Evaluation of Index Tracking Portfolios During the COVID-19 Pandemic

An enhanced GRASP approach for the index tracking problem

“…More recent applications of Ruiz-Torrubiano and Suarez (2009) hybrid GA concern comparison against a nonmonotone projected gradient (Xu et al, 2016) and adapted to solve an index tracking model with class constraints (Wang et al, 2018). A recent study (de Amorim et al, 2021) GA obtained better performance relative to the CPLEX solver for nonrebalancing strategies in both models.…”

“…The hybrid GAs of Ruiz-Torrubiano and Suarez (2009) (GA_ruiz-torrubiano2009) and Sant'Anna et al (2017b) (GA_santanna2017) were selected because they use a general-purpose solver to adjust weights, which is the same for the GRASP that was adopted in this work. Also, these two GAs have been adopted to solve different types of index tracking models (Ruiz-Torrubiano and Suarez, 2009;Wang et al, 2012;Xu et al, 2016;Sant'Anna et al, 2017b;Wang et al, 2018;de Amorim et al, 2021). Thus, this final comparison can give insights on the usability of the adapted GRASP approaches on this problem.…”

“…After a metaheuristic optimizes a tracking portfolio, we can evaluate this portfolio capabilities by verifying, for instance, if its cumulative returns are close to the index over time. The relation among cumulative returns and tracking error can vary according to the rebalancing strategy, index tracking model, and chosen metaheuristic, which has been explored by de Amorim et al (2021). Although the purpose of this work has been to assess how suitable would be a GRASP approach to already established formulations of the index tracking problem, future work will explore in detail the in-sample and out-of-sample performance according to the rebalancing strategy, different model configurations, and most suitable metaheuristics.…”

“…The study of heuristics for the index tracking problem is an open problem because the cardinality constraint, intrinsic for index tracking models, makes it difficult to compute optimal solutions in practical time by an exact solver. Nature-inspired algorithms have been widely adopted for portfolio optimization models, especially for the construction of tracking portfolios (Ruiz-Torrubiano and Suarez, 2009;Wang et al, 2012;Scozzari et al, 2013;Sant'Anna et al, 2017b;Kalayci et al, 2019;Strub and Trautmann, 2019;de Amorim et al, 2021). Baykasoglu et al (2015) provided evidence on the potential that greedy randomized adaptive search procedure (GRASP) has against standard portfolio optimization algorithms, including nature-inspired, to solve instances of a problem containing the aforementioned cardinality constraint, named cardinality-constrained meanvariance optimization problem.…”

“…The study of heuristics for the index tracking problem is an open problem because the cardinality constraint, intrinsic for index tracking models, makes it difficult to compute optimal solutions in practical time by an exact solver. Nature‐inspired algorithms have been widely adopted for portfolio optimization models, especially for the construction of tracking portfolios (Ruiz‐Torrubiano and Suarez, 2009; Wang et al., 2012; Scozzari et al., 2013; Sant'Anna et al., 2017b; Kalayci et al., 2019; Strub and Trautmann, 2019; de Amorim et al., 2021). Baykasoglu et al.…”