Factors are one of the most useful ways to understand what is really driving a portfolio’s behaviour. They can reveal exposures that are not obvious from the holdings alone — and show when an intended exposure is being offset elsewhere in the portfolio.
This guide explains what factors are, how factor sensitivity works, how to use factors on a map, and how to set them up in I-Maps.
The foundation
What are factors?
Factors are any market forces which may be driving a portfolio’s performance or alpha. They typically include market indices, exchanges, commodity prices, or style factors, such as growth, momentum, value and quality. However, in I-Maps, anything with a return series can be viewed as a factor, including securities.
Measuring exposure
Factor sensitivity charts
I-Maps factor analysis helps us understand how a portfolio is likely to respond if a factor moves up or down. We call this factor sensitivity. More technically, factor sensitivity represents the likely impact on a portfolio’s performance for a one standard deviation move in a factor. You can read this number off the left axis of the factor charts.
The right axis of the chart shows factors’ volatilities, or annualised standard deviations. We show these because equally likely factor moves can be of very different magnitudes. For example, a one standard deviation move in the oil price is much larger than a one standard deviation move in bonds.
Factor charts show the likely impact of an increase in the factor. However, a decrease is equally likely to occur, and this would probably have the proportionately opposite impact on portfolio performance.
Relative to the benchmark
‘Difference From’ factor charts
Often, what we really want to know is whether a portfolio is more or less sensitive to a factor than its benchmark. ‘Difference From’ factor charts show us exactly this. The charts are calculated by subtracting the benchmark’s factor sensitivity from the portfolio’s factor sensitivity. The interpretation then becomes: the likely impact on a portfolio’s alpha for a one standard deviation move in a factor.
Visualising sensitivity
Factors on the map
Like all other statistics in I-Maps, factor sensitivity is visually represented on the map. To understand this, we first need to know the formula for factor sensitivity:
Factor sensitivity is made up of two components. Firstly, correlation: if a portfolio is more correlated with a factor, it is more sensitive to it. Secondly, portfolio volatility: a portfolio’s volatility will heighten or dampen its sensitivity to all factors.
On a map, factor sensitivity can therefore be thought of as the portion of a portfolio’s risk that is in the same direction as the factor. If a portfolio is more correlated with a factor, represented by a smaller angle, more of its risk is in the same direction as the factor. Similarly, if a portfolio is more volatile, represented by a greater distance from the origin, more risk will be in the same direction as the factor.
Practical examples
Factors in practice
Revealing a hidden exposure
Consider a multi-manager who believes the property market is set for a downturn and is therefore considering purchasing a no-property fund. While this fund does not hold property shares, it predominantly holds shares which are highly correlated with the property market, such as banks, retailers or other interest-rate-sensitive shares. A factor analysis with the property index would immediately show that the fund, despite holding no property, is still highly exposed to the property market.
Revealing the absence of exposure
Consider a fund manager whose fund holds a large portion of gold shares because she believes the gold price is about to go up. The fund also holds a number of rand-sensitive shares, which are negatively correlated with gold. These positions would cancel each other out. A factor analysis with the gold price would instantly show the manager that her fund was not particularly sensitive to gold, even though she may have intended it to be.
Positioner Wizard
Setting up factors in I-Maps
The simplest way to set up factors in I-Maps is to add them to a Positioner Wizard. This means the factor charts will automatically use the fund and benchmark pairs from the Positioner Wizard, and the factors will automatically be included when drawing a map.
Go to the Wizards tab of your Positioner Wizard and, under Select Factor Wizard, tick Add Wizard. Select your chosen Factor Wizard from the drop-down list. To edit an existing Factor Wizard, or create a new one, click Open….
Step by step
Creating a Factor Wizard
- Click New.
- Give the Factor Wizard a name.
- Click the Edit… button and select up to 12 factors from the available list.
- Optionally, give the factors clearer names and order them using a Convert File:
- Text file: Create a text file with the original names in the first column and the desired names in the second column, in the required order. Then browse for and select this file from within the Factor Wizard.
- In-system: In the Convert File block, click Edit. A premade convert file will be set up for you to edit. Rename factors in the second column and drag the rows to reorder them.
- Recommended: Export the edited table to a text file so that it is not overwritten. I-Maps will reference that file and ask whether it should be updated after future changes.
- If you opened the Factor Wizard from the Positioner Wizard and want to use it there, click Save and Select.
Once created, the Factor Wizard can easily be added to other Positioner Wizards using the Factor Wizards drop-down on the Wizards tab.
Technical note
For the nerds
By default, I-Maps uses Factor Sensitivity for factor analysis. To see why, it helps to compare the three available measures:
- Correlation
- Beta
- Sensitivity
Correlation
Correlation measures the extent to which the factor and portfolio simultaneously experience high or low returns.
All three measures include correlation. The more correlated a portfolio is with a factor, the more sensitive the portfolio will be to changes in the performance of the factor.
Beta
Normally, a portfolio’s beta is measured relative to a benchmark. Here, we replace the benchmark with a factor. In I-Maps, beta is usually shown as ‘Beta %’, or ‘Beta × 1%’, because this is easier to interpret than beta itself. If a portfolio has a beta of 0.8, then its beta % is 0.8%. Beta % is the likely change in the portfolio’s performance for a one percent increase in the factor.
Beta is the slope in the regression-line formula:
If the return of the factor changes by 1%, then, all else being equal, the return of the portfolio changes by beta × 1%. Beta × 1% is therefore the change in the portfolio’s performance for a 1% change in the factor.
Sensitivity
Sensitivity is the product of the portfolio’s own volatility and its correlation to the factor. The portfolio’s own volatility is included because:
- If a portfolio has very low volatility — for example, if it consists mostly of cash — its performance is not going to be sensitive to many factors.
- Conversely, highly volatile portfolios are likely to react more to a factor’s moves than low-volatility portfolios.
Sensitivity measures how much of the portfolio’s volatility is in the direction of the factor. It is therefore on the same scale as portfolio volatilities: standard deviations of returns.
Beta (1%) versus sensitivity
From the two definitions:
It can be seen that:
Since ‘Beta × 1%’ is the change in portfolio performance for a 1% change in the factor, sensitivity can be interpreted as the change in portfolio performance for a one standard deviation change in the factor. This is because Vol(Factor) is one standard deviation of the factor’s returns.
Sensitivity and beta are similar. However, sensitivity takes into account the magnitude of likely moves in the factor. This is why sensitivity is usually the most useful measure in I-Maps: it tells us not only whether a portfolio moves with a factor, but also how large the portfolio impact is likely to be, given the typical size of that factor’s moves.