Progressive Corporation Peter Lewis felt a sense of relief. Bruce…

salah_ismail Progressive Corporation Peter Lewis felt a sense of relief. Bruce…Progressive Corporation Peter Lewis felt a sense of relief. Bruce Marlow, Progressive’s COO, had just handed him the preliminary fourth-quarter results, and 1993 looked like a turnaround year. Lewis, who in 1965 had become head of the auto insurance company his father had founded in a Cleveland garage in 1937, had seen ups and downs during his 38 years at Progressive. However, 1991 and 1992 had been two of the roughest years ever. In 1991, profits had dropped precipitously from $121 million to $57 million causing return on equity to plunge from 27.8% to 11.7%. Recovery in 1992 had only been modest: net income climbed to $84.6 million; its weakest result, behind 1991, in the last six years (Exhibit 1). Analysts quipped that Progressive had become similar to its target customers: high risk. With revenues of $1.8 billion, Progressive had grown from being the 48th largest auto insurer in the US in 1980 to the ninth largest in 1993. With an annual growth rate of 22% over the last ten years, Progressive had greatly outpaced the industry’s annual growth of 8.7%. Progressive had defied the conventional wisdom that automobile coverage was the “wasteland of the insurance industry.” Moreover, while most carriers shunned drivers with a bad driving record, Progressive eagerly embraced them. Progressive’s clients—in the insurance parlance the “nonstandard” segment—were drivers who could not get insurance with another carrier. Until recently Progressive’s growth had been highly profitable. Over the last ten years, Progressive had achieved an average underwriting profit for its core automobile business of 6.4% when most other insurers had considered auto insurance to be a profit drain. The industry average underwriting profit over the period had been -6.4%. Progressive had earned an average return on equity of 25.3% over the past decade, and Progressive’s shareholders had been remunerated handsomely: Over the last ten years, Progressive’s stock had generated a return, assuming dividend reinvestment, of 28.7% as compared to an S&P 500 performance of 14.9% (Exhibits 2, 3). Changes in the automobile insurance industry, however, were threatening Progressive’s historical position. The large, standard automobile insurers, such as State Farm and Allstate, had traditionally avoided the non-standard segment. In the mid-1980s, however, Allstate had started to move aggressively into the non-standard segment by underwriting more non-standard clients. Lewis wondered how he should respond. The Automobile Insurance Industry. Under state law, most drivers of motor-vehicles in the US were required to have insurance. An automobile insurance policy consisted generally of three parts: liability coverage which protected the policy holder against bodily injury and property damages to third parties in an accident caused by the policy holder; collision coverage which paid for the damages to the policyholder’s car; and medical coverage which paid for the policy holder’s medical expenses stemming from accidents while driving his/her own or other vehicles. Each state had issued regulations specifying minimum liability coverages and other mandatory coverages, such as uninsured motorists coverage, which paid for damages caused by drivers with no auto insurance. At the same time, insurers offered many different options with respect to levels of liability coverage, the amount of collision coverage and the size of deductibles. In 1975, A.M. Best Co. had estimated the total size of the auto insurance market to be $21 billion. By 1993, the market size had increased to $109 billion. Approximately 300 insurance companies were competing in the automobile insurance industry (Exhibit 4). In most years industry average operating- and claims expenses exceeded premiums. Over the period 1984-1993, for every $100 that property and casualty insurers had earned in premiums, they had spent $106.4 on claims and operations. Hence, industry profits were due solely to investment income, which resulted in an average industry return on equity of 9% over the last ten years. In general, broader-line insurance companies often offered automobile insurance as a loss leader to win customers to whom they could sell more profitable policies such as homeowners’ coverage. Two segments had emerged in the automobile insurance industry during the 1950s, a time in which the market had grown rapidly due to the general rise in living standards and automobile ownership. Many insurance companies, such as State Farm and Farmers Insurance, had realized that 80-90% of all customers were easy to price based on only a few standard criteria, such as gender, age and miles driven to work. This group became known as the “standard” segment. To deliver a consistent message to their agents with respect to underwriting procedures and to reap scale economies in policy processing, standard insurers avoided the remaining fringe of 10-20%—the “nonstandard” segment. This segment comprised not only drivers with poor driving records but also drivers older than 65, drivers who did not speak English, drivers of high-performance cars, and young, first-time drivers with no insurance history. The Standard Segment Competition in the two market segments was quite different. Since the large standard writers considered their auto