As I mentioned in the other thread where this was posted, if you used the values he proposes (PE10 and q ratio) to determine when prices were low, i.e. below average, you wouldn't invest very often in the past two decades. Since 1990 there exists two time frames when the S&P index was below average based on these metrics. You'd miss out on a whole lot of gains if that were your strategy.
I wonder if you are right here from a statistical or mathematical perspective. If you invested at certain ratios but at that point threw all your cash into the market how would it turn out.
So you made me curious and I decided to check.
These are the assumptions I made:
Interest was done on a monthly basis.
Interest for buy and hold was simply the monthly delta on S&P500 adjusted (with dividends).
Interest for market timing was (1 year t-bill rate + 0.5%)/12. The rate was constant for that year (i.e. all of 1990 had the same rate). The .5% fudge factor was due to current rates being very low and I was trying to approximate a high yield savings rate on the assumption that the market timer would have to have their assets liquid to be able to jump into the market when the opportunity presented itself.
Each investor contributed $1,000 a month starting Jan 1, 1990 going until Nov 7, 2013.
Buy and hold investor simply put the $1,000 in and let the market do its thing.
Market timer only put money in the stock market when the Shiller PE ratio was below the
current average Shiller PE ratio. Note that use of the current PE ratio means I may have invested in a couple places that the actual user wouldn't have since the mean has raised every so slightly recently. I don't think this should make any significant difference given the mean includes over 100 years of data.
If the market timer was not putting money in the stock market, they earned the "safe" interest rate.
Once the market timer put money in the market, they did not sell. The rational here is that by using the Shiller PE as a metric they are buying when they expect the returns to be high, so they will hold the stocks to obtain the expected returns. Using a simple metric such as selling when above average wouldn't tell us anything useful (the market timer would not make good money), since they would only hold a few months.
One could argue for a sell metric such as sell when the Shiller PE is 1, 1.5, 2 SD above normal. I may check this in the future, but right now I don't have the time. I would argue against any "sell at peak" rational as that requires omniscience on behalf of our investor.
The market timers investments ended up looking like this:
From Jan-1990 to July-1990 investor was earning safe interest.
From Aug-1990 to Jan-1991 the market timer invested all his accumulated money and all further contributions into the S&P index. This money was left in the S&P index until the end (Nov-2013). (group 1)
From Feb-1991 until Sep 2008 the market timer was investing only in safe assets.
From Oct-2008 until Jun-2009 the market timer put his second accumulated stash and all subsequent contributions into the S&P index. (NB he did lose money here since he did not enter at the bottom. As before I would argue against entering at the bottom as that requires as much omniscience as exiting at the top). This money was left in the market. (group 2)
From July-2009 until the end the market timer invested in safe assets. (group 3) The total of these 3 groups of money were added up at the end.
Results:
Buy and hold: $633,365.84
Market Timer: $616,359.67So the market timer achieved 97% of the amount of the buy and hold strategy. There may be a chance to exceed it by using some "exit at X standard deviation above mean" strategy, but I don't have time to run that right now.
Psychologically, the buy and hold is much easier to maintain. It would require great discipline to follow the market timing rules without thinking you can beat them and missing the opportune times to invest. That work is also for a lower return.
If anyone has problems with my safe interest rate assumption, I understand, but there's only so much time I was willing to spend looking for data I could use. Ideally I'd use a savings account or MMA interest rate, but I couldn't find the level of granularity that I needed. If someone else can, it shouldn't take me long to rerun the numbers as it's just changing 1 column in the excel sheet.