At the 1996 Berkshire Hathaway shareholder meeting, Warren Buffett and Charlie Munger were asked about GEICO (at the time, the seventh largest U.S. auto insurer). While Berkshire owned ~50% of GEICO by the mid-1990’s (for which they’d paid $45.7 million, with a lot of help from share repurchases over the years), a January 1996 deal gave them control of the other 49% - at a cost of $2.3 billion. As Buffett wrote in 1995, “That is a steep price.”

The question at the 1996 meeting related to that “steep price”: what led Buffett and Munger to pursue the GEICO deal, and how would the auto insurer impact the growth / value of Berkshire’s broader insurance business over time? In response to that question, Buffett said the following:

“We paid a good price, but GEICO is a terrific company. It has outstanding management. It has a low-cost method of distribution, which is very difficult [to compete with]; everybody wants to have that, but very few come close. GEICO management is focused on bringing costs down further and widening that competitive moat. I personally think GEICO’s growth rate is likely to be greater in the future than where it’s been in the past… There are some advantages to being part of Berkshire; there are costs to bringing new business on the books, and we don’t care at all about quarterly earnings. GEICO was relatively insensitive to those concerns before… but they had more pressure on them in respect to reported earnings than they will have as part of Berkshire. And I think there’s some really big opportunities, in terms of what can be done as part of Berkshire. So, I think five years from now, you’ll be very happy that we own 100% of GEICO… As marvelous as GEICO was independently, it will flourish even more as part of Berkshire.”

As he so often does, Buffett captured what would be critically important for this business over the next 10-20 years in a few sentences. Specifically, there are two comments about GEICO that stand out to me: (1) it’s low-cost method of distribution; and (2) the removal of incentives that likely impacted management’s decision-making when GEICO was a publicly traded company.

On the first point, there are a few important characteristics about the auto insurance industry: (1) it’s a relatively expensive product (premiums per average policy were ~$1,250 per year in 1995); (2) it’s a mandatory purchase (you have to have auto insurance to own / drive a car); and (3) it’s an offering where customers do not generally assume a meaningful difference in product or service quality among the established competitors like State Farm, GEICO, Progressive, etc. (Tegus interview with a former Marketing Director at State Farm: “It moved from a value-driven category to a price-driven category”)

As Buffett discussed at the 2004 meeting, those conditions informed the strategy that would be most effective in this industry: “Saving significant money makes a real difference in a lot of household budgets, so low cost is going to win… It’s a better system, and better systems win over time… We have a very tough competitor in Progressive because they’ve seen how well our model works, and they, in effect, shifted over [from an agency model to a direct model]… It’s always a good idea to go with a low-cost producer over time. You can mess it up in other ways, but being a low-cost producer of something essential to people is usually going to be a very good business.”

The following chart shows GEICO’s underwriting expenses as a percentage of earned premiums. As you can see, GEICO’s cost structure has significantly improved over the past 25 years (that wasn’t the case in the late 1990’s and the early 2000’s, which I’ll discuss in a moment). As Buffett noted at the 1996 meeting, sustaining and improving GEICO’s low-cost position was key to widening its moat. Given that GEICO’s customer base (voluntary auto policies in force, or PIF’s) has grown by >7x from 1995 to 2022, it’s clear that this has been a successful strategy that resonated with consumers.

This brings us to the second noteworthy point from Buffett’s 1996 comments (“there are costs attached to bringing new business on the books, and we care not at all about reported quarterly earnings”). Growing the business came with two headwinds for GEICO: first, they have to spend on marketing to attract new customers (expensed immediately through the P&L); second, once they add a new policyholder, it takes some time before reaching the level of profitability that is expected for seasoned customers (as Buffett mentioned at the 2017 shareholder meeting, loss ratios in year one for a new policyholder are nearly ten points higher than on renewals). When GEICO was a public company, even if they knew the next cohort of customers would deliver attractive long-term economics, this was a constraint given its short-term impact on earnings (EPS). As a subsidiary of Berkshire, GEICO was no longer constrained by this short-term financial consideration.