lines to be loss leaders, competition in the standard segment was fierce (Exhibit 5). Moreover, since pricing was relatively easy, competition in the standard segment focused mainly on reducing cost. Besides automobile insurance, many standard insurers offered other lines of coverages, such as life and homeowners’ insurance. Once a customer had signed up for a whole insurance package, standard insurers generated their profits through customer retention. Thus, the pricing of automobile policies took renewal streams into account because customers tended to remain with each provider for a long time. Automobile insurers had to choose the channels through which they sold their policies. Some insurers, the so-called direct writers, relied on an internal sales force (e.g., State Farm and Allstate), others on a network of independent agents (e.g., ITT Hartford). A third, low-cost way was to dispense with sales agents altogether and offer policies mainly via remote selling methods such as mail or phone (e.g., GEICO and USAA). One could observe that the leading standard insurers each focused on one distribution channel. In general, insurance agents had the authority to offer insurance coverages subject to company-mandated underwriting guidelines. These guidelines prescribed the kinds and amounts of coverage that could be written and the premium rates that were to be charged for specified categories of risk. Independent agents worked on a straight commission basis, while internal agents were paid a salary coupled with a lower commission. Total compensation for internal agents tended to be lower than for independents. Tellingly, in the last years independent agency writers had been on the retreat. Over the period 1989-1993, agency writers had increased their premiums by a compounded annual rate of only 0.4%, while the direct writers had increased their premiums by an annual 7.1% (Exhibit 5). The Non-Standard Segment xxx In terms of growth, the non-standard segment had outpaced the standard market by a considerable margin. Over the period 1988-1993 the non-standard segment had grown at a compounded annual 10.8%, while the private passenger market as whole had grown at an annual 6.0% (Exhibit 5). The size of the non-standard market was strongly influenced by two factors which subjected the market to large fluctuations: the pricing of Assigned Risk Plans, and the risk-taking behavior of standard insurers. The unwillingness of the large carriers to offer insurance to non-standard clients, coupled with legal requirements to have insurance in order to operate a vehicle, had led state governments to take action starting in the early 1950s. State governments, under whose jurisdiction insurance regulation fell, required standard carriers to underwrite non-standard risks under the so-called Assigned Risk Plans (later called Automobile Insurance Plans). State regulators set premiums for nonstandard risk applicants and assigned applicants to standard insurers in proportion to the insurers’ standard underwriting. For instance, when a standard insurer wrote 20% of a state’s policies, it had to take 20% of all non-standard applicants, which the standard insurers called their “Residual Business.”1 The general template adopted by Automobile Insurance Plans to determine premiums reflected the underwriting style of standard insurers since it utilized only a few criteria. The pricing of insurance in Assigned Risk Plans (ARPs) remained a contentious issue. Social advocates frequently endorsed subsidized rates, e.g. for inner-city markets, while insurance companies opposed any subsidies. In general, ARPs did not pose a great competitive threat to nonstandard providers. Since ARPs had to insure practically everyone, their rates were considerably higher than providers who had the ability to pick and choose. Moreover, those customers who found ARPs to offer a lower premium were very likely to generate big losses for the Risk Plans. As a result, Risk Plan’s (members’) underwriting losses on the order of 20-30% were not uncommon. Whenever rates in ARPs were adjusted, the attractive market for non-standard suppliers expanded or contracted. In the 1980s, for example, the California ARP wrote policies in excess of $2 billion; it had shrunk to $500 million by the mid 1990s. Similarly, 7.8% of all vehicles nationwide were insured under an ARP in 1978—in 1992, this number had dropped to 6.4%, yielding a total Assigned Risk Pool size (residual market) of $6,035 million. One year later, the total residual market had again shrunk by 12% to $5,320 million (Exhibit 6). The size of the non-standard market also fluctuated as standard insurers decided to take on more or less risk on the fringe of their traditional market. Consequently, estimates of the nonstandard market varied from 12 to 19% of the total automobile insurance market. For 1993, industry observers had gauged the non-standard market size to be $18 billion. Industry structure was very fragmented. In 1993, about 125 independent specialty companies and more than 25 subsidiaries of large insurance companies operated in the non-standard market. Of these, approximately 40 operated on a nationwide basis. Unlike standard insurers, non-standards focused almost exclusively on automobile lines and were not engaged in homeowner or life insurance. By the mid 1980s, the non-standard automobile insurance industry had acquired a somewhat “sleazy” reputation. It was