Sn 210-02 Forecasting And Financial Mgt In Health Ins.

Front 1

Difficulties in Pricing Health Insurance

Back 1

 must project claims 18-24 months ahead, b/c:
 state approval is required
 claim runoff is long
 claims uncertainty is as high as the profit margin
 Ins Cpy cannot raise rates, because of:
 rate guarantees
 competitivity
 antiselective lapses
 surplus is needed, to:
 grow
 have good rating (AAA)
 satisfy regulators
but pricing competitively limits surplus.

 Underwriting cycles occur
The stages of an “underwriting cycle” (every 6 years):
1. Claim costs rise unexpectedly  underwriting loss.
2. pricing becomes more conservative.
3. Claim cost trend reverses, but the ins cpy fails to forecast it in time.
4. overpricing leads to underwriting gain.
5. claim costs unexpectedly rise again. Repeat.

 underwriting gains/losses are unpredictable, b/c:
 They vary by market segment
 cyclic nature has disappeared lately.

Front 2

Solutions to Pricing Difficulties:

Back 2

 Improve forecasting
 Consider competitors’ financial experience
 Take a multi-year view
 (for an MCO) Use Risk-Sharing with providers
 stabilizes finances.

Front 3

THE PHASES OF THE FINANCIAL MANAGEMENT PROCESS

Back 3

1. Business Planning
 business goals
 corporate strategy (e.g. target markets)
 Choose some actions to test.
 Build a forecast model:
 Initialize the model (based on recent history)
 use simulations
 determine which actions would achieve the objectives.

2. Implementation

3. Reporting & Monitoring emerging results; comparing with the original plan
 do monthly
 Watch for unintended consequences
 Must report results on the same basis as the original forecast

4. Take corrective action

Front 4

CHARACTERISTICS OF A GOOD FORECAST MODEL

Back 4

 Simple
 to input data
 to execute / test assumptions
 to verify reasonability
 to update
 Self-documenting
 Not time-consuming or resource-consuming
 Flexible; shows different levels of detail
 Output should look like a financial statement
 Serves a wide range of purposes, such as:
 experience monitoring; comparison to a benchmark
 reasonability checking (quality control)
 detailed review

Front 5

PROJECTION ELEMENTS: HOW TO DEVELOP A FORECAST MODEL

Back 5

| base period | intm per. | projection period |
Time: 0 12 mo. 18 mo. 5 yrs

Front 6

1. Definition of the Projection Cells

Back 6

 projection cells can be:
 Market segments
 Product lines
 differently rated populations
 projection cells must correspond with the cpy’s reporting capabilities.
 Each cell must be homogeneous:
 same trends
 same rating methodologies.
 adjust for differing ph renewal months

Front 7

2. The Base Period

Back 7

 12 months long, to eliminate seasonality
 include 3 months runout
 should correspond with the company’s fiscal year.

 Must reconcile the Base Period data to the Financial Statements
Data items on financial statements but not on IS reports:
 manual claims adjustments
 bad debt
 reinsurance recoveries
 provider settlements

 Ideally, use a restated financial statement.

Front 8

3. The Projection Period

Back 8

 2 to 5 years
 Short forecasts are based on actions that have already been implemented.
 Long (5-year) forecasts project actions that aren’t yet solidified. (Scenario testing)

Front 9

4. Membership Projections

Back 9

 s/b realistic
 premium income UP  membership DOWN

Front 10

5. Premium Income Projections

Back 10

Seriatim basis vs. Cell Basis
Disadvantages of the seriatim basis:
 complicated (each ph within a cell has its own formula)
 Inaccurate, because:
 rate changes are usually computed by class
 rates subject to state laws
 rates affect the lapse rate

Moral: Use the Cell basis.

Front 11

Two ways to project premium income from month to month:

Back 11

The Renewal Month method (assuming 1-year policies)
Total Premium IncomeMar 2005
= Total Premium IncomeFeb 2005
+ (number of new members entering in Mar 2005) * (their prem rate)
+ (number of March 2004 members renewing) * (their rate increase)

 Used for large group cells, where data by renewal month is available.

The Aggregate method
Prem PMPMMar 2005 = Prem PMPMFeb 2005 * (1 + rate incr %)(% of income subject to renewal)
 Used for small group cells, where renewal month data is not available.

