Progressive Corporation (The)

Good morning, and thank you for joining us today for Progressive's fourth quarter investor event. I am Doug Constantine, Director of Investor Relations, and I will be moderator for today's event. The company will not make detailed comments related to its results in addition to those provided in its annual report on Form 10-K, quarterly reports on Form 10-Q, and the letter to shareholders, which have been posted to the company's website. This quarter, we will have a presentation on a specific portion of our business, followed by a question and answer session with members of our leadership team. Introductory comments by our CEO and the presentation were previously recorded. Upon completion of the previously recorded remarks, we will use the balance of the 90 minutes scheduled for this event for live questions and answers with leaders featured in our recorded remarks as well as other members of our management team. As always, discussions in this event may include forward-looking statements. These statements are based on management's current expectations and are subject to many risks and uncertainties that could cause actual events and results to differ materially from those discussed during today's event. Additional information concerning those risks and uncertainties is available in our annual report on Form 10-K for the year ended December 31st, 2021, where you'll find discussions of the risk factors affecting our businesses, safe harbor statements related to forward-looking statements, and other discussions of the challenges we face. These documents can be found via the investor relations sections of our website at investors.progressive.com. To begin today, I am pleased to introduce our CEO, Tricia Griffith, who will kick us off with some introductory comments. Tricia?

Good morning and thank you for joining us today. As I stated in my letter, 2021 was a year like no other. We were forced to confront the new normal imposed on us by the pandemic. We faced business challenges unlike those we've ever previously seen, all while continuing to serve our customers at the level they have come to expect from Progressive. Our people are flexible. They can see the challenges coming and react appropriately to ensure we meet our business objectives. Our ability to respond like this is supported by our corporate culture, which is built on our four cornerstones. Who we are, which is based on our five core values. Why we are here, which is our purpose. Where we are headed, which is our vision. And how we will get there, which is our strategy. This construct guides us so we're all headed in the same direction, regardless of the bumps in the road. To achieve our vision to become consumers and agents, number one choice and destination for auto, home, and other insurance, we need to execute on our four strategic pillars. In our quarter two call, I spoke in detail about our pillars and their importance to our business, and I continued that discussion in my annual letter to shareholders. While each pillar is equally important to support our strategy, today we're going to focus on a single pillar, competitive pricing. Competitive pricing does not mean having the lowest rate all the time, although we strive to do that as much as possible. Rather, it means having the correct rate to match the risk we are trying to ensure, while at the same time delivering more value to the customer. We believe that if we can out-segment our competition, If settled claims accurately and managed expenses appropriately, providing attractive rates to our customers will come naturally. It's no secret that the insurance industry is going through a period of immense change brought on by catastrophic weather events and macroeconomic headwinds. Our business is no exception. The changes in personal lines results, which is our largest line of business, were especially pronounced. Personal lines results continued to benefit in the first quarter of the year by lower frequency brought in by the pandemic. However, the last three quarters delivered challenges. Frequencies started to rise, returning to near pre-pandemic levels. While the increase in frequency did not surprise us as people started driving more, what was a bit of surprise were the increases in severity, especially the cost to repair and replace vehicles. It was these rising trends that forced us to increase rates. Once we realized these developments were not temporary, our product managers acted quickly and in the second quarter started taking personal auto rate increases, which were in advance of many in the industry. We've continued to increase rates to the fourth quarter with a total of 8% taken in calendar year 2021. In addition to rate changes, we also had to make some difficult decisions to reduce expenses and increase underwriting scrutiny to hit our profitability goal of a 96 combined ratio, which is the expected impact of reducing growth. While these decisions are tough to make and the short-term results can be difficult to acknowledge, we believe we have built a better mousetrap. History has shown that following times of industry disruption, we have grown our personalized business faster than our competitors with better profitability. Further, it has been after times like this that we have made some of our greatest strides to achieving our vision. Today, we will spend some time talking about the mousetrap that allows us to react faster and, we believe, more accurately than the industry as a whole. Our first topic today will be on the science of personal auto pricing. We will discuss how we take massive amounts of incoming data and apply tools to make risk-based decisions and provide an accurate view of rate adequacy. John Curtis, our national product development leader, will be presenting this topic. John currently leads our personal auto product R&D and pricing groups and has held numerous roles throughout his 19-year career with Progressive. Next on the agenda is Conak Varma, who will demonstrate the real-world execution of the science used by our product managers. Conak has been with Progressive for over 20 years and was recently appointed as the general manager of the West region. Connick currently leads a team of about 12 product managers and draws on many years of direct product management experience to lead the team. Connick will wrap up the presentation by talking about Progressive's deployment capabilities. Following the presentations, we will of course have our Q&A session. Again, thank you for joining this morning. I will now hand it over to John Curtis. John?

