Concerning the society membership count: an analysis with external factors.

Choudhury, Askar ; Jones, James

INTRODUCTION AND BACKGROUND

The Chartered Property and Casualty Underwriters (CPCU) society is a non-profit organization and primarily consists of CPCU designees. The CPCU examinations and designation is the most recognized certification system in the area of property and casualty insurance. This Society is a community of property and casualty insurance professionals who promote excellence through continuing education and knowledge. The Society's more than 25,000 members from every region in the United States, and also other parts of world, such as, Europe, Japan, Korea, and Bermuda hold the Chartered Property Casualty Underwriter (CPCU) designation. This designation requires taking and passing rigorous college level courses, meeting experience requirements, and also agreeing to be bound by a strict code of professional conduct.

In general, the CPCU designation is attained by completing eight college-level equivalent courses. Once a professional has earned the CPCU designation, they are automatically enrolled in the CPCU Society for a brief introductory period. Following that, usually within six months, new designees are then invoiced membership dues. The primary function of the Society is to facilitate networking among industry professionals and provide continuing education venue. The CPCU Society offers tremendous opportunities to its members that help them to excel in their career. The Society also promotes the value of the CPCU designation to the insurance industry and to the community. A few of the benefits of being CPCU society member:

a. Continuing educational opportunities through seminars, workshops, and symposia help to improve member's skills and knowledge.

b. Professional development programs such as, leadership training, public speaking and other courses to help enrich career objectives.

c. Global Networking helps to interact on a global level by joining one of 148 chapters in the United States, Bermuda, Europe, Japan, and Korea.

d. The CPCU Society's job network to post resume and review job openings. This is also a critical resource for insurance industry employers.

[GRAPHIC 1 OMITTED]

In recent years, the CPCU Society membership numbers has been in a declining phase (see, Graph 1). Therefore, the objective of this study is to understand the cause of this decline of membership and to understand the possible factors affecting this attrition. The decline of the society membership may be the result of both internal and external factors. It may also be a direct consequence of declining CPCU designees. In general, trade associations, membership societies, and other similar not-for-profit groups are no different than any for-profit organizations during economic downturn. In tough economic condition, all these organizations are trying to implement necessary steps to sustain and maintain their current business level. However, for both the members and the organizations there are some cost associated with the process. Those members who are committed to these organizations usually find the capital needed to spend for the membership dues. However, when economic conditions get tough, membership dues payments may become difficult because of financial constraints of members and their employers. Therefore, we expect an opposite relationship between membership fees and society members' count even after adjusted for inflation.

H1: There is a negative relationship between membership fee (inflation adjusted) and society members' count.

At the external level, we would like to explore what aspects of economic buoyancy were most important in influencing the trend of membership's growth or decline. Here we study factors responsible for the decline such as growing interest in other insurance areas. There may be salary differences between job categories within the insurance industry, which may provide insights into the causality of higher attrition rate. Increased opportunities in these other fields collectively, could reduce the likelihood of people taking CPCU exams and ultimately affect the attrition rate. Understanding these relationships and their directional effect can prove valuable for strategic management of the organization in increasing memberships. We will explore the impact and significance of CPCU exam takers on the society membership count. This will provide us some hierarchical relations between external factors and the membership trend itself. This study explores ideas and identifies factors that are associated with the CPCU society membership trend. Understanding these factors is helpful in the managerial decision making process for long-term strategic planning.

[GRAPHIC 2 OMITTED]

Factors internal to the CPCU Society, such as, the privileges of being a part of the Society might not be viewed as valuable enough for members (or their employers) to continue paying for their memberships in the Society, or possibly similar privileges are provided by other associations. A company environment and its policy that supports and encourages society membership will influence the decision to become or remain a member. Other factors such as salary level, tenure level with the company they work, job category, age, and gender may also play important role in the membership decision. These factors aggregate into important differences in the number of membership over time. Depending on the direction these changes take place, changes in the distribution of the work-force across different job categories within the insurance industry will then affect the trend in the aggregate membership of the society. In addition, it is also apparent that a variety of institutional policy changes can also affect the trend in society membership. To aid in managerial decision-making we explore ideas and identify factors that are primarily external and associated with the Society membership trend.

