Customers’ Satisfaction and Insurance Intermediaries: An Ordered Probit Analysis Samiya Chattha* * Assistant Professor, Khalsa College Amritsar, Punjab, India. Email: samiya_chattha@yahoo.com Abstract The essence of providing sophisticated services has helped the insurance sector to become increasingly competitive and market oriented. In such a competitive milieu, survival and success of Indian life insurers depends on the quality of services rendered by distribution channels. So, the present study examined satisfaction of customers towards distribution channels of insurance industry in India. For the purpose of analysis, customers’ satisfaction has been regressed with various benefits received as well as problems faced by them respectively, while dealing with intermediaries, along with demographic variables. For this, sample of 617 policyholders has been taken from three cities of Punjab (India) and data have been analysed using Ordered Probit Regression. The research findings revealed some significant relationships of customers’ satisfaction with respect to benefits, problems and demographic variables. But the magnitude of effect of customer relationship management is more prominent than other benefits whereas magnitude of effect of difficulty in availing services is strong predictor among other problems faced by the policyholders. This paper offers insight to distribution channels that can help them in designing insurance services for the policyholders, in which they would emphasize on enhancement of said benefits and curb addressed problems to intensify the level of satisfaction. Keywords: Policyholders, Satisfaction, Intermediaries, Insurance Industry, India, Ordered Probit Model Introducton The ordered probit model serves as an appropriate framework for statistical analysis whenever survey responses are ordinal or categorical in nature (Daykin & Moffatt, 2002; Long & Freese, 2006). The ordered probit model is commonly used model for the analysis of ordinal categorical data and comes from the class of generalized linear regression models (Armstrong & Sloan, 1989). It is a generalization of a regression model when the response variable has ordinal categories (Williams, 2006). The model is used to estimate the odds of being at or below a particular level of the response variable each estimating the probabilities at or below the level of the outcome variable. In categorical response models, point estimates of the dependent variable cannot be only used for understanding the reasons. The application and use of marginal effects is usually preferred (Long & Freese, 2006). The marginal effects represent the change in the probability of a dependent variable with one unit change in the independent variable (Maddala, 1986). The sign of the slope coeffcient determines the directional change of the probability. A positive coeffcient means that increase in associated variable tends to raise the probability, if all other variables are held constant, while a negative slope value indicates that the increase is more likely to reduce the probability (Daykin & Moffatt, 2002). Keeping in light the aforesaid discussion, the present study used such model in determining satisfaction of customers towards intermediaries of insurance in International Journal of Banking, Risk and Insurance 7 (1) 2019, 56-70 http://publishingindia.com/ijbri/ Submitted: 20 August, 2018 Revised: 05 October, 2018 Accepted: 20 October, 2018