I need a better title for this post but I can’t come up with one, so hopefully you all will catch on to the purpose shortly. What I am attempting to do is share some insight into how I used to do my financial analysis different from most other investors. I figure it’s been sufficiently long enough that I can now start sharing a few of my old tricks.
Also, as an aside, I shared some brief thoughts on the Marsh corporate name change here for those who may have missed it.
Setting the Stage
Let me start with a simple observation. Let’s assume a reinsurer has an expected cat load of $300M for third quarter. When we have a year like this one where there are no hurricanes, we can all agree earnings go up by $300M (let’s call it $250M a/t).
Most analysts stop there. They increase book value by $250M and call it a day. The point of this note is to explain why that’s lazy analysis. Make sense?
The Missing Cash
If you stop your analysis at raising your earnings by the amount of the cat load, you have left a lot of cash on the table. Literally!
That $250M we added to book value is cash the company doesn’t have to pay out to customers. That’s pretty important!
If you look at a financial model for an insurer, most analysts estimate investment income by looking at the total investments and multiplying by their best guess of the portfolio yield (usually something very close to last quarter’s result).
However, few of those analysts try to actually model out how the invested assets will change over time. Sure, they have it grow over time, but usually at a similar clip each year, ignoring what the actual cash flows in or out might be.
For example, if the combined ratio for the year is exactly on target and investment yields didn’t change much, most people would expect their investment income forecast to also be on target.
However, that’s a pretty bold assumption as it ignores a major driver of invested assets – claims paid. If the combined ratio was as expected, but paid losses were elevated, then invested assets would be lower than expected and so would investment income.
Believe it or not, most analysts only loosely tie out paid losses to invested asset projections (if they do it at all). This takes us back to the missing $250M of cash.
If earnings were higher because cats were lower, that means claims paid will be lower in the coming quarters by that $300M of avoided cat losses. This means invested assets will be higher (minus the cash tax paid on the higher earnings).
Thus, going back to your Accounting 101 T account, the book value increase of $250M is matched by an invested asset increase of $250M.
You would think this is elementary stuff, but you would be surprised at how few analysts model balance sheets and cash flow statements. Most only model the earnings and then fudge the balance sheet.
Reinvesting the $250M
You might think this is minor stuff, but it adds up over time. For simplicity, let’s assume the company reinvests the $250M at 5% pre-tax, so 4% after-tax. That’s an extra $10M/yr of earnings out into the infinite future.
We can debate how to value that (put a P/E on it, discount the value of an annuity, etc.) but let’s for argument’s sake say this company had $10B of surplus and $1B of “normalized” earnings.
Thus, a $10M annual earnings increase is only a 1% increase in earnings and thus worth 1% to the stock beyond the 2.5% increase in book value ($250M/$10B). Still, as a percentage of the total gain in value, those investment earnings are almost a third of the contribution of the absent cats alone.
Alternative Uses Of Cash
But you don’t have to keep the windfall cash in boring bonds. These extra earnings are, by definition, excess capital. A company can deploy them at higher returns as it sees fit.
The easiest way to show this is to assume it can write more cat business the next year. If it deploys an additional $250M of capital at say a 12% a/t return, it now has produced $30M of extra annual earnings, rather than the $10M from the bond portfolio.
Thus, recurring earnings have increased 3% and the value creation from deploying the capital now exceeds that from the increase in book value from the one-time gain.
Now, you might push back and say “but if all the insurers used their one time gains to grow their cat books, it would depress pricing and lower the returns below that 12%”. And you’d be correct.
That’s why my default assumption was always to invest excess capital into stock repurchase. Let’s assume the company trades at 1.25X BV, so the market cap is $12.5B. A $250M repurchase would reduce the share count (and increase earnings) by 2%.
So, it’s not as valuable as reinvesting in the business, but as you, the astute reader, noted, it may not have been practical to find $250M of organic growth without depressing pricing.
What About When Cats Are Worse?
While in the scenario I described today, paid losses were better due to the lack of cats, don’t forget this works in both directions!
When there are larger than expected cat losses, few analysts raise their future paid losses sufficiently and thus lower their future investment income forecasts. They additionally forget about the impact on excess capital and often don’t lower their future repurchase activity enough (more accurately, they’ll wait for management to tell them that they won’t be doing repurchase for the next X quarters while capital cushion rebuilds).
Now, sometimes they luck into the right answer because the reduced repurchases come close to the higher paid losses, so the investment income doesn’t change much, but this is coincidence rather than good analysis.
Always Tie Off Loose Ends
What’s the lesson in all this? Most people are sloppy! Financial Modeling 101 teaches you that your income statement should always tie out to your balance sheet and cash flow statement. Yet, few practitioners actually do it that way.
I always made sure changes in one area tied out to the other statements. And no, not with circular loops! I could write a whole post on that topic, but most analysts who lazily attempted to “tie out” the balance sheet did it through circular loops which led to errors that were hard to catch.
Other lessons…paid losses matter! Very few analysts pay attention to paid loss trends, let alone paid/incurred trends. Unexpected changes in paid trends were one of the best sources of information in the entire financial statements.
Not only did it help you better model investment income, but it also gave you an early warning system on reserve problems. Often, I would be the only person asking management why the paids looked poor and that would unlock information about things that had gone wrong on the underwriting side.
More telling was when management couldn’t answer why the paids were worse. That was always a great confirmation that there was an upcoming reserve problem, especially if I had other reasons to be skeptical.
Finally, it is important to understand how all this fits into capital adequacy. For some companies, higher than expected cats just means reduced repurchases for a year. For others, it means they need to buy more reinsurance, or even issue capital, which dilutes earnings beyond the simple math I did here.
If everything I described today sounds pretty basic and obvious, well, it should be. This isn’t rocket science. It’s just attention to detail! But you’d be surprised how few people actually make the effort to get the detail correct.
mbellinger@orangeinsure.com