Thanks, Tricia. Today, Connick and I will give an update on the rate-making process in our personal auto business. We have structured our discussion in a Q&A format based upon common questions we receive from our investors. There are many facets to managing rate level in our personal auto business, and today we're going to focus on two key aspects. First, we will share how we determine our rate need to support our operational goal to grow as fast as we can at or below a 96 combined ratio. This can broadly be categorized into two areas, segment level and aggregate level. Segment level pricing is largely the responsibility of our product R&D group. While very important, we are not going to discuss this in detail today. You might remember that Pat Callahan and Sanjay Vyas discussed our approach to product segmentation back in our third quarter 2018 investor relations call. Today, we're going to focus on aggregate rate level. ensuring we are collecting enough premium in total to cover our expected future loss costs and expenses to achieve our 96 combined ratio goal. We will provide an overview on two key aspects of how we do this. The first topic that I'll cover is the role of our auto pricing team, who is responsible for the pricing indication process and providing product managers with an accurate estimate of rate need to ensure they hit rate revision profit targets. The second topic that Connick will cover is how our product managers leverage these pricing indications, their broader toolkit, and local knowledge and expertise to set the pricing strategy for their respective states to ensure they hit not only rate revision targets, but also our calendar year goals. We will close the presentation with an overview of our deployment capabilities. In our personal auto business, it is very common for us to do hundreds of rate revisions per year, which involves many groups at Progressive. This is an area we've invested significantly in over the years to ensure we have industry-leading speed to market and the agility to adapt our resources to meet the dynamic needs of the marketplace. So let's start with the first topic on how pricing and product management work together to determine our aggregate rate needs. Before I go into details on our pricing tools, I will introduce six key concepts that we will be covering today. The first is compliance. Our product and rates need to comply with all state regulations, which can vary in each jurisdiction and often include limitations in our ability to use specific variables to rate policies. The second is that we price to cost. Our primary goal is to match the rate we charge with the expected cost to ensure our customers. To determine our aggregate rate need, we need to consider loss and loss adjustment expenses, operating and acquisition costs, and our profit load. Third, insurance pricing is perspective. Insurance is unique, as we do not know our cost of goods sold when we write a policy. As a result, we need to estimate future losses and expenses and ensure we charge enough premium to cover these and earn our profit load. Fourth, our rates need to comply with actuarial standards. The three primary considerations are to ensure rates are not excessive, inadequate, or unfairly discriminatory. And for clarity, when I say unfairly discriminatory, I mean not tied to underlying loss costs. Fifth, when managing our business, product managers need to consider both accident year and calendar year results. Our pricing decisions are based on accident year data, which is determined by the actual timing of a loss or accident, as our product offering provides coverage for accidents we expect to occur during a rate revision. Our calendar year results consist of all the losses and premiums for the year, which include the current accident year plus any runoff from prior accident years. Finally, we will discuss our premium earnings convention. Unfortunately, we can't change rates on our entire book of business overnight. Rather, rate changes start earning in when a rate revision goes live, either upon filing or after approval by a Department of Insurance, and that depends on the state. While new customers see the rate changes immediately, existing policyholders' rates will change at their next renewal event. As a result, it takes time for rate changes to earn in. Fortunately, Progressive writes mostly six-month policies, which allows us to earn rates faster than many competitors who issue a higher mix of annual policies. Let me start with the auto pricing group. Pricing's goal is to deliver best-in-class tools to enable product managers to make risk-based decisions and to provide an accurate view of rate adequacy for our auto business. Think of this as the science behind the rate-making. Given our beginnings as a non-standard writer and evolution to become a leader in segmentation and product innovation, we've honed our pricing sophistication for nearly 85 years. We continue to invest in attracting and retaining high-quality talent, building systems and processes to make sure we can accurately