This was my take on things back in 2007, when it was published. As of 2010, both my opinion and the math have sharply changed. I revisit the issue in a new post.
State-administered “529 plans” for education savings are another in the series of tax dodges doled out by the Bush administration (thereby further and regressively lowering the effective rate of taxation on the higher-income people most likely to avail themselves of such dodges). (Sticklers will observe that 529s predate Bush; true, but their extra tax favorability is a post-2001 invention.) On the bright side, they offer quite a good deal if you can find an investment that keeps pace with college tuition (and you don’t need to worry about also beating the tax rake, since 529 gains are tax free when used for qualifying tuition etc.). You do have to nominate a beneficiary when you set one up, but you as the controlling owner can change the nominee at any time to any blood relation (and it can even be yourself).
There are two types of 529 plans: one is like a 401k plan and involves picking a retail investment or mix of assets; I’m sure someone, somewhere has done the analysis to pick out correlates of tuition costs, so if you can find that, maybe you should look at the first form. The second 529 plan is somewhat more interesting. In the second form, the 529 plan is actually selling you “prepaid tuition.” This kind of thing is normally a terrible deal: fronting money for something way in the future is generally a sucker bet, one made against the collected wisdom of armies of actuaries by not-sophisticated-enough retail investors with both informational and scale disadvantages. But I have a few reasons for believing that the Washington State 529 plan (known as the GET: see their web site) offers an overlay in certain circumstances.
1. Each state may offer either or both types of 529 plans to its residents. As it happens, a couple of states that have offered guaranteed tuition programs have run into trouble with the plans being underfunded and unable to meet obligations, according to this article. Although this might raise an alarm in some folks’ minds, to me it says that tuition costs have been rising faster than can be achieved by even the professional money managers hired by state 529 plans specifically to meet that hurdle rate. Therefore, I see it as a sign that tuition is in general an expensive thing to guarantee and that if you can get a reliable guarantee (see below), you are getting the best of it (until or unless a mean reversion on tuition growth rates vs. inflation occurs).
2. Washington’s GET program is (ostensibly) backed by the full faith and credit of the State, unlike other states’ 529 guarantees. In my mind, combined with the evidence of other states’ underfunding difficulties, that means the WA GET is a good bet to get bailed out by the State at some point in the future. (In general, any time you see something other than a plain vanilla bond backed by the full faith and credit of an American government, someone in the government got snookered or corrupted, and the public purse is about to make its counterparties rich: see e.g. the PBGC, the S&Ls, the Federal Housing Enterprises, etc.)
3. The index for the GET tuition price is the most expensive state university in Washington, invariably the University of Washington, located in Seattle, a thriving and growing metropolis which is well positioned to weather many economic threats and therefore in which prices and incomes are likely to remain high enough to induce the University to charge steadily increasing sums for tuition. (Although I certainly don’t think that the outlook for Seattle is monotonically ever rosier, it seems a much better bet than somewhere like e.g. Montana or Idaho, where relatively small disturbances to nondiversified industrial bases could result in stagnation at all state universities.) The University also has competition in town from a number of private schools charging full freight to their students, therefore establishing the viability of increased tuition at UW.
5. The WA GET program has a scam built into its marketing strategy: rich people can buy the tuition units at the “buy-in price” (see below), but they try to sign up poor people for a payment plan where they charge them 7.5% interest on top of the buy-in price. Whether this is morally appropriate or not is a separate issue; the fact that we can spot the fish (and we ain’t it) is good for us.
