Progressive Corp (PGR) — Q2 2025 Earnings Call Transcript

Good morning and thank you for joining us today for Progressive's Second Quarter Investor Event. I am Doug Constantine, Director of Investor Relations and I will be the 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 includes a presentation on a specific portion of our business, followed by a question-and-answer session with members of our leadership team. Introductory comments in 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 31, 2024, as supplemented by our 10-Q reports for the first and second quarters of 2025, where you will 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 section of our website at investors.progressive.com. To begin today, I'm pleased to introduce our Personal Lines President, Pat Callahan, who will kick us off with some introductory comments. Pat?

Good morning and thank you for joining us today. Through the second quarter, 2025 continues to be one of our best years on record by all objective measures. We delivered strong profitability while simultaneously growing at an incredible pace, adding over $5 billion in premiums written and nearly 2.4 million additional policies in force during the first half of 2025 compared to the first half of last year. We've previously said it's easy to grow an insurance company if you're not focused on underwriting profit, and conversely, relatively easy to generate profit in insurance if you're not looking to grow. But simultaneously doing both, while operating in the highly competitive U.S. property and casualty marketplace, requires the rare combination of strategic focus and relentless execution that we consider key sources of Progressive's competitive advantage. Full statutory industry results released in June show that Progressive gained more than 1.5 points in personal auto market share in 2024 while outperforming the industry combined ratio by more than 7 points. Our 2024 market share increase was the largest share gain of any carrier in the past 15 years. This rare combination of profitable market share growth isn't an isolated event for Progressive. Over the past 15 years, we've increased our auto premium almost fivefold while simultaneously running close to 9-point wider underwriting profit margins. Our performance is the direct result of executing against our four strategic pillars: people and culture, product breadth, brand and competitive prices. Today, we'll focus on competitive prices and provide you with some insights into how our ability to predict and price future loss costs and to rapidly deploy products and pricing to match our premiums to our costs remain key ingredients of our continued success. The compounded effect of doing this with each successive product model has enabled us to continue to make great progress towards achieving our vision of becoming the number one destination for consumers, agents and business owners for insurance and other financial needs. Building on our exceptional underwriting profit performance, we continue to invest to drive continued growth. Despite seeing greater competition now than we did at the beginning of the year, through the second quarter, we continued to see strong demand for our personal auto products across both of our distribution channels. The independent agent channel is a great barometer for the competitive environment, where available coverage options across carriers are presented via comparative raters, which provide agents and their clients real-time comparisons of how our products compare to those offered by other carriers. Every indication is that our auto products have continued to outperform on a relative basis as evidenced by strong year-to-date double-digit growth in new applications, premiums written, and policies in force. On the direct side, our marketing engine remains highly effective, generating high-quality prospects at near-record levels and our conversion rates indicate that despite the increasing marketing spend, our prices are still highly competitive and provide consumers a good value relative to other in-market options. Year-to-date, we have spent $2.5 billion on marketing, an increase of about $900 million compared to this time last year. And as our volume denominator grows, achieving year-over-year new application growth becomes more challenging. But so far, we have continued to leverage our scale in identifying new opportunities to refine where and how we invest our marketing spend to drive profitable growth. Similar to Personal Lines, we continue to rapidly grow market share in our Commercial Lines business while consistently beating industry combined ratios by 8, 10 and as much as 20 points over the last 20 years. This consistent profitability is particularly impressive when looking at how U.S. commercial auto continues to struggle with profitability, producing its 14th consecutive unprofitable calendar year in 2024. Our success can be attributed in part to the intense focus on commercial auto as a core line of business in our Commercial Lines offering. This auto focus, in conjunction with introducing the segmentation of commercial auto into business market targets more than 10 years ago, has enabled us to capitalize on meaningful and actionable differences between resulting vehicle types and usage that we can operationalize across all aspects of the business. This granular focus has allowed us to quickly develop segment-level insights and execute proactive rate and underwriting actions at the business market target level to deliver strong division level performance, as we did late last year when we adjusted rates and underwriting across business market targets to drive written premium growth across all our business market targets other than for-hire transportation. We're now extending those capabilities to our expansion product lines like the business owner policy product. To support long-term positive contributions from these newer businesses, we leverage established pricing and product delivery capabilities to ensure we have the necessary monitoring and insights in addition to leading capacity and the delivery agility to bring rapid rate and underwriting adjustments to market. Leveraging these proven capabilities should help our expansion commercial businesses deliver the profitable growth we expect from all of our underwriting businesses. A significant contributor to delivering our continued profitable growth is the ability to quickly and decisively respond to changes in loss costs to ensure we can remain open for business when inflationary pressures create hard markets and elevated shopping levels. This is no small feat, especially given the lack of historical precedent for some of the drivers of recent loss cost increases. Fast forward to today and we're similarly working to model first, second and third order effects of global tariffs and potential supply chain disruptions to determine the appropriate future rate levels for these emerging macroeconomic events. Our pricing teams are responsible for translating highly complex and dynamic internal and external data into timely rate level recommendations. Each of our product teams are supported by a group of these talented individuals who leverage complex actuarial methods to understand and trend our future costs towards our goal of collecting the right premium to cover both the costs we'll incur and our target profit margin. Today, we'll be diving into the details of our personal and commercial lines pricing theory and practices. To guide us through this, we have two senior pricing leaders with us. First, Brad Granger, who has been our national auto pricing leader for the past 17 years and who will be celebrating his 25th Progressive anniversary later this month, will discuss some of the technical theories behind our pricing methodology. Following Brad, Jen Kubit, who has been with Progressive for 21 years and recently took on the leadership of our Commercial Lines pricing organization, will explain how these theories are applied in practice at Progressive. Once again, thank you for joining us today. I'll now turn it over to Brad.

