement contributions, a large portion of employees holding TDFs don’t actually follow their glidepath or stick with TDFs. We constantly get requests from employers trying to figure out why their employees are so irrational. There are 25-30 year olds who hold 0%-20% equity. Fewer hold 100%. Never mind the fact that the vast majority of employees will never save enough for retirement. CARES Act led to a majority withdrawing the max and lowering their contributions, only to increase them after the market recovered. It’s sad, but makes me happy to be far ahead of most people.

    1. Back in the 1980’s I worked in HR at a company which matched up to 6% of salary dollar for dollar. And the company match vested after one year. I sold that deal hard in new employee orientation sessions and still couldn’t get more than about 50% participation.
      There were some constraints on how hard I could sell to the groups in orientation, but in one on ones with young HR people I pointed out that if one were to contribute $100 the company would match it and you would have $200 which could be withdrawn one year later at the cost of a 10% penalty to the IRS. So you could turn $100 into $180 in a year, surely the best investment deal on the planet.
      Despite that there were a couple of young HR people who simply could not defer so much as a dollar. And these were not scantily paid people.

  9. Thanks for taking time to write! I’m one of the people who started out with no concept of FIRE and used a target date fund in my taxable account 15 years ago. It’s a decent amount of my portfolio, about 30%. Its a 2040 fund and my original plan was to spend it down first and use the rest of my portfolio to counteract the automatic changes. But ick.

    Your article makes me think about changing that plan and instead paying the capital gains (or harvesting losses if I can find any) over time while I’m still working to divest myself of it before the auto rebalance starts kicking in.

    Time to do some modeling.

  10. In my 401k, the TDFs had lower ERs than the other options. However, I too noticed they were too conservative for someone on a FIRE path, whose savings rate alone can erase the effects of a correction. My solution was to just add 15 years to my expected retirement date and select that TDF!

    Incidentally, my 401k has outperformed my online brokerage accounts for the last few years. I attribute this to my own behavior – trading too much and being less diversified in my brokerage accounts. [accepts flogging] So for real world idiots like myself, the TDFs may deliver value, especially if you misuse them.

    1. OK, then stick with the TDFs! Good point. Not everyone has a good selection of index funds in the 401k. But it still means that it’s best to replicate the TDF with the underlying in a non-401k, say, in a Fidelity IRA/Roth/HSA, etc.. 🙂

  11. Thanks for the post. As always you’ve made me think again about increasing my stock allocation (at 72 and 10 years fully retired). I’m loathe to do that with the market at such heady P/E ratios. but inflation has to eventually present a real risk even though so many who claim to be economists seem to pretend that governments can literally spend as much as they want for as long as they want. Pondering this may put me over the top on the issue.

      1. hello ERN.
        Thank you for the article.
        And the 3% yield CAPE now – real or nominal? and why. This is important!
        thanks

  12. >> Dislike #1: An additional layer of fees

    Boom. A few clicks just saved me around 8 basis for the exact same portfolio at Vanguard.

  13. I find the idea of the bond tent problematic. Like most saving for early retirement, I consider the actual date to be entirely discretionary. Since that date is flexible, does reducing my equity allocation before actually committing to retirement make sense?

    Intuitively it seems like I would have the best average outcome by holding close to 100 percent equities right until I decide to retire, and then shifting to a more conservative allocation concurrent with that decision. I could always punt a year or two and keep working if the market is way down.

  14. Great article as usual. I’m wondering about dislike #2. My impression of how the Vanguard Target Funds work is different. Specifically, I believe they use inflows and outflows to keep target weights at desired levels. This allows them in turn to be very tax efficient and not pass along capital gains in the form of distributions (certainly not a definitive source for this information, but a reference: https://www.bogleheads.org/forum/viewtopic.php?t=75881 and a look at some historical distributions https://www.dividendinvestor.com/dividend-history-detail/vforx/). My impression is this might be an advantage of TDFs. That is assume I end up retiring with $1,000K 100% in equities in a taxable account where 500K is actually long term capital gains. To start retirement I want to have 40% bonds. So I would end up selling the $400K and having a 200K capital gain, which would incur a significant tax hit. In this scenario you probably have a mix of taxable and tax advantaged accounts so you may be able to use that to create the asset allocation you want without the capital gains. But I’m just curious how you look at this.

    1. During the really large market swings, it’s impossible to rebalance through new contributions alone. So, the TDFs most definitely cause capital gains along the way.
      Also, the tax liability from the distributions is not the real concern. You’d also have distributions if you hold the underlying funds yourself.

  15. As a ‘math wonk’ I definitely found your optimizations of glidepaths to be the most interesting part of your post!

    I have long felt that log utility isn’t sufficiently risk averse but have never looked at more generalized CRRA models. I am curious how you pick ‘gamma’=3 (or if there is any recommended reading?)

    1. If you pick gamma=3 or 3.5 and you do a one-time optimization, you’d get a ~60-70% equity share, assuming real returns 6%/1% and zero correlation.
      So, in the absense of any new contributions, as a buy-and-hold investor if you pick 60-70% equities it sounds like you have a decent and realistic gamma.

