Abstract
This chapter presents and illustrates a method to identify distinctive population segments in retirement-related risk management. Its focus is the conduct of multiple risk management activities, as measured by a composite indicator. A multivariate prediction model generates the conditional probability of the targeted behavioural outcome. The condition comprises a specified combination of attributes. These represent important causal factors in the outcome. Large multidimensional population segments with unusually high (or low) average probabilities are targeted when applying the method. Such population segments are usually represented by small subsamples in surveys, thus eliminating cross tabulations for estimating the required conditional probabilities. The method illustrated here is ready for application with suitable data sets anywhere in the world. Considering together the contrasting profiles of the high-performing and low-scoring key demographics, we find support for our theoretical position that both high and low potentials to achieve effective retirement-related risk management arise from networks of variables linked via causal chains in which no single variable is dominant. The data suggest that efforts to bring assistance to those who are not well positioned in risk management will be confronted with challenging heterogeneity among the relevant distinctive population segments.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
This remark assumes that the GSS was not affected by serious sampling bias.
- 2.
The authors are deeply grateful to the Desjardins Financial Security authorities who provided access to the microdata files of their 2007 Retirement Survey.
- 3.
The SAS code and the associated substantive documentation of item categories are available from the author.
- 4.
A methodology paper that covers the entire procedure, including the development of the model, and provides detailed supporting tables will be available on the Internet as a free download to purchasers of the book.
- 5.
The expression “causal influence” between X and Y means that we hypothesize that X has an impact in the real world, by either a direct or an indirect route, on the pattern of variation shown by Y. We set out to see what lessons might be learned by fitting to the data a model that assumes certain kinds of causal dependency among the predictor variables.
- 6.
There are no questions about wealth in the General Social Survey, except those regarding home ownership and whether the respondent had a workplace pension plan.
- 7.
The phrase “unusually large concentration” is used for convenience because the model predicts a probability distribution for each respondent. Our process has been to aggregate those who share a relatively (by comparison with the whole sample) high probability for one of the extreme levels of the indicator.
- 8.
The analysis has been done within each of three broad income classes because these classes were powerful contributors to the results of the simulation of DFS-based indicator scores among GSS respondents, making a strong direct control of income essential.
Reference
Friendly, M. (1991). Visualizing categorical data. Cary: SAS Institute Inc.
Goodman, L. (1978). The analysis of multidimensional contingency tables when some variables are posterior to others: A modified path analysis approach. In L. Goodman (Ed.), Analyzing qualitative/categorical data: Log-linear models and latent-structure analysis. Cambridge, MA: Abt Books.
Helman, R., Copeland, C., & VanDerhei, J. (2011). The 2011 Retirement Confidence Survey: Confidence drops to record lows, reflecting “the new normal” (EBRI Issue Brief, No. 355). http://www.ebri.org/pdf/surveys/rcs/2011/EBRI_03–2011_No355_RCS-11. Accessed 15 July 2011.
Hildebrand, D. K., Laing, J. D., & Rosenthal, H. (1977). Prediction analysis of cross classifications. New York: Wiley.
Schellenberg, G., & Ostrovsky, Y. (2008, September 9). The retirement plans and expectations of older workers. Canadian Social Trends (Statistics Canada Catalogue no. 11–008-X). http://www.statcan.ca/english/freepub/11–008-XIE/2008002/article/10666-en.pdf. Accessed 15 July 2011.
Society of Actuaries. (2007). Key findings and issues. Understanding and managing the risk of retirement: 2007 Risks and Process of Retirement Survey report. http://www.soa.org/files/pdf/research-key-findings.pdf
Society of Actuaries. (2011). The impact of the economy on retirement risks. 2009 Risks and Process of Retirement Survey report. http://www.soa.org/files/pdf/research-key-finding-impact-econ.pdf. Accessed 27 May 2011.
Statistics Canada. (2009). General Social Survey: Family, social support and retirement (GSS). http://www.statcan.gc.ca/cgi-bin/imdb/p2SV.pl?Function=getSurvey&SDDS=4502&lang=en&db=imdb&adm=8&dis=2. Accessed 17 July 2010.
Stone, L. O., & Nouroz, H. (2006). New frontiers of research on retirement: Technical annex. Ottawa: Statistics Canada. http://www.statcan.gc.ca/bsolc/olc-cel/olc-cel?catno=75–512-XIE. Accessed 15 July 2011.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2012 Springer Science+Business Media B.V.
About this chapter
Cite this chapter
Stone, L.O. (2012). Distinctive Population Segments in Multi-mode Risk Management. In: Stone, L. (eds) Key Demographics in Retirement Risk Management. Springer, Dordrecht. https://doi.org/10.1007/978-94-007-4044-0_6
Download citation
DOI: https://doi.org/10.1007/978-94-007-4044-0_6
Published:
Publisher Name: Springer, Dordrecht
Print ISBN: 978-94-007-4043-3
Online ISBN: 978-94-007-4044-0
eBook Packages: Humanities, Social Sciences and LawSocial Sciences (R0)