Solvency evaluation of the guaranty fund at a large financial cooperative 3038 Conclusions
This study faced the challenge of estimating the solvency of the guaranty fund of
523 million CA$ insuring a network of 570 credit unions totalling some 79 billion
CA$ in assets invested in a broad spectrum of personal, commercial, agricultural and
institutional loans. As a preliminary step, the array of possible approachesaccording
to the aggregation of data, the estimation technique, the depth of historical data and
the projection horizon was examined. After analysis, two polar approaches in terms
of the aggregation of data were selected for implementation. Under the aggregated
approach three statistical distributions were fitted to twenty five yearly observations
of the total rate of subsidy. Under the disaggregated approach, an elaborate Monte
Carlo simulation model was set up whereby the financial statements of each credit
union and those of the guaranty fund itself were generated, integrating four types of
risks and using more than 7500 risk parameters, mainly representing credit risk at
a very segmented level. The Monte Carlo simulation was also used to evaluate the
impact of stressed values of the risk parameters and of a relaxation of the policy for
the premiums charged by the fund. Overall, both approaches converged to similar
estimates of the solvency of the fund, thus reinforcing the level of confidence in the
results. Accordingly, the solvency of the fund could be considered as excellent, being
well within an implied triple A rating under the base case scenario, and still qualifying,
although marginally, for this rating under stressed hypotheses. The detailed analysis
of the solvency of the fund and the good evaluation it brought had three significant
financial impacts: the target capital ratio of the fund was revised downward, the
premiums charged to credit unions were reduced and the Mouvement Desjardins
itself obtained a sizable reduction of the premium it pays to the public deposit insurer.
Needless to say, management of the guaranty fund was quite satisfied with these
outcomes. Finally, it was decided to update the study every five years. From this
perspective, several improvements and extensions, namely regarding a more refined
modelling of the risk factors other than credit risk and a more adaptive premium
policy, are already envisaged.
Acknowledgement.I would like to thank Miguel Mediavilla, M.Sc., of Globevest Capital for
programming the simulation model.
References
- Bennett, R.L.: Evaluating the Adequacy of the Deposit Insurance Fund: A Credit-Risk
Modeling Approach. Federal Deposit Insurance Corporation (2002) - Kuritzkes A., Schuerman T., Weiner S.: Deposit Insurance and Risk Management of the
U.S. Banking System: How Much? How Safe Who Pays? Wharton Financial Institutions
Center (2002) - McIntyre M.L.: Modelling Deposit Insurance Loss Distributions – A Monte Carlo Simu-
lation Approach. Confidential report to Canada Deposit Insurance Corporation (2003) - Roy J.: Funding Issues for Deposit Insurers. International Association of Deposit Insurers
(2001)