Simˈpōzēəm
The Solvency II Extrapolation has become a central element of the Long-Term Guarantees package. But what started as a purely technical concept to value ultra long-dated liabilities was hijacked and has become a tool for managing pro-cyclicality.
Introduction
An economic extrapolation is a method of using known market values to estimate, or ‘extrapolate’, future values for which there is no accurate market data. In Solvency II it is used to smooth market volatility at the long-end of the interest rate curve by providing a more stable regulatory risk-free rate against which insurers’ long-term liabilities can be discounted.
Origins of macro-economic extrapolation
The Solvency II Directive text (Article 76) requires that technical provisions are consistent with market prices and financial market data. Extrapolation was originally proposed only as a theoretical technique to place a market-consistent value on a non-traded item, namely ultra-long-dated cash-flows past the term of any traded instruments, for example after 50 years in Euro (EUR). Following the global financial crisis, stakeholders started to question the reliability of market data and the potential impact of market-consistent valuation on macro-economic stability. Extrapolation was identified as a possible tool to manage pro-cyclicality. [caption id="attachment_119108" align="alignright" width="389"]
Extrapolation in Solvency II
In July 2009 CEIOPS (EIOPA’s predecessor) issued advice on the risk-free rate (Advice from CP40) in which it acknowledged the pro-cyclical effects of volatile long-term rates. It required that, next to a requirement for “realism”, extrapolation techniques should be used to manage the “effect on financial stability”. [caption id="attachment_119113" align="alignright" width="498"]
Finding the Last Liquid Point
At this stage the extrapolation debate was still regarded as technical, albeit with an expanded objective. However, as I previously observed, over time the Solvency II debate became increasingly political. The industry, particularly in Germany, pushed for a shorter-dated LLP – ideally 10 years – and much quicker convergence. [caption id="attachment_119110" align="alignleft" width="233"]
Convergence of extrapolation techniques
As interest rates on core European government bonds fell further during the sovereign debt crisis, the Dutch, Danish and, most recently, Swedish regulators all adopted Extrapolation for current Solvency I reporting purposes. Inspired by this regulatory endorsement, insurers have used the UFR for the calculation of MCEV, where it has been adopted by more than half of listed insurers. The Smith-Wilson technique and the parameters proposed for Solvency II are also being widely adopted as generally accepted, if not necessarily best, practice. The market impact of adopting these methods has so far mainly been to simply transfer pro-cyclical distortions on the curve from the ultra-long end to around the key 20 year entry point – as noted above this is a relatively artificial choice of tenor. Interestingly the Swedish and Dutch regulators have both recently challenged aspects of the Solvency II Extrapolation, including in the Netherlands challenging the fixed UFR.Choice of UFR – lessons from history
Surprisingly there has been little debate over the fixed 4.2% UFR despite this effectively being a “collective bet on long-term economics”. CEIOPS derived 4.2% from the sum of a long-term expected inflation rate of 2% and expected short-term return on risk-free bonds of 2.2%. It is questionable whether this choice of UFR is sustainable. CEIOPS based its assumptions on the previous 15-20 years of inflation targeting by central banks, rather than the high inflation rates of much of the 20th Century. But can we rule out further fundamental changes to monetary policy in the next 15-20 years?1st lesson from history: USA 1994
Testing the proposed Solvency II Extrapolation against historical scenarios demonstrates how exposed it can leave insurers. In the USA we saw a scenario of a sudden rise in rates caused by even small changes to monetary policy in 1994 when, following a prolonged rate cutting cycle, the Fed implemented a surprise 25 basis point rate hike. This was intended to signal its determination to control inflation and anchor long-end rates. Unfortunately the rate hike had the opposite effect, causing long-end rates to rise by 200-300bps in the US, Europe and the UK. [caption id="attachment_119111" align="alignright" width="233"]
2nd lesson from history: Japan
The chosen UFR is deliberately higher than current market rates, but this bases regulation on the implicit assumption that current yields are artificially low and actual realised forward rates will exceed market rates. The lesson from the Japanese scenario is that this may be a dangerous assumption. [caption id="attachment_119112" align="alignright" width="233"]