An interview I gave a few months back was used in a piece published at the Broker News Blog today. It's the final part of a very well written 6-part series on kaChing and other innovative startups in the financial services industry. Definitely worth reading if you want to learn more about what I do. Special thanks to Dan for a unique interview and a great write-up!
« February 2010 | Main | May 2010 »
March 2010
03/16/2010
03/04/2010
Life Partners Part III: The Model
In part I of this analysis, I introduced readers to LPHI and gave some background on the operating environment and competition. In part II, I dug into LPHI a bit more, focusing on some of the risks LPHI faces and the criticism it’s received from the investment community. In part III, I promised to share my model, but before I get started it pays to review some of the key stats just to get some perspective.
With these numbers, we can roughly estimate that with no revenue growth and future retained earnings invested exclusively in life settlements generating a 10% yield, LPHI shareholders will also see a 10% yield on their investment, half of which is coming right back to us in the form of a dividend every year. With that as a baseline, we can build a more complicated analysis.
Methodology
LPHI is difficult to model because it’s a service company with limited tangible assets. We can’t simply assume that they will continue to earn some rate of return on the stores they’ve built or technology they’ve patented. To continue making money, LPHI employees have to go to work every day and facilitate transactions. With revenue (and growth) so disconnected from any sort of tangible invested capital, it’s difficult to estimate what revenue will look like in the future. What we do know is that LPHI’s ability to sustain or grow revenue is a function of the size and wealth of the investor base and the availability of policies. We have a rough sense for the number of investors currently holding LPHI brokered policies, and we also have a rough sense for the growth rate of that investor base, but unfortunately, short of just extrapolating recent trends, that doesn’t do much to help us model continuing growth.
The most reasonable way for us to model LPHI is to assume that revenue will grow at a rate proportional to the yield end-investors see on life settlements. High returns on life settlements allow LPHI’s customers to invest more every year. While it seems unlikely that even the most satisfied investors will reinvest 100% of their life settlement payouts into new life settlements, strong returns will attract new customers to the life settlement industry, at least offsetting the reinvestment shortfall. LPHI also generates a return on it’s own retained earnings and investments in partnerships which benefit from higher yields as well. What’s good for the goose is good for the gander.
The simplest assumption I can make is that the revenue growth rate will roughly equal life settlement yields. In the most optimistic scenario I was willing to construct, I assumed a 15% yield on life settlements, the IRR LPHI targets during the underwriting process. On the top line, 15% growth in the face value of policies is less than half the growth LPHI has seen in all but one year since 2002. The result of this assumption is that management won’t hit their $1 billion in transaction face value per year until 2014. This makes me confident that even my most aggressive estimates are fair.
The Model
In all scenarios, I assumed LPHI generates revenue equal to 15% of policy face value. 15%, the fee rate earned in 2009, is the lowest rate earned since 2002. In the LTM, LPHI is back up to 20%. I then drop these numbers down to the EBIT line with fairly conservative but 100% variable (meaning costs vary as a function of revenue) margin assumptions. 48% goes to brokerage fees, 13% to G&A, and about 3% to settlement costs and premium advances.
Interest income is generated off cash, investment in securities, and investment in policies again at a rate equal to life settlement IRR. I’m guessing LPHI will slowly shift more retained earnings into life settlement policies as the cash balance grows. I also have LPHI generating a variable return on its investment in the partnership based on the leverage we saw in the 2009 10-K ($131mm in cash and life settlements and $110 in debt) and an assumed 7.5% interest rate. For example, if the return on life settlements is 10%, LPHI should earn approximately $969k on its 20% ownership. With this partnership investment on the books for $6.5 million, LPHI sees a 15% rate of return. Within the range of yields I’ve modeled, LPHI cannot take a loss on this portfolio in the long-term.
Using a discount rate of 15%, my model estimates EPS and the dividend in 2014, pays 6.67x the dividend per share, and some multiple of earnings dependent on growth. At a 15% growth rate, I would be willing to pay an additional 6.67x for earnings. At 10% I would be willing to pay 2/3 x 6.67, or 4.5x earnings (on top of the dividend). So at 15% growth, with a 50% dividend payout rate, I would be willing to pay approximately 10x earnings ($2 EPS = 6.67x earnings + $1 dividind x 6.67 = $20.01/share). At 10% growth I would be willing to pay 7.8x earnings. And as always, the price I would be willing to pay in 2014 is discounted back to today at 15%/year, plus the value of dividends paid out annually along the way. Needless to say these are cheap valuations, providing a wide margin for safety.
Valuation
Even having built this model myself, I don’t believe I’m accurately predicting what the future will hold for LPHI. That’s one of the reasons why I look at how a range of policy yields might affect valuation. More than predicting future earnings, my model has helped me understand the mechanics of how this business operates. But the uncertainty of future cash flows, and the timing of those cash flows concern me. To increase my comfort, I checked my valuations against the simplest valuation model of all, the Gordon Growth model.
