Berkshire's Diversification Offsets Insurance Losses

Despite a difficult third quarter, in which wide-moat-rated

We use a sum-of-the-parts methodology to derive our fair value estimate for Berkshire. Following our updates to the different models we use to value the company, our valuation of the insurance operations decreased 3% to $94,200 ($63) per Class A (B) share, primarily because of the impact of this year’s large catastrophe losses. Our estimate for Kraft Heinz KHC, which we value separately from the insurance operations, is $14,900 ($10) per Class A (B) share.

Following a strong recovery in rail volume (especially for coal) during 2017, our estimate for BNSF increased 3% to $63,100 ($42) per Class A (B) share. As for the company’s utilities and energy division, our fair value estimate increased slightly to $26,000 ($17) per Class A (B) share after we backed out assumptions for the Oncor acquisition and adjusted the model for year-to-date results. While our value for Berkshire’s manufacturing, service, and retail operations increased slightly to $82,600 ($55) per Class A (B) share, our estimate for the finance and financial products division remains in place at $9,200 ($6) per Class A (B) share.

Looking more closely at Berkshire’s insurance operations, we’ve adjusted our near-term underwriting forecasts for each of the company’s four insurance subsidiaries--Geico, General Re, Berkshire Hathaway Reinsurance Group, and Berkshire Hathaway Primary Group--to account for the catastrophe-related losses incurred in the third quarter. While we continue to forecast the potential for large natural catastrophe losses, we have pushed out the timing of these events to the tail end of our projection period. Our revised forecast has total earned premium growth expanding at a 9.2% compound annual growth rate during 2017-21 (versus 8.9% previously), with Berkshire’s overall combined ratio for its insurance operations averaging 98.0% (compared with 96.5% previously). We’ve also incorporated stronger investment returns this year for the segment’s equity and other investments portfolios, given the ongoing gains being recorded in the markets as well as the investment portfolio itself. The net result is a 3% decrease in our value for the insurance operations to $94,200 ($63) per Class A (B) share from $97,400 ($65).

We continue to be concerned about Geico’s relentless pursuit of growth, given that it has come at the expense of profitability the past several years. The earned premium growth of 16.5% during the third quarter was the strongest quarterly result we can ever remember Geico putting up (which is what we also said after the second quarter, when the auto insurer put up a 16.0% earned premium growth figure). Given that Geico is starting to lap the more aggressive underwriting path it started on during the fourth quarter of last year (when earned premium growth tacked up to 13.3% from a quarterly run rate of 11.7% in the previous year and 11.0% over the previous three years), we don’t envision future earned premium growth to be as strong as we’ve seen the past several quarters. Over the next year, we expect to see earned premium growth of 12%-16%. It should then drop down to 11%-13% in 2019 and average closer to 11% during the final years of our five-year forecast (closer to the 10.7% rate of annual growth Geico put up on average during 2007-16).

While we’re glad to see Geico taking advantage of opportunities to pick up share and increase the scale of its operations, it has come at an added cost, with the company seeing a marked increase in its loss ratio the past couple of years. Although the company’s loss ratio during the third quarter was hurt by catastrophe losses related to hurricanes Harvey and Irma, it was still at an astonishingly high 85.3% after the $500 million in hurricane-related losses was backed out. Over the past four calendar quarters, Geico’s average loss ratio (excluding hurricane-related losses) was 84.1%, which compares unfavorably with the average quarterly loss ratios of 80.1% and 78.3% during the three- and five-year periods preceding the fourth quarter of 2016, when the company ramped up its underwriting efforts. Berkshire CEO Warren Buffett noted during this year’s annual meeting that first-year business, which comes with acquisition costs and a higher loss ratio, tends to run almost 10 points higher than renewal business, which would explain the rise in Geico’s loss ratio during the past year, but with the insurer continuing to hit the gas pedal on underwriting, it gets harder and harder to determine when we might see loss ratios return to more normalized levels.

Even with this uptick in the loss ratio, Geico has for the most part been able to keep its (non-hurricane-affected) underwriting profitability in the black, with its combined ratio (excluding hurricane-related losses) averaging 99.0% during the past year. While this is worse than the 96.3% and 95.5% average combined ratios reported during the three- and five-year periods leading up to Geico’s more aggressive underwriting push, it is impressive given the 410- and 580-basis-point increases in its loss ratio at the end of the third quarter, respectively, when compared with the same periods. Much of Geico’s ability to keep its (non-hurricane-affected) underwriting profitability in the black has been tight expense controls, as well as the benefits of increased scale. The company’s 13.6% expense ratio during the third quarter of 2017 was the lowest we remember seeing, and its average loss ratio of 14.9% during the past year represented a meaningful improvement over the average quarterly loss ratios of 16.2% and 17.2% put up during the three- and five-year periods preceding the fourth quarter of 2016.

