Faire la différence entre ces terminologie :
- Risk Classes
- Risk Characteristics
- Risk Classification System
- Risk Classes : Group of risks with similar characteristics that are expected to have similar insurance costs. Grouping risks this way allows insurers to apply consistent pricing across the group, which is why it’s sometimes called class rating. For example, one class might include drivers between the ages of 20 and 29, if age is shown to affect the likelihood or cost of claims
- Risk Characteristics : Is a measurable or observable factor that influences the expected cost of a claim. These characteristics are used to assign each risk to an appropriate risk class within a classification system. Examples may include : age, location, vehicle type, or prior claims history
- Risk classification system : a risk classification system is a framework used to assign individual risk to risk classes based on their risk characteristics. The goal is to group risks in a way that reflects differences in expected costs, considering both frequency and severity
Décrire le process de la spirale of adverse selection
- Departure : Low-risk insured realize they are paying more than their fair share and begin to leave the risk pool. With fewer low-risk insureds, the average risk level of the pool increases
- Premium Increase : The insurer raises premiums to account for the higher expected claim costs
- Cycle : The premium increases cause even more insureds to drop their coverage, and continue to deteriorate until high risk insureds remain
- Death : Eventually, premium reach unsustainable levels. The insurer may be forced to with draw the product or exit the market entirely
What is one of the ongoing challenges with adverse selection
Lag between spotting the issue and getting a fix in place. This can be slow down by :
- Time needed for management to review and make decision
- Waiting for regulatory approval
- Delays in IT making system changes
Décrire les 7 considérations lors de la sélections de caractéristiques de risque
- Relationship of Risk Characteristics and Expected Outcomes : clear link to expected claims, supported by data and expert judgment
- Causality : important for public acceptance of risk classification, but proving a cause-and-effect relationship between a rating factor and expected costs is often difficult or impossible
- Objectivity : Risk characteristics should be clear, objective, and based on observable, verifiable facts that can’t easily be manipulated. Should be structured so that the risks fit only one class
- Practicality : how easy a characteristic is to measure, the cost and effort asses and maintain it, and how acceptable it is to the public and regulators + verifiability
- Applicable Law (statutes, regulations, or accounting standards)
- Industry Practices : the practices often align with legal requirements
- Business practices : IT capabilities, staffing, financial resources
Nommez les 4 data considerations (for claims), autant pour les trend que pour la tarification segmentation
- Catastrophes and Large claims
- Development (to ultimate)
- Loss trending
- One-time Changes
Nommez les 2 data considerations (for premiums)
- On-Levelling
- Premium Trend
Décrire les deux raisons pourquoi la méthode ‘one-way analyses’ peut ne pas bien fonctionner.
- Distributional Bias : Occurs when exposures are not uniformly distributed across risk characteristics (plus d’exposition dans la catégorie âge qui dans la catégorie genre), so one-way analysis mistakenly attributes part of one variable’s effect to another.
- Dependance : Occurs when the true relationship between one risk characteristic and claim costs depends on another risk characteristic. In more familiar terms, this is called an interaction effect, because the effect of one variable interacts with another.
That said, univariate techniques still have their place and are often used for quick adjustments or as a starting point for refining rating variables.
Décrire les deux méthodes qui peuvent utiliser pour définir un ‘base rate’
- Overall average pure premium : the base rate represents the pure premium that would apply if no classification were used
- Base Class Pure Premium : choose a class as a base rate
Décrire les deux principaux défis en lien avec la classification des territoires
- Granularity of territories : Actuaries must decide how finely to divide regions, balancing responsiveness and stability. In Canada, territories are often based on postal codes, FSA, or census blocks.
- Assignment of locations to territory : Territories don’t need to be geographically connected. Areas with similar risk levels can be grouped together, even if they are far apart. For example, northern Alberta and Manitoba could be assigned to the same territory if their modeled risk is comparable.
Modern insurers often use GLMs or GBMs to help define territories.
Nommez les 4 objectifs des déductibles
- Reduce moral (when the risk of loss increases because the insured might benefit from the loss) and morale (when the insured becomes careless simply because they have insurance) hazard : Sharing in losses gives the insured a financial incentive to avoid careless or risky behavior
- Encourage risk control : insureds are motivated to invest in safety measures, training, or security to prevent losses
- Filter small claims : Deductibles reduce insurer costs for minor claims, including investigation and administrative expenses (ULAE), which helps keep premiums lower
- Limit exposure to catastrophes : Large deductibles can reduce insurer exposure for catastrophic events such as in regions prone to natural disasters.
