Unum: A Core Business Perspective
Summary
- Unum has an attractive core employee benefits business which is highly consistent and has been increasingly profitable.
- The long-term care insurance reserves continue to represent a valuation overhang for the company which impacts multiples.
- Our analysis suggests the company's premium income is closely related to employment trends but not as sensitive as may be expected in the current cycle.
- Our valuation models indicate the potential for significant compound appreciation in coming years even if GAAP loss reserve additions exceed the statutory reserves required by the Maine Bureau of Insurance.
- Unum represents a long-term holding across our portfolios.
Unum (NYSE:UNM) is a major employee benefits insurer serving various markets with a focus on accidental death, disability, group life, supplemental, and related product lines. The company operates primarily in the United States with subsidiaries serving the United Kingdom and Poland. Unum's product lines are, for the most part, attractive businesses in stable markets despite intense competition and moderate growth trends. The employer-based nature of the core employee benefits business results in stickiness of customers.
Our previous article presenting a broad perspective on the company's long-term care liabilities represented an initial take on the company's position. In this article, we set aside the long-term care issue for a moment to examine the company's core operating business, the sensitivity of premium income and operating costs to unemployment rates, and the impact of changing benchmark interest rates on the company's profitability.
We conclude that while Unum's business will be impacted by the economic downturn the company has a record of emerging from recessions in decent condition despite the interest rate challenges. In addition, even if the company were to determine moderate additional reserves for long-term care insurance were required on a GAAP basis, the company would be able to support the dividend (or at worst a reduced dividend) although not share repurchases. In any event, Unum's core businesses represent an extremely stable and increasingly profitable line of business that should warrant a premium valuation relative to peers absent the long-term care insurance reserve issues.
We project forward valuations of between $22.00 and $26.00 assuming moderate annual additions to loss reserves (beyond those recently required on a statutory basis by the Maine Bureau of Insurance) and, in the absence of any material future reserve additions, a valuation between $27.00 and $34.00. The potential compound annual returns, including dividends, thus fall into a range of 10%-15% under an ongoing loss reserve scenario. However, should additional long-term care insurance reserves of any meaningful magnitude ultimately prove unnecessary, total compound annual returns could easily exceed 20% per annum for the next few years and still result in a very modest valuation compared to projected book value and earnings per share.
Unum's shares at the present quotation are undervalued, possibly by a wide margin, though the margin is not as wide as it was when we first began developing this article earlier in the week. Regardless, under any assumption other than severe additional loss reserve requirements for the long-term care closed block, the potential forward returns remain quite compelling.
Core Operating Results
Unum's core businesses are quite attractive. Unum has a longstanding trend of positive operating results notwithstanding three separate charges to earnings in the last several years primarily associated with boosting loss reserves for long-term care insurance liabilities. The company experienced a mild decline in revenues during the financial recession although earnings remained surprisingly resilient. Revenue growth resumed as unemployment declined during the subsequent recovery with consistent annual gains since 2012. Earnings have followed a similar upward trajectory despite period to period volatility due to reserve charges, a trend dating back more than a decade. The increasing profitability of the company contrasts starkly with the generally underwhelming performance of the company's shares over the same period of time.
Source: Winter Harbor Capital
The growth in earnings per share has been driven by improvement in the company's net income margin, as reflected in the following chart, largely driven by slowly declining average benefit expenses in combination more recently with annual gains in revenue.
Source: Winter Harbor Capital
A part of the periodic volatility in net income margin is associated with periodic additions to long-term care insurance reserves, as noted earlier, and other factors which occurred in 2011, 2014, and 2018. The financial recession at the end of the last decade slowed the upward trend in net income margin but, notably, did not reverse the trend which had again begun to accelerate in recent years before the advent of the coronavirus. The disruption to employment will almost certainly impact net interest margins going forward to some degree. However, it's notable how well the company's net income margin held up in the face of a deep recession especially considering the company's focus on employee benefits. Unum's historical ability to sustain the net income margin, even in the face of challenging economic conditions and a very low interest rate environment, is a quite appealing attribute from an investment perspective.
Indeed, there is much to like in the company's operating results and consistency in revenue, profitability, and earnings growth beyond the long-term care insurance block. Unum's operating results reflect the company's solid position in its market and incremental pricing power especially during periods of sustained economic and employment growth.
