Market Timing vs. Buy & Hold...

I've previously mentioned that I experimented with market timing to determine when to buy/sell stocks. See my posts on the topic - here, here, here and here.

That theory, and those indicators, was based on some simple backtesting I did with a few large indices, such as the S&P500 and FTSE, for which I was able to easily plot and get data for using Google Finance. Sadly, Google seem to have removed all their lovely charts and data, presumably under pressure from financial companies who earn money for providing data that Google was providing for free.

In any case, the backtesting I did at that time was manual - i.e. I would follow the graph and make a note of the trades once the indicators were hit.

I wanted to automate this process to make it easier to test different timings, so I made an epic spreadsheet (and learned some new Excel skills along the way):

Here's an extract of an exponential moving average and SSTO (slow stochastic) to indicate when to buy and when to sell the Dow Jones (DJI).

The reason I was so interested was that this method avoids the large market drawdowns seen in the recent 2000 and 2007 crashes. You can see in the screenshot that this strategy gets out of the market near the top and gets back in near the bottom. It also beats a buy and hold strategy, generating a 9.9% annual return vs. 9.4% for buy and hold.


Sadly, I also found that the timings are specific for every market. When I applied the timings that worked on the DJI, they didn't work on the S&P (another large US index) or the MSCI World or Gold.

Each market had its own very specific timings that were needed to beat simply buying the market and holding it through all the ups and downs.

I had hoped that because:

  • markets ultimately reflect human sentiment and
  • the majority of the market is made up of large players who need to move their money more slowly than us small fish
I would be able to find a decent long-term indicator (e.g. the 144 week exponential moving average) that would allow those of us with small holdings who can sell in an instant avoid the long bear market period and buy back in near the low.

Sadly, this proved not to be the case.

I should have known better - like these guys.

They found that, out of more than a million market-timing strategies, 0.2% beat buy & hold.

On the plus point, I learned some new Excel skills, such as the OFFSET function and the use of the Solver add-in.

So where does that leave my investment strategy?

20% Gold / Precious Metals (gold and silver coins of the country in which you reside)
20% Cash (bank savings and physical notes)
20% Property (the equity in your own house, or buy-to-let, but for most people, the former will mean that they cannot afford the latter)
20% Stocks (a single global index that ideally includes some exposure to emerging markets)
20% Bonds (a mixture of short and long dated)

  • Modified by a valuation metric:

Using inflation-corrected (also known as real) prices using figures from Nationwide, I calculate quarterly value for each asset class. If prices are high relative to their historical average, then I assign a smaller % to them. At the last quarter, that puts asset weights as follows (for UK-based investments):

  • 20.2% Gold / Precious Metals
  • 20% Cash (this is always fixed, as this provides the buffer to buy assets when they are cheaper)
  • 22% Property (this applies only for Northern Ireland - if UK-wide, this drops to 20.4%)
  • 20.2% Stocks
  • 17.6% Bonds (UK Gilts)

A way to visualise this is shown in the graph above.

This is the real (inflation-corrected) prices of the (UK-based) assets. It's a log scale just for presentation purposes.

One could interpret this as (relative to historical norms):

  • UK House prices are expensive
  • UK Gilts (government bonds) are expensive
  • The FTSE 100 is fairly cheap
  • Gold is about average

Using a valuation metric means that during the stock market crashes that happen from time to time, you would slowly (monthly or quarterly) transfer cash into stocks because they would:

  • be getting cheaper (better value)
  • form a smaller portion of the portfolio
This is known as rebalancing, but is probably best done during the accruement (i.e. saving) phase rather than rebalancing the entire portfolio. During the accruement phase, you are feeding money in every month / quarter, and therefore actively buying stocks as they become cheaper during a bear market. When the stock market is riding high, you are buying less, as they may be relatively expensive versus property, bonds or gold.

However, even rebalancing has its critics.

I'm an insatiable tinkerer though and would not advise this strategy for most people. I also do not currently hold any gilts or any other forms of debt, whether government or corporate, despite the Permanent Portfolio requiring it. That's a subject for another post...

99% of people would be better off concentrating on the simple things outlined in the previous post

  • Pay off debt
  • Save more
  • Spend less
  • Invest in assets (buy and hold)


Popular posts from this blog

UK House Price Update and Attempts to Model it...

7 years on... Real world natural muscle gains