**Insurance Abstract**
A method for optimizing insurance estimates utilizing Monte Carlo
simulation includes the steps of ascertaining the total number of
potential insured units and obtaining a quote for full insurance
based on the total number of potential insured units. The method
further includes creating a model of total costs of self insurance
for the potential insured units, obtaining data distributions for
all variables in the model of total costs of self insurance and
running a Monte Carlo simulation on the model a preselected number
of iterations. A range of range of possible total costs of self-insurance
and the probabilities of such costs is then obtained facilitating
a selection between full insurance and self-insurance.
**Insurance Claims**
What is claimed is:
1. A method for optimizing insurance estimates utilizing statistical
simulation comprising the steps of: ascertaining the total number
of potential insured units; obtaining a quote for full insurance
based on the total number of potential insured units; creating a
model of total costs of self-insurance for the potential insured
units; obtaining data distributions for all variables in the model
of total costs of self-insurance; running a statistical simulation
on the model a preselected number of iterations; and obtaining a
range of possible total costs of self-insurance and the probabilities
of such costs.
2. The method for optimizing insurance estimates of claim 1 further
comprising the step of: comparing the range of possible total annual
costs of self-insurance to the quoted cost of full insurance to
determine possible savings.
3. The method for optimizing insurance estimates of claim 2 further
comprising the step of: selecting a type of insurance based on the
possible savings.
4. The method for optimizing insurance estimates of claim 1, wherein:
said potential insured units include individual and family potential
insured units.
5. The method for optimizing insurance estimates of claim 4, wherein:
said variables include the administrative expenses to administer
a self-insurance plan, the cost of stop-loss insurance at specific
cap levels, broker commissions, demographics of the group of potential
insured units and the location of employer.
6. The method for optimizing insurance estimates of claim 5, wherein
the statistical simulation is a Monte Carlo simulation.
7. The method of optimizing insurance estimates of claim 6, wherein
the preselected number of iterations is about 10,000 iterations.
8. The method of optimizing insurance estimates of claim 1, wherein
said data distributions for all variables in the model of total
costs of self-insurance are either preexisting data distributions
or are generated from a maximum and minimum value for a variable.
9. A method for optimizing insurance estimates utilizing Monte
Carlo simulation comprising the steps of: ascertaining the total
number of individual and family potential insured units; obtaining
a quote for full insurance based on the total number of potential
insured units; creating a model of total costs of self-insurance
for the potential insured units; obtaining data distributions for
all variables in the model of total costs of self insurance, said
data distributions being either a pre-existing distribution or generated
from a maximum and minimum value for a variable; running a Monte
Carlo simulation on the model a preselected number of iterations;
and obtaining a range of possible total costs of self-insurance
and the probabilities of such costs.
10. The method of optimizing insurance estimates of claim 9, wherein
the preselected number of iterations is about 10,000 iterations.
11. The method for optimizing insurance estimates of claim 10,
wherein: said variables include the administrative expenses to administer
a self-insurance plan, the cost of stop-loss insurance at specific
cap levels, broker commissions, demographics of the group of potential
insured units and the location of employer.
12. The method for optimizing insurance estimates of claim 9 further
comprising the step of: comparing the range of possible total annual
costs of self-insurance to the quoted cost of full insurance to
determine possible savings.
13. The method for optimizing insurance estimates of claim 12 further
comprising the step of: selecting a type of insurance based on the
possible savings.
14. A method for optimizing self-insurance estimates utilizing
Monte Carlo simulation comprising the steps of: ascertaining the
total number of individual and family potential insured units; obtaining
a quote for full insurance based on the total number of potential
insured units; creating a model of total costs of self-insurance
for the potential insured units; obtaining data distributions for
all variables in the model of total costs of self insurance, said
data distributions being either a pre-existing distribution or being
generated from a maximum and minimum value for a variable; running
a Monte Carlo simulation on the model a preselected number of iterations,
said preselected number being about 10,000 iterations; obtaining
a range of possible total costs of self-insurance and the probabilities
of such costs; comparing the range of possible total annual costs
of self insurance to the quoted cost of full insurance to determine
possible savings; and selecting either self-insurance or full insurance
based on the possible savings.
15. The method for optimizing insurance estimates of claim 14,
wherein: said variables include the administrative expenses to administer
a self-insurance plan, the cost of stop-loss insurance at specific
cap levels, broker commissions, demographics of the group of potential
insured units and the location of employer.
**Insurance Description**
CROSS REFERENCE TO RELATED APPLICATIONS
[0001] This application claims the benefit of U.S. Provisional
Patent Application Ser. No. 60/503,543 filed on Sep. 17, 2003, entitled
"USE OF MONTE CARLO SIMULATION TO PREDICT RESULTS ON HEALTH
INSURANCE," herein incorporated by reference in its entirety.
FIELD OF THE INVENTION
[0002] The present invention relates to a system and method for
optimizing insurance estimates. Specifically, the present invention
involves a system and method for calculating the costs of self-insurance
and the probability of such costs using Monte Carlo simulation to
assist employers in selecting an appropriate type of insurance.
BACKGROUND OF THE INVENTION
[0003] Employers obtain health insurance funding in one of two
ways. Employers may be either fully insured or self-insured. Fully
insured employers pay a monthly premium to an insurance carrier
to cover their employees' medical expenses. Being fully insured
offers employers several benefits including known premiums that
may be included in a budget, minimal financial risk and ease of
plan administration.
