Lender's Mortage Insurance (LMI) provider Genworth Mortgage Insurance Australia (ASX: GMA) is a company I've been interested in for some time. It's now trading at a discount to its Net Tangible Assets (NTA) of $3.86 per share, and I wrote about the company previously here.
Further investigation and yesterday's half-year results both ameliorated some of my concerns, and raised new ones. First, readers should understand that Genworth insures the lender, someone like Commonwealth Bank of Australia (ASX: CBA), against borrower defaults on their CBA loan. In the event of default, the lender can repossess the house, sell it, and if it is unable to recoup their losses, Genworth will cover the difference.
This means that Loan to Valuation Ratio (LVR) is crucial. A house with a 90% LVR (e.g. the loan is $450,000 and the house is worth $500,000) could easily fall short of its full value in a forced sale, resulting in Genworth covering the difference. This means the LVR of the houses that Genworth insures is very important. The composition of Genworth's insurance portfolio is also important – e.g. is it lending to owner-occupiers, or investors? Are there many 'low-doc' loans? Let's take a look:
*IIF = Insurance In Force, i.e., policies that are currently active
As we can see, 74% of Genworth's insured homes are owner-occupied (sometimes perceived to be lower risk than investment homes) and 26% investment properties. Just 5% of the properties are 'low doc' loans, which appears a fairly insignificant percentage, but depending on its composition it could potentially pack a wallop if they all go bad.
Yesterday's presentation was chock-full of quality information for investors:
We can see that Genworth writes a lot of high-LVR insurance – because lower LVR loans don't need as much Lender's Mortgage Insurance. We can also see there are super long 'tails' on Genworth's policies- with some policies dating back to before the GFC. We'll get to those in a sec.
Now, Genworth's weighted average LVR right now is around 53%. This is pretty reasonable, and allows for significant wiggle room when mortgages go delinquent and are forcibly sold. Earlier loans, especially pre-2007, have even lower LVRs – of 33%. Genworth has been writing larger amounts of higher-LVR loans in recent years though, which could be a concern. Now, another chart – the last one!
For me there are a couple of takeaways from this chart, first of which is that the long 'tail' (policies written pre-GFC) appear pretty safe both in terms of low delinquencies and low LVRs. Second is that delinquencies take some time to peak (around 3-4 years), and peak delinquencies from policies written in the past four years are less than half of the level of delinquencies seen during the GFC. In between lower interest rates and the relatively prosperous post-GFC years, I wonder if the recent level of delinquencies is below 'normal' levels.
So What?
All in all, Genworth's whole portfolio looks pretty good. I won't force any more charts on you, but there are other ones that show the size of the average delinquency claim (about $75,000) and other important information such as delinquencies by state, with Queensland and Western Australia the worst performing.
That said, Genworth is a small company and there are a couple of concerns. The company's comments about rising delinquencies in mining areas suggest a different pattern of claims – likely the houses are worth only a small fraction of their boom-time value, leaving Genworth with much bigger bills. This differs to areas like Sydney where houses hold their value better. As we saw yesterday, it doesn't take a lot of bad loans to hurt profits.
Likewise Genworth's ability to recover the LMI value from mortgage holders has got to be pretty slim – after all,people whose mortgages go 'delinquent' (the industry word for it) lose their house.
The other thing to note about Genworth's low average LVR is that – as the graphs above show – it is generally not the individuals with the lower LVRs that have trouble making their repayments. When people cannot meet their repayments, as long as property values hold reasonably constant, it is the higher LVRs that are more likely to result in a claim on Genworth.
The big question is, how much of the portfolio is likely to go bad?
This is hard to answer.
Genworth also commented that it was in negotiation with its second-largest customer over whether that customer would continue to use Genworth's LMI services or find an alternative (such as self-insuring). With losses due to delinquencies expected to move even higher over the full year, the departure of this customer could have a big impact on the amount of business Genworth writes. That could smash profits, dividends, and the share price. At the very least I would not buy until this situation becomes clearer.
Foolish takeaway
The reduction in the amount of high-LVR business written by the banks recently has hurt Genworth, no question, but it is also likely to result in higher quality insurance policies and a better claims ratio over the long term.
Yet it's the next 2-3 years that will prove a crucial time for the company, as this is when delinquencies from policies written in the past four years will be peaking, and potentially there could be interest rate rises as well. If there was a lot of loose lending going on in those years, we will certainly find out about it.
The worst case scenario for shareholders is pretty bad as Genworth has 1.46 million policies in force, worth about $270 billion, while it has $2.2 billion in assets. I find the company intriguing, and reckon there could be an opportunity in there sometime in the next two years. The risks are significant, however, and this is one of those times when it is far better to be too late to the party than too early. I'm staying on the sidelines for now.
