AGB 2022.12 - Verisk Analytics (VRSK)
Refocusing on the Core Insurance Vertical
Subscribe to AGB - One analysis of a good business every three weeks.
Verisk Analytics
Verisk reminds me of Constellation Brands in that there is a prized asset that the company has owned since inception, which generates a lot of excess cash flow. For Verisk, it’s the core insurance business built on the consortium data model. For Constellation Brands, it’s the perpetual import rights for the Grupo Modelo beer brands into the U.S.
However, the management teams have a questionable track record for capital allocation. Excess cash flow has been used for making acquisitions that are lower in quality. For Verisk, the list consists of the mortgage, healthcare, financial services and energy acquisitions. For Constellation Brands, it’s the wine, craft beer acquisitions and the Canopy Growth investments. Even with this drag, these companies have grown their intrinsic values over the years because the prized assets generate such high returns.
The difference going forward is that Verisk has decided to sell the non-insurance businesses (partially due to the urging of an activist investor) to refocus on the core insurance vertical. This could mean a more consistent return + growth profile and higher margins. Furthermore, Lee Shavel was instated as CEO in May 2022 (he joined Verisk as the CFO in November 2017), and he seems to be more returns focused than his predecessor.
Verisk is the leading data and analytics solutions provider to Property and Casualty (P&C) insurance industry. The company was founded by a collective of the industry in 1971 as a non-profit advisory and rating organization that became the central repository for industry data. Members of the P&C insurance industry (now all customers of Verisk) initially leveraged the data to define coverages, issue policies and price risk. As Verisk became an independent for-profit organization, the company expanded its data capabilities into claims processing, fraud detection, catastrophe modeling, etc. through a series of acquisitions.
The company’s P&C insurance data is gathered in consortium model, meaning Verisk’s customers provide their own individual data and then the company sells access to the collective data back to its customers. In 2021, customers of Verisk submitted 2.6B detailed data points of insurance transactions such as premiums and losses. Then the company scrubs the data to increase the reliability and accuracy of the data set. Verisk’s database has 29.7B statistical records (8.2B commercial and 21.5B personal) and details on over 1.5B claims.
It's almost natural to view Verisk as a utility for P&C insurance given how reliant the industry is on the company’s data and services. As the company explains it, the consortium model makes sense for the P&C insurance industry because of (1) the regulatory complexity in the U.S., (2) the high number of data points that need to be considered when pricing risk and adjudicating a claim, and (3) the fragmented industry structure.
In a typical claims process, the insurance company will take all the pertinent information related to the claim and send the information over to Verisk to analyze in real time. Verisk matches the information against similar claims in the database to determine the likelihood of fraud occurring. Not only does the process help prevent fraudulent claims, which are a direct cost that insurance companies want to minimize, but it also helps facilitate a speedy payout of a legitimate claim, which is also more cost efficient.
Verisk also helps its customers meet compliance requirements with regulators. The company’s lawyers and experts on staff review any changes to state insurance rules and regulations. This process can be cumbersome for a smaller insurance customer to do in-house. P&C insurance providers are required to report statistical data about premiums and losses to their state. As a “statistical agent” in all states and territories, Verisk aggregates the statistical data and reports them to regulators on behalf of their customers.
As mentioned above, Verisk has generated substantial excess cash flow from its core insurance business but has historically reinvested into data repositories for other industries like mortgage, healthcare, financial services (5% of revenues in 2021) and energy (22% of revenues). While it’s not obviously wrong for Verisk to assume that the success from the insurance business could be replicated into other industries, history has proven that the quality of these data assets weren’t up to the level of the insurance business.
The company exited the mortgage business in 2014 and the healthcare business in 2016. The company recently sold the financial services business and part of the energy business in 2022. The remaining energy business is still under strategic review and will likely be sold or spun out later this year.
Verisk can call most of the top 800 P&C insurance providers in the U.S. as customers with a penetration rate of 97%-98%. But that doesn’t mean that the insurance business lacks growth opportunities. Revenue growth comes from added services and new data subscription offerings as well as annual price increases through escalators built into service contracts that average 3-5 years in length.
Organic revenue growth consistently has been in the 6%-8% range for Verisk. In recent years, the company’s growth has fallen below that range due to pockets of weakness in energy and declining revenues in financial services. With plans to divest the remainder of the energy business and a refocused effort to growing the core insurance business, the company is expecting margins to improve by 3%-5% by 2024 off of the baseline of 50%-51% adjusted EBITDA margins.
Why is it a good business?
Verisk benefits from intangible assets and its customers’ increasing switching costs. The company has been the industry standard for data when it comes to pricing and rating risk since its founding. And because of the consortium model, the data set that is collected from the individual P&C insurance customers grows each day. Verisk has over 40 petabytes of data that is unique to the company.
