Thanks for that link. I don't have time to browse it at the moment but
must get on it later.

This subject was something I was just looking at last week. My
employer (U. of Washington) doesn't seem to have a plan for college
savings other than the state's "prepaid college tuition program"
www.get.wa.gov

As a Washington resident, is the state's prepaid plan the only one I
can participate in on a tax-favored basis? Washington in-state tuition
probably will go up pretty fast, but I don't know if it's expected to
match a typical investment program. OTOH, the return is basically
guarenteed to match the cost of education, at least here. What is the
range of opinion on this type of plan vs. actual portfolios?

Mike

Tad Borek wrote:
- quote -

cporro wrote:
- quote -

Just an update on this thread...the pension bill from last week included
a rider that extends the tax-free nature of 529 plans. See
www.savingforcollege.com

It's hard to beat "tax free."

-Tad

thanks bill. good stuff again.

cporro wrote:
- quote -

There were funds that did EXTREMELY well over that time frame. Despite
9/11 and the tech bubble, real estate was extremely profitable.
Vanguard's REIT index, VGSIX, had an annualized return of nearly 20%
from 2001-present.


- quote -

It sounds like you're looking for some kind of calibration. We can get
a sense of how our computation matches up with Vanguard's (since they
run the Iowa 529) by looking at Vanguard's reported return for VFINX.
If you look here, you'll find the performance page for VFINX:

http://flagship4.vanguard.com/VGApp/...BarChart=false

I'm going to go off the 10-year return. Vanguard reports the
since-inception return, but Yahoo's data doesn't go back that far.
Vanguard reports the 10-year annualized return from 7/31/1996 to
7/31/2006 as 8.8%. I looked up the prices on Yahoo and used the same
formula as before and got 8.43%

OK, that's a little shy of the 8.8% reported by Vanguard. But recall
what I said about compounding periods. The fomula from before assumes
continuous compounding. Let's modify the model to account for discrete
compounding periods.

r = n * ((F / P)^(1 / (n * t)) - 1)

where the variables are the same as before except n is the number of
compounding periods per year.

(Now you know why I prefer the continuous compounding model :-p)

I put this in Excel so I could easily vary n. If you set n = 1, you'll
get r = 8.79%. That's close enough. Personally, I don't think annual
(n = 1) compounding is a good model for the stock market. But
mathematically speaking, all compounding periods are equivalent so I
guess it doesn't matter.


- quote -

It sounds like you're getting close to doing something like a
least-squares fit. Unfortunately, that's WAY past the scope of this
post. If you like, I can cook up a spreadsheet to do this and sent it
to you.


- quote -

Technically, the largest portion of the 529 fund is VINIX. VINIX is
also an S&P 500 index fund but for institutional investors. So it has
a lower expense ratio. And yes, the remaining portion of the 529 fund
would have to have a really high return. But like I said above, that
was certainly possible.

--Bill

cporro wrote:
- quote -

There were funds that did EXTREMELY well over that time frame. Despite
9/11 and the tech bubble, real estate was extremely profitable.
Vanguard's REIT index, VGSIX, had an annualized return of nearly 20%
from 2001-present.



- quote -

It sounds like you're looking for some kind of calibration. We can get
a sense of how our computation matches up with Vanguard's (since they
run the Iowa 529) by looking at Vanguard's reported return for VFINX.
If you look here, you'll find the performance page for VFINX:

http://flagship4.vanguard.com/VGApp/...BarChart=false

I'm going to go off the 10-year return. Vanguard reports the
since-inception return, but Yahoo's data doesn't go back that far.
Vanguard reports the 10-year annualized return from 7/31/1996 to
7/31/2006 as 8.8%. I looked up the prices on Yahoo and used the same
formula as before and got 8.43%

OK, that's a little shy of the 8.8% reported by Vanguard. But recall
what I said about compounding periods. The fomula from before assumes
continuous compounding. Let's modify the model to account for discrete
compounding periods.

r = n * ((F / P)^(1 / (n * t)) - 1)

where the variables are the same as before except n is the number of
compounding periods per year.

(Now you know why I prefer the continuous compounding model :-p)

I put this in Excel so I could easily vary n. If you set n = 1, you'll
get r = 8.79%. That's close enough for me (and I'm not even
stastician!).



- quote -

A more robust way to go about this is to use a least squares fit.
Unfortunately, that is WAY past the scope of this post. If you're
interested in how that works, I can cook up a spreadsheet and send it
to you.

- quote -

Technically, the largest portion of the 529 fund is VINIX. VINIX is
also an S&P 500 index fund but for institutional investors. So it has
a lower expense ratio. And yes, the remaining portion of the 529 fund
would have to have a really high return. But like I said above, that
was certainly possible.

update. i did some looking into the other funds held by the iowa 529
agrro growth fund. and i think you're right bill. but i still like my
spidey sense.

hmm, thats some good stuff bill. really simple easy to remember
formula. but i have to say... my spider sense was tingling a little. i
mean any fund that's got 9.7% return including expenses from
2001-present time seems too good. what i would like to do is run the
same math on both options (vfinx and iowa 529) using maybe a weekly or
monthly average for the Past and Future variables (just to get rid of
some aberrations). unfortunately i can't find any price per share type
values for the 529 plan. thier site has a link for this, but no actual
values.

another red flag for me: a large portion of the 529 is the vfinx.
right? 60-72%. that means the other holdings had to do very very well.
the difference in rate of return was around 3%. so it seems the 30-40%
portion of the 529 would have to get returns of say....13-18%.

and the expense is higher for the 529, about .5% yeah....

i don't know. maybe you're right. you post was very helpful. but i'd
feel better if there was a strict apples to apples comparison. letcha
know if i find one.

cporro wrote:
- quote -

You can make this very simple - the difference in pre-tax returns should
just be the difference in annual expense levels for the index fund
within the 529 plan (about 0.6%?) and in a taxable account (varies,
depends on the Vanguard fund). Note that for a fair comparison, you
shouldn't compare the plan to the 500 fund. The 529 plan doesn't hold
the 500 index fund so the returns of the two should be different - for
example if you compare Iowa's Aggresive Growth option, or whatever they
call that all-stock one, to the 500, you'll see not just expense
differences but also (and more significantly) the results of adding
smaller-company and international stocks to your investments over the
chosen time period.