price a very complex product in a high-quality manner, and advancing the science of pricing to improve our assessments of rate adequacy. The pricing indication is the key tool we use to determine our indicated rate need. It helps product managers answer the question, how much do I need to raise or lower rates in my next rate revision to achieve the target combined ratio or better on an accident year basis? The indication is based upon two primary loss ratio calculations. The numerator is the projected loss ratio. This is our best estimate of the loss ratio we expect for policies that we will write in the upcoming rate revision at our current rate level. The denominator is the target loss ratio. This is the loss ratio we need to achieve to have enough premium left over to cover our prospective non-loss costs and profit load. The difference between these two ratios is the indicated rate need for the next revision. Let me provide more detail on how we determine the target loss ratio. Our goal is to price policies in a way that we will collect enough premium to cover our expected loss costs, loss adjustment expense, operating and acquisition costs, and have enough left over to hit our 4% profit margin. The chart on the slide depicts the various components to determine this loss ratio. And in this example, as you can see, our target loss ratio is 66%. It's very important to note that changes in our non-indemnity cost structure can result in changes in our target loss ratio. And thus, having a low cost structure is very important. All else equal, if expenses are lower, our target loss ratio will be higher, and we can charge customers less for the same protection. The other key input into our indication is the projected loss ratio. This is our best estimate of the loss ratio we would expect for policies we are going to write in an upcoming rate revision at our current rate level. As I said earlier, we do not know what our loss cost will be when we write policies, and given our 4% profit target, we need to be very accurate in our projection. So how do we get there? This requires us to adjust our historical accident year losses and premiums to reflect our best estimates of the future. Let me talk a little bit about the adjustments we make to our losses first. The first consideration is loss development, which is the process of estimating the ultimate frequency and severity of our claims. Since rate making is perspective, we need to know how much will recent claims change from what we know about them today. We also need to know how many more claims will be reported from this time period and how much they will cost and change over time. To develop our historical frequency and severity, we analyze historical development patterns using a variety of methodologies to help us achieve accurate estimates. Second, we also need to adjust for weather, which can be highly volatile over time. Our goal here is to price to a longer-term average to make sure that we are not over or underreacting to recent weather events. Once losses are fully developed, both frequency and severity are trended to the midpoint of the rate revision period. The midpoint represents the average cost to get sold of losses during the revision. There are many factors that can impact our loss trends from macroeconomic variables such as inflation and gas prices to improve safety technology to law changes and even our mix of business. We monitor these to understand where frequency and severity may be headed and we update our trends on a monthly basis. On the premium side, we make two important adjustments. First, we adjust historical premium levels and bring them to current rate level to reflect our most recent pricing. Second, premiums also have trends, which are largely a function of changes in our mix of business. We need to account for the fact that our premiums can change due to the segments we write. We do not want to change rates simply because our mix of business changes over time. Our indications contain many inputs to help us project the future, and as a result, there is pricing risk. Let me share a few examples of the strategies we employ to manage this. First, when possible, we like to rely on the most recent accident periods, which allow us to be responsive to our most recent experience and limits the trending period. Here, our data scale is a big asset as it affords us increased precision and prevents us from reacting to noise. The second is our high frequency of rate revision. This allows us to price to a shorter rate revision length and to be more responsive to changes in our indications. You have probably heard us say that we take smaller and more frequent bites at the apple. As a result, it is common for product managers to complete multiple revisions in a year. This is true in general, but not always, as we are currently taking much larger bites in the current rapidly escalating loss cost environment to ensure rate adequacy. The third is our high mix of six-month policies. As I mentioned earlier, once a rate revision goes live, it takes