Front 12

How to choose rate increase assumptions (for use in the above formulas)

Back 12

 Early projection months: Based on recent increases.
 Later projection months: Based on target LR

Front 13

Miscellaneous considerations in projecting premium income:

Back 13

 Retrospective rating
 product mix
 demographics changes
 antiselective lapses

Front 14

6. Claims Projections

Three methods of incorporating claims assumptions into the forecast model:

Back 14

A) Different for each provider contract.
Advantage:
 Accurate.
Disadvantages:
 too complex.
 Requires a database of providers
 providers must be mapped to the projection cells

B) Different for each service type.

C) Only split into fee-for-service and capitated
 Easiest. Commonly used.

D) Loss Ratio Method
ClaimsNov 2002 = (projected premium)Nov 2002 * Anticipated LRNov 2002
Disadvantage:
 Assumes that claims will match original pricing assumptions.

Front 15

The model should incorporate monthly claims seasonality.

Back 15

 Claim costs are seasonal, due to:
 High January claims
 calendar-year deductible
 Seasonality is most crucial if the forecast will be used monthly.
 Seasonality is least important if:
 The model will only be used quarterly
 The company is an MCO with lots of capitation contracts

Front 16

Ways of incorporating seasonality

Back 16

Annual Trend method
 apply annual trends directly to the monthly base period values
ClaimsNov 2003 = ClaimsNov 2000 (1 + annual trend factor)3

Monthly Trend method
 use average base period cost and project it monthly
ClaimsJan 2001 = (avg claims pmpm2000) x (Dec_to_Jan trend fact.)
ClaimsFeb 2001 = ClaimsJan 2001 x (Jan_to_Feb trend factor)
etc.
Seasonal Trend method
 use average base period cost, project annually, and apply seasonality factors
ClaimsNov 2001
= (Claims2000 / 12) x (1 + annual trend)3 . . .
x (seasonality factor) x (seasonal normalization factor)

Front 17

Miscellaneous considerations in projecting claims:

Back 17

 Split into Cost and Utilization
 some trends are monthly; some are annual
 provider contract changes
 risk-sharing

Front 18

7. Administrative Expense Projections

Back 18

 historical expenses compared to budget
 project expenses in the aggregate; prorate by cell

Front 19

8. Other Items to include in a forecast model

Back 19

 Taxes
 Commissions
 Reinsurance
 Investment Income

Front 20

9. Comparison of Interim Period Forecasts with Actual, Reported Values

Back 20

 interim period = 6 months.
 fine-tune the model.

Front 21

THE FORECASTING PROCESS

Timing

Back 21

 Update the base period once a year
 Coordinate forecasts with:
 Re-rating dates;
 Annual Statement;
 provider contract renewals;
 budget preparations.

 The forecast should be completed midway through the year.
 Too late  no time to act on the forecast’s output.
 Too early  actual experience might diverge.
 Ideal: Run the forecast mid-year; use output for preliminary planning.
 Prior calendar year c/b used as base period, since it is still recent.

Front 22

Updating the Base Period data:

Back 22

 annually.
 need to re-reconcile to financial statements
 decide what adjustments to make.

Front 23

Coordination with Senior Mgmt and other Departments

Back 23

 The forecast must be suitable for each business segment to use:
 Sr. Mgmt
 Finance (reconciliation)
 Underwriting (re-rating assumptions)
 Marketing (membership assumptions)

Front 24

THE MONITORING PROCESS: ANALYSIS OF RESULTS
1. Restate Reported Results

Back 24

Financial statements don’t provide an accurate picture, because of:
 reserve estimates
 cashflows that apply to prior periods
 anomalies
 risk-sharing settlements

Reasons why Financial Statements should be Restated:
 To correct inaccuracies;
 To show actually-incurred results

Example: Restated claims reserve

 Restate operating statements quarterly
 margins must be consistent

Front 25

2. Compare Actual (Restated) Results with the Forecast

Back 25

 A/E underwriting gain, split up by Enrollment, Premium Income, Claims, Other

Enrollment:
 rerun the forecast substituting actual enrollment for the assumptions.

Premium Rate Increases:
consider changes in:
 product mix
 demographic mix

Claims:
 split by major categories.
 examine seasonality

Other:
 compare expenses, investment income, and taxes to the forecast

Front 26

3. Look at Financial Implications of the Divergence

Back 26

 re-forecast based on actual experience
 revise the assumptions
 Quarterly

Front 27

4. Make Business Plan Corrections

Back 27

 preemptive planning.

Done.