The volatility adjustment: from poisoned gift to dependence
Introduced to protect insurers from artificial balance-sheet volatility, Solvency II’s volatility adjustment has become an established component of European insurance capital management. A decade later, the question is no longer simply who uses it, but how much reported solvency depends on it.
When Solvency II entered into force in January 2016, the volatility adjustment was part of an uneasy compromise.
European regulators wanted insurance liabilities to reflect market conditions without allowing temporary movements in bond spreads to generate equally temporary deterioration in reported solvency.
The problem was particularly acute for life insurers.
Their liabilities could extend decades into the future, while the bonds held against them were exposed to movements in market prices. A widening of credit spreads could reduce the market value of assets even where the insurer intended to hold those securities to maturity.
Without an offsetting mechanism, such movements could create pressure to sell assets at precisely the wrong moment.
The volatility adjustment, or VA, was intended to reduce that risk.
But its introduction raised an uncomfortable question: could a mechanism designed to prevent artificial volatility become a permanent source of reported capital?
A regulatory compromise becomes part of the balance sheet
The mechanism operates through the discount rate applied to insurance liabilities.
By adding an adjustment to the relevant risk-free interest-rate curve, the VA can reduce the present value of technical provisions and consequently affect an insurer’s eligible own funds and solvency ratio.
The size of that effect depends on the insurer’s liabilities, asset-liability position and regulatory methodology.
It is not a simple capital subsidy. Nor is its impact identical across insurers.
Nevertheless, the mechanism creates a measurable difference between reported solvency and the counterfactual position calculated without the adjustment.
Solvency II disclosure requirements recognise this distinction explicitly. Insurers applying the VA must quantify the effect of setting it to zero on their technical provisions, capital requirements and own funds.
These disclosures provide an unusually direct way of examining the importance of a regulatory mechanism to individual balance sheets.
They also raise a question that becomes more interesting with time.
A temporary stabilisation measure may be straightforward to justify during exceptional market conditions. What happens when its use becomes part of ordinary financial management?
The lesson of 2020
The Covid-19 market shock provided an early test of the mechanism.
Credit spreads widened abruptly as investors reassessed economic risks and liquidity conditions.
For insurers holding substantial fixed-income portfolios, the widening affected asset valuations. At the same time, movements in the VA could partly offset the effect of spread volatility on the valuation of liabilities.
This was precisely the type of market environment the adjustment had been designed to address.
The experience also demonstrated how important the relationship between asset spreads, liability discounting and solvency capital had become.
An insurer’s reported solvency ratio could not be understood solely by examining the market value of its assets.
The behaviour of the liability discount curve mattered too.
The VA was no longer merely a technical provision in the Solvency II rulebook. It had become part of the explanation of how insurance balance sheets responded to market stress.
From using the adjustment to managing around it
Over the following years, insurers became increasingly familiar with the interaction between their asset portfolios, liability characteristics and the VA.
This created the possibility of a more deliberate approach to its management.
Asset allocation, credit quality, duration and asset-liability matching can all influence the economics of an insurer’s balance sheet. The regulatory adjustment adds another dimension to those decisions.
There is, however, an important distinction between optimising a balance sheet under a regulatory framework and deliberately maximising the benefit of a particular adjustment.
The former is an ordinary part of insurance capital management. The latter raises questions about incentives and the extent to which reported solvency remains sensitive to changes in regulatory methodology.
The most revealing measure is not simply whether an insurer applies the VA, but how its solvency position would change without it.
A company whose solvency ratio moves only marginally when the adjustment is removed occupies a different position from one for which the effect is substantial.
Both may be complying fully with the same regulatory framework.
Their sensitivity to that framework is nevertheless different.
A dependence that needs to be measured
The public data make this question unusually accessible.
SFCR disclosures provide the information needed to examine the effect of removing the VA from individual insurers’ reported solvency positions.
A systematic comparison over time could establish whether its use has become more widespread, whether the average capital benefit has increased and whether particular national markets or business models are more sensitive than others.
It could also distinguish between two phenomena that are often confused.
The first is widespread use: a large proportion of insurers applying the mechanism.
The second is dependence: the extent to which their reported solvency positions would deteriorate without it.
An adjustment can be widely used without creating substantial dependence. Conversely, a relatively small number of insurers may account for a disproportionate share of its aggregate capital effect.
That distinction is essential before drawing conclusions about the European market.
It is also why comparisons must separate the VA from transitional measures and the matching adjustment. These mechanisms have different purposes, regulatory conditions and balance-sheet effects.
The rules are changing again
The question has acquired renewed importance as Europe prepares to implement its revised Solvency II framework.
The new methodology for risk-free interest-rate term structures, including changes affecting the VA, will apply from 30 January 2027.
EIOPA has published updated technical documentation and calculation examples, while its revised supervisory opinion on internal models introduces additional considerations for the dynamic volatility adjustment.
The reform is therefore not merely an administrative update.
For insurers whose solvency positions are materially affected by the VA, changes in methodology may alter the relationship between asset spreads, liability valuations and regulatory capital.
The consequences will differ between companies.
Some may experience limited effects. Others may find that a familiar component of their capital management behaves differently under the revised framework.
The significance of the reform cannot be established from the regulation alone. It requires a comparison of balance sheets under the existing and revised methodologies.
The paradox of stability
The volatility adjustment was introduced to address a genuine problem.
Long-term insurance liabilities should not necessarily force institutions into procyclical investment decisions when financial markets experience temporary dislocations.
A mechanism that reduces artificial volatility can therefore support financial stability.
But the more deeply such a mechanism becomes embedded in capital management, the more consequential any change to it becomes.
The paradox is that a tool designed to make solvency less sensitive to market movements may increase its sensitivity to regulatory design.
A decade after Solvency II entered into force, the VA is no longer an unfamiliar technical adjustment.
It is part of the financial architecture within which European insurers operate.
The question for the next decade is whether a mechanism designed to protect insurers from market volatility has also made some of them dependent on the rules intended to contain it.