[GRAPHIC 3 OMITTED]

In investigating the causes of attrition of society members, we have conceptualized a few categories: socio-demographic factors, educational/skill, and employment in the insurance industry. CPCU chapters play an integral part in the development of the society and thus the growth /decline of the society membership count (Marks 1994). In recent years, the number of insurance industry designations has continued to grow. Although, most of these other designations may not compete directly against CPCU in terms of curriculum offered, they may compete in scarce time people have to devote for professional development. Difference in the times required to complete these programs along with the expected value earned for their completion may affect society membership count through reduced number of CPCU exam completion. According to the 2007 Society of Insurance Trainers and Educators Designation Handbook, there are over 200 designations and certifications. Comparison between educational alternatives, such as, MBA or bachelor's degree in business with respect to CPCU has been explored by Choudhury and Jones (2011). A recent study shows that51 percent of schools offering MBA stated that they had special outreach efforts for females and in 2006 public universities saw a 55 percent increase in female applicants for MBA programs (GMAC 2006). A study done by the CPCU Society revealed that 59 percent of their members were over the age of 50, indicating even higher attrition rate possibility in the near future. Choudhury and Jones (2011) explored the effect of GMAT takers as a substitute to MBA enrollees on the CPCU society membership. They have found that gender influence on the process of society membership construct is age dependent.

[GRAPHIC 4 OMITTED]

Although, previous studies have focused on various factors contributing to CPCU attrition and are diversified in nature, very little attention has been given to factors, such as, industry trend. In this study, we will attempt to focus on these factors to observe a different view as opposed to previous studies. Thus, this research may contribute towards the improvement of the CPCU society membership trend by providing vital information to the managers for making a better long-term strategic plan.

DATA AND RESEARCH METHODOLOGY

The sample period is a time series of yearly data beginning 1999 and ending 2009. The data on number of CPCU society members were collected from the CPCU society, CPCU exam takers were collected from the American Institute for Chartered Property and Casualty Underwriters (AICPCU); and the inflation and industry employment data were obtained from the Bureau of Labor Statistics (BLS). As a part of our exploratory study we have considered factors by taking into considerations of the pattern of CPCU exam takers, society membership fee, employment trend in the total insurance industry, and also the employment trend in the property and casualty insurance sector. An initial approach into the analysis is to identify a direction in the trend of the number of CPCU exam takers to link them with the society members' attrition as part of a possible reason. Thus, the relationship of CPCU exam takers with the Society membership count may provide one of the links in understanding the Society members' attrition. Specifically, we identify if the number of students (i.e., the pipeline) increases/decreases over time, such that, students who take the CPCU exam, in conjunction with the time required completing the CPCU program is discouraging. For example, if majority of the students pass the exam in a single attempt as opposed to making several attempts it would likely encourage more to join and complete the program. Thus, designation completion time may be one reason for a decline in enrollment for CPCU exams and consequently in the number of new designees joining the Society. Examining CPCU exam takers is designed to test the hypothesis of declining society membership trend is due to fewer number of potential CPCU designees. This will help us to understand at least one of the core causes of society members' declining trend. Thus we formulate the hypothesis that,

H2: There is a positive relationship between CPCU exam takers and society members count.

To observe the relationship between number of society members and four possible factors; two separate analyses were performed. First, correlation analysis was done (results not shown) to examine the direction of the association between factors. Second, society member counts (number of CPCU members) was regressed on these factors to observe the degree of association. In addition to the primary independent variables, we have also explored time delayed factor to observe the effect of certain factor's length of time on the membership trend. As for example, the number of CPCU exam takers (more/less) may affect society members' count one or two years later.

Two external factors were considered in this study to explore the relationship with the Society members' count. They are: total insurance industry employment and property & casualty insurance employment. These factors are interrelated among themselves, as some of these are sub-set of others. Thus, regression models in this study include each factor individually to observe the effect on society members' count without any interaction or confounding effect of other factors. Therefore, four separate regression models were estimated and analyzed in this study. Regression analysis was applied to assess the significance and magnitude of the relationships between these factors over time. Hence, autocorrelation scenario invariably arises. The primary objective of this paper is to understand the dynamics of society members' count with these four factors that we have considered in this paper through single factor models.

