The Solvency II approach and calibration form the base of our Solvency Assessment and Management (SAM) expense risk capital requirement calculation. The data used in the calibration came from a study by Watson Wyatt in 2004 on the 99.5% confidence level assumption over a 12 month time horizon for increases in the level of ICAS expenses in the UK. This study indicated a potential increase in the level of expenses between 5% and 50%, with an average increase of around 26%. Later UK studies showed about a 10% increase in the level of expenses.

In addition to the increase in the level of future expenses, there was also a need for the expense risk capital requirement to cover the risk of increases in expense inflation. For the Solvency II calibration, UK studies concluded on around a 1-2% per annum absolute increase in the expense inflation rate. This was in line with the 1% movement in nominal interest rates proposed for the interest rate risk component of the market risk module, assuming interest rates are relatively stable. (We can compare this to our Financial Soundness for Insurers (FSI) 4.1 requirement to “apply a minimum absolute change in interest rates of one percentage point (+1% in the upward stresses and –1% in the downward stresses)”.)

The SA Quantitative Impact Study (QIS) 2 draft (13 July 2012) proposed stresses of an increase of 10% in future expenses compared to the best estimate assumptions, and an increase of 1% per annum in the best estimate expense inflation rate assumption. At the time there were some discussions on whether the stresses proposed by QIS 2 were sufficient for South Africa. For QIS 3 the expense inflation shock was increased to the current formula set out in FSI 4.1, i.e. the greater of an absolute addition of 2% in the best estimate level of expense inflation and a 20% increase in the best estimate level of expense inflation. The shock to the level of future expenses was kept at 10%.

The current expense risk calibration assumes a macroeconomic environment where inflation is roughly under control and inflation targeting is successful. Arguably this is not always the case for South Africa. Consider for example the recent increases in inflation, not only in South Africa but worldwide, due to various global events. The expense risk component of the SCR may be inadequate for environments where inflation is not under control, driven upwards by expenses relating to staff costs, commissions, IT infrastructure, building occupation, etc. Under such circumstances, the shock percentages may have to be increased to provide adequate capital cover.

Additionally, concerns were raised regarding the calibration not taking the risk associated with lower new business volumes than expected into account. Lower new business volumes together with the natural attrition of existing business due to lapses, would lead to higher per policy expenses if fixed overheads are to be spread across a fewer number of policies. This may increase the expense risk going forward. It is suggested that in their Own Risk and Solvency Assessments (ORSAs), insurers perform stresses that include lowering the levels of new business.

Another part of the SCR where expenses come into play, is under the mass lapse stress. FSI 4.2 only allows for expenses directly linked to assets under management (AUM) to decrease in the first year. Realistically, one would expect some other expenses to also decrease if business volumes decrease. Examples are staff bonuses and outsourcing agreements where the cost is a function of, for example, number of policies or annual premium income. In this regard, a strict reading of FSI 4.2 may lead to capital requirements being on the prudent side.

The expense risk component of the SCR may be inadequate for macroeconomic environments where inflation is not under control. Additionally, the calibration does not take into account the risk of new business volumes being lower than expected and as a result the expense risk may be understated (especially if the insurer has large, fixed overheads). Insurers should assess the risk of lower new business volumes in their ORSAs.