dominated by small local operations whose reputation for not settling claims often put them at the top of consumer complaint lists. The high premiums and modest entry costs in this segment (low fixed costs) had traditionally attracted numerous competitors. Some independent agents were willing to represent providers even if they had weak financial ratings (such as a B- from A.M. Best, the industry’s leading rating agency2). Unreasonably low premiums could allow a company to garner quickly volume with price-sensitive customers. Since payouts due to accidents always trailed incoming cash from premiums (especially larger claim payments that could be dragged out for some time), uneconomic pricing was only apparent with a delay. Many small nonstandard insurers eventually became insolvent, however, because they had not set aside sufficient reserves to cover losses. All non-standard insurers relied on independent agency forces to sell their policies. Hence, their underwriting costs tended to be higher than those of standard insurers who either employed internal agency forces or underwrote policies over the phone (Exhibit 4). At the same time, the typical non-standard insurer did not try to build up a brand name, but relied on the efforts of its independent agents to promote its services. Non-standard drivers had higher accident risk and generated a larger number of claims. They tended to miss payments more frequently, moved more frequently, and canceled their policies more often. Policy duration was under 18 months on average, and often only half a year. Since premiums were much higher than in the standard market, customers also normally chose lower liability levels (often the legal minimum) than in the standard segment. Thus, liability coverage levels of $25,000 were common in the non-standard segment, while standard customers usually chose a coverage of $300,000. Lastly, the non-standard customers differed in self-perception. Until the actual contact with an agent, a customer frequently did not realize that he/she was a non-standard risk. Furthermore, experience had shown that in shopping for automobile insurance customers seldom solicited more than three bids from different companies. Hence, it was important for an insurer to get onto the mental “bid list” of customers who were searching for a new insurance carrier. Activities of a Typical Automobile Insurer The set of activities involved in competing in automobile insurance is shown in Exhibit 7. The most important activity in the category of inbound logistics was the gathering of information about customers, their risk characteristics and other factors that correlated with the likelihood of accidents. This data served as “input” for the pricing decisions and the new product development processes described below. The operations category encompassed three main groups of activities: rating, underwriting, and—not directly related to the insurance function—investing. First, the gathered data had to be analyzed and converted into rating schemes. These rating schemes allowed the matching of a particular customer profile with a premium. It was the job of the actuaries to find which of the collected data variables best predicted the likelihood of future accidents and to establish a commensurate premium for each possible customer profile. The actual matching process, i.e. the evaluation, classification and pricing of an application, was called the underwriting process. Lastly, since insurance companies received premiums before they had to pay out loss redemptions (for any given polity), they were able to invest the premiums in the interim. For most auto insurers the investment income was a crucial element of profitability, as noted above. Support activities for the operational side of the business included the training of actuaries, the establishment of actuarial methods (how to elicit the best predictions from the collected data), and investment practices that would frequently determine the profitability of the entire business. Outbound logistics consisted of two main sets of activities: policy issuing, and billing and collections. Once the underwriting process had been completed, i.e. the correct premium computed, the actual policy had to be printed and sent to the insured party; a further copy was usually sent to the originating sales agent and a copy was retained in a branch or central office. Billing and collections was concerned not only with collecting monthly premiums, but also with keeping track of payments which would arise from deductibles that were part of policies. The marketing and sales category spanned activities dealing with new policy sales, policy renewals, and advertising. As noted before, there were different channels through which policies were sold (internal agents, independent agents, and direct selling via the telephone). Since independent agents usually offered policies from many carriers, agent management was a vital undertaking for agency writers. Once a customer had been obtained, customer retention became a crucial factor. Thus, shortly before policies expired, customers had to be informed and policy renewals initiated. Similarly, whenever the risk characteristics of the customer changed (e.g. the customer bought a new car) the policy needed to be repriced. Finally, this category included advertising activities which raised brand awareness with customers. The sales activities were supported by