Gyroscopic Investing

Hello,

I have read both Craig's book and HB's Failsafe Investing, and I think I recall that HB said not to reinvest dividends into the stock market, and just add them to the cash position.  Is that correct?  Also, what am I supposed to be doing with the interest I am earning from my 30 year treasuries.  Should I reinvest that into more treasuries, or should I just add it to my cash position, and not deploy the cash it unless my cash exceeds 35% of my holdings?

Thanks

I have never reinvested the cash from bonds, with the exception of TLT and EDV. Those were in tax-deferred accounts, so I didn't worry too much about all the cost basis calculations. And, the reinvested dollars were not huge, so it wouldn't throw my portfolio off balance. It was convenient to let them automatically reinvest and buy fractional shares, especially in small accounts that would take forever to build up much cash.

I hope I'm not misquoting: I think Craig said on more than one occasion that he uses treasury interest for some living expenses.
I don't do anything with the interest except let it accumulate in my money market account. It's not much on its own. When enough cash finally piles up from that plus cash from stock dividends (say, $5,000), I decide whether to plow that back into the pp, into the vp, or just over to the bank account for spending.

The HB-orthodox answer is to have all the stock dividends, bond interest, and cash interest deposited into the cash allocation. Then follow the 15/35 rebalancing rules as usual.

KevinW wrote: The HB-orthodox answer is to have all the stock dividends, bond interest, and cash interest deposited into the cash allocation. Then follow the 15/35 rebalancing rules as usual.

Thanks!  That's exactly the information I was looking for.

In practice, does it change anything if you automatically reinvest dividends ?

k9 wrote: In practice, does it change anything if you automatically reinvest dividends ?

Dividends are not large for a broad index fund, so i don't see why it would. In a taxable account, of course there'd be much more paperwork.

k9 wrote: In practice, does it change anything if you automatically reinvest dividends ?

I think I read over in the boglehead forum somewhere you lose around 0.2% in opportunity cost.

I prefer that dividends and coupon payments be reinvested in stocks and bonds, respectively.  I don't think that rebalancing should occur just because I receive one of these payments.  But what do the numbers say?

Below I have drawn the CAGRs for PPs starting on January 1 of each year since 1972 and ending today (December 2, 2013)—one set with reinvesting, the other without (all data from peaktotrough; 35/15 rebalancing bands; years 2011 to 2013 omitted for clarity).  Reinvesting appears to be a generally better strategy, winning all but nine times over 42 years.  The average difference in CAGR between the two approaches has been 0.16%.



Here is the plot of the difference in CAGR between the two approaches (including years 2012 and 2013):



Conclusion?  Reinvesting has been a better strategy in general, but not always.  The advantage is small, but will add, on average, about 5% to a PP's returns over a 30-year period.

That downward-sloping CAGR graph is not encouraging.

Pointedstick wrote: That downward-sloping CAGR graph is not encouraging.

We need a graph of real CAGRs after subtracting inflation.  I think (hope) that would not have the downward-sloping trend.

stuper1 wrote:

Pointedstick wrote: That downward-sloping CAGR graph is not encouraging.

We need a graph of real CAGRs after subtracting inflation.  I think (hope) that would not have the downward-sloping trend.

I'll upload one tomorrow.

Here's the plot of real CAGRs.  Gratifyingly, it displays no downward trend, and reveals that the PP has provided a real return of between 4 and 6% for every portfolio begun on January 1 since 1972—except for the 2011, 2012, and 2013 portfolios, which are suffering from this year's negative return.



I used CPI-U data to determine the rates of inflation.  (Note that I used the CPI-U value for October 2013—the latest available—as my current value as of December 2, 2013, so all of my charts actually present returns as of October 31, 2013.)