time to earn into the book because we need to wait for policies to renew into the new rate level. As you'd expect, six-month policies earn in roughly twice as fast as annual policies. Our high mix of six-month policies dramatically increases the speed with which new rates earn into our results. As you can imagine, COVID was a massive shock to our business and has forced us to adapt our pricing indications in multiple ways. Today, I'm going to use collision coverage as a quick case study to highlight the impact on our loss trends and to share our responses. This slide contains a time series view of our collision coverage frequency, severity, and pure premium through the end of 2019. It represents a pre-COVID period. The solid line is monthly data. You will notice that this data can be bumpy, which is largely due to seasonality. The dotted lines represent the trailing 12-month average. As you can see, during this time, our trends are relatively stable. Peer premiums were increasing on average by slightly more than 3% per year, with frequency gradually decreasing and severity growing faster than the rate of inflation. During this time, we were tracking a variety of variables that could help inform where we thought trends might be headed in the future. One thing we did not expect was a once in a lifetime pandemic. On this slide, the data has been updated through April 2020 to reflect the onset of COVID. As you are aware, a massive decrease in driving significantly reduced our collision frequency. This also dramatically changed our data and challenged us to think of new ways to project trends. The key question we needed to answer was, how quickly will frequency rebound? To help us answer this question, we leveraged a variety of data elements to inform future projections of frequency. This data came from a variety of sources. First, we closely monitor our claims frequency data to understand both monthly changes and comparisons to pre-COVID levels. Second, we leveraged our extensive snapshot data, which we know is highly correlated with frequency. This data provides a daily view of vehicle miles traveled and important segment level data on driving patterns such as time of day and day of week. Third, we utilized input from product managers who brought local knowledge about changes to driving in their respective states. By analyzing this data over time, we could start to quantify with what likelihood, by how much, and when frequency might return to pre-COVID levels. This analysis is conducted at a state, channel, and coverage level, and we update it very frequently. This slide shows the same three graphs through the end of 2021. As you can see, our collision pure premiums have rapidly accelerated to well above historic levels. While frequency has rebounded, it is still below where it was prior to the pandemic. However, collision severity has dramatically increased, driven by supply chain disruptions and the soaring prices of used cars. Like frequency, our response for severity has been to leverage new data sources like the Mannheim Index to explain the causes of these increases and where they might be headed in the future. In collision, more than half of our lost dollars are from total losses as we pay to replace totaled vehicles damaged beyond net cash value. This plot is showing the average value of used cars from Mannheim and Orange. and the estimate of actual cash value, or ACV, of the vehicles we write estimates on in blue. Not surprisingly, what we see is a very strong correlation between these two metrics with ACV lagging Mannheim by a few months. Given this, signals of upward movement and car values are a leading indicator of ACV estimates and severity. With this data, we are better able to evaluate scenarios of where severity may be headed in the future with the cones representing the fact that there is still a high level of uncertainty. Given there's a high level of uncertainty of where our trends and therefore our rate need might be headed, I want to share with you how we ensure we are getting our latest views of rate adequacy to product managers in a timely manner. As I mentioned earlier, our goal is to frequently update our trend analysis and pricing indications so that product managers can quickly respond to changing market conditions. Our indications are updated quarterly for each state, channel and coverage combination, which generates thousands of pricing recommendations per year. As I mentioned earlier, we update our loss trends and premium trends on a monthly basis. Given the rapidly changing trend environment due to COVID, we are now able to update the trend portion of our indications monthly, which provides product managers even more up-to-date data on changes to their rate indications. This allows product managers to adjust their plans as the needs of their business change. And now I'll hand the presentation over to Kanak, who will provide an update on how product managers leverage our pricing science to deliver our operational goal of growing as fast as possible at or below a 96 combined ratio.