6. The big catch to the WA GET is the difference between the buy-in price and the pay-out value (the bid-ask spread, if you will). For the past few years, that has looked like so:
Year Purchase $ Purchase YOY % Purchase CAGR 1999 $35.00 NA NA 2000 $38.00 8.57% 8.57% 2001 $41.00 7.89% 8.23% 2002 $42.00 2.44% 6.27% 2003 $52.00 23.81% 10.40% 2004 $57.00 9.62% 10.25% 2005 $61.00 7.02% 9.70% 2006 $66.00 8.20% 9.48% 2007 $70.00 6.06% 9.05%
Year Payout $ Payout YOY % Payout CAGR 1999 $33.75 NA NA 2000 $35.19 4.27% 4.27% 2001 $36.42 3.50% 3.88% 2002 $38.64 6.10% 4.61% 2003 $45.31 17.26% 7.64% 2004 $48.63 7.33% 7.58% 2005 $51.81 6.54% 7.40% 2006 $55.05 6.25% 7.24% 2007 $58.80 6.81% 7.19%
Year Purchase $ Payout $ Payout Ratio 1999 $35.00 $33.75 96.43% 2000 $38.00 $35.19 92.61% 2001 $41.00 $36.42 88.83% 2002 $42.00 $38.64 92.00% 2003 $52.00 $45.31 87.13% 2
$57.00 $48.63 85.32% 2005 $61.00 $51.81 84.93% 2006 $66.00 $55.05 83.41% 2007 $70.00 $58.80 84.00%
Observations on these numbers:
- 2003 saw a big bump in tuition and hence in both purchase price and payout values. This jibes with 2002 news reports of the same.
- The payout ratio (payout $ / purchase $) has been declining steadily, but has recently hovered around 84%.
- That said, WA GET is raising the price of purchase by 9% annually, while tuition has been increasing at 6-7% annually. If these two don’t converge, the payout ratio will get worse.
- This supports a 6-7% tuition increase rate, especially because that makes the 7.5% interest on the payment plans accretive to GET’s situation.
7. The time-lag spread between purchase and payout makes sense only if you assume that college costs rise faster than the rate of return you can make elsewhere, adjusted for tax treatment. Assuming a fairly conservative 5.25% risk free rate and a 28% tax rate, the line crosses with 7% annual rising educational costs after about 6 years:
Years Held After-tax return risk free GET Return GET to risk-free % 0 $70.00 $58.88 84.11% 1 $72.65 $63.00 86.72% 2 $75.43 $67.41 89.37% 3 $78.36 $72.13 92.05% 4 $81.45 $77.18 94.76% 5 $84.69 $82.58 97.51% 6 $88.11 $88.36 100.29% 7 $91.71 $94.55 103.10% 8 $95.49 $101.17 105.94% 9 $99.48 $108.25 108.82% 10 $103.67 $115.83 111.72% 11 $108.09 $123.93 114.66% 12 $112.73 $132.61 117.63% 13 $117.62 $141.89 120.63% 14 $122.77 $151.82 123.67% 15 $128.18 $162.45 126.73% 16 $133.88 $173.82 129.83%
As you can see, given these numbers, it doesn’t make any sense to buy units for a teenager. You’re just going to get hammered down by the time-lag spread if you hold less than 6 years before redemption. Of course, it gets a lot worse if risk-free interest rates rise a lot compared to college tuitions, or if you’re paying more than 28% in tax.
However, if you think that college tuitions will maintain their higher growth rate relative to the risk rate, and if you can hold for well over 6 years, then you should strongly consider buying tuition in the GET. Why do this if you can get a higher return in a normal 529 in, say, the stock market? Well, remember that you need to beat UW’s tuition growth rate, and do so without a great deal of volatility. I’d be surprised if normal 529 plans let you use sophisticated tools like options to hedge against volatility.
I, like numerous others, believe that we’re entering a period of increased market volatility, and that if you can offload the risk of matching investment returns that are linked to a tax-free, inflating expenditure requirement to a full-faith-and-credit backed State obligation, you should seriously consider it.
Serious risks in this strategy include the possibility that Washington state politics will result in tuition increases that do not track inflation. Also, general mean reversion in higher education could threaten this strategy. I don’t think that there’s a lot of risk from falling behind a big run-up in equity prices that sustains for 15 years without a concomitant rise in inflation and tuition, but if equity returns beat my expectation and whoop up on tuition increases, you could stand to lose relative to a more traditional asset allocation. Also, waiting is not advisable; the bigger the gap between the purchase price and payout value, the longer you have to spend invested to catch up to risk-free.
Of course, remember that WA GET isn’t something you invest in for strict performance; it’s a way to cover a known expenditure requirement with lower risk. As far as I can tell, for periods well over 6 years (and ideally ~ 18 years, since you can purchase GET units naming yourself as a beneficiary and then transfer them later to your as-yet-unborn children), WA GET makes good sense.