agent training that familiarized agents with the insurers’ underwriting guidelines and rating procedures. To speed up the application process, communication links between the agent and the insurers’ underwriting departments became critical. Moreover, in order to compare premiums with competitors, many insurers had build up considerable market research capabilities. Lastly, expertise had to be developed for new product introductions. New products included the underwriting in new geographic areas and of new customer groups. The activities in the after-sale service category brought the insurance company in close contact with the customer. In the eyes of the customer, the processes of claims response and loss settlement were the most important functions of the insurance carrier. Each insurance company had to establish a system of claim reporting by the customer. It had to be determined whom the customer was supposed to call; e.g., the agent or the insurance company. The role of the contacted party had to be specified; for instance, was the contacted person authorized to send a tow truck, a rental car, or an adjuster (the person who would put a value on the damage)? In addition, the nature of the next response of the insurance company had to be determined; e.g., over the phone or face-to-face. Finally, the process and speed of settling claims had to be managed and control mechanisms put into place. Support activities in the service category included the recruiting and training of adjusters. In addition, company procedures had to specify the role and the competencies of the adjusters; e.g., whether the adjuster was authorized to compose check on the scene of the accident. On the level of the firm infrastructure, several activities spanned the entire value chain. As the insurance industry was highly regulated, regulatory compliance played an important role (e.g., premium increases had to be submitted and approved by state insurance commissioners). Similarly, since many insurance claims involved lawsuits, every insurance carrier had built up an elaborate legal support system that would, for instance, negotiate out-of-court settlements. Lastly, each insurance company had to decide on a particular organizational form (e.g., organized along geographic lines, or strictly along functional lines) and on contractual arrangements for its employees and top managers (e.g., incentive contracts that would emphasize growth or profitability). Competitors State Farm xxx State Farm was by far the biggest player in the automobile insurance industry with a market share of almost 20% (Exhibit 4). The State Farm Group, led by State Farm Mutual Automobile Insurance Company, offered multiple lines of property/casualty, life and health insurance throughout the US and Canada through a cost-efficient in-house agency force. The company enjoyed an A++ (superior) rating by A.M. Best, which reflected State Farm’s high capitalization and its conservative operating strategies. State Farm’s exclusive agency force numbered about 17,500. Agents were capable of cross-selling all products, including commercial lines for smallto medium-sized accounts. Business with policy holders was handled via regional offices that were located throughout the US. In addition, approximately 1,000 claims offices had been established in the principal cities throughout the country. Almost 28,000 full-time salaried claims employees and supervisors were working out of the claims and regional offices. As a mutual company, State Farm did not pay stockholder dividends, which had helped considerably in generating policy holder’s surplus (the insurance-industry’s equivalent to equity). Over the period 1990-1993 State Farm had been able to increase surplus from $17.9 billion to $21.3 billion. A frequently used measure of capitalization of insurance companies was the premium leverage (premiums/surplus). State Farm’s premium leverage had increased only slightly from its very low level of 1.37 in 1990 to 1.42 in 1993. With a highly liquid $13 billion stock portfolio and strong operating cash-flows averaging $2.5 billion annually, State Farm’s liquidity position was excellent as well. Allstate xxx The Allstate Insurance Group, led by Allstate Insurance Company, was primarily engaged in property/casualty and life insurance. These two business segments made up 85% and 13% of total revenues. Allstate Insurance Company had been established in 1931 by Sears, Roebuck and Co. and developed into the second largest property-liability insurer and the 20th largest life insurer in the U.S. In preparation for an eventual spin-off by Sears, Allstate was incorporated in 1992. In June 1993, Allstate sold 19.9% of its outstanding common stock in the largest recorded initial public offering. The IPO provided net proceeds of $2.29 billion. Sears continued to hold the remaining 80.1% of the stock, which it planned to sell by mid-1995. Allstate was rated A- reflecting its strengthened balance sheet after the IPO capital infusion which helped to offset the $2.5 billion loss that was incurred in the wake of Hurricane Andrew in 1992. Its policy holders’ surplus increased from $4.7 billion in 1990 to $7.1 billion in 1993. The capital infusion also helped to reduce the traditionally aggressive premium leverage from 3.00 in 1990 to 2.25 in 1993. Like many other property/casualty insurers, Allstate recorded underwriting losses in most years and relied on its