Historical Combined Ratio Trends

[bsilly]

I have been looking at some long term valuation metrics for Fairfax, one of which is the combined ratio. I plotted the reported combined ratios since inception - 1985, and then overlaid a ten year trailing average. I like the ten year average because it eliminates the noise of annual fluctuations and will also tend to average out the ups and downs of insurance cycles (it will include a few hard market years & a few soft market years). Note that the arithmetic average differs from the weighted average. A weighted average will tend to skew results towards years of higher net premiums written. Accident year results are also an excellent way to look at reserving, as Prem did in the most recent annual letter, but I am looking at it differently - on an as reported basis, which includes reserve deficiencies/redundancies in the year reported. Over 10 years, issues of over/under reserving should more or less come out in the wash.  This year is of note, because, as you can see from the image below, Fairfax' ten year average 'as reported' combined ratio, barring a major disaster between now and year end, will go below 100 for the first time ever. I think this is an interesting indicator. The trend reflects the effect of reserve surprises, and the fact that Fairfax has gone from a company that tended to report reserve deficiencies (negative surprises), to one that is tending to report redundancies (positive surprises). You can see that after 2001, there has been a steady long term trend towards stable underwriting profitability. Part of this is due to industry dynamics. Even though I did not plot industry averages, I think the trend would compare favorably. Fairfax is improving it's underwriting results relative to the industry. I now feel justified, perhaps for the first time (and I have owned and followed Fairfax for a long time), in projecting average combined ratios for Fairfax of 100 or better going forward. Comments?

 

http://i.imgur.com/4xRm1Pw.jpg

[Jurgis]

I wonder how much this is due to the fact that we had very few cats in recent years. You average over 10 years, but last 4 years are still .4 of the weight...

 

OTOH, few cats kinda equals soft pricing, so it's still good that they dropped below 100. :)

[bsilly]

Low cat activity the last few years is definitely part of it. And in the late 90's to 2001 we saw a soft market punctuated at the end by massive losses from 9/11 - historically bad years. We also had falling investment yields since then.

 

I'd say generally that industry wide trends are behind most of the decline. But I did manage to cobble together some industry data that supports my assertion that Fairfax' average 10 year combined ratio has come down relative to the industry. In 2001 FFH had a 10 year trailing average CR of 110, versus industry average of about 107. This year, it looks like FFH will be 99 versus 100 for the industry. So it went from an industry laggard to slightly better than average. I may try to plot it up for y'all when I get some more time.

 

With investment yields where they are, I think the industry will write around 100 or below for the for the foreseeable future, like they did in other low yield eras such as the 1940s and 50's. I looks to me like FFH will write a least as well if not better. I don't see the trend reversing.

[bsilly]

As soon as I posted that, I found the following presentation from Munich Re. Slide 44 somewhat contradicts what we thought about Cat losses. Last 2 or 3 years are maybe a bit lower than the previous few, but on a ten year average - not a huge effect.

 

http://www.munichre.com/site/mram/get/documents_E-1959049670/mram/assetpool.munichreamerica.wrap/PDF/2014/MunichRe_III_NatCatWebinar_01072015w.pdf

[A_Hamilton]

As soon as I posted that, I found the following presentation from Munich Re. Slide 44 somewhat contradicts what we thought about Cat losses. Last 2 or 3 years are maybe a bit lower than the previous few, but on a ten year average - not a huge effect.

 

http://www.munichre.com/site/mram/get/documents_E-1959049670/mram/assetpool.munichreamerica.wrap/PDF/2014/MunichRe_III_NatCatWebinar_01072015w.pdf

 

This understates how small the losses were though as you've had price declines in each of the last several years, so smaller denominator in your combined ratio. Meaning revenue has declined but cost in combined ratio points has also declined suggesting that actual $'s of catastrophe loss are down dramatically as well.

[Phoenix01]

Fairfax is improving it's underwriting results relative to the industry. I now feel justified, perhaps for the first time (and I have owned and followed Fairfax for a long time), in projecting average combined ratios for Fairfax of 100 or better going forward. Comments?

 

Thanks for the excellent post.

I think a big portion of the increase is due to Andrew Bernard.  He has done a great job with the insurance side of the business.