Thanks, Pat, and thank you, everyone, for joining this morning. What is our product? A property and casualty insurance company such as Progressive doesn't actually manufacture a physical product. Our product is not tangible. Instead, it is a transfer of risk from insured to insurer, exchanging the small probability of an adverse financial event in exchange for the payment of a premium. That transfer of risk is outlined in our policy contract. And that transfer of risk is both dependent on when accidents occur or the accident date and time bound within our short-term contract lengths of 6 or 12 months. Car accidents are not corn flakes. As opposed to a tangible product like corn flakes, where costs are largely known before the product is priced and sold, Progressive doesn't know for sure what our costs are until long after our product is priced and sold. That is precisely the transfer of risk from insured to insurer I just described only now from the perspective of the insurer, that risk is now ours. And as we will discuss in a few minutes, we do many things to mitigate that risk within both our pricing science and our operationalization of getting the right rate to market quickly. Pricing to expected cost. First, when we talk about pricing for costs, we are referring to all parts of our premium dollar: losses, loss adjustment expenses, or LAE, general expenses and profit. According to the Casualty Actuarial Society statement of principles on ratemaking, 'A rate is reasonable and not excessive, inadequate or unfairly discriminatory if it is an actuarially sound estimate of the expected value of all future costs associated with an individual risk transfer.' A few things to highlight there. First, it references expected value of future costs where actual outcomes could be lower or higher. Second, it is always forward-looking or prospective as it is an estimate of future costs. Third, today's presentation will focus primarily on not excessive and not inadequate as opposed to not unfairly discriminatory, which is largely in the domain of our product development departments. This is analogous to the field of economics and the distinction between macroeconomics and the focus on changes in whole economies and aggregate variables, and microeconomics and the focus on changes in individual firms and consumers. We will focus today on macro level pricing as opposed to the individual relative risk rating and segmentation of micro level pricing. In addition to our approach being aligned with actuarial principles, it is also well aligned with state regulation as nearly all states legally require that rates are not excessive, not inadequate and not unfairly discriminatory. And it also fits quite well within Progressive's vision to be consumers' number one choice via competitive rates. But to do so by growing profitably and adhering to our core value of profit. The fundamental pricing questions. There are two questions that completely underlie our approach. One, if we don't do anything to current rates, what loss plus LAE ratio should we expect in the upcoming rate revision for the policies we are about to write? And two, what do we need to do to current rates in order to hit our combined ratio target in the upcoming rate revision for policies we are about to write? We refer to these as the fundamental pricing questions. And in the coming slides, we will build the fundamental pricing equation to answer these questions. Actuaries develop predictions that minimize bias and variance. An enormous competitive advantage that Progressive has is the quality, granularity and quick availability of data. Our growth and scale have only deepened that advantage. But if you don't think you have a blind spot, then you have a blind spot. And with all of this quality data, if we are not careful, we could slice the cake and slice the cake and slice the cake until all we have are crumbs. This is where we, in our pricing organizations, come in. Actuaries are experts in matching rate to risk, balancing responsiveness and stability and maximizing accuracy and precision. And as we will see in the coming slides, as we try to price this accident year promise, our data is trying to fool us. There is no single perfect piece of data. Our work centers around correcting for bias or for accuracy and spread and variance or precision and finding the signal through the noise. And bias, as used here, refers to statistical bias. From Wikipedia, 'In the field of statistics, bias is a systematic tendency in which the methods used to gather data and estimate a sample statistic present an inaccurate, skewed or distorted depiction of reality.' Every month, Progressive releases financial information publicly describing our underwriting performance by line of business. To align with generally accepted accounting principles, this is predominantly a calendar year or month view that includes incurred activity on all losses and LAE regardless of date of occurrence. For the purposes of ratemaking, however, we are attempting to price for the accident year, that is ensuring that we have the correct rates at the time the accidents occur, which I'll explain further in the next few slides. Calendar year incurred losses are a combination of paid losses and change in reserves. They can also be subdivided to show contribution to that calendar year of current accident year versus all prior accident years. We will now use a historical example derived from a Progressive personal auto state comprehensive coverage to demonstrate how we can effectively answer the fundamental pricing questions. All amounts are in thousands of dollars. The first piece of data we readily have is trailing 12 calendar year incurred losses of $43.24 million. As we just stated, calendar year incurred losses can also be subdivided to show the contribution of the current accident year versus all prior accident years, as can be seen here as we see the same paid loss plus change in reserves pattern for both the current accident year and all prior accident years. We will now use our simple example to fill in values for each element of the formulas. In this example, when we isolate the contribution of this accident year to the current calendar year incurred losses of $43.24 million, we see that the current accident year incurred losses are $43.56 million. The other element of calendar year incurred losses is the contribution of prior accident years, also known as prior accident year runoff, which in this case is negative $327,000. The $43.56 million current accident year losses is what we need to start to answer the fundamental pricing questions. Unbiased ultimate accident year losses equal accident year incurred losses times loss development factor. Loss reservings' goal is to set financial reserves to be adequate with minimal variation from the date of loss until final settlement. We examine the development over time of historical accident year losses as claims are reported and settled across the columns of the loss development triangle, as can be seen in the upper right. That is known as a loss development factor. In this case, the reserves set by claims and loss reserving have historically been accurate. In our example, that loss development factor is then slightly less than 1, at 0.99. At this point, we can also bring in another piece of data that we readily have, trailing 12 calendar year earned premium of $61.14 million. Loss adjustment expenses are correlated with losses. Loss adjustment expenses can be divided into Defense and Cost Containment or DCC, which is defense, litigation and medical cost containment expenses, whether internal costs or external fees and Adjusting and Other or A&O, which is adjusting and other overhead expenses, whether internal costs or external fees. Both can change in the short term and long term depending on a number of factors, including attorney representation rate, statutory and regulatory changes, changes in efficiency in our claims organization among other possible causes but in general, tend to move with losses. In this example, we have selected LAE to be 12% of losses. A portion of our costs are mean reverting. Forecasting elements of the fundamental pricing equation deals with two distinct forms of time series. First, time series with stationarity, the graph on the left. This series tends to revert to a historical mean and requires a longer experience period to provide an effective future forecast. Weather is a prime example of where this approach is warranted. Care must be taken to decide whether such historical mean needs to be slightly adjusted going forward due to environmental changes in our future pricing period. Time series without stationarity, the graph on the right. That series does not revert to a historical mean. It is dominated by trend and seasonality. We will discuss this further when we examine frequency trend, severity trend, and premium trend. In this example, the last year contained $14 million of wind, flood, hail losses, well above our long-term expected average. Therefore, restating the long-term expectation implies a weather factor of 0.926. For Commercial Lines, an additional area where we must consider this paradigm of reversion to the mean is in the treatment of large losses. As with weather, care must be taken to decide whether such historical mean needs to be slightly adjusted going forward due to environmental changes in our future pricing period. Frequency and severity of losses change over time. What changes each over time? Essentially, this is time as a segmentation variable. It helps to separate the multiplicative components of losses as the drivers of each can be different. Frequency is the probability of having a claim, severity is the dollar amount of the claim itself. Factors