Investment Process

 How we select strategies for inclusion in our investment process.

      Objective  |  Elimination Procedure  |  Methodology  |   Forecasts  |  Valuation  |  Qualifications 

 

 

 

 

Our investment process is an odds-based betting system, like a form book for horse racing.

It is designed to maximise the chances of picking winners and minimise the odds of picking losers.

Investment Process

The diagram below shows how investment opinions are generated by Investors RouteMap, using the Shares RouteMap as an example. The specific strategies will vary between RouteMaps, but they all incorporate the same Valuation, Sentiment and Trend concepts and are designed to generate opinions that similarly minimise risk and maximise rewards.

 

 

 

Strategies target 6-12 months holding period

Objective

The market timing strategies developed for Investors RouteMap have been designed to provide a high level of investment performance that is compatible with low portfolio maintenance requirements. These portfolio measurement tests apply for portfolios invested in long-term government bonds, foreign currency or stock markets, globally and are shown in terms of total returns combining capital gains and income. Results are shown both in absolute terms and relative to peer group benchmarks.  Specifically they contain the following design features intended to make them suitable for long-term investors, who may not wish to review the markets on a daily basis.

  • Once a month portfolio review, after each month-end
  • Six month average holding period for each of the investing strategies
  • An average of only two transactions each year per investment. 



Many popular investing strategies do not stand up to close examination

Elimination Procedure

These market timing strategies are the result of an extensive weeding-out process to identify strategies that offer the highest returns for the least maintenance cost - whether it be time invested or portfolio turnover. In this process a wide range of investment concepts have been examined and rejected for a variety of reasons. These include those where: -

  1. Results were inferior to those of the winning strategies
  2. High turnover made the strategy unsuitable for long-term investors
  3. Data is only available after unacceptable time lags
  4. Comparable data does not exist for  many countries in our database 
  5. Records of forecasts in the past are not available
  6. Testing is not meaningful because markets are only rated by ranking list
 

 

 

We do not just mine the data that is easily available, but test back over 35 years, where possible

Methodology

The greatest risk of back-testing is selective recall, as small sub-sets from a database can generate misleading conclusions. Typically this risk occurs when the sub-set only includes the most recent years or is limited to only one country, or a small handful of countries. Several factors are employed to minimise this risk.

  • Consistency - This is a what-if exercise. That is to say, what would have happened to investment performance if the same portfolio strategy had been simulated for all markets for the whole time. Our strategies do however differ between bonds, foreign currencies, stock markets and investing styles or sectors.
  • Testing to Destruction - These market timing strategies were tested in 50 countries and regions over 35 years of monthly data, to include all possible environments. The tests include periods of inflation and deflation, boom and bust and markets of widely differing characteristics.
  • Large Database - The sample size in each asset class consists of 6,500 to 15,000 monthly observations per RouteMap so as to satisfy the most rigourous demands for testing in a wide variety of market conditions. The only exclusion from the database, other than short runs of historical data, is in relation to hyperinflation, owing to its effects on the exchange rate and valuation of securities.
  • Odds of Success - The percentage change of a profitable outcome has been analysed to reduce the possibility that a minority of big successes hide a majority of small losses. These odds are expressed as an average of all monthly returns. NB Increasingly higher odds are generated when aggregating monthly returns into holding periods and holding periods into long run chains for individual countries.
 

Spines on these hedgehog charts show the direction of year end predictions in Financial Forecasting

These are the basis for our exchange rate forecasts

Investing Strategies based on Forecasts

While the first four reasons have led to the exclusion of many popular market timing strategies from the Investors RouteMap, the latter two reasons merely make it inappropriate to include them in these tests. The difficulty with strategies based on economic forecasting is that, while effective models can be built to explain past relationships based on actual variables, and, while we can project into the future by making our "Best Guess" based on consensus forecasts for key economic variables, these consensus forecasts may turn out to be wrong. All that can usefully be done is to show how well our models work on the assumption of perfect forecasts by providing a couple of key examples.

Look first at the amount of spare capacity in each economy, alias the GDP Output Gap, as this is a key variable in our investment models for all four RouteMaps, bonds, foreign currency, shares and equity investing styles.

  • For bond markets, this affects the level of interest rates.
  • For foreign currencies this affects interest rate differentials between countries.
  • For stocks and shares this affects profit margins.
  • For investing styles this affects the key indicators.

In our example we have taken the aggregate global figure that we calculate as the most representative of overall trends. You can see how closely our estimate for Global Spare Capacity tracks that of official organisations on the subject in Econometric Modelling.

In order to illustrate how these investment models apply to financial markets, we show the most important exchange rate as an example - that of the US Dollar to IMF Special Drawing Rights (SDR), rather than any single major trading partner. This eliminates bilateral bias. We use a Hedgehog chart, as shown alongside, to illustrate this in Financial Forecasting. This type of chart breaks the forecast series up into annual slices and rebases each one at the appropriate place against the actual series to highlight predictive value, assuming perfect forecasting ability.

 

 

Generally speaking, signals have been on the right side of the big moves - at the risk of short-term whip-sawing

Market Timing Strategies based on Valuation

As regards valuation strategies, while it is often easy to see what is cheap and what is dear, it is much harder to formulate a meaningful simulation test as it is difficult to define a rule about how much cheaper or dearer a market can become before it turns. This is because the valuation ratios seldom fluctuate between the same highs and lows. This restricts the ability to formulate a mathematical rule that can be tested across many markets.

The most dramatic recent example of this problem was the technology bubble. A long list of valuation strategies generated sell signals for Wall Street far too early, sometimes years too early and even at only half the final peak level in 2000. Indeed this problem occurs during every great investment boom.

Valuation strategies do have their uses, and are equally weighted with trend strategies in the RouteMaps but cannot be effectively back-tested because such mean-reversion techniques do not lend themselves to precise market timing.

Please read the financial health warning
Please read the financial health warning

 

 

Government policies are specifically what the Liquidity strategy is designed to exploit.

Qualifications

Before looking at the tables for performance simulation, please consider this discussion about some of the main qualifications and review the Financial Health Warning:

  • These are tests using indexes. They are not actual results for any specific investment or investment management firms. - However fund  managers either change their styles or themselves get changed, so that also doesn't guarantee consistent investment performance.
  • These tests exclude transaction costs. - However our investing strategies are specifically designed to keep turnover low, at less than two deals a year on average, and anyhow there is now a rich choice low-cost execution-only dealing facilities, and specialist closed-end funds.
  • Cynics believe that once investing strategies becomes popular, they cease to work. - However that has not been the case for our selected strategies. If so, this would show up as flattening curves on the fan charts used fot back-tested simulation of our Composite Trend Indicators, and that does not appear to be the case. Remember also that the measurement tests relate to whole markets that are too big for anyone other than a government to manipulate.
  • These performance simulations are not audited. - However, subscribers can check the complete record for any country as all past signals are shown on each chart. Visitors can check the record of signals in any of the samples shown in the Chart Library.
  • These tests assume the courage to act on one's beliefs decisively. - True, many investors water down their decisions with restrictions on divergence from benchmark weightings. However, Investors RouteMap is not in the closet index-linking business. There is more to risk than volatility, and even if volatility is the chosen measure of risk, most of our strategies add less than 1% to the standard deviations recorded for benchmark indices.