We’ve talked about the concept of portfolio replication, and the many benefits and uses that it represents. But you may be asking:
"How does it work?"
Most replication processes are based on linear regression.
Regression analysis is a process that measures the relationship between a dependent variable and one or more independent variables. It statistically estimates observed relationships among a group of variables. Without getting lost in the weeds, regression is usually better at explaining the past than forecasting the future. Its limited ability to “see into the future” explains why replication is not a common investment process, and not more widely used. When the number of factors (securities) surpasses 8 to 10 holdings, the degradation in performance rapidly increases. You’re left with a less than optimal portfolio with a daily tracking error so high, that it doesn’t make sense.
New quantitative statistical methods allow the portfolio manager to dynamically create and implement a tradable portfolio with a practically unlimited number of factors.
New methods, allow the manager to view replication as an optimization problem, and to optimize the quality of the “fit” of new securities to make the replication work properly. Known as “Advanced Dynamic Modeling”, this process maximizes the potential for better portfolio calibration using publically available data. While publically available fund holding data is at times sparse, Advanced Dynamic Modeling has the ability make the most of that data. In contrast, linear regression would require much more data to maintain stability of portfolio regression.
Is the bus running on time? The answer is yes.
With a simple list of portfolio holdings and their respective weightings, Advanced Dynamic Modeling allows the portfolio manager to accurately measure the portfolio’s ACTUAL RETURN versus its EXPECTED RETURN. In addition, the forward looking optimization process can identify changes in holdings since the last public reporting, and offer optimal combinations of new replacement securities for a future period of time. And in real time it gives the portfolio manager quantitative feedback as to the predictive quality and value added.
Get ready. The Bus is arriving, and the ride is going to be a good one.