that can affect frequency include vehicle technology, safety laws, product mix, for example, deductibles or tier, statements, new business growth, retention, weather, seasonality, underwriting and billing. Apart from the magnitude of the trend itself, the number of months that we need to trend is important, too. That is a function of time to price, file, get approval and elevate, policy term and rate revision length. Remember the bull's-eye slide from earlier? The precision of our estimates declines, meaning we have a greater spread, the further into the future we have to estimate. That is of particular importance in commercial auto as they have a preponderance of annual policies which increases the number of months into the future we must trend. In our example, we have estimated frequency trend to be plus 1% annually. That needs to be applied for approximately 16 months from the midpoint of the historic period to the future average accident date of our prospective pricing period. The faster we can analyze our data, the shorter we can make our policy terms and the more frequently we can elevate rate revisions, the fewer months we need to trend and the more we can reduce the spread of outcomes and increase precision in our forecasts. The process is similar for severity. Factors that can affect severity include vehicle technology, safety laws, product mix, for example, limits, state mix, medical inflation, used car values, car part inflation, body shop labor rates, weather, seasonality, claims staffing. In our example, we have estimated severity trend to be plus 7% annually. As with frequency trend, that needs to be applied for approximately 16 months from the midpoint of the historic period to the future average accident date of our prospective pricing period. The further into the future we have to estimate, the wider the spread of future possible outcomes. Progressive changes rates frequently. Remember, the fundamental pricing equation is working towards answering the fundamental pricing question of determining what we need to do to current rates. Progressive changes rates a lot. Consequently, any recent historical period of earned premium will include premium written at varying rate levels. This phenomenon is further exacerbated by the presence of varying policy terms, 6 months and 12 months. So we need to adjust this historical premium entirely to today's current rate level. In our example, we have elevated two rate decreases in the last year. That means our current rate level factor is 0.958. Controlling for the effect of product mix shifts on loss trend. A frequency or severity trend that is caused by a product mix shift will not necessarily mean our rate adequacy position, as seen by our fundamental pricing equation, will shift. For example, Progressive has been shifting to Robinsons for several years. In isolation, what would we expect to happen to frequency and severity of losses? Frequency would decrease significantly due to more lower-risk drivers in our book. Severity would increase for liability coverages due to higher purchased limits of liability. Overall, losses per exposure would decrease as the frequency decline would overwhelm the severity increase. But we would also be collecting less premium per exposure as we charge less for Robinsons per exposure. We control for the effect of product mix shifts on our frequency and severity via premium trend, which measures changes in average earned premium at current rate level over time. We must put all premium at a common rate level as only looking at changes in average earned premium over time would be confounded by Progressive's frequent rate changes. The graphs of bodily injury earned premium on this slide illustrate this. While Progressive's average earned premium per exposure has increased due to our rate increases in recent years, the average earned premium per exposure when controlling for that, i.e., at current rate level, has declined. The net effect of the rate increases and product mix shift is still to have increasing bodily injury average earned premium per exposure. And therefore, a product mix shift will not necessarily mean our rate adequacy position would shift. It would only change if we are shifting into a part of our book that has a different relative level of profitability. If we are shifting into a part of our book that is less profitable, the rate need, as indicated by the fundamental pricing equation, would go up as frequency/severity would rise more than premium, and we would need more rate. If we are shifting into a part of our book that is more profitable, the rate need, as indicated by the fundamental pricing equation, would go down as frequency/severity would rise less than premium, and we would need less rate. In our example here, in contrast to the graphs of Progressive bodily injury, annual premium trend is actually positive at plus 5%. And like frequency and severity trends, that needs to be applied for approximately 16 months from the midpoint of the historic period to the future average accident date of our prospective pricing period. What target loss plus LAE ratio would meet our underwriting profit target? First, we determine a forward-looking estimate of expense ratio. Second, we work backwards to determine our target loss ratio, which is equal to 1, minus the expense ratio minus profit. In our example, the expense ratio is 17.7% of premium, together with our profit target of 4%, our budgetary loss plus LAE ratio is 78.3%. Some elements are correlated. While we have detailed the elements of the fundamental pricing equation individually, we do not assume the correlations do not exist between elements. Some examples are losses and LAE, premium trend or product mix and frequency trend, premium trend or product mix and severity trend, LAE and severity trend, loss development factors and severity trend, premium trend or product mix and expenses via the budgetary loss and loss adjustment expense ratio. Understanding these patterns can inform our predictions of each element and improve the accuracy and precision of the fundamental pricing equation. Balancing responsiveness with stability. Remember the bull's-eye diagram from earlier; we always seek truth but observe data which is truth plus random noise. Credibility, also known as the crown jewel of casualty actuarial science, helps us ultimately deliver the best minimum variance unbiased estimate or as close to the bull's-eye pattern in the lower right-hand corner as possible and allows us to slice the cake optimally to balance responsiveness and stability and achieve forecasts that minimize both bias and variance. We want to emphasize recent data such that random noise is kept to an acceptable level. And we want to emphasize that recent data because it is closest to what we can expect in our future pricing period in terms of our book of business and the external environment. But there can be a trade-off there as that data can be thinner and noisier and we need to make adjustments to account for that noise. Credibility is a number between and including 0 and 1 that we use to weigh our data. The higher credibility is the more weight we attach to our data in the fundamental pricing equation. A sample of experience reaches full credibility, so credibility equals 1. If we have enough claims that 90% of the time our experience is within plus or minus 5% of the true value. That standard for full credibility for each coverage is determined through complex actuarial formulas and increases as the variability of experience increases. In our example, the standard for full credibility is 4,559 claims. We have over 27,000 claims in our experience period. So we have full credibility and credibility equals 1 or 100%. Our growth and scale have significantly enhanced our credibility and our ability to best react to changes in our environment. And in general, for personal auto, we have achieved that full credibility with one year of data in the overwhelming majority of our state channel coverage combinations. We now have developed the fundamental pricing equation where we can answer the first fundamental pricing question. If we don't do anything to current rates, what loss plus LAE ratio should we expect in the upcoming rate revision for the policies we are about to write? So that's our experience loss plus LAE ratio. We can also answer the second fundamental pricing question, what do we need to do to current rates in order to hit our combined ratio target in the upcoming rate revision for policies we are about to write? That's our experienced loss plus LAE ratio divided by our budgetary loss plus LAE ratio. And one final step is to weight our estimate with a complement via credibility. So the experience loss plus LAE ratio divided by the budgetary loss plus LAE ratio times credibility plus a complement of credibility times 1 minus credibility. The complement is an alternate estimate of rate need apart from the fundamental pricing equation that augments the fundamental pricing equation with an estimate of future net trend. In our example, we used a complement of credibility of plus 2.4%. The final product is what we refer to as a credibility weighted rate level indication. Here that is plus 1.3%. This concludes the theory of pricing. As you can see, it's complex with many variables and considerations and thus many ways to go astray. This is really difficult to do successfully. We have been at this for decades, and combined with the availability, quality, scope and size of our data, it is really challenging to replicate this at our scale. And to be as accurate and precise as possible when we apply this theory in practice, we have many additional considerations, which Jen Kubit will discuss in the next section. Jen?

