click below
click below
Normal Size Small Size show me how
FIL 250 Chapter 2
Risk assessment
Question | Answer |
---|---|
What are four requirements when compiling data when analyzing loss exposures? | relevancy, complete, consistent, and organized |
What are four measures of central tendency? | expected value, mean, median, and mode |
What are common measures of dispersion? | variance, standard deviation, coefficient of variation, and tail percentiles (such as the 5th or 95th percentiles, called value at risk) |
When forecasting outcomes using the normal distribution, what percentage of outcomes is predicted to be within one, two, and three standard deviations of the mean? | 68%, 95%, and over 99% |
What methods are used to identify loss exposures? | Document analysis, compliance review, inspections |
List some examples of documents that are reviewed when identifying loss exposures | financial statements (balance sheet, income statement, cash flow statement), checklists/questionnaires, organizational policies/documents, organizational chart/flowchart, loss history, contracts, insurance policies |
theoretical probability | likelihood of an event is based on the known structure of possible outcomes, such as the roll of a die or spin of a roulette wheel. In risk management, we rarely know the structure of the event so we don't know theoretical probabilities. |
empirical probability | estimate of the likelihood based on a sample of data |
law of large numbers | when increasing the amount of historical data for loss exposures (assuming the risk environment is similar) improves the accuracy of forecasting. (e.g., remember the accuracy of rolling the dice in class) |
probability distribution | presents estimates of the probability of all possible outcomes. these can be presented as a graph/chart or as a table |
discrete probability distribution | outcomes have only a countable number of potential outcomes, often used for the frequency of events (e.g., the number of hurricanes in a year) |
continuous probability distribution | outcomes have an infinite number of possible outcomes, often used for severity (e.g., the total amount of damages caused by a hurricane) |
expected value | using the theoretical probability distribution, the expected value is the weighted average of the outcomes, where the weights are theoretical probabilities |
central tendency | a single number that attempts to represent a distribution of outcomes. (e.g., the "middle" of the distribution) |
mean | it's the average that you and I tend to think of. To determine the mean, add up all outcomes and divide by the number of observations. without additional information, it is likely the best "guess" at what an outcome would be (i.e., the predicted outcome) |
median | 50th percentile of a distribution, or the midpoint of sequenced data. 50% of outcomes are larger than the median and 50% are less. |
mode | the outcome occurring most frequently in a dataset |
dispersion | a measure of how spread out a probability distribution is. If the dispersion is high, measures of central tendency are less reliable since some outcomes are far away from the middle of the distribution |
variance | the "average" squared deviation from the mean. first, you must calculate the mean. next, subtract the mean from each data point and square this result. finally, add up these squared deviations and divide by the number of observations minus one (n-1) |
standard deviation | the most common measure of the dispersion of a probability distribution. to determine the standard deviation, take the square root of the variance. |
coefficient of variation | a measure of dispersion that relates the standard deviation of a distribution to its mean. useful for comparing the dispersion of two distributions. |
loss frequency | the number of losses that occur in a given period of time. or the probability that a loss will occur |
loss severity | the size of a given loss |
total dollar losses | the cumulative value of all losses that occur during a period of time |
maximum possible loss (MPL) | the total value of loss exposures that can be caused by any one particular event |
risk map / matrix | a two dimensional graph that depicts both the frequency of losses (horizontal access) and the severity (vertical access) |
Prouty approach | a type of risk map assisting risk managers in identifying risk management techniques. includes four categories of frequency (zero, slight, occasional, and definite) as well as three categories of severity (slight, significant, and severe) |
data credibility | the weight (or confidence) attached to historical data when forecasting future events |
collectively exhaustive | in a probability distribution, all possible outcomes should be illustrated - an outcome not listed is not possible |
Mutually exclusive | when only one outcome is possible at a time (like heads vs. tails on a flip of a coin) |
percentile | indicates the likelihood of falling below a specific value |
tail percentile (value at risk) | percentiles that are close to the minimum and maximum of a dataset. tail percentiles give a measure of the worst case scenario with a particular probability (like the 5th or 95th percentile) |
cumulative probability distribution | based on a ranked sample of data, illustrates the probability of being less than or equal to observed values (see percentiles) |
What are four dimensions of loss exposures? | frequency, severity, total dollar losses, and timing |
balance sheet | a financial statement that reports assets, liabilities, and owners equity as of a specific date |
income statement | a financial statement that indicates revenue, expenses, and profit over a period of time |
hold harmless agreement | a specific provision of a contract that releases one party from liability, such as a retailer selling the finished goods of a manufacturer |
hazard analysis | a detailed investigation of the conditions that may increase the amount of loss experienced by an organization |