Premiums and Unemployment
A pressing question for investors and shareholders going forward, given past experience, is whether it's possible to develop an assessment of the company's earnings power and resiliency in the current coronavirus driven economic situation. In particular, how might falling interest rates and rising unemployment impact the company's net investment and premium income going forward and, consequently, earnings per share? In order to address this question, we look at the last fifteen years of economic and financial data to determine whether there is a functional relationship between premium income and unemployment rates which would, hypothetically, allow a reasonable prediction of future operating performance. In particular, we examine whether unemployment rates, or the periodic change in unemployment rates, is a meaningful indicator of premium income, or the periodic change in premium income, for a company specializing in employment based benefits.
The following chart reflects Unum's historical net investment income and premium income in comparison to the average unemployment rate. The average unemployment rate is calculated as the simple average of the January and December unemployment rates for a given year reported by the Bureau of Labor Statistics.
Source: Winter Harbor Capital
A simple visual assessment of the graphical data suggests that there may indeed be a relationship between premium income and average unemployment rates. In order to assess the sensitivity of premium income to changes in the average unemployment rate, we developed several statistical models to determine the degree of correlation and whether such correlation was sufficiently strong to provide a reasonably reliable basis for projecting forward operating results. We excluded net investment income from the assessments since net investment income would be, at best, only tangentially related to unemployment levels through the relationship between unemployment levels and benchmark interest rates and instead focused on the connection between premium income and unemployment rates.
Our models considered a number of different potential relationships, including relationships between the absolute levels of premium income and unemployment rates, average unemployment rates versus year end unemployment rates, the relative change in those values, etc. Unsurprisingly, we found that while premium income and average unemployment rates are indeed correlated, the strongest correlation occurred between the change in the average unemployment rate and the change in premium income for the corresponding period. In addition, the change in average unemployment rates rather than year end unemployment rates resulted in a stronger correlation. We believe the stronger correlation of the change in average annual unemployment rates results from an inherent lag between a decline in employment and the resulting impact on premium income.
A summary of the distribution and result is presented in the following chart. In the last recession and subsequent recovery, Unum's change in premium income over the course of a given year was quite strongly correlated with the change in the average unemployment rate for the same year, resulting in a parabolic function:
Source: Winter Harbor Advisors
Indeed, with an R-squared value of 0.88, substantially all of the change in premium income is described by the change in average unemployment rates. A correlation coefficient of 0.94 (the square root of the coefficient of determination) is quite high for any statistical analysis with perfect correlation being 1.0. Alternate approaches also resulted in quite high correlation coefficients, typically around 0.7 or above, so while the correlation may not be as strong based on other factors, there is still a clear and strong relationship between premium income and average unemployment rates (or the changes in the two factors), providing a basis for projected premium income under various economic conditions.
Nonetheless, it's important to note a few considerations. First, while the statistical result means the data sets are correlated it does not necessarily mean that the dependent variable actually directly causes the change in the independent variable. A relationship exists, but it's still possible that an undefined third factor is simultaneously acting on both variables in a highly similar way. The high degree of correlation, though, would imply that any third factor is highly correlated to both variables such that it would yield a similar result.
Second, we only have one or two data points for periods with large negative changes in average unemployment rates which could skew the relationship. The tendency of unemployment to rise quickly at the beginning of a recession before more slowly declining during the subsequent recovery inevitably yields a relative dearth of data points at the low point in the data set. The lack of a larger quantity of data points may result in the apparent correlation being stronger than in reality.
Still, the correlation between premium income and unemployment rates makes sense from both a contextual and logical standpoint as reflected in the curve. Unum's premium income declines during periods of rising unemployment on an almost linear basis. In contrast, premium income growth tends to be strong even if positive changes in the average annual unemployment rate are rather modest.
The differential relationship between rising and declining unemployment rates makes intuitive sense in the context of the company's business. In periods of declining employment, employees lose their benefits along with their jobs such that premium income declines in concert with job losses. In periods of rising employment, Unum gains premium income as jobs are added and, as the labor market tightens and unemployment rates flatten, can still drive premium growth through rate increases. In fact, this differential relationship should be one of the factors investors find attractive about the company since recessions tend to be short and expansions tend to be long. Unum may suffer short-term premium income losses when unemployment initially spikes but tends to make up that ground - and much more - over the subsequent recovery and growth cycle.
