A Methodology of Discovering Comparable Models. The Case of Investing in Retirement Accounts when Considering Age, Main Residence and Education before 1989 vs. Globalization Cover Image

A Methodology of Discovering Comparable Models. The Case of Investing in Retirement Accounts when Considering Age, Main Residence and Education before 1989 vs. Globalization
A Methodology of Discovering Comparable Models. The Case of Investing in Retirement Accounts when Considering Age, Main Residence and Education before 1989 vs. Globalization

Author(s): Daniel Homocianu
Subject(s): Economy, Micro-Economics
Published by: Editura Universităţii »Alexandru Ioan Cuza« din Iaşi
Keywords: investing in retirement accounts; ex-communist vs. non ex-communist countries; data mining; binary logistic regressions with average marginal effects; statistical script for generating models’ tables;

Summary/Abstract: This paper provides a way to discover strong individual influences on investments in retirement accounts. Data are from SHARE-ERIC (Wave7). Principal residences in ex-communist countries or not and full-time education before 1989 served as filters. Two particular models with good classification accuracy resulted based on data mining, variable selection methods, and logistic regressions. A statistical script generated tables with comparable coefficients (average marginal effects). Common influences from the same financial category as the outcome emerged (having life insurance or ever investing in mutual funds or stocks). The younger respondents, those with computer skills or exposed to high stress, are more likely to invest in retirement accounts regardless of the presence of the communist heritage. Specific influences (personality traits and life experiences) also resulted despite the increasing globalization, which, in the case of people over a certain age, was not able to erase some behavioral differences reflected until today.

  • Issue Year: 67/2020
  • Issue No: Special
  • Page Range: 19-31
  • Page Count: 13
  • Language: English