2015-12-31

Added · Updated

Applying the volatility adjustment in internal models

Insurers using internal models under Solvency II may apply either a constant or dynamic volatility adjustment, provided they satisfy specific modelling, risk management, and disclosure conditions. DNB assesses these applications case-by-case, requiring insurers to justify estimates, model sovereign risk exposures fully, and validate their approaches against granular source analysis. Insurers must also demonstrate that their risk management includes incentives for good practice and report the sensitivity of technical provisions and own funds to volatility adjustment assumptions.

De Nederlandsche Bank logo

Netherlands

De Nederlandsche Bank

Click to view thumbnail

Q&A

Read aloud

Question:

Should insurers use a volatility adjustment (VA) in their internal models that does not change over time, or are they allowed to model a dynamic VA?

Published: 31 December 2015

Latest update: 23 December 2022

Answer:

When using an internal model to calculate the Solvency II capital requirements, insurers must estimate their technical provisions one year after the reporting date. They may apply a VA in the relevant interest rate term structure, which must then be modelled in an internal model. DNB allows the use of a constant as well as a dynamic VA for modelling purposes. However, for dynamic a VA, which moves with the relevant spreads, certain conditions apply.

Naturally, when an insurer asks DNB for permission to use an internal model, DNB will assess, on a case-by-case basis, whether that insurer satisfies all requirements and applies all good practices with regard to internal models as listed in the Solvency II Directive.

Concerning the VA, DNB will generally assess whether all risks and uncertainties surrounding the modelling, the source data and the parameters relating to the VA and the resulting capital requirement (SCR) are adequately addressed in the modelling process and risk management. When applying for permission to use an internal model, an insurer must demonstrate that it has explicitly addressed these risks and uncertainties or that it has, at least, implicitly done so, for example by limiting the VA (or its application ratio). It must address all model components in relation to the VA.

In assessing the VA, DNB will also consider the modelling of risks arising from government bonds (sovereign exposures). An insurer must make allowance for sovereign exposures in its internal model, taking into account how it has modelled the VA. If an insurer opts to apply an internal model with a dynamic VA, it must incorporate all sovereign risk exposures and other spread-sensitive investments in full.

In particular, DNB will assess whether an insurer has given appropriate consideration to the following issues in its modelling, risk management and public disclosure in relation to the VA.

Modelling

The insurer must justify and quantify the estimates it applies in the VA expectations under the various scenarios (including the credit risk adjustment – CRA).

The insurer must adequately include in its modelling all elements of a dynamic VA as required by the internal model's use test and statistical quality standards and, more particularly, the following elements:

the correlations used between credit risk and other risk factors in the various scenarios, such as spread narrowing and widening; and

the representative portfolio of bonds and other fixed-income securities and the reference portfolio of market yield indices for calculating the VA, allowing for differences compared with its own portfolios and possible changes to them (flight to quality) that might occur in exceptional circumstances.

Risk management

The insurer must adequately demonstrate that its VA modelling includes incentives for good risk management and explain how it considers these incentives in its risk management.

The insurer must consistently consider, as part of its risk management, the risks arising from its own investments and from the representative portfolio and the reference portfolio used to calculate the VA.

The insurer must validate its VA modelling on the basis on a sufficiently granular source analysis, including quantification of the difference between a dynamic VA and a constant VA.

The insurer must consider a change to the modelling of the VA as a major change within the meaning of Article 114(2)(b) of the Solvency II Directive.

In managing its risks, the insurer must explicitly take into account the situation where reducing the VA to zero would result in non-compliance with the SCR, and the insurer must have an analysis of the measures it could apply in such a situation to re-establish the level of eligible own funds covering the SCR or to reduce its risk profile to restore compliance with the SCR.

The insurer must assess and report the sensitivity of its technical provisions and eligible own funds to the assumptions underlying the VA calculation and the possible effect of a forced sale of assets on its eligible own funds.

Public disclosure

In its public disclosure, the insurer must include a statement on:

the impact of the VA, including its effects on the level of the technical provisions, the SCR, the MCR and the eligible own funds; and

the method and underlying assumptions used to calculate a dynamic VA.

Pursuant to Article 37(1)(d) of the Solvency II Directive, DNB may consider to set a capital add-on if it concludes that the insurer's risk profile deviates significantly from the assumptions underlying the volatility adjustment.

Base law

Richtlijn 2009/138/EG (Refers to an external site)

Related legislation and regulations

Article 37 (Capital add-on), Solvency II Directive

Article 44 (Risk management), Solvency II Directive

Article 51 (Public disclosure), Solvency II Directive

Article 114 (Changes to internal models), Solvency II Directive

Article 120 (Use test of internal model), Solvency II Directive

Article 121 (Statistical quality standards for internal model), Solvency II Directive

Discover related articles

Q&A

Capital

Capital requirements

Insurers

Share:

Share on LinkedIn

Share on X

Share on Facebook

Share via Email

Interesting articles

Dutch insurers and pension funds have been investing more in private assets in recent years

15 July 2026

News item supervision

Dutch insurers and pension funds are investing more and more in private assets, such as private equity and private credit. By 2025, they had a combined total of €276 billion worth of these investments on their books.

Read more Dutch insurers and pension funds have been investing more in private assets in recent years

News item supervision

15 July 2026

De Nederlandsche Bank publishes ‘Integrity Supervision in Focus 2026’

25 June 2026

News item supervision

In the third edition of ‘Integrity Supervision in Focus’ (ISF), we share the key insights from our integrity supervision.

Read more De Nederlandsche Bank publishes ‘Integrity Supervision in Focus 2026’

News item supervision

25 June 2026

DNB email on technical adjustments

25 June 2026

News item supervision

This week, you may receive an email from De Nederlandsche Bank (DNB). This email concerns technical adjustments required to continue corresponding with DNB by email.

Read more DNB email on technical adjustments

News item supervision

25 June 2026

Update FATF-warning lists June 2026

23 June 2026

News item supervision

FATF released an update of its ‘grey’ and ‘black’ lists.

Read more Update FATF-warning lists June 2026

News item supervision

23 June 2026

Necessary cookies

To ensure the proper operation of the website, De Nederlandsche Bank (DNB) uses functional cookies and analytics cookies, and has taken measures to ensure that these cookies have little or no impact on the privacy of website users.

Optional cookies

Some pages include embedded content from external websites. These websites may use proprietary (tracking) cookies. This allows third parties to track visitor statistics, show personalised content and display targeted ads, for example.

You can make your choice about allowing these optional cookies both when you first visit the website and when you navigate to a page with embedded content.