investment income to generate profits. In 1993, Allstate’s total underwriting losses summed up to $2.39 billion. Yet due to its investment income of $3.54 billion, Allstate achieved a total after-tax operating income of $1.15 billion. In its 1993 Annual Report, Allstate identified as its “core competencies” its brandname, its exclusive sales force, its experience with other distribution channels and its proprietary data base. Allstate had more than 20 million customers and enjoyed high recognition for both its brand name and its slogan “You’re in Good Hands.” Products were marketed through a variety of distribution channels, with the core of its distribution system being a broad-based network of approximately 14,600 full-time Allstate agents in 9,300 locations in the US and Canada. The agents operated out of neighborhood offices and 130 sales offices located in Sears retail stores. In addition to its full-time exclusive sales force, Allstate’s policies were sold by some 1,900 independent agencies in rural areas. Allstate wrote 93% of its business through its own agency force, which was characterized by a low turnover of 5% (in 1992) and an average tenure of 13 years. In 1993, Allstate increased its spending on education for its agents, stressing in its training programs the need to move towards long-term customer-agent relationships. Agents were capable of cross-selling other products to policy holders, including homeowners insurance, commercial lines and life insurance. Allstate’s life insurance policies were also distributed by financial institutions, specialized brokers and direct marketing techniques. Its proprietary database of 58 million Sears/Allstate households enabled Allstate to profile and market directly to additional potential customers. GEICO xxx The Government Employees Insurance Company (GEICO) was founded in 1936 by Leo Goodwin based on two insights. First, federal, state and municipal employees had fewer accidents than the general population (in addition to having a stable income). Second, by cutting out the middle-man, i.e., the policy-selling agent, one could sell auto insurance at a significant discount. Hence Goodwin started selling policies directly to his target customers via mail and telephone. The concept proved to be successful and GEICO grew rapidly. In 1976, however, because of miscalculations of its claims and weak cost accounting leading to persistent mispricing, GEICO was on the brink of insolvency. Learning of GEICO’s woes and believing that its basic business concept was still sound, Warren Buffett began investing $50 million. Over the years, Buffett’s Berkshire Hathaway increased its stake in GEICO to 48%.4 GEICO continued to grow and achieved excellent operating results. For instance, its 1993 pre-tax return on revenue of 11% was more than 2.5 times better than the comparable industry return. As a result, GEICO’s stock appreciated greatly. In 1993, Berkshire’s stake in GEICO was worth more than $1.5 billion. By 1993, the GEICO Group had set up four companies: GEICO, which was still selling policies only to “preferred,” i.e., low-risk government employees and military personnel (70% of all premiums); GEICO General Insurance Co., which sold to preferred-risk applicants other than government employee and military personnel (25%); GEICO Indemnity Co., which wrote standardrisk private passenger automobile and motorcycle insurance; and finally, GEICO Casualty Company, a subsidiary of GEICO Indemnity, which wrote non-standard private passenger automobile insurance. GEICO still relied primarily on its direct-response marketing methods, through which customers could buy policies directly through the mail or by calling a toll-free number. Military and military-base civilian personnel were served primarily through 94 commissioned contract agents. These agents were a major source of new business for GEICO Indemnity and GEICO Casualty, acquiring nearly 64% of new auto business for these divisions. In total, these few agents generated 13% of GEICO’s new auto business. While GEICO’s growth as a whole had stalled from 1992 to 1993—it actually saw its total volume of premiums shrink—GEICO Indemnity’s (including GEICO Casualty’s) net premiums had grown in both years by 14.3% and 15.1% respectively. Moreover, in GEICO’s 1993 Annual Report, CEO O.M. Nicely singled out these two divisions as future engines for growth. He promised to invest in training and information systems so that new business for these divisions could be handled on the phone and not only by its commissioned contract agents. Since its direct response channels gave GEICO underwriting expenses which were among the lowest in the industry, GEICO had become a feared competitor in every market (Exhibit 4). Progressive Corp. Early History. The Progressive Insurance Company was founded on March 10, 1937 by Joseph M. Lewis and Jack H. Green. The company originally focused on offering automobile insurance to blue-collar workers and property insurance on cars financed by local finance companies. Between 1945 and 1955, Progressive’s net premiums written increased from $300,000 to $4,500,000. By that time, Progressive had developed into a broader underwriter, relying on a small network of independent agents. An apocryphal story has it that in the mid-1950s, Peter Lewis and some of his colleagues discussed several times over lunch the many calls they received from customers whose policies had