Thank you, Brad, for explaining the theory behind the calculations to the two fundamental pricing questions. If we don't do anything to current rates, what loss plus LAE ratio should we expect in the upcoming rate revision for the policies we are about to write? And what do we need to do to current rates in order to hit our combined ratio target in the upcoming rate revision for policies we are about to write? So now let's discuss how we answer these questions at Progressive and how the answers influence customers' rates and future profitability. This ratemaking process at Progressive slide was last shown in the 2021 Fourth Quarter Investor Relations call with John Curtiss and Kanik Varma. This process is utilized in all our lines of business at Progressive. The product R&D, pricing and product management teams collaborate to determine the rate need to support our operational goal to grow as fast as we can at or below a 96 combined ratio. Additional teams join these three to deploy the final rate changes to the marketplace. Once deployed, a consumer's quoted premium reflects the revised rates. Let's focus in on the three teams that collaborate to determine our rate need. There are two broad areas of rate need, segment level and aggregate rate. In this presentation, Brad went in depth into the best actuarial science we use in pricing to evaluate the aggregate rate need in the future revision to hit the target of budgetary loss and LAE ratio. Product R&D calculates the segment rate need for the relative rate need across variables used in the product design and product managers leverage these aggregate and segment level rate needs, along with their knowledge of the local market dynamics to set the pricing strategy for their respective states. They decide if the segmentation, aggregate rate level or both need to be revised. A product manager strategy ensures that we seek to not only hit rate revision targets but also our calendar year goal to grow as fast as we can at or below a 96 combined ratio. At Progressive, we believe that our ratemaking process is effective because of the knowledge at the local level combined with the advanced and accurate view of segment level and aggregate rate need. And we strive to maximize its effectiveness by analyzing the rate need often and bringing revised rates to market quickly. In personal auto pricing, we frequently evaluate the expected loss and LAE ratio in the future revision. We complete rate level indication analyses on 51 states with each agency and direct distribution channel done individually. Also, there are 12 to 15 coverages analyzed separately. For example, the bodily injury liability and comprehensive physical damage coverages each have their own indicated aggregate rate need because the data patterns are different, such as frequency and severity trends. These analyses are completed three or four times per year and discussed with the product manager of the specific state and channel and their team. The numbers on this slide show that we're calculating the fundamental pricing equation roughly 4,000 times in a year. And each of those includes all the complex actuarial methodologies, analyses and selections that Brad discussed. So that's 4,000 times our pricing teams are analyzing how losses will develop to ultimate costs. And 4,000 times we're discussing with product managers how losses will trend into the future rate revision period. In commercial auto pricing, our dataset is smaller. However, we're still analyzing the aggregate rate need quarterly and discussing with the product management teams. We aggregate most states together to improve the credibility of our Progressive data. The largest four states are analyzed individually. Rate revision and calendar year combined ratio results for smaller states are monitored. Rates are revised at the state level because of insurance regulation. And for rate revisions, we complete a rate level indication that uses state-specific data. We evaluate our aggregate rate needs separately for our five business market targets. Differences by BMT and how losses present and how they develop and frequency and severity trends over time, affect the estimated loss ratio in the future revision. And finally, there are nine coverages analyzed separately in commercial auto rate level indications. Overall, we're analyzing and discussing at least 900 fundamental pricing equations in a year. The regular cadence of these complex analyses in both personal lines and commercial lines auto is so important in understanding the aggregate rate need and being responsive to changes in the data. But it's not enough to analyze the rate need often. To truly respond to the changing data, we need to deploy rate changes to the market quickly. This chart shows the number of rate revisions that are deployed each year over the past five years in our Personal Lines Auto and Commercial Lines auto products. We deploy many rate changes to market to adjust either the segment level, or aggregate rates, or both. And we have the capabilities at Progressive to increase the number of revisions deployed. For example, in 2023, both the personal lines and commercial lines auto rate revision teams responded quickly and often to increasing loss costs and loss development in our underlying data. These rate revision capabilities are very important to our success at Progressive. Moving quickly is important because of the time needed for some states' regulatory approval of rate changes and because of the premium earnings cadence a policy written today at the new rate level will have earned premium in each month that it's in force or for the next 6 or 12 months depending on the policy term. So it takes many months for the earned premium in the denominator of our calendar year combined ratio to fully reflect the new rate level. We are confident in our pricing teams' aggregate rate level recommendations using our robust data sets and we have an efficient process to deploy rate changes to the insurance market. However, there are times we must respond to changes and intervene in our fundamental pricing equation because historical data may not be informative. Recall that car accidents are not corn flakes. We are selling a promise to pay for claims in the future policy term. For example, changes in tariffs on imports may impact the payments on claims