Abstract
In order to satisfy the inter-temporal securities investors, we discuss the inter-temporal securities portfolio model with adding transaction fees in this paper. Meanwhile, for the purpose of improving the practicality of the model, we take the transaction fees, the trading volume and the minimum trading unit into account in founding the model. We divide the main content of this paper into three parts. First, Obtaining the yield and variance buy least squares in a specific period of time. Then, founding the inter-temporal securities portfolio model with adding transaction fees. Finally, obtaining the integral solution of the model. The proposed of this model provide a valuable reference standard for inter-temporal securities investors.
Introduction
Traditional portfolio models, such as the Markowitz mean-variance framework, often struggle with real-world application due to two primary factors: the assumption of infinitely divisible assets and the omission of transaction costs. In actual markets, securities are traded in "round lots" (e.g., blocks of 100 shares), and "frictions" like commissions and taxes can significantly impact multi-period returns. This paper introduces a multi-period (inter-temporal) portfolio selection model that incorporates these constraints alongside Uncertainty Theory to better model expert belief in volatile markets.
Methodology
The model leverages Uncertainty Theory to treat stock returns as "uncertain variables" rather than random variables following a fixed probability distribution. This approach is particularly useful when historical data is unreliable or unavailable, allowing for the integration of expert qualitative judgments.
To quantify these judgments, the authors employ Least Squares Estimation to determine the parameters ($\mu$ and $\delta$) of the uncertainty distributions based on expert surveys. Furthermore, the model utilizes Integer Programming to ensure that decision variables—representing the number of "hands" or blocks of shares ($n_{ij}$)—are non-negative integers, reflecting actual trading practices. The Inter-temporal Budgeting component ensures that capital available in each period accounts for returns, liquidations, and transaction fees from the previous period.
Case Study and Data
The model's efficacy is validated through a hypothetical case study involving four securities over a four-month investment horizon. Ten experts provided independent judgments on expected returns and variances, which were then used to establish the input parameters for the model.
As shown in the table above, the uncertain returns are modeled as normal distributions $N(\mu, \sigma)$, and unit prices are established based on expert consensus.
Key Results and Discussion
The primary objective of the case study was to identify a strategy for an investor with 100,000 yuan aiming for a minimum profit of 5,200 yuan over four months while minimizing risk. The model successfully generated an actionable trading matrix, specifying the exact number of hands to hold for each security across the four stages.
For the target return, the model identified a minimum cumulative risk (variance) of 9723.864. The strategy often involves concentrated holdings in specific securities to meet return thresholds while minimizing total variance. By explicitly modeling transaction fees (entrustment, commission, and transfer fees), the model discourages excessive churning, promoting a more stable investment path.
Conclusion and Future Work
This integer programming approach provides a practical framework for inter-temporal investing. However, several areas for improvement remain. Future research could explore weighting expert opinions based on historical accuracy rather than assuming equal authority. Additionally, as integer programming is computationally intensive, scaling the model to larger portfolios may require heuristic algorithms like Genetic Algorithms. Finally, incorporating "slippage" or market impact for large-scale institutional trades would further enhance the model's realism.
Summary updated