As we noted in an early October presentation on the auto insurance industry, “Technology Will Continue to Drive the Auto Insurance Industry, but Will Eventually Crash It: Outlining Winners and Losers,” Geico’s primary competitive advantage comes from its scale, with management squarely focused on maximizing growth in order to maintain and expand that advantage. While we view this as the right course of action over the long run, as evidenced by the impact that greater scale has already had on its overall expense ratio (a key metric for tracking an auto insurer’s scale advantage), it has been difficult to chalk it up as a victory given the deterioration in the combined ratio. With Geico expected to maintain a more aggressive growth posture, we project a combined ratio of 100.0% for 2017, which should eventually improve to 96.1% by the end of our five-year forecast, with most of the improvement coming at the tail end of our projection period as earned premium growth tempers.

This does nothing to alleviate our long-term concerns about Geico, highlighted in the auto insurance report that went out last month as well as our company-level presentation, “Berkshire Hathaway Will Survive the Departure of Buffett and Munger,” published in September 2016. The biggest issue we have is the company’s reticence about using telematics in its underwriting operations. As long as the company trails the industry and its closest rival, narrow-moat-rated Progressive PGR, from a loss ratio perspective, we expect there to be calls for Geico to adopt some of these technologies. We don’t believe Geico’s lack of a telematics program is a big issue today, as the insurer continues to compensate for its higher-than-average loss ratio by scaling its underwriting expenses, but it could be playing a dangerous game if it waits too long, as it could open the company up to an adverse selection problem, with competitors that have been collecting and analyzing telematics data for longer periods being able to pick the best drivers, leaving the rest to those that have come too late to the telematics market.

Unlike our aggressive growth outlook for Berkshire’s auto insurance business, we’re not expecting much in the way of underwriting growth for the company’s two reinsurance arms--General Re and BHRG--given that the industry continues to struggle with a weak pricing environment driven by excess capacity created by the entrance of alternative capital (including hedge funds and catastrophe bonds) and a general lack of large-scale catastrophes (which has made it harder for reinsurers to justify raising their rates). While we’ve long believed that a string of major catastrophes like we’ve seen this year could serve as a catalyst for some of the nontraditional capital to leave the market, which would allow pricing to return to more normalized levels, Buffett said in September that he doesn’t expect pricing to improve materially in the aftermath of this year’s hurricanes.

He also expects the reinsurance business to remain unattractive for at least the next decade. Given such a poor pricing environment, we believe that reinsurers with more diverse portfolios, serving multiple distribution channels--much as General Re and BHRG do--are much more likely to offset the negatives in the market and should be able to target profitable opportunities as they arise. It should also be noted that General Re and BHRG (unlike many of their peers) have the luxury of walking away from reinsurance underwriting when an appropriate premium cannot be obtained. With Berkshire’s reinsurers also dealing with a number of runoff contracts in the near to medium term, the ability of the two companies to generate float (aside from one-off transactions like the deal with American International Group AIG that BHRG initiated earlier this year) will be limited. More than two thirds of Berkshire’s insurance float is generated by General Re and BHRG, which are long-tail businesses. This means that future increases in the insurance float (which stood at around $113 billion at the end of the third quarter) will come predominantly from Geico and BHPG, which are short-tail businesses, with the float generated by these businesses tending to be invested in less risky and more liquid investments with smaller return profiles.

While General Re has generated an abnormal level of earned premium growth this year, with year-to-date growth of 10.3% through the first nine months driven by a combination of new business and increased participations for renewals, we don’t expect to see that same level of growth in future periods. The company’s earned premium growth was basically flat over the past five and ten calendar years, with General Re posting a negative 0.5% annual rate of earned premium growth on average during 2012-16 and a negative 0.7% annual rate during 2007-16. Part of the growth we’ve seen so far this year is probably the result of Ajit Jain’s push to get General Re to pursue international opportunities that he believes it has overlooked in the past, which we believe will be difficult to replicate in future years. Our current forecast for 2017-21 has earned premiums growing at a 0.2% annual rate on average, but backing out the contribution from this year’s stronger-than-average earned premium growth leaves average annual growth at negative 2.0% during the remainder of our forecast.