or loss payments in the future. Our role in the pricing teams is to answer the two fundamental pricing questions for the policies we are about to write in the upcoming rate revision. But we have no previous loss payment experience with the changes to tariffs. Consistent with the Casualty Actuarial Society's statement of principles on ratemaking, we answer these two questions with an actuarially sound estimate of the expected value of all future costs, including changes in tariffs. Calculating the expected value of the future loss payments from changes to tariffs is not straightforward because auto claims are not all the same. Broadly speaking, we split the loss payments into claims for damaged vehicles versus injuries. And these two categories can be further split into similar types of costs. Starting with damaged vehicles. If it is repairable, the loss payment includes the cost of labor and parts and materials. If the repair is more costly than the value of the vehicle less anticipated salvage recoveries, then it is declared a total loss and the claimant has paid the value. The expected value of the future loss payments from changes to tariffs is calculated at this more granular level of data. For example, the labor, parts, and materials, the value of a vehicle and salvage recoveries. This is because the costs in this granular level of data are more similar. Also, we consider the insurance coverage to the calculation of the expected value of the future loss payments from changes to tariffs. Collision and comprehensive coverage for our customer's damaged vehicle usually has a deductible for the first $500 or $1,000 but the maximum payment is unlimited. Compare this to property damage liability coverage for a claimant's damaged vehicle; it has a limit to the maximum payment. For example, $25,000 on a personal lines auto policy. For injuries to claimants or our customers, the loss payment includes the cost of medical treatment. The loss payment may also include general damages, such as pain and suffering. Again, we analyze how the changes to tariffs may affect these two different types of costs included in injury loss payments, and we consider the insurance coverage. Insurance coverage for injuries usually has a limit to the maximum payment per claim. For example, $50,000 per injured person and $100,000 for all injured claimants on a personal lines auto policy's bodily injury liability coverage. A commercial auto policy often carries significantly higher bodily injury and property damage liability coverage. For example, $1 million combined single limit. But the pricing team is not working alone. We believe our analyses are improved by collaborating across functions at Progressive. We leverage subject matter expertise to calculate the expected value of the future loss payments from changes to tariffs. The economics team within Progressive's capital management function provides interpretation on the federal government's actions and timing for implementation. They also share insights and external economic indices related to tariffs and imports. Members of the claims functions, process and control teams provide subject matter expertise on the categories of loss payments described in the previous slide to evaluate if they are impacted by tariffs. They provide the necessary granular data. The pricing team aggregates this information from the economics and claims teams to calculate the expected value of future loss payments, including changes to tariffs. We then incorporate this expected value of loss payments within the fundamental pricing equation to calculate the expected loss and LAE ratio in the upcoming revision for the policies we will write. Progressive has multiple pricing teams within the personal lines and commercial lines functions. Personal Lines is also split by auto, home and special lines. In addition to collaborating with economics and claims, we also collaborate among the pricing groups to share and debate methodologies and assumptions. Despite all the best actuarial methods for evaluating the aggregate rate need, communicating with product managers and deploying quickly to market, predicting the future is impossible. So our initial expected value of future loss payments, including changes in tariffs, will be wrong. The fundamental pricing equation requires us to estimate future loss payments and other costs. All our future predictions are wrong to some degree, and the probability of being wrong increases with changes that have minimal historic precedent, such as inflationary impacts when an economy emerges from a pandemic or the frequently changing impacts of tariffs. Pricing, claims and economics teams are working together to monitor the actual change in loss payments from implemented changes to tariffs. The scale and quality of our data at Progressive allow us to quickly identify the change in the actual loss payments and compare it to our expected value. This leads to refining our calculations. Our robust dataset at Progressive and the expertise of multiple teams contribute to our speed through this iterative cycle of estimate, monitor, and refine. We are moving quickly to the center of the bull's-eye from Brad's slides, where our expected value of future loss payments from changes in tariffs are accurate and precise. Producing the rate-level indications often provides the updated aggregate rate level advice to product managers so they can recognize the differences in the refined expected value of the future loss payments. And we have the capability to deploy necessary rate changes quickly. The speed of estimating, monitoring, and refining our rate level indications and deploying the right rate change is important to deliver on our profit targets. And we want to write more business when we have accurate rates relative to the expected costs. Changes in tariffs are not our first intervention to the fundamental pricing equation. Our combined ratio results show that we have moved quicker than the industry to recognize increasing costs, raise rates, and deliver on our operational goal of at or below a 96 combined ratio. In Personal Auto, the standard deviation of our annual combined ratios in this 11 years from 2014 through 2024 is half of the average standard deviation from the other top 10 carriers. So we are responding quicker and making smart interventions.