The GGM equates the current share price to the coming year’s dividend divided by the discount rate minus sustainable long-term growth. This is technically an accurate model, but also highly sensitive to our assumptions, and it falls apart if the growth rate exceeds our discount rate. My model is technically flawed in many ways, but it produces a conservative valuation we can actually use in practice over a full range of expected growth rates. In no scenario does my model valuation exceed the GGM valuation.
The GGM valuations below use a 15% discount rate and the terminal growth rate is taken from 2014 of my model. The resultant valuations, as a function of life settlement yields, are charted below:
This chart shows just how sensitive my model is to assumptions with possible valuations ranging from $11.11 to $21.28/share. The GGM is even worse, ranging from $17.97 to $177.18, before the equation falls apart implying literally infinite value! As I mentioned above, the GGM is technically more accurate than my model, and it only falls apart because I failed to make assumptions about how the growth rate will slow down over time. It would be crazy to assume that any company could grow at even a 10% rate indefinitely. So in this sense, the GGM is probably the roughest of all models as even the lowest terminal growth rate is almost 10%. For me, the GGM isn’t useful as a valuation technique, and I don’t mean to imply that shares are worth anywhere near $177. It’s just a check against my own model.
Discussion
Trading at around $21 today, LPHI appears to be fully valued according to my model (ie shareholders should only expect a 15% return if life settlements also generate a 15% yield). I wouldn’t really call this a worst-case-scenario for valuation, but I do believe this is really the lower limit of where LPHI could trade. Especially considering that this company operates with no leverage, operating or financial, we would really need to see a drastic slow-down in business for LPHI to take a hit, and in the meantime it’s continuing to grow and generate cash. With reasons to be optimistic about the operating environment, I don’t think a slowdown in business is terribly likely.
One thing my analysis hasn’t considered is what LPHI might be worth if growth is front-loaded into the near-term future. For example, if LPHI continues to grow at 30% for the next 3 years, and then flattens out to only 5% growth forever, shareholders benefit enormously, and this is a real possibility. It’s been happening for the last decade and LPHI still has the securitization upside. Obviously these scenarios are sensitive to assumptions, so I’ll leave it up to my readers to play with my model or build their own and draw their own conclusions.
- Revenue has grown at a CAGR of 56.2% since 2002
- EPS has grown at a CAGR of 56.0% since 2003
- LPHI has a market cap of $300MM and an enterprise value of $270MM
- In the LTM, LPHI has generated earnings of $30.5MM, and FCF of $26.1MM
With these numbers, we can roughly estimate that with no revenue growth and future retained earnings invested exclusively in life settlements generating a 10% yield, LPHI shareholders will also see a 10% yield on their investment, half of which is coming right back to us in the form of a dividend every year. With that as a baseline, we can build a more complicated analysis.
Methodology
LPHI is difficult to model because it’s a service company with limited tangible assets. We can’t simply assume that they will continue to earn some rate of return on the stores they’ve built or technology they’ve patented. To continue making money, LPHI employees have to go to work every day and facilitate transactions. With revenue (and growth) so disconnected from any sort of tangible invested capital, it’s difficult to estimate what revenue will look like in the future. What we do know is that LPHI’s ability to sustain or grow revenue is a function of the size and wealth of the investor base and the availability of policies. We have a rough sense for the number of investors currently holding LPHI brokered policies, and we also have a rough sense for the growth rate of that investor base, but unfortunately, short of just extrapolating recent trends, that doesn’t do much to help us model continuing growth.
The most reasonable way for us to model LPHI is to assume that revenue will grow at a rate proportional to the yield end-investors see on life settlements. High returns on life settlements allow LPHI’s customers to invest more every year. While it seems unlikely that even the most satisfied investors will reinvest 100% of their life settlement payouts into new life settlements, strong returns will attract new customers to the life settlement industry, at least offsetting the reinvestment shortfall. LPHI also generates a return on it’s own retained earnings and investments in partnerships which benefit from higher yields as well. What’s good for the goose is good for the gander.
The simplest assumption I can make is that the revenue growth rate will roughly equal life settlement yields. In the most optimistic scenario I was willing to construct, I assumed a 15% yield on life settlements, the IRR LPHI targets during the underwriting process. On the top line, 15% growth in the face value of policies is less than half the growth LPHI has seen in all but one year since 2002. The result of this assumption is that management won’t hit their $1 billion in transaction face value per year until 2014. This makes me confident that even my most aggressive estimates are fair.
