Why is full replication impractical for bond indices, and how does stratified sampling solve this?

Full replication of bond indices is impractical because fixed-income markets have far more individual securities, many of which are illiquid or unavailable for purchase. Stratified sampling provides a practical alternative by selecting a representative subset of bonds that matches the index's key risk characteristics.\n\nWhy Full Replication Fails:\n\nThe Bloomberg US Aggregate Index contains over 13,000 individual bonds. Compare that to 500 stocks in the S&P 500. The challenges are:\n\n- Illiquidity: Many bonds trade infrequently or only in large institutional blocks. Buying one bond of each issue is often impossible.\n- Minimum lot sizes: Corporate bonds typically trade in $1M+ blocks. Replicating 13,000 positions would require an enormous portfolio.\n- Turnover: Bonds mature, get called, and new issues constantly enter the index. Monthly reconstitution creates persistent transaction costs.\n- Odd lots: Buying fractional amounts of illiquid bonds incurs wide bid-ask spreads.\n\n`mermaid\ngraph TD\n A[\"Bond Index
13,000+ securities\"] --> B{\"Replication Approach?\"}\n B -->|\"Full Replication\"| C[\"Buy all 13,000
Impractical: illiquidity,
lot sizes, cost\"]\n B -->|\"Stratified Sampling\"| D[\"Define Risk Cells\"]\n D --> E[\"Duration Buckets
0-2, 2-5, 5-10, 10+\"]\n D --> F[\"Sector Buckets
Govt, Corp, MBS, ABS\"]\n D --> G[\"Quality Buckets
AAA, AA, A, BBB\"]\n E --> H[\"Select 200-400 bonds
matching cell weights\"]\n F --> H\n G --> H\n H --> I[\"Track index with
low tracking error\"]\n`\n\nStratified Sampling Method:\n\n1. Define cells: Create a matrix of risk dimensions. For example, 4 duration buckets x 4 sectors x 4 quality tiers = 64 cells.\n2. Calculate cell weights: Determine the percentage of index market value in each cell.\n3. Match cell exposures: Select 200-400 liquid, representative bonds such that the portfolio's allocation to each cell matches the index.\n4. Minimize tracking error: Within each cell, choose bonds with characteristics (coupon, maturity, callability) close to the cell average.\n\nWorked Example:\n\nBrookfield Index Fund targets the Bloomberg Aggregate with 300 bonds:\n\n| Cell | Index Weight | Portfolio Weight | Deviation |\n|---|---|---|---|\n| Govt / 5-10yr / AAA | 18.5% | 18.3% | -0.2% |\n| Corp / 2-5yr / A | 7.2% | 7.4% | +0.2% |\n| MBS / 5-10yr / AAA | 22.1% | 22.0% | -0.1% |\n\nBy keeping each cell deviation within +/-0.5%, Brookfield achieves tracking error below 15 basis points annually using only 300 bonds instead of 13,000.\n\nTrade-Offs:\n\n- More cells reduce tracking error but require more bonds and higher transaction costs\n- Fewer cells are cheaper to implement but sacrifice precision\n- The optimal number of cells balances tracking error tolerance against portfolio size constraints\n\nExplore passive fixed income management in our CFA Fixed Income course.