Risk theory
Risk theory: 3 technical terms explained – definition, synonyms and legal basis.
Central Limit Theorem
Synonyms: Zentraler Grenzwertsatz
The central limit theorem states that the distribution of the standardised mean of a large number of independent, identically distributed random variables can be approximated by the standard normal distribution, and it underpins risk pooling within a collective.
Statement of the theorem
For an independent sequence of random variables with identical distribution, finite expected value μ and finite standard deviation σ, the distribution of the standardised sample mean converges to the standard normal distribution as sample size grows – regardless of the original distributional shape of the individual variables.
Application in insurance
In insurance, the central limit theorem provides the theoretical basis for risk pooling within a collective: the larger the number of independent, similarly distributed risks in a portfolio, the better the aggregate loss distribution can be approximated by a normal distribution. This considerably simplifies the calculation of safety loadings and ruin probabilities.
Limitations
The approximation requires independence and roughly similar distribution of the individual risks. For accumulation risks – for example natural catastrophes affecting many policies simultaneously – this assumption is violated, requiring alternative models such as copula approaches or simulation.
Random Fluctuation Risk
Synonyms: Process risk
Random fluctuation risk is the component of an insurer's technical risk arising from the random dispersion of actual losses around the calculated expected loss, and it cannot be eliminated even with correctly estimated parameters.
Distinction
A portfolio’s technical risk can be decomposed into several components: random fluctuation risk (dispersion around a correctly estimated expected value), parameter risk (misestimating the parameters) and change risk (subsequent shifts in the loss distribution, for example through inflation or legal change). Even with full knowledge of the true distribution, there remains a positive probability that the periodic aggregate loss exceeds the funds available from risk premium and risk capital (technical ruin).
Quantification
Random fluctuation risk is typically measured through the probability of loss or ruin. Unlike parameter and change risk, it can be estimated relatively well using probabilistic methods, for example through the variance of the aggregate loss distribution across the collective.
Practical relevance
Random fluctuation risk decreases as portfolio size grows (law of large numbers) but can never be fully eliminated. It materially determines the required safety loading on the net premium and the risk capital that must be held.
Random Risk (Loss Variable)
Synonyms: Loss random variable
In risk theory, random risk is the non-negative random variable describing the loss amount of an insured risk, forming the basic mathematical building block of actuarial models.
Formal definition
In risk theory, a risk is formally modelled as a random variable X with P[X ≥ 0] = 1: X describes the loss amount that a given insured risk may cause within a period, excluding negative values. This definition is the starting point of the individual and collective risk models used in actuarial science.
Use in models
Building on the random risk, distributional assumptions (for example Poisson, gamma or Pareto distributions for claim frequency and severity) are made, from which the aggregate loss distribution of a portfolio can be derived. That distribution in turn underpins premium calculation, reserving and capital requirements.
Practical relevance
Accurately characterising the distribution of the random risk – particularly in the tail – is decisive for setting deductibles, probable maximum losses (PML) and reinsurance structures. Misjudging the underlying distribution is among the most common causes of under-reserving.