Average Yields on Investments
A second key driver of the company's revenues and profitability is the rate of return on portfolio investments. Unum's average yield on investments has declined over time in the face of low benchmark interest rates (for which we use the federal funds rate as a proxy) although the long period of near zero benchmark interest rates over the last decade has established something of a floor for average yields. The decline in average yields thus began to flatten around 2012 and, while the declining trend persists, the company has seen a deceleration in the rate of decline, as reflected in the following chart:
Source: Winter Harbor Capital
The increase in benchmark interest rates in the last few years, though modest and temporary, may slow the erosion in average yield in the short term. The portfolio likely gained somewhat from reinvestment of some maturing funds (and investment of new funds) at higher average rates during the period the impact of which won't yet be fully reflected in average yields.
In terms of forecasting future average investment yields, Unum has already experienced an extended period of time with near zero benchmark interest rates. The company's average yield on investments declined early in the last decade as maturing securities in the investment portfolio were replaced with lower yielding securities but the declines began to flatten around 2012 - four years into the zero interest rate cycle. A return to near zero benchmark interest rates will result in further erosion in average yield but there is a limit to the potential decline. The real risk is that near zero interest rates prove not to be the floor, i.e., that interest rates actually turn negative as has happened in other parts of the world.
We assume negative interest rates won't occur and there is an effective floor average yield of about 4.0%. A decline to this level would represent an unprecedented decline inconsistent with the company's experience during the previous period of low benchmark interest rates and we consider this to be at the extreme of likely outcomes at least within the next few years. A reduction of this magnitude would reduce investment income by some $224 million to $336 million annually partially offset by growth in the investment portfolio associated with additional reserve contributions required in the reserve agreement made with the Maine Bureau of Insurance.
Benefit and Operating Ratios
The final component of any projection is understanding the company's operating ratios and the impact that declining premium income due to rising unemployment has on those ratios. Unum's benefits and operating ratios, whether as a percentage of premium income or a percentage of premium and net investment income, have been very consistent over the last 15 years. The following chart reflects the company's benefit expense and other operating expenses as a percentage of premium income over that time period:
Source: Winter Harbor Capital
The inclusion of investment income in the calculation along with premium income results in an essentially identical chart except that the percentages are lower and the lines somewhat flatter since investment income has been highly consistent over the last several years. The depressing effect is caused by the lower growth in net investment income - premium income grew 19.5% over the last 15 years while net investment income grew only 12.8%.
The jumps in benefit and loss reserve expenses are, as noted before, associated primarily with additions to the company's long-term care insurance reserves in 2011, 2014, and 2018. However, absent these additional reserves, Unum's benefit and loss reserve expenses as a percentage of premium income have been trending slightly downward over the last decade, supporting the incremental growth in net income margin observed earlier.
A breakdown of the company's non-benefit related operating expenses shows similar consistency over time with some variation in the company's expense ratios during recessionary periods. Commission expense, for example, has risen rather consistently as a percentage of premium income since about 2006. A decline in other expenses has generally offset a good part of the increase in commissions. Compensation expense is most sensitive to declines in premium income presumably since compensation expense is rather inelastic and represents something of a fixed cost compared to more variable premium income.
Source: Winter Harbor Capital
The consistency in the company's benefits and operating expenses, in combination with the potential predictability of the impact of changes in unemployment on premium income, provides a decent basis for developing a projection of the company's forward results based on these metrics.
Forward Projections
We developed forward projections for the company based on the above historical data which included a base case (assuming no additional reserves are taken for long-term care insurance liabilities) and a modified case which assumes the company incurs an additional $350 million in annual benefit expense to boost long-term care insurance reserves over time. The $350 million annual addition to loss reserves would increase total reserves over seven years (the time frame which the Maine Bureau of Insurance has specified for augmenting the statutory reserves) by approximately $2.9 billion on a pre-tax basis ($2.3 billion after tax) using our projected forward average yields.
