Financial Education

When a stock or bond index contains many illiquid securities it is difficult to fully replicate the index, and doing so would likely lead to significant underperformance due to transaction costs. One alternative method for building an index-tracking portfolio is to use stratified sampling.

Stratified sampling starts with dividing the index securities across a variety of characteristics. These could include industry, growth, value, yield and market capitalization, among many other possibilities. Stocks are categorized within a matrix of all such categories (for example, software stocks below $5 billion in market cap and growing more than 10% per year) , and each cell would be assigned a weight based on the weight of that group of stocks in the index. Then, a stock would be randomly chosen from each cell and weighted according to that cell’s weight in the index.

Stratified sampling allows the portfolio to match the basic characteristics of the index without requiring full replication. The more dimensions within the matrix, the more closely the portfolio will match the index (but the greater the possibility of underperformance due to excess transaction costs.)

Source: Managing Investment Portfolios: A Dynamic Process (CFA Institute Investment Series)

For more information, see all articles on: Asset Allocation, Investing in Stocks, Investment Returns, Portfolio Management, Security Selection, Valuation

See also:

Creating an Index Portfolio Using Optimization

Enhanced Indexing Strategies for Fixed Income Portfolios

Passive Investment Vehicles

Creating an Index Portfolio Using Total Return Swaps

Creating an Index Portfolio Using Full Replication