Verisk’s customers have high switching costs because there is no viable alternative to the consortium data. Even if a large P&C insurance provider were to try to replicate the data independently, because of the fragmented nature of the U.S. P&C insurance industry, the company would only have access a mid single digit percentage of the data generated going forward. According to the National Association of Insurance Companies, in 2021 the top 5 U.S. and Canada based P&C insurers commanded just over 31% of the direct written premiums (DWP) and the top 25 commanded just over 66%. Furthermore, this P&C insurance provider would have to do more of the analytics and service work in-house like reporting statistical data to state regulators.
As Verisk develops more software offerings for the insurance industry, there is a high likelihood of successfully selling into their existing customer base, which also increases customer switching costs.
Verisk only represents ~25bps of the industry’s DWP. This number has moved higher from just 10bps in 2005 because the company’s revenues have outpaced the industry’s DWP growth by a factor of almost 2x. At such a low cost, it doesn’t make much financial sense for Verisk’s customers to try to circumvent Verisk’s data offerings.
And lastly, Verisk has a high portion of its revenues under 3-5 year subscription arrangements. Customers often pay annually in advance of services rendered, which helps with the company’s cash flow margin. With 81% of the company’s revenues under subscription agreements in 2021 (this includes financial services and energy) and the mission critical nature of the data for customers, Verisk is also recession resistant. During the great financial crisis, Verisk still reported 5% organic growth in 2009.
Returns on incremental capital?
Over the past 10 years, Verisk has spent 22% of its capital on capex, 4% on R&D and the remaining 73% on M&A, offset by 11% sales of assets. Capex is spent on internally developed software (similar to IQVIA, this expense is categorized as capex and not R&D) and hardware for its hosted data and software products. Internally developed software expenses have contributed to most of recent the capex increase as the company has increasingly needed to capture and deliver data to its customers through software networks.
The company’s cloud migration has contributed to Verisk’s increasing capex intensity in recent years. Verisk started its cloud migration in 2019 from mainframes to AWS. From a cost perspective, the move lowers capex, reduces operational complexity and allows the company to eventually shut down its two data centers. Other benefits include (1) quicker product development cycles, (2) faster client onboarding, and (3) the ability for Verisk to expand to new geographies with lower upfront capex investment. The company is more than 80% complete with its cloud migration.
On the acquisition front, Verisk bought 41 companies from 2002 to 2015 for a total of $4.9B or an average deal size of $120M. Many of these acquisitions were in adjacent verticals like mortgage, healthcare, financial services and energy. As mentioned previously, these segments have all been divested except for energy, which should be divested in the near future. Per the company’s IRR calculations, these deals generated returns of 19% for the larger deals (greater than $100M) and 12% for the smaller ones (less than $100M). Just as a caveat, this IRR analysis was done in 2018. With the declining revenues in the financial services business prior to its sale, these IRR calculations now should be lower.
The larger deals from 2015 until 2018 did not work out well for the company. In the same analysis, Verisk estimated that these deals totaling $3.8B in capital only returned 4%. Included in that calculation were the acquisitions of Wood Mackenzie, Sequel, PowerAdvocate, G2 and LCI but most of the underperformance was from Wood Mackenzie. The company explained that this had to do with lower commodity prices and negative currency impacts.
While some of that is true (energy companies have capital cycles of their own depending on commodity prices), Verisk likely overpaid for the asset (close to near-term peak earnings) and mispriced how integral the data from Wood Mackenzie was for their customers’ businesses. In 2015, the company paid 17.3x EBITDA in a competitive bid to acquire Wood Mackenzie for $2.8B. In 2022, part of the energy business was sold (3E sold for $950M with earn outs) and the remainder should be divested shortly. Some analysts are estimating a ~$3B sale price for the remaining energy business.
There is also some evidence that Verisk mismanaged Wood Mackenzie after the acquisition. When current CEO, Lee Shavel, joined his first earnings conference call in 1Q 2018 as the company’s new CFO, he went over the EBITDA margin differential pre- and post-acquisition for Wood Mackenzie, which had declined from 47% to 27% in short order. Part of the analysis pointed to a 3% impact from Verisk allocating overhead costs to Wood Mackenzie, which shows that Verisk added costs on top of the asset without finding sufficient operational efficiencies to offset the cost.
Verisk’s entry into the mortgage data business also eventually resulted in a divestiture. The company acquired what became Interthinx in 2005 and bolted on many subsequent mortgage data related acquisitions until 2010. After the housing bubble ended and as mortgage originations fell, Interthinx was viewed as being too volatile and reliant on transactions revenues. The entity was then sold in 2016 for $155M to First American Financial.