We're reasonably confident that the increase in unemployment will erode premium income, per our financial models, while reversing the company's longstanding trend of rising net income margins. The extent remains to be seen based on the severity of the impact on premium income though our models suggest a rather mild reversal. The primary driver is comparatively fixed costs deleveraging over a quicker initial decline in premium income coupled with essentially stagnant net investment income excluding gains and losses.
The greatest uncertainty from a projections standpoint is the average annual unemployment rate and the impact of the nature of unemployment in the coronavirus induced recession on premium income. The model currently projects an average annual unemployment rate for the current year of 7.9% peaking next year at 10.9% (due to the way the average is calculated) based on the expectation that the unemployment rate at the end of the calendar year will be closer to 12% after the initial spike. The model projection is a little above the 10% year end forecast from Bloomberg Economics. Also, given the improvement in the unemployment rate in May to 13.3%, our year end projection may prove too high, resulting in lower earnings per share projections than actual results.
It's also necessary to realize that the current economic situation is dissimilar to prior recessions and may result in a lower correlation between average unemployment rates and premium income. Unemployment in the current economic landscape is more heavily concentrated (for the moment) among employees who often don't receive employee benefits of the kind offered by Unum. The manager and assistant manager of a hotel or restaurant, for example, may be eligible for traditional employee benefits but the largely part-time wait staff often isn't covered by the same (or any) employee benefit plan. In essence, the present spike in unemployment is driven by factors that are somewhat different than in past recessions and is comprised of a different mix of workers. The consequence is that it's less likely that the earlier prediction models will quite as accurately reflect the potential impact on the company's premium income in the current cycle.
Investment income is projected based on an expectation that average yields will continue a slow decline although our projections do not include any incremental benefit from investments made while interest rates were higher in the last couple of years. We also have incorporated a slight acceleration in the decline in average yield. In light of these considerations, we actually think our projection may be a bit on the conservative/pessimistic side.
Ultimately, our projections indicate Unum will remain quite profitable despite the recession. The base case scenario indicates forward projected earnings per share around $4.30 (ranging from about $3.70 to $4.80 depending on inputs), with net income margin declining to around 7.5% to 8.0%.
Source: Winter Harbor Capital
The modified case, as discussed above, results in a lower earnings per share range. The forward projected earnings per share in this scenario is closer to $3.20 within a range of $2.90 to $3.30. The net income margin declines to around 5.5% roughly in line with earlier periods in which the company added to loss reserves.
Source: Winter Harbor Capital
Our projections are generally below those of sell side market analysts. We utilize our projections to develop a valuation for the shares in a moment.
Dividend
Unum's dividend yield - more than 6% - may suggest a level of risk to the company's dividend. The increase in statutory reserves and the additional cash flow commitment required to boost reserves will pressure the company's cash flow available for dividends and share repurchases. Unum's suspension of share repurchases, disappointing as some shareholders may find it given the differential between book value and market value, is a prudent and necessary decision.
Unum hedged somewhat in the last quarterly conference call and didn't provide a definitive statement in support of the current dividend. Instead, the company provided qualified statements such as "we anticipate maintaining our current shareholder dividend" and "our modelling suggests that we can maintain the current shareholder dividend…". The mild equivocation is understandable as the company has no way of predicting future reserve adjustments beyond the ranges the company already uses for its reserve calculations.
A deterioration in the company's reserve position with respect to long-term care insurance contracts could impact the company's ability to pay dividends. However, the deterioration would need to be somewhat larger than the anticipated statutory reserve additions. Our models suggest that while the company will be precluded from any meaningful repurchases of shares in the near future, the company should be able to sustain the present dividend - though with reduced margin - even should additional annual loss reserve accruals approaching the $350 million per year incorporated into our models prove necessary.
Unum's dividend is as a result not as secure as we'd prefer in light of the long-term care insurance reserves issues but also does not appear to us to be at imminent risk.
Valuation
Unum can be difficult to value despite the relatively straightforward nature of the business even excluding any consideration of the long-term care insurance reserve concerns. The present valuation, at less than half book value and a multiple of forward analyst earnings expectations below 4.0 suggests a great deal of concern about the company's loss reserves despite the recent determination by the Maine Bureau of Insurance.
