Optimized portfolio construction using risk analytics
Picking an individual stock or even selecting a Mutual Fund is important and one needs to do careful analysis while doing so. But, credible research backed by empirical evidence has shown that this contributes to just a small percentage of overall returns experienced by an investor. Instead, asset allocation and portfolio diversification drive the bulk of returns in an investor’s portfolio over the long run.
Asset allocation relies on the notion that different asset classes offer returns that are not perfectly correlated with each other and diversifying a portfolio across multiple asset classes will help deliver better risk-adjusted returns. Short articles on portfolio diversification and asset allocation can be found here.
To illustrate the process of portfolio construction and the benefits of diversification, let’s pick a top performing Indian mutual fund and see if we can construct portfolios that include other mutual funds and/or stocks along with it to come up with even better risk/return performance outcomes.
Let’s select “SBI Blue Chip Fund — Regular plan growth”, a large-cap scheme as a reference mutual fund, which is rated as one of the overall best performing mutual funds in India and has a history of over a decade. Its performance can be reviewed by going here — http://vespanalytics.com/mf_select
Below Table in the analysis section gives a summary of its risk/return characteristics.
Now, let’s try to construct a portfolio comprising of a few mutual funds and stocks. Below image shows how to select and add a mutual fund or stock to a portfolio.
For this illustration purposes, add the following mutual funds to the same portfolio “Diversified Portfolio” (there is a small glitch with the UI for now, success message stays on the popup for subsequent additions if the mutual fund page is not refreshed, but nonetheless, the fund will still be added to the portfolio)
SBI BLUE CHIP FUND-REGULAR PLAN GROWTH
SBI Dynamic Bond Fund — REGULAR PLAN — Growth
Kotak Emerging Equity Scheme — Growth
While we are at it, let’s add two stocks as well to the same portfolio (ensure you provide the same portfolio name) from http://vespanalytics.com/stock_select.
Sun Pharmaceutical Industries Ltd
Maruti Suzuki India Ltd
After adding all the portfolio components, now go to portfolio page @ http://vespanalytics.com/portfolio.
It takes a few seconds to load, once it loads, click on “show available portfolios” button and select the newly created portfolio from the dropdown.
At this point, if we like, we can delete any portfolio components by selecting them and hitting delete button. But in this case, there is nothing to delete, so let’s move forward.
Once, we are satisfied with the portfolio composition, click on the “Analyze portfolio component correlations” button. This will bring up an initial analysis of the portfolio components. The left chart shows the risk/return distribution of the different components. The right heat-map shows the correlations among the portfolio components. Correlation scores is the key to construct a diversified portfolio, they could range between -1 to 1 (but most of the time they tend to be within the range of 0 to 1). Feel free to click on any points on the risk/return distribution chart or on the heat-map grid to see their values over time. The more grids with lower correlation scores, the better it is for portfolio diversification.
Now, let’s move to the next step in portfolio construction. There are several strategies to it, let’s first use Mean-Variance portfolio optimization that came out of Modern Portfolio Theory (though no longer considered modern, but still very much in use that has stood the test of time). Click on “Calculate optimal portfolio allocations” button to calculate portfolio allocations for different risk tolerance levels.
In the “Mean Variance Optimized Portfolios” chart, risk is represented on X-axis and its corresponding return is on the Y-axis (this curve is also called as Efficient Frontier). Select any portfolio on this chart to view the weights for the different components in the portfolio and also its comprehensive Risk/Return characteristics below those charts.
Now, just to compare, let’s pick one of the portfolios at the bottom of curve. Below are the summary performance metrics for just the SBI Blue Chip Fund followed by that of the portfolio. Its pretty clear that the portfolio has much better risk/return properties. This holds true even for all the other portfolios along the efficient frontier curve.
Let’s look at their performance through their returns distributions.
Returns distribution for the portfolio are within a much tighter range than that of the portfolio (there are some big outliers). Feel free to compare and analyze other charts.
Before we wrap this article, one might question the fairness of this comparison because we have the benefit of hindsight (this is all based on historical data and performance and there is no guarantee that it will hold in the future as well) and also the advantage of selection bias (cherry picking mutual funds and stocks just to illustrate this example). But, the following reasons should put one at ease on these areas of concerns:
This is based on Modern Portfolio Theory, which has stood the test of time and has gained universal acceptance based on several decades worth of research and real world portfolio management practice
Regarding the above example, even though its all historical data, the diversified portfolio has better risk/return properties than that of the standalone mutual fund across most time periods, certainly over longer time horizons. Thereby, the odds of such superior performance holding into future are much better.
For this illustration, we picked the components of the portfolio as a quick exercise. The main factor in picking a well diversified portfolio is that the components correlation matrix scores must spread out over a wide range. Given this condition, its pretty easy to construct portfolios comprising of other mutual funds or stocks that exhibit similar or even better risk/return properties. Vesp Analytics platform gives the ability to do such an analysis without too much difficulty.