The harder question is the tax benefit you'll get year to year, to see
if it justifies those slightly higher expenses. You'd be avoiding taxes
on the dividends in the 529, and allow them to be reinvested and grow
free of tax. But depending on the investment choice that might be just
1.5%-2% per year of dividends, taxed at a 15% federal + ??? state rate,
if you held in a taxable brokerage account.

It's not just that annual tax drag though. The more interesting tax
"event" comes up as you approach college age. You'll probably want to
start shifting from stocks to bonds/cash gradually as your beneficiary
approaches freshman year. That will trigger capital gains taxes on sale
of your what? ten-year-old holdings of these stock index funds? and
you'd lose the use of the money that went to pay the taxes at that time.
In the 529 you make those shifts and reinvest every dollar in the
bond/cash fund, where it can earn interest for another few years.

To complicate things further, the taxes you'll pay on distribution from
the 529 plan are uncertain at the moment. If the bill creating 529s is
renewed, then the gains will all be tax-free if used for education; if
not renewed, the gains after 2010 will be taxed to the beneficiary. If
the bill is renewed it's pretty easy - tax-free beats taxed, and it's
probably true even with that small additional expense paid each year.
Especially when you factor in the tax impact of rebalancing the
investments towards bonds/cash in a taxable account.

If the bill isn't renewed there's no easy answer. You would need to make
assumptions about your beneficiary's tax rates many years from now, as
well as the tax rate you'd see in that taxable account, which is of
course impossible to do with any accuracy. At the moment the best answer
might be to hedge your bets by saving in both a 529 and a taxable
account, at least until the 529 tax-law renewal happens (or doesn't).

Comment on that Slate article - the criticism of 529s centers on
broker-sold plans, not the low-expense "direct-sold" plans like
Vanguard's in Iowa, Nevada, etc. (or other low-expense direct-sold
plans). It's a valid exercise to compare the tax benefit to the slightly
higher expense levels, but some of the broker-sold 529 expense numbers
are much different (5.75% initial load, plus 1.5%+ annual expenses,
which could easily flush away the tax benefit).

-Tad

cporro wrote:
- quote -

Here's how I would compare these 2 investments. But before I get in to
that, I should nail down which of the Iowa 529 funds I'm comparing.
Since VFINX is 100% stocks, the only fair comparison is against the
"Aggressive Growth Portfolio". The rest of the funds in the Iowa 529
plan are blended funds.

- quote -

I'm seeing different numbers than you, but you might have been looking
at last month's numbers. Since today is 8/2, it's possible they just
recently updated them. Here's what I see:

Average Annual Investment Returns as of 07/31/2006
Since Inception Inception Date
9.71% 9/20/2001

This tells us that the annualized return of the Iowa 529 fund between
9/20/2001 and 7/31/2006 was 9.71%.

Now we need to figure out what the annualized return of VFINX was over
the same period. So I went over to Yahoo Finance to find the prices.
On 7/31/2006, it was $117.70. On 9/20/2001, it was $83.93 (I'm using
the "adjusted close", assuming that you reinvest dividends and capital
gains).

Finally, we put these numbers through the ringer. The annualized
return is given by:

R = (ln(F) - ln(P)) / T

where R is the annualized return, F is the future value, P is the
principal value and T is the time difference in years. (This assumes
continuous compounding, which is the simplest case. I have no idea
what compounding period they used to estimate the performance of the
529 plan.)

For our specific case, I get

R = (ln(117.7) - ln(83.93)) / 4.86 = 6.96%

So the 529 plan beat the pants off of VFINX over that period. Not to
mention the tax advantages. Of course, you could realize similar tax
advantages with VFINX by investing through an IRA. And I must add the
caveat that past performance is no guarantee of future results.

As for the expense ratio, it says on the Iowa 529 FAQ that the computed
returns are net of the management fee. I believe this is always the
case. Certainly the NAV reported on Yahoo Finance is net of the
expense ratio.

--Bill

i'd like to compare the returns of 2 investments. one is a 529 savings
(iowa) the other the van guard s&p index fund. i've done some math at
the bottom of this post comparing what i know.

i'm not sure what the chart means....1yr (12%) 2yrs (5%)....since
inception (8%).

the only way i figure you can make a real comparison is if you compare
the same intervals. but to me it looks as if 1yr means one year from
inception. this makes these numbers pretty meaningless to me.

am i right? is there no way to compare these 2. i already figured that
if the vanguard can beat the 529 by around 1.13% then there is no point
in me doing the 529.

here is my down and dirty math real quick: i try and calculate expense
as a yearly percentage. i assume an optimistic 8% return. high returns
should hurt my brokerage more then the 529 i figure.

brokerage with vanguard s&p 500:
maintenance fees/ .18%
losses due to long term capital gain as an annual expense/ .08 * .2 =
.016 = 1.6%
total annual expense/ .18% + 1.6% = 1.78%

iowa 529:
maintenance fees/ .65%
no capital gains.
total annual expense/ .65%

not to mention the tax free withdrawals on the 529 may change in 2011,
the limited investment option, and the fact that you need to spend it
on higher ed or you get slapped HARD.