The Model
In all scenarios, I assumed LPHI generates revenue equal to 15% of policy face value. 15%, the fee rate earned in 2009, is the lowest rate earned since 2002. In the LTM, LPHI is back up to 20%. I then drop these numbers down to the EBIT line with fairly conservative but 100% variable (meaning costs vary as a function of revenue) margin assumptions. 48% goes to brokerage fees, 13% to G&A, and about 3% to settlement costs and premium advances.
Interest income is generated off cash, investment in securities, and investment in policies again at a rate equal to life settlement IRR. I’m guessing LPHI will slowly shift more retained earnings into life settlement policies as the cash balance grows. I also have LPHI generating a variable return on its investment in the partnership based on the leverage we saw in the 2009 10-K ($131mm in cash and life settlements and $110 in debt) and an assumed 7.5% interest rate. For example, if the return on life settlements is 10%, LPHI should earn approximately $969k on its 20% ownership. With this partnership investment on the books for $6.5 million, LPHI sees a 15% rate of return. Within the range of yields I’ve modeled, LPHI cannot take a loss on this portfolio in the long-term.
Using a discount rate of 15%, my model estimates EPS and the dividend in 2014, pays 6.67x the dividend per share, and some multiple of earnings dependent on growth. At a 15% growth rate, I would be willing to pay an additional 6.67x for earnings. At 10% I would be willing to pay 2/3 x 6.67, or 4.5x earnings (on top of the dividend). So at 15% growth, with a 50% dividend payout rate, I would be willing to pay approximately 10x earnings ($2 EPS = 6.67x earnings + $1 dividind x 6.67 = $20.01/share). At 10% growth I would be willing to pay 7.8x earnings. And as always, the price I would be willing to pay in 2014 is discounted back to today at 15%/year, plus the value of dividends paid out annually along the way. Needless to say these are cheap valuations, providing a wide margin for safety.
Valuation
Even having built this model myself, I don’t believe I’m accurately predicting what the future will hold for LPHI. That’s one of the reasons why I look at how a range of policy yields might affect valuation. More than predicting future earnings, my model has helped me understand the mechanics of how this business operates. But the uncertainty of future cash flows, and the timing of those cash flows concern me. To increase my comfort, I checked my valuations against the simplest valuation model of all, the Gordon Growth model.
The GGM equates the current share price to the coming year’s dividend divided by the discount rate minus sustainable long-term growth. This is technically an accurate model, but also highly sensitive to our assumptions, and it falls apart if the growth rate exceeds our discount rate. My model is technically flawed in many ways, but it produces a conservative valuation we can actually use in practice over a full range of expected growth rates. In no scenario does my model valuation exceed the GGM valuation.
The GGM valuations below use a 15% discount rate and the terminal growth rate is taken from 2014 of my model. The resultant valuations, as a function of life settlement yields, are charted below:
This chart shows just how sensitive my model is to assumptions with possible valuations ranging from $11.11 to $21.28/share. The GGM is even worse, ranging from $17.97 to $177.18, before the equation falls apart implying literally infinite value! As I mentioned above, the GGM is technically more accurate than my model, and it only falls apart because I failed to make assumptions about how the growth rate will slow down over time. It would be crazy to assume that any company could grow at even a 10% rate indefinitely. So in this sense, the GGM is probably the roughest of all models as even the lowest terminal growth rate is almost 10%. For me, the GGM isn’t useful as a valuation technique, and I don’t mean to imply that shares are worth anywhere near $177. It’s just a check against my own model.
Discussion
Trading at around $21 today, LPHI appears to be fully valued according to my model (ie shareholders should only expect a 15% return if life settlements also generate a 15% yield). I wouldn’t really call this a worst-case-scenario for valuation, but I do believe this is really the lower limit of where LPHI could trade. Especially considering that this company operates with no leverage, operating or financial, we would really need to see a drastic slow-down in business for LPHI to take a hit, and in the meantime it’s continuing to grow and generate cash. With reasons to be optimistic about the operating environment, I don’t think a slowdown in business is terribly likely.
One thing my analysis hasn’t considered is what LPHI might be worth if growth is front-loaded into the near-term future. For example, if LPHI continues to grow at 30% for the next 3 years, and then flattens out to only 5% growth forever, shareholders benefit enormously, and this is a real possibility. It’s been happening for the last decade and LPHI still has the securitization upside. Obviously these scenarios are sensitive to assumptions, so I’ll leave it up to my readers to play with my model or build their own and draw their own conclusions.
Disclosure: I hold LPHI shares in my kaChing portfolio.
Recent 10 Companies
Shipmates
Disclaimer: This blog is an outlet for my work. This is not an individualized recommendation. Please do not buy, sell, trade, or make any financial decision whatsoever based on the material presented here. If you have any questions regarding personal finance, please seek the advice of a professional who is trained to analyze your situation.