To this end, several regression models were run using SAS software (see, SAS/STAT User's Guide, 1993) on four different factors; namely property and casualty insurance employment, total insurance industry employment, membership fee (inflation adjusted), and number of CPCU exam takers. Insurance industry employment is to measure the effect of insurance industry as a whole on the CPCU society membership trend over the years. Thus, a declining industry trend would coincide with the declining society membership trend. On the other hand, an increase in insurance industry wide employment cannot explain the declining trend in society members' count. Thus, we formulate a hypothesis that,

H3: There is an opposite relationship between insurance industry employment trend and society members count.

However, if the property and casualty insurance employment trend and total insurance industry trend do not match, then the reason may lie with this specific (property & casualty) insurance sector. Hence, we will expect that one of the reasons for declining society members' count is due to the declining trend in the property and casualty insurance employment. The hypothesis that we have formulated is that,

H4: There is a positive relationship between property and casualty insurance employment trend and society members count.

In an effort to better disentangle the effects of these factors on the society members' count, regression model included these variables independently. Additionally, Durbin-Watson statistic of ordinary least squares (OLS) estimates indicated the presence of autocorrelation in one of the models. One major consequence of autocorrelated errors (or residuals) when applying ordinary least squares is the formula variance [[[sigma].sub.2] [(X' X).sup[.-1]] of the OLS estimator is seriously underestimated (see Choudhury, 1994) and affects statistical inferences. Durbin-Watson statistic is not valid for error processes other than the first order process (see Harvey, 1981, pp. 209-210).

Therefore, we have evaluated the autocorrelation function (ACF) and partial autocorrelation function (PACF) of the OLS regression residuals using SAS procedure PROC ARIMA (see SAS/ETS User's Guide, 1993). After evaluating the ACF and PACF, the residuals' model of the "total insurance industry" regression model is identified as purely second order autoregressive model: (1 - [[phi].sub.2][B.sup.2]) [v.sub.t] = [[epsilon].sub.t] (see Box, Jenkins & Reinsel, 1994). The final specification of the regression model thus takes the form in equation (3) below.

Specification of the regression models takes the following form:

Society_Membership = [[beta].sub.0] + [[beta].sub.1[Membership_Fee + [[epsilon].sub.t].... (1)

Society_Membership = [[beta].sub.0] + [[beta].sub.1]Exam_Ta ker s + [[epsilon].sub.t].... (2)

Society_Membership = [[beta].sub.0] + [[beta].sub.1]Total_Industry_Trend + [v.sub.t] and [v.sub.1] = [[phi].sub.2][v.sub.t-2] + [[epsilon].sub.t].... (3)

Society_Membership = [[beta].sub.0] + [[beta].sub.1]Casualty_Industry_Trend [[epsilon].sub.t].... (4)

Where:

Society_Membership: Number of CPCU society members

Total_Industry_Trend: Number of employees in the total insurance industry

Casualty_Industry_Trend: Number of employees in the property & casualty insurance industry

Exam_Takers: Number of CPCU exam takers (yearly data).

Membership_Fee: Inflation adjusted dollar amount

EMPIRICAL RESULTS

Descriptive statistics for the various measures of dependent and independent variables are presented in Table 1. Relatively smaller standard deviation (1636.50) of CPCU member count with an average of 26,629.36 members does not indicate much fluctuations in the aggregate membership from year to year. However, Graph-1 depicts a disturbing declining trend in the CPCU society membership. Table-1 reveals that CPCU exam takers spiral down much faster than the society members by comparing minimum and maximum values. Number of exam takers went down almost fifty percent in ten years. This is causing a larger impact on the number of society members than any other factors. This suggests that due to some unobservable factor(s) exam takers number is in decline and thus prompting the society membership count to decline. Thus, the idea of this exploratory analysis is to observe the association between CPCU society members count and possible potential relevant factor(s) that are affecting it.

