Working Paper: NBER ID: w18709
Authors: Yongyang Cai; Kenneth L. Judd; Rong Xu
Abstract: We apply numerical dynamic programming to multi-asset dynamic portfolio optimization problems with proportional transaction costs. Examples include problems with one safe asset plus two to six risky stocks, and seven to 360 trading periods in a finite horizon problem. These examples show that it is now tractable to solve such problems.
Keywords: portfolio optimization; transaction costs; dynamic programming
JEL Codes: C61; C63; G11
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Transaction Costs (D23) | No-Trade Regions (NTR) (F13) |
| No-Trade Regions (NTR) (F13) | Trading Behavior (G40) |
| Transaction Costs (D23) | Trading Decisions (G11) |
| Transaction Costs (D23) | Size and Shape of NTRs (R12) |
| Time (C41) | Sensitivity of NTRs to Transaction Costs (D23) |