  16. Your comments here regarding the bond tent (rising equity glide path after retirement) brought me back to a discussion I had yesterday on another financial blog. My correspondent claimed you advocated the use of the rising equity glide path and I responded that you had analyzed it and found it to be theoretically valid but did not actually advocate its use in practice. Now I’m not so sure. Do you actually recommend retirement investors follow Kitces’ rising equity glide path model?

    1. In Parts 19/20 of my series I indeed propose a rising GP, in retirement.
      It is theoretically the optimal thing to do if you want to commit to a specific path ex ante. Of course, if the equity market doesn’t crash early in retirement and your portfolio in on target, you might as well keep the asset allocation.
      Another example where the fixed GP is different from the path-dependent GP that satisfies Bellman’s Principle.

  17. I agree with much of what you say, and they are reasons I’ve not gone down that route. That said, I think there are other advantages to Target Date (or the more fixed LifeStrategy) style funds. By being all-in-one, they conceal some of the decline when the stock market craters. Seeing one’s stock fund fall 30% may cause worse behaviors than seeing your all-in-one fund fall 20%. And people may be more hesitant to rebalance back into stocks in that scenario, which the all-in-one fund will do for you. I also like that it gets people into Int’l stocks, less excited about it getting into int’l bonds. But that may be a feature or a bug depending on what your opinion is of those asset classes. I think another positive all-in-one feature is that it discourages paying too much attention (e.g. you don’t need to watch values to rebalance as needed), and discourages tinkering. Both good traits for most investors.

    1. Yeah, you bring up the B word. Behavioral. I concede that TDFs might be useful for the very meek, scared, and uninformed investors. But most of my readers are OK with the vol during accumulation. We’re all Bogleheads here. For people who are still scared of recessions, I recommend JL Collins’ book.
      Also, I just have a huge problem advising people to do suboptimal things because they “feel” better. I’d much rather educate people about what’s right. Then, eventually, the mathematically optimal path will “feel right”! 🙂
      See this one about a similar “feel-good” issue: https://earlyretirementnow.com/2017/06/14/good-advice-vs-feel-good-advice/

      1. Exactly. I come from both a stats and psych background so I’ve seen both sides and at some points struggled to reconcile them. At the end of the day if an investor is rational then they would have the same fear of lower total return as the crowd has of sudden corrections. This assumes the market will always rebound but I probably have a higher chance of dying in a car crash than the economy not recovering. The behavior argument never had merit.

  18. I’ve got to tell you that I’m also so afraid of losing money that I don’t even invest heavy in bonds but short-term only. I’m 31 years old and my allocation is 75% JPST and 25% ITOT.

    1. Very strange. This is already the second commenter in a week or so with a similar fear of stocks problem.
      My recommendation: from now on put 100% of the new contributions into stocks and don’t open any statements for 10 years. Your bank account will thank you! 🙂

      1. It seems there are a lot of millennials out there who got burned in 2008 and never got back to the market. A post about how to overcome that fear would be very welcome

        1. Very true. It’s on my “maybe list”. But: Others have written about the psychological stuff already; how to overcome the fear of a market drop. Bogleheads, JL Collins. Not sure why I should write more about that.

          1. OK Thank you. Maybe you should write about it with a different perspective, not the psychological one.
            Idk, maybe prove to us that mathematically we can trust the board market will always go up in the future no matter what happens in the world (use Japan’s stock market case as a cautionary tale) and that investing in stocks is safer than investing in bonds and gold.

            1. Please see previous posts:
              Stock market always goes up: https://earlyretirementnow.com/2017/08/09/us-equity-returns-history-and-10-year-forecast/
              Especially this chart:
              https://i2.wp.com/earlyretirementnow.com/wp-content/uploads/2017/08/sp500-cumulative-returns.png?resize=863%2C626&ssl=1

              Bond risk:
              https://earlyretirementnow.com/2016/05/19/bond-vs-stock-risk/
              Especially this chart:
              https://i0.wp.com/earlyretirementnow.com/wp-content/uploads/2016/05/bondsriskierthanstocks_022.png?resize=863%2C629&ssl=1

              I never said that gold is inferior. See SWR Series Part 34. Small shares allocated to gold might not be a bad idea in retirement.

              The US market has never been as overvalued as Japan in late 80s. It’s not a good comparison.

  19. Thanks Big Ern for another great article ! What do you think of this other dislike : every dollar present in the savings account will not be invested for the same amount of time before being taken out as a retirement revenue. The income perceived at 85 wil get to capitalize 20 years more than the one perceived at 65.

    1. Very true. That’s another fly in the ointment of TDFs. They are usually calibrated/optimized to a certain final value at retirement. But that final value has to be transformed into a cash flow again, potentially with another GP during retirement.

  20. Hi Big ERN,
    The saying (&joke): “Great minds run in the same gutter” seems applicable in this case of the TDFs with you and Vanguard. Vanguard is now working on setting up different Glide Paths in retirement per this 20 Nov 2020 article. See this link:

    https://vanguardinstitutionalblog.com/2020/11/20/building-better-plans-with-a-little-help-from-our-friends/

    And see this other previous link:
    https://vanguardinstitutionalblog.com/2020/09/03/how-retiree-friendly-is-your-retirement-plan/

    Glad you are continually feeding us more good ideas.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.