This concludes the previously recorded portion of today's event. We now have members of our management team available live to answer questions, including presenters Brad Granger and Jen Kubit, who are available to address inquiries about the presentation.

Yes. Just first question on quote volume growth. I was just looking at quote volume growth and it looks like direct quote volume increases have really taken off while agency quote volume has not seen exactly the same acceleration. So I'm curious if the actions you're taking in the property book are limiting agency quote volume? And would you expect to see a tailwind in agency quote volume as you wind down those actions in the property book?

Thanks, Rob. Our direct volume growth is a result of increased advertising. When considering property quotes, there is a distinction between agency and direct channels. Through an agency, you have a specific offering with Progressive, namely Progressive home. In contrast, our direct channel allows us to collaborate with various unaffiliated partners, enabling us to meet their appetite needs even if a particular quote doesn’t align with ours. Given the improvements we've made over the last few years, we anticipate positive momentum in both agency and direct channels, as our property portfolio is now positioned in less volatile areas and our combined ratio looks favorable.

That's helpful. And then I just wanted to follow up on Florida. Can you help us think through the potential size of the Florida refund related to the excess profitability? And how are you thinking about pricing moving forward in Florida given where the profits are?

I'll address the latter part of the question first. We have lowered rates in Florida twice over the past year, by 8% in December and another 6% in June. Our commitment to Florida is strong as we are the leading insurer in the state. The 2023 insurance reform, House Bill 837, gave us hope for reduced loss costs, and it has indeed made a positive impact. I want to highlight that these reforms have been beneficial for Floridians, and I commend Commissioner Yaworsky and Governor DeSantis for their efforts in this regard. Florida has an excess profits statute that operates on a rolling three-year basis, covering 2023, 2024, and 2025. Without data from half of 2025, I can't provide a precise estimate, especially with hurricane season approaching. However, if our profits during these periods exceed the statutory limit, we will comply with the regulations and return that money to policyholders. We're monitoring this closely and have our internal estimates, but they could vary significantly depending on hurricane season.

Looking at your 10-Q, you talked about policy life expectancy for personal auto declined 5% due in part to business mix shift. Is this the same business mix you talked about in the other parts of the 10-Q where you're shifting towards the Robinsons and the Wrights? Intuitively, I thought those should have higher policy life expectancy. Can you help us think about that?

Yes. The mix shift has changed dramatically because of what happened with inflation in 2023. And so we closed down our underwriting appetite and brought in a lot more preferred business mix. Since we've opened up, as you can see, especially in both channels, we are writing a lot more Sams, which is lower policy life expectancy. We expect that. We know that. We have a history of that as long as we make our profit target margins on those Sams, it is great. So a couple of different things on auto policy life expectancy that I would talk about. This is probably a little bit redundant from last quarter, but there's a lot of shopping going on. So in a hard market, that's going to happen. Is that necessarily bad? No, because if the price isn't right in our book and people shopping can get a lower price, we believe, as you just heard from both Brad and Jen, that we price pretty accurately and we have a rate revision machine. So it could be adverse selection. It could be going to someone who hasn't got the right price on the street. And secondly, the mix shift that you're seeing in the policy life expectancies is with Sams. So it's compared to the base of the preferred business we've put on the books a few years ago. And then lastly, when people do shop, we talked about this last quarter, we will look at their policy, do a policy review, look at build plans and different things that could cause us to rewrite with us starting the clock ticking over. So although we keep that customer, you'll see the decline in policy life expectancy. I have reasons to believe that, that will start to turn around. But again, that we'll see as the data comes out. I would point you to, and we're not going to actually share this all the time, but we have an internal measure of household life expectancy, which gives a 30-day ability to rewrite and our household life expectancy is up. So that's kind of my reason to believe that PLE could follow. Again, a lot of PLE has to do with mix, how long the hard market continues or the shopping behavior of consumers, which could maybe dramatically change in this last 2 or 3 years based on all the inflationary measures. So a lot of data going into that but hopefully, that gives you a little bit more color.

Great. Second question is on tariff. Again, this is something that I think the right way to think about it is it introduces uncertainty. But maybe on the personal auto side, if we remove the tariff as a headwind, is it fair to say that you should be able to grow much more aggressively or reduce your pricing significantly? Is that essentially the only thing that kind of keeps you away from reducing pricing further?

We aimed to be cautious due to the uncertainty surrounding tariffs. However, with each passing day, we gain more clarity on the situation. Our approach involves closely examining every state, both for new and renewal policies, assessing various products, evaluating margins, and considering our growth potential. We strive to expand as quickly as possible while maintaining a 96 margin. In states requiring small rate increases, we will adjust accordingly, while in locations with strong growth potential, like Florida, we may lower rates. We aim to maintain a balanced approach that keeps our rates both stable and competitive for our customers, which is a fundamental aspect of our strategy. If we see opportunities for growth while holding appropriate margins and gaining more certainty, we will act to grow and do so profitably.

My first question is about personal auto. We've noticed that new business is increasing while policy life expectancy is decreasing, which I think you've addressed. Looking ahead, how do you anticipate policies in force growth to trend considering these factors, especially since you've mentioned a slight rate decline in personal auto this quarter and are still boosting ad spend? I'm trying to gain insight into the expected growth in policies in force with all these elements in mind.

Yes. I mean it's hard to compare 2025, which has been incredible already to 2024, which was the best year in the history of Progressive. But the fact is we grew over 5 million policies in force year-over-year and 1 million in personal lines just in this quarter. So we believe there's an opportunity to continue to grow. We believe we're in a really great position. And I think where we feel like we're in even a better position is to now grow that Robinson book. So we feel like we're in a different position in our property and have a lot of plans to continue to work out the blueprint and ultimately open up a bit. And because we have all those auto policies, we have those future Robinsons or Robinsons that have an auto and home but not home with us. So I think the opportunity really lends itself to grow more preferred. We have a lot of opportunity in that area and that will be our focus. Again, we'll be strategic. We're not going to swing the pendulum the other way. We've certainly learned a lot about the property book and the volatility across the country in the last 5 to 10 years. But we're really well positioned. And the fact that we have all that auto business on the book, I think, is really important. There's a lot of market share for us to capture. So I remain bullish. Comparisons are more difficult when you're comparing to the best year in the history of Progressive.

I have a second question, Tricia. In your letter, you mentioned comments on capital and how holding capital can negatively impact returns. Could you expand on that? Are you considering additional capital returns, knowing there's a balance between using capital for growth? Does this relate to your typical special dividend later in the year, or are you also contemplating incremental share repurchases? I'm looking to clarify that comment a bit more.

Yes. We require significant capital for growth. We have a regulatory base, as well as a contingency layer and extreme contingency. We continue to assess our needs in this regard. There are three primary methods for returning capital. Our preferred method is to invest in business growth, which we have been doing consistently. Additionally, we repurchase shares to mitigate dilution from stock-based compensation, and we plan to buy back more stock if it falls below our intrinsic value. We continuously evaluate this strategy. Usually, in December, the Board of Directors decides, if applicable, on a variable dividend. Jon Bauer, John Sauerland, and I have begun outlining our approach for presenting this to the Board. Ultimately, the decision rests with them. Given that we still have significant time left in the year, particularly with upcoming storm seasons, this could present another opportunity if the Board chooses to return capital via a variable dividend.

I wanted to follow up on the discussions around policy life expectancy and the numerous Sams. I'm curious if you haven’t made any specific changes to enhance the company's policy life expectancy and just allowed the additional Sams to mature naturally over their regular timeline. Would retention improve on its own over time? If that's the case, when can we expect to see that shift, especially with all the business added in 2024?