Décrire les types de déductibles
- Per-event deductible ; straight deductible
- Per-event deductible ; split deductible (similar to a straight deductible, except a different deductible applied for claims arising from different perils/coverages)
- Per-event deductible ; percentage deductible (common bases for the % include the total insured value, the amount insurance, or the amount of loss)
- Per-event deductible ; franchise deductible (the insurer pays the full loss amount if the claim exceeds the franchise deductible amount)
- Per-event deductible ; disappearing deductible (le montane payé est linéaire jusqu’à que l’insurer paie au complet)
- Aggregate deductible : fixed amount that applied over the entire policy term, usually a year
Quelles sont les différences entre un ‘self-insured retention (SIR)’ et un ‘large deductibles’
The key difference is how they’re handled. With a deductible, the insurer pays the full claim first and then collects the deductible amount from the insured. With an SIR, the insured is responsible for paying and managing claims up to a set limit before the insurer gets involved. The large deductible is preferred by insurers for maintaining control over liability claims early. The SIR offers greater control to insured, but limits insurer oversight until needed
Nommez les 4 principaux objectifs des limites sur les polices
- Cap the insurer’s financial responsibility
- Accommodate different risk and financial needs
- Help insurers manage risk and stay within their financial or reinsurance capacity
- Limit exposure without fully excluding coverage
Nommez les 2 principaux types de limites
- Single Limit : Sets the maximum the insurer will pay for one claim. It may apply per occurrence or per person, depending on the policy. In some cases, policies apply both per-occurrence and aggregate limits
- Compound Limit : Splits coverage into two or more parts. In some cases, policies apply both per-occurrence and aggregate limits.
Quel est le principal objectif de la coassurance.
To encourage insureds to maintain adequate insurance relative to the value of their property. Behinds the scenes, the coinsurance penalty reduces the claim payment so that the premium paid matches the expected payout.
Quels sont les deux principes derrière ‘experience rating’
- Credibility of Experience : Larger insureds (usually based on premium size) have more credible claims experience, so their own history carries greater weight in rating. Smaller insureds may have some credibility, but often must meet a minimum premium threshold to qualify.
- Emphasis on Frequency : Claim frequency is generally a better predictor of future losses than severity. For this reason, claims are often capped or split into primary (more predictable) and excess (more volatile) portions when used in the rating formula. This will be shown shortly.
Décrire la méthode ‘prospective experience rating’
Prospective expérience rating is an enhancement to manual rating that adjusts an insured’s premium based on their own past claims experience. Instead of relying only on an exposure measure, the insurer incorporates the insured’s claim experience to estimate whether future claims are expected to be better or worse than average
Décrire la méthode ‘schedule rating’
Schedule rating is a method used to adjust manual premiums up or down based on insured’s specific risk characteristics that are not fully captured in the manual rating process. Commonly used in commercial lines.
Nommez les 4 caractéristiques qui sont généralement utilisé dans la méthode ‘schedule rating’
- The presence of a safety program
- The company’s financial strength
- Employee qualifications training, and supervision
- Management experience
Décrire la méthode ‘retrospective experience method’
Retrospective rating is a way of calculating premium where the final cost is based on the insured’s actual losses during the policy period. This differs from experience rating, which adjusts premiums up front using past experience as a predictor of future claims.
Autre que prospective dans retrospective rating plans, nommez 3 autres individual risk rating plans
- Dividend Plans : these plans usually require the insured to meet a minimum premium level, and the amount of dividend depends in part on the insured’s claim results
- Safety Groups : let employers in the same industry pool premiums and claims. dividend are based on the group’s overall results
- Self Insurers : most programs use experience rating to balance fairness with cost stability
Nommez les 2 principaux objectifs associer avec la ‘self-insurance’
- Stability in Allocations From Year-to-Year. Some options are available in experience rating systems to reduce contribution volatility (claims capping, using counts instead of amounts, credibility weighting)
- Simplicity : Participants need to understand how they can lower their own costs
Quels sont les limitations de la méthode du GLM
- GLM assign full credibility to the data : Although GLM provide standard errors for each coefficient, allowing for confidence intervals, they will still produce a relativity estimate regardless of how little data supports it.
- GLM assume the randomness of outside is uncorrelated : minor violations/correlation aren’t usually a problem, but large violations/correlations can cause the GLM to produce inaccurate predictions
Two common situations that can violate this assumption are : policy renewals (the same policyholder can appear multiple times) and correlated events (policyholders can experience losses from the same event, such as severe storm)
Dans le modèle GLM de tarification, quel sont les avantages et désavantages de modéliser la prime pure (vs freq et sev)
- Advantages : simplified modeling process, since only on model needs to be created
- Drawbacks : The Tweedie distribution assumes that any predictor in the model will cause frequency and severity to move in the same direction
- Drawbacks : Some effects might be “lost” : if predictor has a strong negative effect on frequency, and an offsetting strong positive on severity, they can cancel each other out