As for BHRG, we don’t expect much earned premium growth during our five-year forecast period but do recognize that the company’s $10.2 billion retroactive reinsurance deal with AIG in January will lift earned premium growth well into positive territory this year. This only sets up BHRG for a steep decline in earned premium growth next year, with the company returning to a more normalized level of low-single-digit annual declines in earned premiums during the final years of our forecast period. As for profitability, the best we can hope for longer term is tight expense controls and a lack of extremely adverse events, which should allow both companies to keep their combined ratios below 100%. While it is extremely unlikely that this will be the case this year, given the $2.3 billion in catastrophe-related losses recorded by Berkshire’s reinsurance businesses in the third quarter, we do expect both companies to strive for that goal over the remainder of our forecast period, even as they are likely to face another year of catastrophe losses before the end of our five-year projection period.

The good news for Berkshire is that it has more than one insurance business that is growing, with BHPG generating earned premium growth of 15.4% through the first nine months of 2017, with a combined ratio of 86.8% after excluding the impact of catastrophe-related losses. While not quite as homogeneous as the three other insurance subsidiaries, BHPG has been Berkshire’s fastest-growing business during the past five calendar years, relying on both acquisitions and organic growth to expand earned premiums at a double-digit rate annually. Our forecast for 2017 calls for earned premium growth of 15%-16%, which should gradually wind down to 12%-15% annually by the time we get to the end of our five-year forecast. We envision the company’s combined ratio averaging 89.3% after 2017, slightly above its 87.2% and 87.7% averages during 2012-16 and 2007-16, respectively, but reflective of the higher costs associated with growing business like its Berkshire Hathaway Specialty Insurance unit.

While Berkshire’s insurance operations have struggled with underwriting profitability this year, its equity and other investments portfolios have produced a decent gain, expanding from $139.3 billion at the start of the year to $161.0 billion at the end of the third quarter. Although some of the gains in the portfolios can be tied to new capital being put to work, a fair amount of it reflects market gains during the year, with big contributions from Apple AAPL, Bank of America BAC, and American Express AXP more than offsetting the weakness at Wells Fargo WFC, IBM IBM, Delta Air Lines DAL, and United Continental UAL. We don’t envision the portfolios doing as well in the future, though, expecting more normalized market returns in a 5%-7% level during the remainder of our forecast period (which is what we forecast for the active equity portions of the assets under management at the U.S.-based asset managers we cover). Looking more closely at Kraft Heinz, which we value separately from the insurance operations (despite being held on its books), our fair value estimate increased slightly to $14,900 ($10) per Class A (B) share.

We’ve increased our fair value estimate for BNSF 3% to $63,100 ($42) per Class A (B) share following a strong recovery in volume (especially for coal) during 2017. This puts the rail’s current market capitalization at $97.0 billion, slightly above the market capitalization for Union Pacific UNP, its closest peer among the publicly traded Class I railroads. While Union Pacific is usually a good proxy for BNSF, given both companies’ focus on the Western U.S. market, the competing railroad was negatively affected by Hurricane Harvey, including the temporary closure of seven plastics plants (which hurt industrial product volume) and the loss of 1,700 miles of roadway (which were out of service after the storm and dropped average train speed); the impact of the storm reduced Union Pacific’s operating ratio by 70 basis points during the third quarter.

Over the past couple of years, BNSF has been beset by not only a shortfall in coal volume that started in the first quarter of 2016, with volume falling 21.1% overall last year, but also a falloff in industrial product shipments (driven primarily by the decline in crude oil prices), with volume falling 5.9% and 7.8%, respectively, during 2015 and 2016. Things have looked much better so far during 2017, with coal and industrial product volume increasing 12.4% and 2.2%, respectively, during the first nine months, even as volume moderated in the third quarter. While management is projecting overall volume, which was up 6.0% during the first three quarters of 2017, to moderate during the fourth quarter as well, full-year volume appears to be trending toward our full-year projections.

While we forecast carload volume to increase 5.3% overall during 2017--driven by 6.0% annual growth in consumer product volume, a 2.0% increase in industrial product shipments, a 10.0% improvement in coal volume, and flattish agricultural shipments--our longer-term expectation is for 1.0%-2.0% annual volume growth overall during the remainder of our forecast period. We expect coal to continue to be the biggest detractor from volume growth, as natural gas and renewables, which are significantly cheaper than coal, continue to be preferred by both utilities and industrial companies. We forecast negative single-digit volume growth for coal in the remaining years of our forecast. Shipments of consumer products, meanwhile, should grow in a 2%-3% range annually going forward, while industrial products and agricultural volume growth should be closer to 1.0% annually during the remainder of our forecast period.