I hope I understood your question. I want to mention that Sams have always had a lower policy life expectancy, and we actually further segment different Sams, though I won't go into all the details. If Sams left because they were shopping around, our retention would indeed improve. However, we prefer to keep as many Sams as possible since they are a vital part of our growth. We value having Sams on the books as long as we achieve our calendar year and lifetime profit targets. Did I understand your question, Josh?

Yes, I'm just wondering if in 2024, you might capture a larger share of the Sams market than usual, and your business mix is more focused on Sams. This could explain why our retention is down. However, if we return to a more typical mix of Sams, Dianes, and Robinsons, it seems that policy life expectancy would naturally increase since your new customer acquisition wouldn't be overly concentrated on Sams. Is that an accurate way to interpret it?

Yes, that's a good way of thinking about it. However, a lot can happen. I mentioned earlier in response to a question about our desire to grow Robinson. It really depends on the other segments we introduce, the rates from our competitors, and how our customers respond. Additionally, there is some noise in our data, likely from customers shopping around but still staying with their current provider, which may reset the timeline for those customers without them actually leaving; we just conducted a policy review. So, while your analysis suggests a stable scenario, I don't think all factors will remain stable.

You're currently very profitable and investing significantly in advertising. If those ads are attracting many new customers, is the spending on short-term or low-value policies justified? Are you anticipating that these policies will only last for six or twelve months, and is everything proceeding as expected?

Yes. We will not acquire any customers in our segments unless we believe it meets our target profit margins. As long as our cost per sale is under pressure, which it is, we will continue to invest and grow in every feasible segment. Some of the initial customers will eventually become long-term clients. They may start with renter's policies, which we are expanding, and eventually purchase homes, and the chances of them staying with Progressive for their home insurance are high. Therefore, we aim to attract all potential customers, but we evaluate that based on the targeted acquisition cost for each segment.

I'd love to hear more about Sams turning into Robinsons in the future. So we'll stay tuned.

I'll do that at some point, Josh. I worked on something called Imagine Diane, which I believe we started considering back in 2014. That's one of the reasons we began exploring the concept of different product horizons that our customers need. John Sauerland delivered a compelling investor relations presentation, emphasizing that we want customers to approach us and ask if we have specific products, like life insurance or jewelry insurance. We want to be able to respond affirmatively, even if not all products are under our brand, to provide a comprehensive portfolio that adapts to each customer's changing insurable needs. We can revisit this concept because certain aspects may have evolved. We often think about how someone like Diane starts as a renter, then gets engaged, requiring insurance for her ring, and so on. We can even estimate the expected lifespan of such policies. I'll definitely add this as a topic for a deeper discussion in the future. Thanks, Josh.

I'd like to return to the pricing theory section of the presentation. I understand it was theoretical, but you mentioned a factor of 12% for LAE. I'm interested in that LAE number and would appreciate some insight into how LAE has trended for your business over the past couple of years. More importantly, I'm curious about the technology you might be using to reduce those costs in relation to earned premium. Could you discuss the potential for further improvement in LAE as we look ahead to the next 24 to 36 months?

I'll begin addressing the question, and then Brad can provide additional insights. Our Loss Adjustment Expense (LAE) has consistently decreased over the past 10 to 15 years. In this quarter, we noted that our Non-Acquisition Expense Ratio (NAER), which excludes acquisition costs, declined by approximately 0.3%. We will keep working to reduce costs overall, though we will not cut back on acquisition costs as we want to invest in growth. My team and I regularly discuss different technology, process, and personnel changes that can help us further decrease our expense ratio across the board, which is crucial for maintaining competitive pricing for our customers. I am pleased with the reductions we've seen not just in LAE but in our overall NAER over the last decade, and we will remain focused on this area. We believe there is significant potential for improvement, particularly through technology. The 12% figure Brad referenced was an example, and I'll let him explain that further.

Yes. Thanks, Tricia. Yes, the 12% was an example. It's actually considerably lower when you measure the cost of LAE in relation to premium. But to add to what Tricia said, we also are very careful to both look at his recent historic LAE performance but also to take a future forward-looking view of it to ensure that we are ahead of the curve for any changes, any efficiencies that the business creates.

Okay. I guess the other question I had, just as I was listening to the presentation and your comments about the rate cuts in Florida brought up a concept, and I'm not sure it's valid, so I thought I'd ask you for your opinion. Normally, when you get to price increases because of inflation and other factors, that can be disruptive to your retention ratios. And I'm curious if price decreases can also be disruptive to retention, triggering shopping. I'm curious about your perspectives on that.

I believe that shopping behavior has been quite unstable in recent years due to various changes, including the significant inflation we experienced in 2023 and prior to that. Generally, a drop in prices doesn't automatically lead to increased shopping, although there are also many external influences like advertising. Therefore, I wouldn't make that assumption. There are numerous factors at play, particularly the pace of changes within the industry. As Jen mentioned, our goal is to respond promptly to provide accurate rates. When prices decrease, it often coincides with new business growth, which can vary based on whether individuals are comparing offers from different carriers or reviewing their own policies to see if they can adjust them for lower rates.

So just first had a question on how you view the competitive environment to be in personal auto. And it seems like everybody's margins have improved, and in most cases, at the sort of upper end of their historical ranges. And more and more companies are talking about wanting to grow as opposed to improve margins. Just wondering if you've seen that in the competitive behavior overall and how that affects your view of margins and growth prospectively for Progressive.

Yes. We definitely have seen the environment become more and more competitive. And we are thankful to get out ahead of the rates. And so we've been able to put on the amount of growth we put on in terms of policies in force. Like I said, comparisons will be difficult because you were comparing incredible numbers in 2024. And frankly, incredible numbers in the first half of 2025. That said, that's our sweet spot. We love that. We love the competitiveness. It's great for consumers. It's great just making all of us better. So our goal will remain to grow as fast as we can and make our target profit margin. We are doing great on both right now. I think I started in my letter talking about that in net premium written as our trifecta, and we'll continue to do that. It's going to get more competitive. But again, that's what makes this business fun.