been canceled and who were looking for a new insurer. It was decided to focus exclusively on this emerging “non-standard” segment. In 1965, Lewis discovered that the president, Jack Green, was trying to sell the company. Lewis organized a group which eventually bought the company for $4 million. In the re-named Progressive Corporation, Lewis became president, CEO, and main equity holder with a 26% stake in the company. Progressive’s Activity System6 How was Progressive able to achieve consistently an underwriting profit with its high-cost, high-risk customer group that no standard insurer dared to underwrite?7 Clearly, premiums with surcharges of 50-300% over the standard premiums were part of the story, yet premiums were bounded by the “default” choice of the state-run assigned-risk insurance pools. (For the effects of higher premiums on the cost and revenue structures of typical non-standard and standard policies, see Exhibit 8.) Inbound Logistics and Operations xxx Progressive had realized that if it were able to pick “better” risks out of the non-standard market, it would be able to offer a lower premium than rivals, attract the particular driver, and yet still incur lower losses. Since the 1950s, the company had invested far more heavily than its competitors in collecting and analyzing accident data. With the help of a database which it had developed on the personalities, lifestyles, and driving habits of highrisk groups, Progressive was able to price policies to match the underwriting risk. Progressive separated drivers into far more risk categories than did its competitors and set a wider variety of premiums. For instance, in motorcycle insurance, most insurers considered only the size of the motorcycle but did not consider the age of the rider. Progressive began with two age categories, under and over 25 years, and subsequently expanded it to 11 age categories. COO Bruce Marlow, a Harvard MBA, elaborated: “We . . . know how this zip code is different from that zip code, how a four-cylinder Ford is different from an eight-cylinder Ford, what’s the difference between a 38-yearold single, married, widowed. We’re looking for opportunities where the market price is by some degree above where the cost structure is.”8 In a typical community, Progressive offered more than 14,000 different premiums based on driving record, vehicle year, make and model, driver age, sex, marital status, residence and other factors. Not surprisingly, in 1992, Fortune magazine called Progressive, “The Prince of Smart Pricing.”9 In addition to using a finer-grained pricing system, Progressive also sought to offer a wider array of payment plans, limits of liabilities, and deductibles than its competitors. With respect to its investment practices, Progressive pursued a very conservative investment strategy. According to analysts, Progressive possessed the most liquid portfolio of all major automobile insurance companies. Its 1993 portfolio consisted mainly of short-term and intermediate term, investment-grade fixed-income securities (76.6%). A relatively small portion was invested in preferred and common equity securities (16.3%). The remainder of the portfolio was invested in longterm investment-grade fixed-income securities (2.8%) and non-investment-grade fixed-income securities (4.3%). (see Exhibit 9) Information and Communication Technology xxx The problem of agents’ mispricing policies was identified as a major cost driver. Once an incorrect premium was quoted, large costs could ensue. One way to decrease these costs was to focus on quicker application processing at headquarters, where policies were double-checked. In 1989, Progressive began to install a $28 million computer system and nationwide voice/data system which allowed faster communication between agents and headquarters and shorter processing times for both applications and claims. Moreover, it decreased operating expenses. For instance, by speeding up the claims processing, the new system enabled local adjusters to process 30% more claims. The claims processing network, dubbed Progressive Automated Claims Management (PACMan), allowed local agents to download information into the central computer from their terminals. Previously, agents had to mail reports between offices and corporate or divisional headquarters, which had been a slow and cumbersome process. Marketing and Sales xxx Progressive relied on a large network of independent agents (over 30,000) which provided broad coverage without the large fixed costs of running an internal sales force. With respect to agent management, Progressive stood out from the rest of the non-standard segment. The company used its own sales force to call on agents—a strategy which management believed was unique in the industry. Sixty sales representatives called on independent agents to explain the company’s complex rating systems.10 During the discussions around California’s Proposition 103 in the late 1980s (see below), Lewis had vehemently defended the free market, arguing that insurance markets were characterized by tough competition. As a response, he frequently heard that even though in every market there were many insurers, the pricing was often so opaque that consumers faced high search costs when they tried to shop for the best offer. Taking this criticism to heart, and realizing a market opportunity, Progressive introduced the Express Quote Service in California in October 199