This impacted the business in two ways: First, it led to a significant increase in marketing spend, which explains the heightened underwriting expense ratios during the late 1990’s mentioned earlier; as Buffett disclosed in the 1999 letter, marketing spend was $242 million that year – nearly 7x higher than the $33 million GEICO allocated to marketing in 1995. Second, and more importantly for the long-term economics, this spend led to a significant acceleration in customer (PIF) growth rates: as shown below, after adding ~150,000 net new customers on average in 1994 and 1995 (before the deal), GEICO added nearly 500,000 net new auto PIF’s annually, on average, throughout the end of the decade (the five-year period from 1996 to 2000).

As you look at the above chart, you’ll notice that the results in 2000 were much weaker than in the years preceding it. As Buffett discussed in that year’s shareholder letter, this reflected a number of issues, including less effective marketing spend (which drove higher policy acquisition costs), ongoing societal considerations (older generations were more reluctant to deal with direct insurers), and some noteworthy decisions by competitors (specifically State Farm, which was “very slow” to adjust its auto pricing in the face of rising loss rates). What’s most interesting about the 2000 letter isn’t Buffett’s explanation of GEICO’s current challenges (which he spoke about directly); instead, it was his thoughts on the structural considerations in the business / industry that would be the long-term wind in GEICO’s sails:

“GEICO will be a huge part of Berkshire’s future. Because of its rock-bottom operating costs, it offers a great many Americans the cheapest way to purchase a high-ticket product that they must buy. The company then couples this bargain with service that consistently ranks high in independent surveys. That’s a combination inevitably producing growth and profitability.”

As we look to the long-term results, that shines through: in addition to significant earned premiums growth, GEICO has generated ~$20 billion of cumulative pre-tax underwriting income over the past 25 years.

The Progressive Problem

While that “inevitability” was on display for the better part of the next two decades, GEICO now finds itself playing catch-up to an aforementioned competitor: Progressive. To put it bluntly, Progressive has ran laps around GEICO over the past few years. As shown below, GEICO had 4.5 million more auto PIF’s than PGR at the end of 2017. That advantage has completely disappeared over the past five years: Progressive now has a larger auto insurance business than GEICO (measured on PIF’s).

The problem doesn’t end there. While the pandemic has had a significant impact on the annual figures, the chart below shows that GEICO’s loss & LAE (the costs associated with covering insurance losses) has averaged ~82% of earned premiums over the past five years, inclusive of a ~93% loss & LAE rate in 2022. While GEICO has largely protected underwriting profitability by continually reducing underwriting expenses (as a percentage of premiums), it’s clear that they have work to do on the cost structure: In 2022, PGR’s Personal Lines business, with Auto accounting for >90% of written premiums in Personal Lines, reported a +4% underwriting margin - 900 basis points better than the -5% underwriting margin reported at GEICO. (2022 10-K: “We currently expect GEICO to generate an underwriting profit in 2023.”)

Simply put, GEICO has been handily beaten by Progressive over the past handful of years on volumes and underwriting / profitability. This reality hasn’t been lost on shareholders or Berkshire management (it has been a frequent topic of discussion at the annual meetings). This comment, from Buffett at the 2012 meeting, is noteworthy; in hindsight, I bet he’d agree that he overstated GEICO’s competitive position, at least relative to Progressive:

“Progressive has probably been the leader in what you described [telematics]. We have not done that at GEICO, but if we think there becomes a superior way of evaluating the likelihood of anybody having an accident [we’ll do it]… Obviously, if you could ride in a car with somebody for six months, you might learn a bit about the propensity [of having an accident], particularly if they didn’t know you were there… But I do not see that as being a major change … We’re always looking for more things that will tell us someone’s likelihood of having an accident in the next year… but I don’t think that this new experiment threatens GEICO… Our marketing, risk selection, and retention are working extremely well. GEICO is a machine.”

Buffett took it a step further at the 2013 meeting: