International portfolios: A comparison of solution methods

Abstract

We compare the performance of the perturbation-based (local) portfolio solution method of Sutherland (2010a, 2011) with a global solution method. We find that the local method performs very well when the model is designed to capture stylized macroeconomic facts and countries/agents are symmetric, i.e. when the latter have similar size, face similar risks and trade assets with similar risk properties. It performs less satisfactory when the agents engaged in financial trade are asymmetric. The global solution method performs substantially better when the model is parameterized to match the observed equity premium, a key stylized finance fact.

Currency Risk Factors in a Recursive Multicountry Economy

“…Investigation of the roles of trading frictions, portfolio composition, and market incompleteness offers other promising directions for future research. In these settings, it is important to explore the accuracy of both local and global approximations, in the spirit of Rabitsch, Stepanchuk, and Tsyrennikov (2015).…”

Evaluating international financial integration in a center-periphery economy

“…26 Households studied here are symmetric across countries. Based on a global solution algorithm, Rabitsch et al (2013) conclude that in a symmetric country setting with standard CRRA preferences and standard parameter values the method of Sutherland (2010, 2011a) performs extremely well. When volatilities are asymmetric across borders, the welfare calculation based on Sutherland (2010, 2011a) might be inaccurate.…”

“…The equilibrium conditions in each stage of financial integration are then subsequently described. 12 See recent discussions in Kim and Kim (2003), Coeurdacier et al (2011), de Groot (2013 and Rabitsch et al (2013). 13 This paper does not consider capital flows induced by differences in marginal product of capital across borders.…”

“…When volatilities are asymmetric across borders, the welfare calculation based on Sutherland (2010, 2011a) might be inaccurate. We then require other nonlinear methods to solve the model, for instance, Rabitsch et al (2013) and Devereux and Yu (2014). 27 When the discount factor is zero, β = 0, the steady state portfolios are degenerated to those in the static model with second-order approximations, which are not reported here.…”

Recursive allocations and wealth distribution with multiple goods: Existence, survivorship, and dynamics

“…Rabitsch, Stepanchuk, and Tsyrennikov () point out that a global approximation is required when countries are subject to asymmetric constraints, such as a borrowing limit, and when their wealth distribution is nonstationary. Since we have a frictionless model with complete markets and a well‐defined ergodic distribution of wealth, a perturbation approach provides a good approximation of the equilibrium.…”