next up previous contents
Next: Camp Scheduling for Young Up: Selected Senior Design Project Previous: Class Scheduler for SMU

L&B REIT Initiative

(Jorge Garcia-Spitz, Joshua Hinkel, David Livengood, Britton O’Toole, 2000)

L&B Realty (L&B) provides investment management services to institutions that invest in office, retail, industrial and multifamily real estate. L&B recently initiated a new strategic initiative called the REIT Initiative (REIT: Real Estate Investment Trust). Using a sampling approach, L&B has constructed a portfolio of over 115 REITs in an attempt to replicate the performance of the Wilshire Securities Index. Our operations research (OR) team was asked to help in the creation of the database and the analysis of a proprietary valuation model built to track the performance and price targets of the securities held.

After examination of the problem by our OR team and L&B Realty it was determined that the portfolio valuation model should be developed in three steps:

  1. Assimilation of data: Approximately 115 companies, with 60 variables each
  2. Statistical Analysis: Mean, Variance, Correlation, Covariance, etc.
  3. Portfolio Optimization: Optimize (Return - Risk) using Markowitz Method

Our OR team then built a preliminary optimization model with a subset of 54 stocks from the original portfolio. The problem was modeled using AMPL and optimized using CPLEX.

Based on criteria and preferences underlined by L&B, we recommend that they use the Ibbotson Encorr Optimizer for the Investment Analysis Software package. A demo of this package has been obtained and can be demonstrated on request.

The economic and stock data has been compiled, but additional data transformations are being examined to provide further insight into the significance of the data. The preliminary Markowitz model that has been developed will serve as a foundation for the completion of the REIT Initiative model.

In developing the final model we recommend that management make a careful analysis of the constraints. The current constraints are very restrictive and do not allow for a range of optimal portfolios. We urge management incrementally restrict the model until an infeasible model is found, and then reduce the degree of restriction to the previous increment.


next up previous contents
Next: Camp Scheduling for Young Up: Selected Senior Design Project Previous: Class Scheduler for SMU

Richard S. Barr
Fri Feb 17 16:09:51 CST 2006