Hey folks, how you doing? I feel like its been aaaaaaaages since we had a good content-rich exploration of some investment issue. But come to think of it, it was only the beginning of January I wrote an article analyzing the difference between lump sum and dollar-cost averaging strategies. That got a Rockstar Finance feature that now takes pride of place on the AoF mantelpiece. So if you are a new reader, or a returning reader, then welcome!
Have you read the important notes before proceeding?
College Saving Investment Strategies
We’re going to look at college investment strategies and I have to say I really reveled in this analysis. Actuaries love a good revel so let’s get to work!
Do you know who cares about college savings? Yep, Moms.
And who is the Chief Mom of college savings? Yep, Chief Mom Officer (CMO).
Chief Mom Officer
Did you see CMO’s article on her decade plus journey to save for her three boys’ college education? It provides a testament to the power of consistency of savings and starting early. No gimmicks, no fluff, just getting the basics right year on year.
So I was delighted to provide her, and her boys, with some analysis that I thought would be helpful.
She has three boys of varying ages and levels of college saving, and I summarize the data below.
Eldest | Middle | Youngest | |
Years until college | 3.5 | 7.5 | 15.5 |
Current balance | $70k | $54k | $15k |
Planned yearly contributions | $6.5k | $3.9k | $3.3k |
Her target is to provide full funding for four years at their in-state College which amounts to a total cost of $110k. So you can see that the eldest child has a substantial chunk saved, amounting to 2.6 years’ worth. That’s a great result from CMO; with 3.5 years to go, my sense is that this is a readily achievable target.
Previous Analysis
If you have seen my previous analysis and case study you will recall that I prefer to convert the dollar amounts to fees for one year, where ‘1 unit’ represents a year of college fees, and ‘4 units’ is the cost of providing four years of school. I’m just going to concentrate on the eldest son in this article, and in those terms we have;
Eldest | |
Years until college | 3.5 |
Current balance | 2.57 units |
Planned yearly contributions | 0.24 units |
Target savings amount | 4.0 units |
Method
To perform the analysis I simulated portfolio returns using historical data going back to 1802. I also made sure I used ‘real’ returns that take account of the loss in purchasing power from inflation.
I regularly get the comment along the lines of “how relevant is the last two centuries of investment data, given we are in a totally different epoch of human development”? Honestly, I dunno. Basically, you take it or leave it. I really don’t want to come up with my own subjective assumptions on the future, or borrow some assumptions from elsewhere, and then have to engage in tedious and interminable discussion of those. At least this method is transparent – you know what it is, and can make your own judgement of relevance.
OMG! Give us the Results!
I know, I know! We’re 500 words in and not a single chart! This is not why you visited, I get it, but hang tight wannabe actuary buddies, we’re getting there.
Let’s first look at running-out the projections with the assumptions above and see what the final amount saved up would be. To recap for the eldest child, we will assume a starting balance of 2.57 units, an annual contribution rate of 0.24 units and a target of 4 units. Let’s also start off with a good-old 60/40 portfolio. i.e. 60% equities and 40% bonds.

The above chart shows all the different simulations with the starting year on the horizontal axis, and the vertical axis shows the final amount saved up. Remember each unit is one year of college costs. You can see that almost all the simulations exceeded 4, and so this means there is a very high probability that CMO can meet her target with this annual contribution. In fact, you can see that in many scenarios the amount substantially exceeded the target of 4, and she could have diverted some contributions to the other two kids. If you wanna get all actuarial about it (and who doesn’t right?) then 83% of the scenarios were a success.
Look at those two dips below the red line. These are the periods where the market did not cooperate and CMO did not meet her target of providing four years’ of college costs. In these periods the minimum balance was around 2.5; so even the worst period in two centuries it would not have been a very bad fail for the eldest. Note that we have seen these two pesky periods in previous posts, and they are the Great Depression and the Stagflation era of the 70’s.
More Results
Let’s show this in a slightly different way. For each scenario I calculated the minimum contribution rate (MCR) that would be required to exactly meet the target of 4 units. A low minimum contribution rate means the market was strong and therefore fewer contributions were required. During periods when the MCR was high the market was poor and so CMO had to plug the gap with additional contributions.

Just for variety I looked at a portfolio with 100% equities. Above you can see the MCR by historical period. It is striking how many scenarios that produced a zero MCR.
Yes, there have been plenty of historical periods where CMO could just sit tight and let the market do its work, without plugging in any more money! Nice work huh?
But look at those periods during the Great Depression and 70’s Stagflation where the MCR soars to over 0.5! That means CMO would have to contribute a half-year college cost each and every year to plug the gap. Urgh!