Simple pair-wise correlation analysis (results not shown) among the variables, reveal that "CPCU exam takers" and "CPCU society members count" are positively related at the 5% significance level with an [R.sup.2] of 42.33% (see Table-3). However, the relationship is negative between 'society membership count' and the 'insurance industry trend' with an [R.sup.2] of 85.25% (see Table-4). It is possible that understanding the importance of becoming a CPCU society member requires experience and maturity. Therefore, the relationship is reverse when considering the insurance industry as a whole and thus supporting our hypothesis #3.

Results of linear regression analysis are reported in Tables 2-5. All these models appeared to fit well in estimating the number of CPCU society members. Reported coefficients of determination ([R.sup.2]) are 0.31, 0.42, 0.85, and 0.92 respectively, with highly significant F values. Results indicate that number of CPCU exam takers in general is responsible for predicting CPCU society membership trend (see Tables 3). Analysis also reveals that, increase in insurance industry trend do not necessarily indicative of increase in CPCU society membership.

Therefore, insurance industry trend may or may not affect CPCU society membership trend. Conversely, it is possible that CPCU exam takers trend does impact the trend in the CPCU society membership. Specifically, if the exam takers are completing the program increases at a higher rate than the non-completers. A number of possible explanations can be explored for this decline in number of exam takers. Nonetheless, considering that the average age of a CPCU enrollee is about 31 years, profession change could be a major aspect. Also, career enhancement through further education and training at this stage of their professional life is another possibility. Thus, this study suggests that society membership trend is primarily dependent upon number of exam takers and may not be dependent on the expansion or contraction of the insurance industry as a whole.

CONCLUSION

This study, examines the effect of CPCU exam takers on the process of joining CPCU society and thus affecting the trend of the society members count. In particular, statistical significance and magnitude of exam takers influence on the "CPCU membership count" is observed. This prognostic power of exam takers on the membership trend is most significant and only increases with time delay. An unexpected result is that, increase in insurance industry trend is found to be not instrumental in affecting the process of CPCU society's membership trend positively. This suggests that total insurance industry influence on the process of society membership trend is independent in this sub-population. Furthermore, insurance industry effect is negatively associated with the society membership trend. Thus, providing valuable insights into the managerial decision making process for the long-term strategic planning in support of society's improvement.

REFERENCES

Box, G.E.P., G.M. Jenkins, and G.C. Reinsel (1994). Time Series Analysis: Forecasting and Control. Englewood Cliffs: Prentice-Hall. Bradford, M. (2005). "CPCU designation offers elite status to risk managers," Business Insurance, 39(43), 4, 40.

Choudhury, A. (1994). "Untransformed first observation problem in regression model with moving average process," Communications in Statistics: Theory and Methods, 23(10), 2927-2937.

Choudhury, A. and J. R. Jones (2011). Society Membership Trend Determinants for Sustainability: An Analysis in the Insurance Industry, Journal of Business Case Studies, 7(2), 61-69.

Cooper, R.W. (2005). "Working To Regain The Public Trust: Considerations For Cocas," CPCU Journal, 58(6), 1-22. Cooper, R.W. and G. L. Frank (2001). "Key Ethical Issues Facing The Property And Casualty Insurance Industry: Has A Decade Made A Difference?" CPCU Journal, 54(2), 99-111.

Gjertsen, L. A. (1997). "CPCU Society focuses on teaching about change", National Underwriter / Property & Casualty Risk & Benefits Management, 101(39), 2-3.GMAC (2006). Application Trends Survey 2006. Graduate Management Admission Council survey, http://www.gmac.com/NR/rdonlyres/5532F610-918D 498D-843E-10D9FA20C890/0/AppTrends2006_FINAL.pdf

Harvey, A.C. (1981). The Econometric Analysis of Time Series, London: Philip Allan.

Hurzeler, D. (2004). "Support of CPCU Society Makes Dollars and Sense", National Underwriter / Property & Casualty Risk & Benefits Management, 108(40), 34.

Jones, J. (2010). "Katie School of Insurance and Financial Services," CPCU Journal, 63(6), 1-8.

Kelly, M. (2009). "The CPCU Society: Diversity is Key to Our Future", National Underwriter / Property & Casualty Risk & Benefits Management, 113(15), 20.

Kensicki, P. R., (2009). "A Guide To Organizational Ethics Policy", CPCU Journal, 62(7), 1-8.

Lowry, J.R., S. M. Avila, and T. R. Baird (1999). "Developing a Niching Strategy for Insurance Agents," CPCU Journal, 52(2), 74-83.

Marks, J. (1994). 'The CPCU Society Story', CPCU Journal, 47(1), 5.

Mogel, G. S. (2003). "CPCU Society Prez Spreads Word on CPCU Designation", National Underwriter /Property & Casualty Risk & Benefits Management, 107(42), 8.

Overman, E. S. (1994). "The Evolution Of Professional Insurance Studies By Industry," CPCU Journal, 47(4), 226-234.

Rice, P. (1997). "Theoretically speaking, CPCU is a big plus", National Underwriter /Property & Casualty Risk & Benefits Management, 101(28), 13-16.

SAS/STAT User's Guide (1993). SAS Institute Inc: Cary North Carolina.

Smith, R., L. Brandon, R. Rudolph, and J. Dusterhoff (1996). "Five Perspectives on Mandatory Continuing Education Requirements for the CPCU Designation", CPCU Journal, 49(2), 116-124.

Askar Choudhury, Illinois State University

James Jones, Illinois State University

Table 1: Summary Statistics for the Period: 1999-2009 (Yearly Data).

 Variables Mean Std Dev Minimum Maximum

CPCU society members count 26629.36 1636.5 24225 29303

Membership fee (inflation 171.0457 6.990027 160.8472 181.4555
adjusted in current dollar
value)

CPCU exam takers 19750 3192.98 15294 26464

Total insurance industry 2262.48 31.08694 2220.6 2306.8
employment (thousands)

Property and Casualty 496.3091 9.442929 479.8 511.9
insurance employment
(thousands)

Table 2: Regression results of society membership fee (inflation
adjusted) on the society membership count.

 Analysis of Variance

 Sum of Mean
Source DF Squares Square F Value Pr > F

Model 1 8383545 8383545 4.10 0.0735

Error 9 18397747 2044194

Corrected Total 10 26781293

R-Square 0.3130 Adj R-Sq 0.2367

 Parameter Estimates

 Pr >
 [absolute
 Parameter Standard value of
Variable DF Estimates Error t Value t]

Intercept 1 49034.00 11072 4.43 0.0017

Society membership 1 -130.98905 64.68182 -2.03 0.0735
fee (inflation
adjusted)

Table 3: Regression results of CPCU exam
takers on the society membership count.

 Analysis of Variance

 Sum of Mean
Source DF Squares Square F Value Pr > F

Model 1 11337288 11337288 6.61 0.0302
Error 9 15444004 1716000
Corrected Total 10 26781293
R-Square 0.4233 Adj R-Sq 0.3593

 Parameter Estimates

 Pr >
 [absolute
 Parameter Standard value
Variable DF Estimates Error t Value of t]

Intercept 1 20043.00 2592.55786 7.73 <.0001
CPCU Exam 1 0.33347 0.12974 2.57 0.0302
Takers

Table 4: Regression results of total insurance
industry employment on the society membership count.

 Analysis of Variance

 Sum of Mean
Source DF Squares Square F Value Pr > F

Model 1 20049150 20049150 23.82 0.0001
Error 8 6732143.45 841518
Corrected Total 10 26781293
R-Square 0.8525 Adj R-Sq 0.7486

 Parameter Estimates

 Pr >
 [absolute
 Parameter Standard value
Variable DF Estimates Error t Value of t]

Intercept 1 131149.00 15379 8.53 <.0001
Total insurance 1 -46.1932 6.7935 -6.80 0.0001
industry
employment

Table 5: Regression results of property and casualty
insurance employment on the society membership count.

 Analysis of Variance

 Sum of Mean
Source DF Squares Square F Value Pr > F

Model 1 24643639 24643639 103.76 <.0001
Error 9 2137653 237517
Corrected Total 10 26781293
R-Square 0.9202 Adj R-Sq 0.9113

 Parameter Estimates

 Pr >
 [absolute
 Parameter Standard value
Variable DF Estimates Error t Value of t]

Intercept 1 -55879.00 8101.47955 -6.90 <.0001
Property and 1 166.24389 16.32077 10.19 <.0001
casualty insurance
employment