one time and continuous changes

Question 1

one time changes and common types

Answer 1

• specific changes implemented or adopted by insurer on specific data in time that impact prem, loss, or expense
• RCs: changes to rates and/or rating algorithm; apply to all policies effective on or after the effective date of RC
• Law changes: stipulate change in coverage, benefits, or rates; can be implemented like RCs or impact all policies starting on given date
• court rulings: similar to law changes but imposed by court decision
• expense changes

Question 2

need to adjust for one time changes

Answer 2

• goal in using past data for ratemaking is to adjust data to be most representative of future policy period that is being priced
• want to adjust or re-state historical prem, loss, and expenses to be reflective of future rate levels, coverage levels, and expense levels

Question 3

direct and indirect effects

Answer 3

• direct effects=direct and obvious impacts to prem, loss, or expenses resulting from change all else being equal
• indirect effects=impacts from changes in human behaviors that are consequences of one-time change

Rate increase would result loss in prem due to changes in retention/close ratios

Indemnity benefit increase might cause more injured workers filing claims and workers staying OOO longer

-indirect are difficult to quantify and are not usually incorporated into adjustments

Question 4

3 ways to calc effect of coverage change on losses

Answer 4

1. restate ind claims at new coverage levels

*WC benefit level changes

2. calculate effect on representative group of claims
3. simulate losses under new coverage levels

Question 5

2 methods to on-level

Answer 5

1. EoE: re-rates all historical policies @ individual policy level using the newest rates and then re-calc EPs for each hist period using newest rates
   - most accurate
   - getting detailed data, computing power it may require, need to make assumptions for new RVs with no hist data, and difficult to incorporate changes in schedule rating guidelines for Comm LOBs
2. Parallelogram method: used on group policy data and adjusts hist prem by average factor for each hist period
   - quicker to calculate
   - assumes policies are written evenly throughout historical period -> bad assumption for seasonal LOBs or growing/shrinking books
   - direct effects of changes are often calc at aggregate level but using aggregate direct effects may not be appropriate for class level RM if effects vary by class

Question 6

OLF

Answer 6

=current cumulative rate level index/(weighted average cumulative rate level index for that time period)

Question 7

continuous changes and examples

Answer 7

changes that occur gradually over time like changes in MOV and socio-economic trends

• inflations can cause exposures to change over time
• average premium can change due to more customers switching to higher deductibles
• rising gas prices can cause people to drive less, lowering freq of claims
• increases in cost of medical care can increase claim severity

Question 8

adjusting hist data for continuous changes ensures

Answer 8

that data reflects MOB and levels of social and economic inflation expected in future period

Question 9

data used for trending

Answer 9

• whatever data is being used, want to make adjs to remove any distortions from true trend
• when using insurer data, adjust for one-time and anomalies
• very common to use quarterly or monthly data to determine trends -> may want to make adj to remove or smooth out any seasonality of data

Question 10

why use most recent data for trending?

Answer 10

reduce duration and thus uncertainty in forecast

Question 11

premium data to use

Answer 11

• generally forecasting EP, but can recognize that WP is leading indicator of EP
• using EP to determine trend is that we are trending EP so makes sense to determine trend on EP
• using WP allows us to use more recent data and changes in avg WP will ultimately show up as changes in avg EP

Question 12

loss data to use

Answer 12

• PP trends can be analyzed directly or split into frequency and severity trends
• typically separated since they change for different reasons
• concept of using latest data is applied by using calendar period paid or reported loss data for short-tailed LOBs

Question 13

using calendar period data for loss assumes

Answer 13

BOB is not significantly growing or shrinking since mismatch of losses and exposures

-for short tailed, this is likely to be reasonable, but not long tailed since bigger distance between when exposures are written and losses are paid or reported

Question 14

adv to using paid data

Answer 14

paid data is not subject to changes in case reserving practices

Question 15

adv to using rptd data

Answer 15

incorporates more recent info since case reserves provide info you might expect to eventually see in paid

Question 16

trends for layers of loss

Answer 16

• if split losses by deductible or limit and apply a severity trend to ground-up losses will notice a different effective trend rate for portion of losses below and above split point
• to determine trend on basic limit losses, need to determine basic limits losses before and after the limits trend is applied to total losses
• to determine trend on excess losses, need to determine excess losses before and after total limits trend is applied to total losses

Question 17

2 reasons excess severity trends are greater than total or basic limits severity trends when trend is positive

Answer 17

1. for losses above basic limits, trend is entirely in excess layer
2. losses under the basic limit are pushed into excess layer by trend, creating new losses for layer
   - when trend is negative, excess trend will be more negative because of #1 and basic and total trends will be closer to 0

Question 18

why do we need trend periods?

Answer 18

need to obtain range of dates for which we expect new rates to be in effect so we can project avg prem and PPs to appropriate levels

Question 19

2 assumptions normally made when selecting trend periods

Answer 19

1. policies are written uniformly over time
2. premiums are earned uniformly and losses occur uniformly over policy period

Question 20

2 step trends

Answer 20

• 1st step=from avg date on historical period to avg date of latest time period for which data is available
• 2nd step=avg date of latest time period to avg date in future period
• should be used when historical trend rate is expected to differ from prospective trend rate

Question 21

2 ways to perfom first step of 2 step trending

Answer 21

1. adjust historical period level = latest level

current trend factor = latest avg WP @ curr/ hist avg EP @ curr

• assumes latest level is a reliable number from which to base future
  2. trend historical period level to latest time period
• performed like one-step trend

Question 22

exposure trends

Answer 22

-for LOBS with inflation-sensitive EBs such as payroll for WC, exposures can change over time because of inflation and applying exposure trend may be appropriate in RM analysis

Question 23

expense trends

Answer 23

inflationary trends also impact expense levels

• trend period for expenses varies based on whether expense is incurred at onset of policy or throughout policy
• incurred at onset will have same trend period as WP
• incurred throughout will have same trend period as EP

Question 24

If the data used in both determining and applying the trend is not adjusted for large events & anomalies

Answer 24

it may be difficult to choose an appropriate trend rate and we may over or under project future values

Question 25

development vs trending

Answer 25

• When projecting prem or losses to future levels for use in a rate indication, there is no overlap with projecting prem or losses to their ult levels, even though both are changes that occur over time
• Development brings the data from each hist period to its ult level, while trending reflects the difference in ult levels from one hist period to the next
• In RM terms for losses, loss development makes sure that the future policy is priced to cover ult loss, while loss trend makes sure that the ult losses are at the cost levels corresponding to the future policy period