Mark-to-Market (MTM)

MTM in theory makes perfect sense - it gives a sense of real-time measurement of value.

In practice, it is a completely different story. First of all, how relevant is MTM on assets when liabilities are not really marked-to-market. What do I mean? Let's talk about 2 aspects of liability MTM: interest rates and mortality tables.

Interest Rates: are the cashflows being discounted off a MTM curve? What curve should that be: a generic corporate bond curve, or could we consider that obligations of the pension are really obligations of the corporation and discount each corporation's pension liabilities off that corporation's credit curve?

While the latter makes intuitive sense (and creates more work), it results in the peversions we witnessed throughout 2008: financials in the US were recognizing huge gains from their liabilities when their credit spreads exploded.

With either choice we may be discounting liabilities using a bond curve that is not achievable for the asset managers in the real market due to duration, issuance size constraints or credit diversification constraints. Derivatives may be employed to circumvent some of these issues, which I will address in a separate post.

Mortality: Anybody who discusses mortality MTM is out of their mind. By its very nature these tables are constructed using time series data which eliminate outliers. This is best illustrated by an example: if too many people died in 2008 versus what we would have predicted by extrapolating historical death rates, we may want an extra few years of data before determining that a particular trend (say declining death rates) has come to an end. Or perhaps we are on the cusp of discovering a cure to a disease; how would we factor that into our mortality tables?

So, maybe MTM is not all it is cracked up to be. Have a look at this post on asset MTM.