Want to be kept upto date with every new missive from Actuary Towers? I bet. Simply use the email sign up form in the main menu or follow me on Twitter @actuaryonfire. I’m also on Facebook somewhere – you’ll have to work that one out for yourself.
Investment Strategies
These charts are telling me that CMO’s planned annual contribution of 0.24 will be more than enough. In most cases she will probably be able to turn off the tap in the next year or two, it all depends on the markets, but she’s in a good spot.
We now need to optimize the investment strategy and balance risk and return. Let’s look at varying the investment strategy and see what impact that has on these factors.

I re-ran the analysis with varying equity/bond allocations and looked at how many scenarios achieved our savings target. The above chart shows the probability of success for these different portfolios.
You can see that equity-heavy portfolios are more successful and I wouldn’t recommend less than about 40-50% equities for the eldest child. What’s interesting is that the probability reaches a peak at 60% equities and then declines after. The impact of adding additional equities beyond 60% is detrimental. In other words the additional volatility that results from these equities outweighs any benefit in return.
I suspect that with the younger children there will be a different picture with a 100% equity portfolio dominating, but the short time period for the eldest child means that year by year volatility is important.
Efficient Frontier
We need a way of combining the risk and expected cost, and as in previous episodes I am going to create an efficient frontier.
- COST. I will measure this by the median MCR – how much do I have to pay each year to have a 50/50 chance of meeting my goal? A lower median MCR means a high portfolio return and we don’t have to pay in much contributions.
- RISK. I will measure this by the additional MCR to be sure of success in 95% of scenarios. In other words, how much extra do I need to pay to be 95% sure?
I measured these risk and return numbers for all the different portfolios we looked at above.

Let me orient you here – each dot is a different portfolio, and you want dots in the top left corner, since they have a low cost and a low risk. You can see that the 100% equity portfolio has lower cost than the 60% equity portfolio, but comes at a higher risk. That is the trade-off.
The 60% equity portfolio has a cost of 0.12 and a risk of about 0.25. What does this mean?
- A cost of 0.12 means that to have a 50/50 chance of meeting her goal, CMO needs to contribute 0.12 a year.
- However there is a 1 in 20 risk that she has to put in an additional 0.25 a year.
It’s worth noting that you would never invest 0-30% equities, since other portfolios are more optimal (i.e. lower cost, for the same risk).
Pause
Let’s just pause a minute to admire this chart.
Sometimes I go down into the actuary mine to dig out some gems for you and emerge with nothing more than dirty hands and weary limbs. But this is a diamond hewed straight from the data rockface. If you look at any efficient frontier in a textbook or Investopedia you will see an identical curve. What amazes me is that we’re dealing with centuries of investment data applied to CMO’s son, and we have found a sparkling example that looks as if it has jumped straight from a textbook. Don’t you feel a little misty-eyed at such an example of cosmic Platonism?
Advice!
Ok, dry your eyes guys, what have we learned?
Firstly, Eldest son is one lucky guy to have a mom like this. The current balance will be adequate in many scenarios even without further contributions.
I might encourage CMO to contribute the median cost of 0.12 a year ($3,300) and use the remainder for the other kids or taxable savings. Also, I would recommed CMO to adopt a 60/40 investment strategy, or perhaps 70/30 if she wants a bit more octane. I think 60/40 gives a great balance between risk and return for the Eldest child. I suspect the answer will be different for the younger children, but you’ll have to wait for a future episode.
Did you follow the analysis? Anything you would like me to do differently? How do you calculate how much you want to save for your kids’ college costs? Do you think my efficient frontier should have a Twitter account for itself? Comment below.
Technical Notes:
Note that I did not allow for college cost inflation to exceed price inflation. This was just to make the modeling a bit easier but it means my results are under-stated. If we assume that college costs rise 2% a year faster than prices then you should probably inflate my figures by around 10%. I was pretty relaxed for the eldest child, but this will be more important for the younger two.
My equity returns are just US equities. If I had used other asset classes like international equities and real estate than I am sure I could have moved the whole efficient frontier ‘north-west’. That would reduce the costs, and counteract the above point.
I also ignored the small tax break that CMO’s state gives for 529 contributions. That wouldn’t move the needle on this analysis. One misconception of actuaries is that they are uber accurate. Not true; we’re too busy attending wild actuary parties to worry about the small things.
Thanks again AOF for doing this analysis for me! It’s a relief to know I’m pretty well on track to hit my goal for the oldest-and even possibly be ahead of the game. If he has extra left over, I’ll roll it to his brothers. I’ve got a ways to go before they’re all through school, especially for the two year old.
Thanks for letting me do the analysis, I enjoyed it!
Very interesting and helpful analysis. CMO may be in even better shape than she realizes. The money won’t be spent in one chunk, on the first day of college. It may only be 3.5 years until Freshman year, but it is still 6.5 years until Senior year. An extra 1, 2, or 3 years of earnings on some of the units will add up.