Insurance companies, though, have tended to trade at exceptionally low valuations relative to book value and earnings in prior years much less under current conditions. In particular, life insurers have often traded at significant discounts given very low interest rates while other insurers specializing in certain segments of the insurance market have tended to trade at higher valuation multiples.
So, how to value a company in an environment where book value and earnings multiples are the most common metrics yet the industry and peers suffer from especially low valuation multiples? In part, it's a combination of an assessment of peers and an assessment of how reasonable those valuation multiples appear to be in the context of the business and associated risks.
A peer comparison valuation in the current market environment is not necessarily especially informative since we believe the insurance sector in general is rather undervalued at this time. In addition, the variety of business mixes within insurance companies, even those considered peers, makes a comparison more challenging as certain lines of businesses are less impacted by low interest rates or unemployment rates than others while many insurers comingle a broad line of insurance offerings under the same corporate parent. A pure benefits insurer or life insurer is difficult to come by as a comparison. However, even on peer metrics, Unum is likely undervalued at the present quotation.
Source: Winter Harbor Capital based on Company Financial Reports and Analyst Estimates
In contrast to the peer metrics, Unum's present valuation is roughly 38% of book value and roughly 3.5 times present analyst forward earnings per share estimates.
In addition, we considered both trailing and forward earnings estimates although there is not a significant difference in the results. The forward estimates considered were based on 2021 earnings estimates in order to provide a somewhat more "normalized" operating environment although extending the projection introduces an added measure of potential error in the projections.
We consider the earnings multiple comparisons to be more meaningful than the book value multiples since the market is heavily discounting Unum's book value (and for that matter most insurers' book values) based on an expectation that some portion of that equity will likely be consumed by reserve adjustments.
We also considered Unum's relative market position, the mix and nature of the company's employee benefit and other insurance offerings relative to peers, and operating metrics in refining our valuations. Unum's return on average equity in the last year, for example, was 11.8% including the impact of net investment gains/(losses) and 12.0% excluding the realized losses. The company's return on average equity in the most recent quarter, on an annualized basis, was 6.5% including net investment losses and 11.1% excluding net investment gains/losses. Unum's return on average equity is thus typically much higher than most of the peer companies and, absent the long-term care reserve overhang, would warrant a higher relative valuation.
Ultimately, we believe a fair value on a forward basis for the company's shares under the modified scenario falls within the range of $22.00 to $26.00 per share under conditions where the company will ultimately be required to increase loss reserves, on a GAAP basis, even beyond the increase in statutory reserves required by the Maine Bureau of Insurance. A valuation in this range would represent a premium of between 20% and 40% over the current market price, resulting in potential compound annual returns including dividends of between 10% and 20% over the next three years.
In the event long-term care loss reserves prove more adequate than anticipated by regulators, even if by a rather modest margin, our valuation range increases under the base scenario to $27.00 to $34.00 per share, depending on the magnitude of the differential and the potential for higher valuation multiples resulting from less concern about future loss reserve adjustments. In this case, compound average annual returns including dividends could well exceed 20% over the next few years.
Our base projection implies a forward return on average equity, including net investment gains/(losses), closer to 9.0% and somewhat higher excluding net investment gains/(losses) as we expect the company to incur losses in the near future. The return on average equity for our modified case, which assumes consistent annual additional loss reserve expenses associated with long-term care insurance liabilities, runs closer to 6.5% including net investment gains/(losses). The base case remains above those of peers while the modified case is somewhat below.
In any case, the company's shares need not even recover to their valuation immediately before the decline in the face of coronavirus related market volatility. In addition, given what we believe to be somewhat conservative projections for average investment yields and unemployment rates, we believe our earnings estimates and thus our valuations may underestimate results providing a margin of safety in the projections.
The downside risk is also ultimately dependent on the adequacy of the company's long-term care reserves. Credit Suisse's target price of $19.00 per share, established before the coronavirus pandemic based on an expectation that the company's long-term care reserves were even less adequate than established by either the company or the Maine Bureau of Insurance in consultation with independent actuaries and other state insurance regulators, probably requires some adjustment due to subsequent events. However, short of a catastrophic misjudgment of the adequacy of the long-term care reserves by multiple parties, it's difficult to develop a model which indicates a fair value below $15.00 per share. The downside risk in the share price at the current valuation, from our viewpoint, is much smaller than and by far disproportionate to our valuation projections.