Okay. On a different topic, what are your thoughts on the gradual shift towards more autonomous vehicles and the increasing use of technology such as sensors and cameras? How do you see this impacting the total addressable market and the overall market opportunity for personal auto companies in the next 5 to 10 years?

Yes, we are currently engaged in that assessment, which we've been doing for the past 10 to 15 years. The first time we addressed it during an Investor Relations call was in 2013, followed by another discussion in 2017. We consistently evaluate our understanding of uncertainty, considering conservative and pessimistic scenarios as well as a middle ground. Our most pessimistic estimates were significantly below actual figures; we did not anticipate the addressable market would expand at its current pace. We are currently revising our projections related to our addressable market. Vehicle safety is an important factor, and the advancements in ADAS systems over the past few decades have indeed reduced the frequency of incidents. Although we've seen a decline in frequency over the past 50 to 60 years, it’s often counterbalanced by severity. Consumers take time to accept the cost implications. Additionally, vehicle lifespan has increased to about 13 years, which means that even with the advancements in safety, it will take some time for the fleet to fully integrate these improvements. We believe that safer cars are important for society, and we strongly support this progress. However, it does take time for severity trends to balance out frequency trends. Overall, we see significant opportunities in the addressable market over the next 5 to 10 years, especially as we diversify across all three horizons. I've previously mentioned our potential to offer bundled services, such as our auto and home packages, which we refer to as Robinsons.

Tricia, I'm interested in discussing the policy life expectancies and their retention. Could you explain how much of the retention is influenced by the mix dynamics versus competition? I would appreciate it if you could share how retention varies by customer segment to help isolate the mix dynamics.

We don't usually disclose that information. Our customers are shopping with us more than ever before, though I think that's starting to slow down a bit due to increased competition. The mix will always play a role in this. At some point, we could provide some indexing to illustrate the differences. Regarding Josh's question about the Sams and future Robinsons, we could incorporate that into the discussion, as it's relevant to the overall formula. However, we generally don't go into all the details because each of our segments has various ways we analyze them based on the channel, such as whether it's auto or home with an unaffiliated partner versus us. There is a lot of detailed information to consider.

Got it. No. Okay. That's helpful. And then maybe if I could just follow up. Just on the frequency. So that continues to come in pretty favorable. It's been almost two years now that we've been in negative territory. Could you just talk about if you're seeing any dynamic? It looks like first-party collision claims are down by more than property damage. Are you seeing any impact just from customers that might not be filing claims and sort of with the deductible, just sort of eating the claims? And I guess I'm trying to just isolate how much is maybe that dynamic versus just greater penetration of ADAS and some of these collision avoidance systems that might also be putting downward pressure on frequency.

Yes. I think collision and property damage are always challenging due to timing, subrogation, and the varying amounts we pay. Jen touched on that a bit. Most of our differences are due to changes in our mix and the ongoing vehicle miles traveled. We are monitoring this closely and continue to observe these trends. This is where we have seen the most significant decrease in frequency.

First, I wanted to ask on the 16-month trend time that you guys used when thinking about the various factors and the pricing adequacy model. How has that time frame shifted coming out of the pandemic? Are you guys assuming slightly more months? Was it potentially higher in the past than it is now? And how might you expect that to trend going forward?

Yes. I will let Brad or Jen add to that. We aim to be flexible based on the current circumstances and needs. If you have anything to add, please feel free.

Yes. Thanks, Tricia. I don't think it's changed that much. It's a function of, as we talked about, how far back you're looking in your data period. So our growth helps there. If we only have to look back one year instead of two years, you don't have to trend as far in the history to today. Also, as we pointed out, too, it's also a factor of how quickly you can do rate revisions, get them priced, get them filed, get them approved, but also how many you can do. So the fact that we are able to increase our rate revision capacity quickly if we need to helps to also reduce the number of months we have to trend into the future.

That's helpful. And then perhaps as a follow-up, I know we spoke about Florida earlier in Q&A, but are there any other states where you may potentially have profits that are exceeding statutory limits and might have to consider issuing a refund to policyholders?

No, I don't think there's any other excess profit statutes other than Florida.

In the letter, Tricia, you mentioned that the 8.9 product model and about 50% of your premiums is demonstrating favorable conversion results and elasticity. Any willingness to kind of unpack that a bit? I don't feel like we're seeing it in some of the 10-Q KPIs, although some of them do have tough comps. That's my first question.

Yes. I mean, I think every product model we look at, we have is very specific segmentation schematic in that and to grow, especially in the preferred business. Pat, if you want to comment on 8.9 in particular because we're already on 9.0, thinking about 9.0 and our other ones. And it takes some while to, I think, build into it.

Yes. The only thing that I would add is there's a lot of moving parts as we elevate product models into states and we are taking rates either up or down when we simultaneously elevate the segmentation. So while we look in aggregate at what the contribution happens pre/post any model change, there's a lot of moving parts in the market as well. But as Tricia mentioned, we recently elevated 9.0 in our first state and continue to bring new segmentation that we think better matches rate to risk to market and creates that adverse selection and more competitive prices for more consumers.

Yes. Lastly, we have our 5.0 property product model representing about 75% of our net written premium across approximately 29 states. This, combined with our ongoing segmentation in the auto sector, allows us to attract bundled customers. Given our current market share and the number of auto policies, we feel that we are just beginning.

That's helpful. My last follow-up is about understanding shopping levels. There were different comments made earlier in this call. I want to know if shopping levels are still significantly above what Progressive would consider the normal long-term trend, or have they approached or dropped closer to normal due to the softer pricing environment?

I would say shopping levels are still high. Ambient shopping remains elevated, and we still view this as a challenging market because people continue to shop. We discussed this a bit regarding our policy life expectancy. But yes, for now, shopping remains high.

And just would shopping normalize just as long as incomes keep increasing and auto rates don't go back up to high singles? Is that how we should think about it, like it takes a couple of years to normalize?

Typically, it has in the past. I don't know. It's easy to shop. So I don't know if the pandemic and subsequent events has forever changed shopping behavior. That will unfold as this next couple of years unfolds, but that has been how it typically has worked in markets in the past.

That appears to have been our final question. So that concludes our event. I'll hand the call back over to you for the closing scripts.