TFM LLP Investment Process: Portfolio Construction
Portfolio construction involves taking the client information and shaping a mix of investments that complement one another in order to provide a good chance of meeting the relevant objectives. Portfolio construction involves Asset allocation and fund selection.
Asset allocation is the process of dividing your investment between different types of assets and markets. Each kind of asset is likely to behave in a particular fashion, and it is the overall mix of assets in a portfolio that determines the overall risk, and the potential returns going forward.
Whilst there is no way of guaranteeing what will happen in future, there is sufficient data available to base decisions on the probable behaviour of assets and markets going forward. Portfolio construction is based on Modern Portfolio Theory.
Modern Portfolio Theory
Modern Portfolio Theory is a method for investment planners to establish a disciplined approach to investing. The actual process used for establishing optimal portfolios uses statistical data and programmes which are not described in this report. The basic portfolio theory was originated by the Nobel Prize winner, Harry Markowitz in the early 1950’s.
Investors before then knew intuitively that it was sensible to diversify, or spread risk (i.e. Not to "put all your eggs in one basket."), Markowitz was among the first to attempt to quantify risk and demonstrate quantitatively why and how portfolio diversification works to reduce risk for investors.
An efficient portfolio is one which has the smallest attainable portfolio risk for a given level of expected return (or the largest expected return for a given level of risk), based on past performance statistics.
The process for establishing an optimal (or efficient) portfolio generally uses historical measures for returns, risk (standard deviation) and correlation coefficients for each asset to be used in the portfolio.
Historical measures are used as expected future returns which may or may not be true, particularly over the short term, but a much better indicator if used over the long term.
Modern portfolio theory makes some assumptions about investors. It assumes they
Will act rationally in making decisions
Make decisions based on maximising their return for their acceptable level of risk.
When making asset allocation decisions based on asset classes it is assumed that each asset class is diversified sufficiently to eliminate specific or non-market risk.
While this process can be performed on any portfolio with two or more assets, it is most commonly applied to asset classes. Asset allocation is the process of allocating funds to each asset class.
Many studies been carried out in this area indicating that this may be by far the most important decision, accounting for up to 90% of the return of the portfolio. Each asset class will generally have different levels of return and risk. They also behave differently.
At the time one asset is increasing in value, another may be decreasing or, at least not increasing as much and vice versa. The measure used for this phenomenon is called the correlation coefficient. Correlation coefficient is a measure of the degree to which two assets (or investments) move together.
The value of the correlation coefficient ranges from -1 to +1. Assets which have a correlation coefficient of -1 are perfectly negatively correlated. i.e. Their values move simultaneously in opposite directions and magnitude.
For a value of +1 they are perfectly positively correlated. i.e. Their values move simultaneously in the same direction and magnitude. A correlation coefficient of 0 indicates there is no relationship at all. In reality, most assets have some positive correlation, although it may be very low.
Return is a term understood by most investors. Total Return is a measure of the combined income and capital gain (or loss) from an investment. This is usually expressed as a percentage which may be annualised over a number of years or represent a single period.
Risk (Standard Deviation of Returns)
While there are many types of risk and different methods of measuring them, the Standard Deviation of (historical) returns is probably the most common measure of the risk of listed securities and portfolios. It is a statistical measure which measures the variability of returns (about the mean or average).
The higher the standard deviation, the more uncertain the returns for an asset class over any period. Standard deviation is very useful in that it enables us to compare the level of risk of different types of investment. For example shares against bonds.
By computer processing the returns, risk (standard deviation of returns) and correlation coefficients data, it is possible to establish a number of portfolios for varying levels of risk each having the least amount of risk achievable.
The optimised portfolio is deterministic rather than stochastic and the model is generated from inputs derived from the Consulting Actuaries Watson Wyatt.
Limits to optimal portfolios
The optimal portfolio models concentrate on asset allocation, and take no account of differing styles or philosophies behind investing. (This is a large area in itself, and not the focus of this report.)
It is important to understand that all optimal portfolios are a function of past performance data, and therefore they are no guarantee of the future returns available from an investment.
Optimal portfolios only provide an asset allocation template, and do not populate the model with individual investments. This is addressed by our fund selection process.
TFM uses independent ratings to assist with fund selection. The OBSR rating system as the main selection criteria for the funds recommended to our clients. OBSR is a respected ratings agency that does not derive income from the asset management houses they rate, unlike most of their competitors.
The OBSR Fund Ratings are determined on the premise that the fund selection process should, whilst taking past performance into consideration, ascribe greater weight to identifying the factors which will affect future performance. This process demands a much stronger emphasis on a qualitative examination of funds.
Information ratio is a statistical measure of the returns generated relative to the risk taken for each fund, and is the key indicator used to select funds from a shortlist where several are available.
Departures from optimal portfolios
Where TFM holds a strong conviction, the firm will recommend alterations to the computer model.
This may be for several reasons, and the firm will document the reasons for our recommendations where relevant.
At the time of writing, amendments to optimal portfolios are largely concerned with asset allocation within the “international equities” sector (Dec 2008)