Verisk Health was formed when DxCG was acquired in 2004 and the company built the business through a series of acquisitions until 2012. This business was eventually sold to private equity for $820M in 2012. Again, the company likely overestimated the quality of the businesses and underestimated the difficulties of the regulatory environment in the healthcare vertical. Verisk Health had margins that were close to half of the what the insurance business was generating at the time and almost 70% of revenues were transactions based (vs. just ~15% for the insurance business).
And lastly, Verisk’s attempt to enter the financial services vertical ended in early 2022 when the company sold the segment to Transunion for $515M or 12.5x EBITDA. Most of that was Argus, which the company acquired in 2012 for $425M or 13.5x EBITDA. This vertical was interesting because it looked like Verisk could actually replicate the success of insurance vertical. Argus’ dataset of de-identified bank and credit card transaction information was collected through a consortium model. Argus’ data was used for competitive benchmarking, business intelligence, and mitigating risks of defaults, fraud, etc.
However, given the meager difference in the original purchase price and the sale price (with other acquisitions included along the way) and the recent decline in organic revenues, it’s likely the financial services business had below average returns for the company. Even if you only attributed Argus to the Financial Services business, over the 10 year period, it would imply that EBITDA only increased at a CAGR of 2.7%.
Post the recent divestitures of the energy and financial services businesses, Verisk is refocusing solely on the core insurance business. Recent M&A activity supports this change. Since 2018, most of the acquisition spend has been in the insurance vertical ($1.05B on 14 deals) and these deals are smaller in size.
We estimate that Verisk generated returns on incremental capital between 12%-20% over the past 5 years. The returns are somewhat lumpy, depending on the size of the acquisitions in a given year. Furthermore, the recent large divestitures could increase the volatility of future returns.
Reinvestment potential?
While Verisk can call almost all the P&C insurers in the U.S. its customers, there is still more room to grow in the U.S. market through increased adoption of the company’s different product offerings. At its 2018 analyst day, Verisk showed that of the 26 product categories in the insurance business, the top 10 customers have adopted an average of 18.3, up from 10.7 in 2013. As you go down the different cohorts by size, the average number of products adopted also decreases, indicating there could be an opportunity to sell more to these customers.
Verisk also laid out the TAM for the insurance vertical. The company estimates that underwriting and rating (the original insurance business) has a TAM of $3.6B, claims $2.7B and international at just $2.5B for a total TAM of $8.8B. With industrywide DWP of $627B in 2017, Verisk is assuming ~1% of that for its TAM. Remember that Verisk estimates that its revenues currently represent 25bps of the industry’s DWP, so there is some more room to grow, even in the U.S. market.
After the company’s decision to refocus on the insurance vertical, the two main areas of future investments are in international markets and the life insurance industry. International wasn’t a top priority in the past because the U.S. market dwarfs other “large” international markets by multiples of DWP. At the company’s 2017 analyst day, Verisk estimated that the U.S. market was $550B compared to other large insurance markets like the U.K. ($66B), Germany ($74B), China ($117B), Japan ($79B) and South Korea ($70B). DWP in developing markets like India were $10B and Brazil $27B.
The company has seen some early traction in the U.K. market, with the region representing 6.8% of revenues in 2021. Verisk made inroads through a few strategic acquisitions like Lloyd’s and Sequel. However, the company’s U.K. insurance business doesn’t benefit from a consortium data model like in the U.S. And since there is currently no other central repository of insurance data in Europe, the company can’t acquire into that position.
Verisk claims that success in the U.K. market doesn’t depend on ever developing a consortium data model because Verisk can embed itself in the software networks of its customers to build out the data gradually. While this is certainly possible, the path to a robust data set may take much longer and the margins would likely be lower than the consortium model. Here is the former CEO, Scott Stephenson, discussing entry into non-US markets at the Sanford Bernstein conference in May 2019:
“The other way you can do it is that you have all the algorithms and all the software associated with the interpretation of the data that you have. And you go to someone that you don't do business with and you don't have their data and you say, I can apply this to your data and here's the benefit you will get from that. And that's the basis for the relationship getting started. And then after the relationship has ripened, over time you say, well, look, we've been able to do these things for you up until now. But if you gave us the permission to make use of your data, we could also do the following additional things. And that's the other way that you can do it where you start out and you don't have as much of a footprint. And so that would characterize some of what we do in the non-U.S. markets.”
The other focus for future investment is the life insurance industry. In the U.S., 18 of the top 25 P&C insurers also offer life insurance products and the DWP of the life insurance industry is actually larger than P&C by ~30%. However, the industry hasn’t historically leveraged advanced data from companies like Verisk for their underwriting and claims process. For the most part, life insurers are still using simple data points like age, sex, credit scores and referencing them vs. mortality tables. The company views the life insurance industry more than a decade behind P&C and with new data sources like social media, mobility, etc. that have come online in the recent past, Verisk thinks the company can add value to life insurance customers.
