Words: 1624
Introduction
Manzana Insurance was founded in 1902 in California and specialized in orchard and farm insurance then. But, a series of acquisitions followed the San Francisco earthquake and by 1953, Manzana was the second largest property insurer in California. In the 1970s, it faced intense competition from Golden Gate Casualty, a new entrant in the home insurance business. Golden Gate could spend on intensive marketing and spark off price war due to the backing of its corporate parent, one of the largest retailers in the world. The high-interest rates during this period also adversely affected Manzana’s property insurance business. Hence, Manzana started venturing into other business areas and began insuring liability too.
From almost no liability policies in the mid-60s, it went to almost a 50-50 split between liability and property insurance in the 1980s and now, 65% is property while 20% is liability insurance. But unfortunately, due to the insurance crisis of the 1980s, settlement claims on liabilities rose and the company posted its first annual loss in 1985. It was subsequently taken over by Banque du Soleil in1989. Manzana, under the new management, adopted a back to basics strategy and concentrated on regaining its market share in the property insurance business.
Less profitable lines of business were discontinued and the operations at the branch offices were geographically reorganized in order to improve the market responsiveness. Organization: The Company operates through a network of semi-autonomous branch offices. It treats each branch as a profit center and a territory is assigned to each branch. Manzana does not deal directly with the customer but has a network of more than 2000 independent agents who bring business to the company. Hence, the loyalty of the agents and their motivation to push through a Manzana policy is critical to the profitability of the company.
Each branch has underwriting teams to support the agent (typically, 20-25 agents under a team). Agents earn a 25% commission on a new policy and a 7% commission on a renewal. In order to retain experienced employees, Manzana recently kicked off the ‘salary/plus’ scheme. Process Flow: The time required for different request types are different for each of the activities.
All requests start with the distribution clerk distributing the request for an insurance policy (RUN, RAIN or RAP) received from an agent or computer (RERUN) to the respective underwriting team.
Distribution also analyses and disseminated published data, researches competitor’s rates and oversees rating. The underwriting team evaluates, selects, classifies and prices the request and then it passes on to the rating department where the premiums are calculated. The policy writing departments do the actual typing, assembly, and distribution of completed policies. In the case of RAP, the quote was transferred to the originating agent from the rating department and once accepted, went straight to the policy writers. Only around 15% of all quotes translated into policies.
The Fruitvale branch followed the FIFO method, but in practice, there were 2 priority classes (within which again, FIFO was followed). The RUNs and the RAPs, being more revenue-generating were tackled first and then the RAIN and RERUN. In the case of policy writers, they seemed to handle the easier tasks first and then the more complex ones. The Fruitvale branch of the company has not been doing well over the last few years as revealed by the financials of the company. Its performance has dropped over almost every metric.
Profitability has declined and the TAT has gone up from about 3 days to 6 days now. On the other hand, the Golden Gate was offering a seemingly impossible 1 day turnaround period. This had the effect of the independent agents pushing through the policies of Golden Gate ahead of Manzana’s. The number of new policies has stagnated for the Fruitvale branch over the years in spite of moderate industry growth. The renewal losses were also very high with about 47% renewals being lost every year.
From the high value of correlation, we can conclude that part of the reason for hi renewals loss is the delay in the operation of RERUNS. Reasons behind the increase in turnaround time and late renewals – In the first quarter of 1990, the renewals lost had risen to 400 from the 219 in the previous quarter as shown in exhibit 6. Renewals lost in territory 1 were 403, substantially higher than 227 in territory 2 and 296 in territory 3.
In 1990, Manzana was reorganized according to geographic lines where each underwriting team was assigned to a specific territory. It can be seen from exhibit 7 that the rate of receiving requests from the three regions is quite uneven. Territory 1 receives a total of 1755 requests, territory 2 – 1578 requests and territory 3 – 1347 requests. This new way of assigning the underwriters based on regions has led to underutilization of the resources available in the department due to which territory 1 underwriters are overloaded.
A policy where there is pooling of the teams and assignment is not specific to regions would lead to better utilization. Another reason for the high turn around times would be the amount of backlog that has built up the system. This backlog affects the processing of every new request as they have to be cleared first before the new requests can be processed and hence the turnaround time for each new request has increased. Analysis of Backlogs and TAT The method for evaluating the turnaround time for the backlog of 82 Requests-in-progress is as follows: 1.
The time taken for completion of an operating step is taken as 95% of standard completion time. 2. The time taken for completion is divided equally among the persons/teams responsible for a step and then this is multiplied with the time taken to complete the step to determine the time taken to complete all the requests in the step in days. 3. This time is determined for all the 4 steps and added together to obtain the total throughput time for the backlog, which comes out to be 8. 2 days. However, there are a few issues with this way of calculating the turnaround time. These are 1.
The time taken for completion of a step is taken to be 95% of the standard completion time. However, it should have been taken as the average as the extra time taken for some of the requests would be compensated by the less time taken in the other few requests, thereby compensating the extra time. 2. The total throughput time is determined by adding the total time for individual time. This analysis ignores the fact that when the work is being done on an earlier step, say Distribution, the work can be done on the requests that are in progress and are in subsequent operating steps, say Underwriting. When this fact is taken into account, the total turnaround time for 82 Requests-in-progress would be less than the determined 8. 2 days. Recommendation for priority: Based on Exhibit 6 and 7, we see that Rerun requests far outnumber the RAP requests. Also, only 15% of RAP requests get converted. Given the average number of days (10) clients take to get back after RAP, we see that RAP requests are not very time sensitive and hence needn’t be given very high priority.
Thus, we recommend that Rerun is given higher priority than RAP. Hence, the priority should be RUN, RERUN, RAP, RAIN. Given the above average of 39 requests per day and the distribution of various kinds of requests, we can create a Gantt chart for the various departments to obtain the throughput of the various requests. Based on this analysis we obtain the following figures: RUN: 6 RERUN: 9 RAP: 1. 1 RAIN: 1. 3 Analysis: From the above figures we see that the RUN and RERUN requests can be finished on the same day.
Some of the RAP and RAIN requests would not finish on the same day, but these can be completed by the Rating and Policy Writing departments the next morning before they receive the new requests from previous departments. This implies that on an average, all request types can be completed within a day. But we also see that variability is very high, esp. in the underwriting team. Thus there is always the possibility of a build-up of the backlog. But on an average, it’s possible to complete the requests in one day. Clearance of backlog
The chart for clearance of backlog as on week ending September 6, 1991, given in Exhibit 3 is shown in the section ‘Analysis of Backlog and TAT’ above. As shown in the calculation, it should take two extra days to clear the whole backlog of 82 processes. In this calculation the priority has been maintained the same as that being followed in the Fruitvale Branch, i. e. , RUN, RAP, RAIN, Rap. Another approach could also be taken wherein you would try to process the shortest request first. So going by this approach, we would follow the order RERUN, RAIN, RAP, RUN for the DC department.
Hence, we recommend that Fruitvale should work on one weekend and complete all the backlogs following which the normal operations as described above can be completed within a day. Staffing of policy and rating departments From the analysis done for the normal operations case, we can see that the policy and rating department are around 70% and 85% utilized respectively. Also In exhibit 4, the average processing time for different departments mentioned company-wide in 1986 is similar to the average processing time of Fruitville branch in 1991.
This implies that the Fruitville branch has not capitalized on the better technology developments that have taken place during the late 1980’s and is still operating inefficiently as their processing time should have decreased in the departments like policy writing and rating where the impact of technology is maximum. Hence we recommend that the staff can be reduced in due time as well as the employees need to be given the incentive to improve their performance.
Completed Policy Request originating from the Agent/computer Underwriting Team Distribution Clerk Policy Writers Raters