BNSF’s year-to-date revenue growth of 8.5%, which was aided by an improvement in volume and average revenue per car/unit, was at the upper end of our mid- to high-single-digit top-line growth forecast for 2017. Much as we see volume moderating at the railroad during the fourth quarter, though, we expect revenue growth to cool, which should leave full-year results closer to our forecast. Going forward, we expect revenue to grow at a low- to mid-single-digit rate, with the net result being a 4.3% CAGR during 2017-21. BNSF’s operating ratio of 66.1% during the first three quarters was an improvement on year-ago levels, and we expect the railroad to close out 2017 with full-year results below the 66.3% it reported in 2016. We believe BNSF’s operating ratio will continue to improve, dropping to 58.2% by the end of our five-year forecast; this would put the railroad’s profitability more on par with that of Union Pacific, which we project to have a 56.0% operating ratio by the end of 2021. This should help fuel the annual dividend BNSF pays to Berkshire, which we have growing from $2.5 billion this year to more than $6.0 billion by the end of our five-year forecast, barring any acquisition activity or debt paydowns during the period.

As for the company’s utilities and energy division, our fair value estimate increased slightly to $26,000 ($17) per Class A (B) share after we backed out assumptions for the Oncor acquisition and adjusted the model for year-to-date results. Normally a beacon of stability, Berkshire Hathaway Energy was able to convert year-to-date top-line growth of 4.7% into a 2.7% pretax earnings gain during the first nine months of 2017, basically in line with our forecasts for the year. We continue to expect BHE’s U.S. regulated utilities--PacifiCorp, MidAmerican Energy, and NV Energy--to receive constructive rate case outcomes and earn near their allowed returns on equity. This would drive 5%-6% annual EBITDA growth for these businesses during 2017-21. For Northern Powergrid, we expect ongoing price reviews to generate mid-single-digit annual EBITDA growth during the period, which is slightly higher than the growth we expect to see from BHE’s pipeline and renewable energy businesses. Our renewable energy growth estimate, in particular, assumes the prospects for additional renewable energy projects in the near term will be limited due to the expiration of federal tax credits.

While Berkshire did miss out on Oncor Electric Delivery during the third quarter, being outbid by Sempra Energy SRE at the last moment for this prized asset within the bankrupt Energy Future Holdings, we don’t expect this to be the end of BHE’s pursuit of other utility- and energy-related assets. We continue to believe that the subsidiary’s cash flow generation should easily support another $5 billion-$10 billion acquisition and that purchase prices could go higher with Berkshire having plenty of cash on hand to fund any larger deals that BHE pursues. We believe Alliant Energy LNT remains a compelling pursuit, albeit an expensive one (currently trading at a more than 20% premium to our own fair value estimate), and would not be surprised to see BHE kicking the tires on either Pinnacle West PNW or NiSource NI.

There is also the slimmest of chances that Oncor becomes available again, as Sempra needs the support of both the bankruptcy court and the Texas regulators, who were much more receptive to BHE’s assurances with regards to regarding Oncor’s independence and financial integrity, as well as BHE’s commitment to long-term investments in infrastructure and security, to close the deal. Given the regulatory roller coaster the Texas commissioners have put Oncor’s past suitors on, we’re not entirely convinced that Sempra’s bid will gain regulatory approval. In particular, Sempra is relying heavily on equity (which accounts for 65% of its bid) to fund the transaction, with the remainder coming from debt. Contrast this with BHE’s proposal, which did not rely on any external funding, a structure we believe the regulators were viewing much more favorably.

The Texas regulators, in our view, have been exceedingly difficult the past few years, creating uncertainty for any deal put before them. This is the main reason we attribute only a 75% probability to the Sempra transaction closing. We can’t help but think that the regulators will have the BHE bid in the back of their minds as they consider Sempra’s offer, and as the last two and a half years have shown, it has been the Texas regulators, and not the bankruptcy court, that have had an outsize role in Oncor’s fate. Our most likely scenario is that BHE’s bid for Oncor stays on the sidelines, but should the company actually end up with the asset, our thoughts about the economics--basically that the deal, although attractive, was expensive and, as such, would be dilutive to earnings--have not changed.

(For more details on utility- and energy-related assets that BHE might be interested in, please see our early October Select presentation, “Buffett’s on the Hunt for Utilities: Should You Be Buying?”)