Verisk likes to reference its voice recognition software as an example of the company’s value add to life insurance providers. Customers that use this software can leverage data and technology to take a voice sample to determine with a high degree of precision whether a person applying for a life insurance policy has vocal cord damage, which could indicate the person has a history smoking.
With a reinvestment rate between 50%-75% and a return on incremental capital between 12%-20%, we estimate that Verisk has increased its intrinsic value between 9%-11% annually over the past 5 years. With the insurance vertical throwing off so much high margin cash flow, the company does need to find areas for reinvestment. We’ll have to see if there are sufficient opportunities to invest in the insurance business’ growth initiatives going forward.
What else is important?
New CEO + Activist Investor
With Lee Shavel being promoted to CEO in May of 2022 and D.E. Shaw publicly requesting operational changes at the company, there is hope that there will be better governance and for more capital discipline at the company. D.E. Shaw spent 5 months privately working with the company prior to sending a demand letter on Mar 2022. Things that have happened since D.E. Shaw got involved are the sale of the financial services business, declassifying the board and separating the CEO and chairman role. Verisk also added a ROIC metric to executive incentive compensation targets in 2022.
Here is Lee Shavel (CFO at the time), discussing capital discipline and culture at the Raymond James conference in March of 2022:
“But what's important about capital is that it's a critical balancing point for growth because you can certainly achieve growth through the injudicious use of capital. And that's probably the biggest risk when you look at M&A. There are businesses that are smaller, growing faster. And if you aren't disciplined about making certain that you're going to get a return on that, then you can actually destroy value. And I think what I was very happy to achieve was a recognition across the organization that while everything isn't driven by returns, it's a very important component and deserves a seat at the table.
And as a CFO, the one area that you probably have the greatest influence over is capital allocation, particularly as it relates to M&A, but also significant investments within the business. And I feel as though that culturally that has been adopted. It's been adopted because I've forced it into our financial reporting internally. So I don't just talk to investors about our return on invested capital discipline. I spend a lot of time talking to the business units about the discipline and also where are they investing capital, what are the potential returns. And I know, within each of our business units, how we are progressing against our objective of generating of good returns. And it balances that perspective and it encourages an investment mentality that I think certainly everybody in this room can appreciate, but in a lot of operating models it doesn't receive the attention.”
Why doesn’t management give guidance?
It’s odd that a company with such a predictable business model doesn’t give formal guidance on its earnings calls. In 2021, 81% of revenues were subscription based, and that should move even higher since the financial services and some of the energy businesses have been divested. Furthermore, insurance customers are very slow to make changes, so Verisk should have a good understanding of where the next quarter’s numbers should land.
The company does have long-term targets for organic revenue growth (greater than 7%) but even commentary around directional trends have been wrong. Some of that was from the weakness in the energy and financial services businesses. But with the company refocused on insurance, there should be no reason that the company can’t give guidance going forward.
The company does guide towards annual capex spend, but even the long-term forecasts for that have been incorrect. At the company’s 2017 analyst day, Verisk forecasted that capex spend as a percentage of revenues would decline after a peak year, down to 5%-6% of revenues by 2022. Fast forwarding to today and that number still remains high near 9%.
Thoughts on buybacks
Verisk generates a lot of cash due to its high margins and favorable cash conversion. Operating cash flows as a percentage of revenues has ranged between 35% and 40% over the past decade. Even accounting for the large acquisitions and the increasing capex spend at the company, there is excess cash left over every year. The company has allocated a lot of that cash to buybacks, accounting for 35%-55% of operating cash flow in most years.
The question is whether these buybacks have been the best use of capital since Verisk trades at a high valuation, which has moved even higher to 20x-25x forward EBITDA in the past 4 years. The company did start to issue a dividend in 2019 amounting to 16%-17% of operating cash flow over the past 3 years.
Verisk has also sold the financial services business for 12.5x EBITDA and 3E at a high teens EBITDA multiple, both lower than Verisk’s multiple. The company committed to using the proceeds from those sales for buybacks, which is questionable given the differences in multiples received for the sales vs. paid for the buybacks.
Optionality
For many companies that we’ve covered, M&A is a big part of optionality, especially for companies that can expand geographically or shore up their product offering to improve their scope. But with Verisk, the hope is that large transformative acquisitions are no longer on the table outside of the core insurance vertical.
The international opportunity will likely be a slow grind with many small acquisitions and customer wins that will gradually move the needle. It may be more binary in life insurance if Verisk can find a way to become the industry’s utility like it has in P&C insurance. Whether it’s an essential data source that every life insurance company can use to improve their underwriting or some software network that makes doing business materially easier for these customers, Verisk should stand to benefit significantly.
If you made it this far, I hope you received some value from reading our analysis. Please subscribe to the newsletter and share with anyone that would find it valuable. Thank you for your support!