In the meantime, Unum will continue to build on its book value of $49.28 per share with shares trading at less than 40% of the book value.
Risks
Obviously, there are risks to our assessments and projections, the largest being the uncertainty surrounding the actual long-term care insurance reserves necessary to support the company's policies. Our models indicate the company could absorb a certain level of additional reserves for losses without significantly impacting the company's cash flows or liquidity but there is a limit to the company's ability to do so while maintaining the current dividend. The reduction or elimination of the dividend, unfortunate though the outcome would be for the share price, would provide a much wider margin for adding to reserves. The long-term care policies have been discussed previously at length, so we will not repeat prior work here, but the level of reserves and impact of additional reserving will require ongoing attention.
In addition, while the company's historical operating metrics have been quite consistent over time it's possible that the current environment, which is unique, may result in these measures deviating from the long-term trends. A clear basis for expecting such a deviation is not apparent to us at this time unless the relationship between changes in average unemployment rates and changes in premium income substantially breaks down leading to larger than expected declines in premium income and additional deleveraging of expenses which tend to be less sensitive to changes in premium income. The same could be said for a sudden and unexpected move into negative interest rates.
The average annual unemployment rate, which is core to forecasting the impact on premiums, may also be somewhat less meaningful in the current context due to the unusual mix of jobs lost, the unusual spike in unemployment, and differences between the actual and expected trend over the remainder of the year. We actually consider it more likely than not that the rise in unemployment in the face of coronavirus will have somewhat less of an impact on premium income than in the prior cycle due to the mix of employment impacted by closures, but this assessment could be inaccurate.
Finally, it's possible the company could run into statutory accounting issues long before generally accepted accounting principles issues, limiting the company's ability to pay dividends to shareholders due to weakness in the statutory capital ratios. We also don't believe this is very likely at this time although it is not a risk which can be simply dismissed by shareholders.
Conclusion
Unum is not quite the bargain basement value opportunity that the company appears to be based solely on book value and earnings multiples. The potential for ongoing reserves charges associated with the long-term care insurance book of business certainly weighs on the shares in addition to the likely impact of higher unemployment and recessionary pressures due to the lingering effects of the coronavirus pandemic. On the other hand, it's possible that more sensitivity to potential short and long-term disability brought about by the coronavirus experience could boost certain of the company's insurance products in the future.
In any case, Unum historically has a very stable and profitable business through both positive and negative economic conditions and has built a franchise that has proven able to improve overall profitability despite the economic conditions. The company's core employee benefits business should prosper in the long term even if premium growth is incremental and modest, as one would expect from an insurance company. The economic headwinds, with the possible exception of long-term care insurance reserves, are quite manageable.
Still, Unum's shares will likely continue to trade at a discount to book value so long as there is substantial doubt about the ultimate future cost of the long-term care policies issued by the company. The result is that, barring an unlikely sale of the book of business by the company, a discount of some magnitude will persist for some time going forward. Nonetheless, the present valuation of around four times trailing earnings is unlikely to persist and even a modest gain to 8.0 times forward earnings would yield a share price closer to $25.00 over the next few years in addition to a decent dividend yield. However, if long-term care reserves prove more adequate than current market expectations, the future potential appreciation increases by an order of magnitude.
The consequent total compound annual return potential, including dividends, we foresee ranging from around 10% to well above 20% is attractive. In the longer term, we believe the company will return to a consistent if slow growth in revenues and profitability and, as the long-term care insurance reserve issues are resolved, will be rewarded with higher valuations on the core employee benefits business.
Unum thus represents a long-term holding in our investment portfolios.
This article was written by
Analyst’s Disclosure: I am/we are long UNM. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Seeking Alpha's Disclosure: Past performance is no guarantee of future results. No recommendation or advice is being given as to whether any investment is suitable for a particular investor. Any views or opinions expressed above may not reflect those of Seeking Alpha as a whole. Seeking Alpha is not a licensed securities dealer, broker or US investment adviser or investment bank. Our analysts are third party authors that include both professional investors and individual investors who may not be licensed or certified by any institute or regulatory body.
Recommended For You
Comments (21)
Otherwise very good analysis. I liked it. Thank you.