Yes you’re right. My calculations actually allow for that now. I assume that the drawdown will happen over a few years, and I allow for the investment return over that period. But I didn’t really make that clear in the article. Thanks for raising it!
I generally feel looking at this money in segregation is a bad idea. After all money is fungible. In general I plan on either saving enough to pay college and pay for retirement. If I underestimate for either then I shift from one to the other.
I agree that you shouldn’t really have different ‘pots’ of money, since it is fungible. But you do have different future commitments or liabilities that need quantifying. My main commitments are kids’ college and retirement. So I need to work out my chance of hitting my goals for both. If I don’t hit one goal, then I can take from one pocket to the other, but then I fail the other goal. So I guess it is a case of quantifying the savings, establishing the investment strategy and ongoing monitoring.
There is a chord on the guitar called the Hendrix chord. It’s E7#9. It is what gives Hendrix his “Funk”. You be funkin with this post. Thank you for introducing the Efficient Frontier and a clear understanding that there are 2 variables to every portfolio Reward and Risk. High risk will kill you dead same as poor reward and it’s nice to see on a graph that there is a curve and a point on the curve where you get the most bang for the buck. It’s not just a guess but a calculation!
I would like to see something on disbursement. The money is disbursed over 4 or 5 years which has its own growth involved during spending. It’s a problem somewhat similar to retirement disbursement but much more finite and compact. It’s hard to get your head around 30-40 years but not so hard 4 or 5.
One safety valve is if you’ve saved for 3 years or the market crashes you can always do a couple years at community college at a dramatically reduced cost. One of my kids decided to start a photography business which is actually growing but not self sustaining yet. It’s as good an education as any. I have filled coffers for her education but if she wants to give entrepreneurial gig economy a whirl I’ got her back. Only thing is she needs to be in school with some forward motion toward a BA or BS. We have a local 4 year state college and she can take classes, live in my house, and work on her business. If it flames out or she decides on a different path she has 120 credit available funding at any university in the state, and the credits from the local college are completely transferable towards a degree from the university. Presently she’s on the Dean’s list even while growing her business. Her college cost is only $1100 a semester. So that’s a way to leverage college if you haven’t quite saved enough. If her business hits she has to get her 4 year degree from the local college and she gets the pile I’ve saved which continues to grow in the mean time.
I think it’s interesting given your negativity toward Monte Carlo that you have effectively become a Human Monte Carlo calculator. Monte Carlo calcs just give a projected normalized distribution of outcomes for given inputs and allow for worse case system analysis on a probability of failure basis. It’s no coincidence given a stock bond risk adjusted broad based portfolio you get a 60:40 mix and no coincidence you risk adjust the size of the inputs to yield a % failure rate. Very powerful stuff!
There is a chord on the guitar called the James Brown chord. It’s D9. You play it on the up beat for max funk effect… Get on the beam like a actuary machine…
oh wow E7aug9, I remember that from my guitar days! I then started a family and the actuarial exams and the guitar became history to me… But have not forgotten that. I think that chord should be the embodiment of my blog – not quite the neat satisfying resolution of Maj5th to Tonic, but ramping up the tension with E7aug9 before resolution to the root.
I like the idea of disbursement – I shall think about that.
Yeah, a Human Monte Carlo simulator – funny! I guess I use that in work all the time, so get a bit sick of it, and I like using the historical periods on the blog. What is utterly fascinating is how a 60/40 portfolio almost always wins, however you cut it.
Calc results depends on the number of truly non correlated assets in your portfolio. The calculator I use bases the risk and return on historical data so it is to some extent historically based. The longer the period of analysis the happier I am. The thing I like is I get a distributed prediction and some insight to tail risk which you don’t get with a mechanical 4% rule type calculator. For me it’s a both/and thing not either/or. Given the events of the past 2 weeks I’m glad to be diversified.
Sometimes I add the 13th when feeling really funky! E7aug9? I bet Jesus is just alright with you, OH YEA
This is great stuff. These kind of analysis are both fun and vary useful. Not being an actuary, I can’t critique your methods. But looks good to me!
Thanks for dropping by DocG and feel free to critique away!
Pingback: Thoughts on Market Drops (And CMO Around Town) ~ Chief Mom Officer
AoF, love these analyses that you do–very insightful. Quick question: on some of the charts showing timeframes back to 1800s, the first of the big dips (like in the first chart) looks like it lines up with 1914-1921 (World War I) more than the Great Depression, unless things are frame-shifted for me?
Also, what software do you use to analyze the data and make these charts?
Hmmm good spot. I need to look at that, I wonder whether the scale got screwed up somehow. Thanks for pointing that out.
I use MATLAB to do the calculations.