While our estimate for Berkshire’s manufacturing, service, and retail operations increased slightly to $82,600 ($55) per Class A (B) share, our valuation for the company’s finance and financial products division remains in place at $9,200 ($6) per Class A (B) share. With regards to Berkshire’s MSR operations, the group overall recorded a 5.3% increase in year-to-date revenue at the end of the third quarter, which would leave full-year results pretty much in line with our near- and long-term forecasts, which call for mid-single-digit annual revenue growth during 2017-21. Operating margins of 7.5% during the first nine months of 2017 were slightly better than our near- and long-term forecasts, which have operating margins expanding 20 basis points annually over the 7.0% the segment reported in 2016, but not enough to get us to alter our long-term target of 8.0%.

At Berkshire’s finance and financial products division, year-to-date results were mixed at the end of the third quarter, with revenue increasing 7.2% but pretax earnings declining 5.0% due to lower interest and dividend income from investments, higher railcar repair and storage costs, and lower earnings from CORT’s furniture rental business. The company should benefit going forward from investments in Oakwood Homes and Home Capital Group, which lend more support to our long-term forecast for mid- to high-single-digit top-line growth, with pretax margins of 25%-27%, during 2017-21. Ultimately, though, the finance and financial products segment is a mere rounding error for Berkshire’s overall results, accounting for less than 6% of pretax earnings on average the past five years.

Berkshire’s book value per Class A equivalent share at the end of the third quarter was $187,435. We believe the company can increase book value per share at a high-single- to double-digit rate, much as we’ve seen since the start of the new millennium. This would leave book value per Class A equivalent share at $192,100 at the end of this year and $211,700 at the end of 2018. Depending on the size and the timing of changes to the U.S. corporate tax rate, Berkshire will also probably see a one-time step up in its book value, as the deferred tax liability associated with its investment portfolio and other assets gets written down to reflect the lower statutory rate. Our best guess is that this would result in a 5%-10% increase in book value per share, depending on the value of the investment portfolio at the time that the tax changes are made. As Berkshire tends to trade on book value per share, with its shares trading at 1.4 times book value on average the past five and ten years (and 1.5 times over the past fifteen years), any upward moves in the company’s reported book value per share would be expected to have a positive impact on the shares.

While Berkshire closed out the third quarter with $109.3 billion in cash on its books, a portion of that capital has already been spoken for, given the company’s commitment in early October to purchase a 38.6% interest in Pilot Flying J, the largest operator of travel centers and travel plazas in North America, for an undisclosed amount; the deal also comes with an obligation to acquire an additional 41.4% of the business in 2023. With an estimated value of around $9 billion, the purchase of the equity stake in Pilot Flying J should reduce Berkshire’s cash balance by $3.5 billion this quarter and require the company to earmark at least another $3.7 billion for its future commitment. Add to that the $20 billion in cash that Berkshire likes to keep around as a backstop for the insurance business, as well as another $5 billion in operating cash for its other businesses, and Berkshire probably has close to $80 billion in dry power available right now for investments, acquisitions, and share repurchases (or dividends).

Based on the book value of $187,435 ($125) per Class A (B) share at the end of September and the company’s stated willingness to buy back stock at prices below 1.2 times book value, Berkshire should be willing to repurchase shares at prices up to $224,922 ($150) per Class A (B) share, implying a floor on the common stock that is about 18% below where Berkshire’s shares are trading right now. That said, the company has not bought back shares since December 2012, when the repurchase threshold was raised from 1.1 times book value to 1.2 times to facilitate the sale of a large position of 9,475 Class A shares belonging to the estate of a longtime shareholder, which we believe was not willing to settle for a price of 1.10 times book value, given that the shares had traded at around 1.17 times book value on average during the 30 trading days that preceded the buyback.

We have questioned Buffett several times during the past couple of years about whether the share-repurchase threshold needs to be raised, given his belief that the company’s intrinsic value continues to exceed its book value (with the difference only widening over time) and the fact that the company’s cash balances continue to expand. He has seemingly been more and more open to this idea as well as the idea of a dividend, noting that it will become increasingly more difficult for Berkshire to continue to sit on a large and growing cash hoard should balances reach $150 billion. Given that the company’s cash balances are likely to continue to grow from here because of what we think is a lack of identifiable investment options (with equity markets at or near all-time highs and values for public and private companies extended), we expect that we’ll be at that benchmark in relatively short order.

As a result, we continue to model a special one-time cash dividend of around $20 billion midway through our five-year forecast. We’re basically assuming that should Buffett decide to pay a dividend, it would probably be one-off in nature and based purely on bloated excess cash balances as opposed to a regular quarterly dividend. From a regulatory perspective, Berkshire’s principal insurance subsidiaries alone could have declared as much as $13 billion as ordinary dividends during 2017 without prior regulatory approval, so we don’t envision there being much pushback from regulators, should the company decide to make a slightly larger one-time cash dividend payout in the near term.