Risk criteria
Risk criteria are the predetermined benchmarks against which an organisation assesses how significant a risk is and whether it is acceptable. They include scales for likelihood and impact, thresholds for acceptance and escalation, and the objectives, legal requirements and context against which a risk is measured. Risk criteria should be consistent with the risk appetite set by the management board.
Source: NIST CSRC Glossary (NIST SP 800-160v1r1; NIST IR 8286B-upd1); ISO 31073:2022; IIA Global Internal Audit Standards 2024 (glossary and Standard 11.5); Dutch Cybersecurity Decree (Cyberbeveiligingsbesluit) Article 7; Dutch Cybersecurity Act (Cyberbeveiligingswet) Article 24; Regulation (EU) 2022/2554 (DORA) Article 6; COSO ERM 2017 (principle 11)
Risk criteria determine which risks reach the board's agenda. Without fixed criteria, one unit may rate a risk as high that another sees as medium, and risks cannot be compared or aggregated. The Global Internal Audit Standards describe risk assessment as identifying and analysing risks, with significance typically assessed in terms of impact and likelihood. If the chief audit executive concludes that management has accepted a risk exceeding the risk appetite or risk tolerance, the matter is discussed with senior management and, if necessary, the board (Standard 11.5).
The concept comes from the ISO vocabulary. NIST takes its definition from ISO Guide 73, which was withdrawn in 2023; the current ISO vocabulary for Risk Management is ISO 31073:2022. As examples of such benchmarks, NIST mentions organisational objectives, the internal and external context, and mandatory requirements such as standards, laws and policies.
The concept also appears in Dutch regulation. The Dutch Cybersecurity Decree (Cyberbeveiligingsbesluit) requires most entities under the Dutch Cybersecurity Act (Cyberbeveiligingswet) to have a Risk Management policy that includes at least a Risk Management methodology and criteria for risk acceptance (Article 7(2)); for cloud and data centre providers, among others, Commission Implementing Regulation (EU) 2024/2690 applies instead. DORA requires financial entities to set a risk tolerance level for ICT risk in accordance with their risk appetite (Article 6(8)(b)). Risk criteria differ from risk appetite, which sets the types and amount of acceptable risk at organisation level, and from risk limits, which allocate that appetite quantitatively. The probability-impact matrix is the tool in which the criteria become visible.
Last updated: · Editor: ONE Risk Advisory
Frequently asked questions
What is the difference between risk criteria and risk appetite?
Risk appetite is the board's choice about the types and amount of risk the organisation is willing to accept. Risk criteria are the benchmarks used to assess an individual risk, such as scales for likelihood and impact and thresholds for acceptance. Good criteria are consistent with the risk appetite, so that a risk the criteria classify as unacceptable also falls outside the appetite.
What does the Dutch Cybersecurity Act require regarding risk criteria?
The Dutch Cybersecurity Decree, which elaborates the Dutch Cybersecurity Act (Cyberbeveiligingswet), requires a Risk Management policy with at least a Risk Management methodology and criteria for risk acceptance (Article 7(2)). The management board approves the cybersecurity measures (Article 24 of the Act). For cloud and data centre providers, among others, the requirements of Commission Implementing Regulation (EU) 2024/2690 apply instead of these provisions of the Decree.
Where does the concept of risk criteria come from?
From the ISO vocabulary for Risk Management. In its glossary, NIST takes the definition from ISO Guide 73, which was withdrawn on 2 November 2023; the current vocabulary is ISO 31073:2022. COSO ERM 2017 puts risk appetite at the centre; there, risks are prioritised by severity in the context of the risk appetite.