[0004] Many employers, however, choose to self-insure rather than
purchase group insurance plans to minimize their expenses. These
employers typically set aside funds from which employees and their
families are reimbursed for their medical expenses. Self-insured
employers usually hire an administrator to process their employees'
claims. While self-insurance is often an excellent cost-saving measure,
it exposes employers to a high level of financial risk. If an employee
incurs unexpectedly high medical expenses, an employer's medical
reimbursement funds may be exhausted. To reduce the financial risk,
employers obtain stop-loss insurance, which reimburses employers
for medical expenses that exceed a certain deductible threshold,
often referred to as a cap level.
[0005] There are two basic types of stop loss insurance; aggregate
and individual/specific stop loss. Aggregate stop-loss insurance
reimburses an employer when all claims exceed an agreed upon threshold
or cap, typically described as a monthly amount per employee and
employee with family. The cap is typically a percentage, e.g., 120%
or 125%, of what the carrier expects the claims will be.
[0006] With specific stop loss insurance, the carrier reimburses
the employer when claims for an individual exceed a specified amount
or cap in a plan year. The carrier reimburses the employer for the
remainder of the plan year. Specific stop-loss insurance has different
rates for single employees and for families. The rates are lower
the higher the cap at which the carrier begins reimbursing the employer.
[0007] While stop-loss insurance reduces financial risk for self-insured
employers, it is an added expense that must be considered when determining
the whether self-insurance is appropriate.
[0008] In light of the above, employers contemplating self-insurance
must perform a detailed analysis to determine whether the costs
of this type of insurance are greater or less than the costs of
full insurance. Performing such an analysis, however, is often difficult
as stop-loss insurance carriers do not provide information regarding
the projected total annual costs of self-insurance to the employer
and do not compare such data to the annual costs of full insurance.
Stop-loss insurance carriers simply quote prices for different cap
levels, e.g., $50,000, $60,000 or $70,000 leaving the employer to
determine whether self-insurance is the best option.
[0009] In light of the above, there exists a need for a source
of the total projected annual costs of self-insurance and the probability
of such costs so that employers can determine whether self-insurance
is appropriate. The present invention fulfills these needs and more.
SUMMARY OF THE INVENTION
[0010] It is an object of the present invention to provide a system
and method of optimizing insurance estimates that offers potential
insurance purchasers information regarding the amount and probability
of total annual insurance costs from which they can make an informed
decision as to an appropriate type of insurance coverage.
[0011] It is another object of the present invention to provide
a system and method of optimizing insurance estimates that provides
employers with projected total annual costs of self-insurance and
the probabilities of such costs through a Monte Carlo simulation
and compares the projected costs to the cost of full-insurance so
that employers may make an informed decision as to the appropriate
type of insurance.
[0012] A method for optimizing insurance estimates utilizing statistical
simulation according to the present invention includes the steps
of: ascertaining the total number of potential insured units; obtaining
a quote for full insurance based on the total number of potential
insured units; creating a model of total costs of self insurance
for the potential insured units; obtaining data distributions for
all variables in the model of total costs of self insurance; running
a statistical simulation on the model a preselected number of iterations;
and obtaining a range of possible total costs of self insurance
and the probabilities of such costs.
[0013] These and other objects and advantages of the present invention
will become readily apparent upon further review of the following
drawings and specification.
BRIEF DESCRIPTION OF THE DRAWINGS
[0014] FIG. 1 is a schematic of one possible system according to
the present invention.
[0015] FIG. 2 is a flowchart illustrating the steps of a method
according to the present invention.
[0016] FIG. 3 is a table illustrating the projected total annual
costs for self-funded plans at various probability levels and the
saving over fully insured plans at these levels.
[0017] FIG. 4 is a spreadsheet illustrating an implementation of
a method of the present invention.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
[0018] As shown schematically in FIG. 1, a system in accordance
with an embodiment of the present invention includes a computer
2, and at least one database 4 containing data representing the
variables used to calculate projected total annual costs of self-insurance
for an employer 4. As discussed in greater detail below, the computer
2 contains software that utilizes Monte Carlo analysis to calculate
projected total annual costs for self insurance, as well as the
probabilities of such costs, from data in the database 4. As will
be appreciated, the database 4 may be resident on the computer 2
or may be accessible via a network such as the Internet.
[0019] The database 4 contains data representing the variables
used to calculate projected total annual costs of self-insurance.
These variables may include quantifiable factors such as administrative
expenses to administer a self-insured plan, the cost of stop-loss
insurance at specific cap levels, broker commissions, demographics
of the group of potential insureds, location of employer and net
work cost savings.
[0020] The database may also include variables such as the number
of expected claims and aggregate maximum and shock claims i.e.,
claims that are 50% or more of the specific limit. The number of
expected claims might be obtained from historical insurance industry
loss statistics. The statistics typically include age, sex, geographic
location, occupation and other relevant statistics of individual
loss incurring insureds at well as the amount of each loss. The
statistics also include whether the loss incurring insured was a
single insured or family insured. Third party companies typically
compile these statistics and they may be stored in a database. As
will be appreciated, the loss statistics used to calculate the expected
claims may be carrier specific or may be general industry statistics
and may be an annual compilation or may represent a greater time
period.
[0021] The software on the computer 2 can include statistical software
programs capable of performing a Monte Carlo simulation. These programs
include statistical software such as SAS.RTM., Systat.RTM. and Crystal
Ball.RTM.. Such software programs also include spreadsheet software
such as Microsoft Excel.RTM..
[0022] As will be appreciated, Monte Carlo simulation is a quantitative
risk analysis technique. Monte Carlo simulation involves repeatedly
executing a model that contains multiple variables to be analyzed.
Each execution of the model is referred to as iteration. Each variable
in the model is represented by a probability distribution of possible
values. During each iteration, values from the probability distributions
are randomly selected. A simulation provides a range of possible
outcomes that could occur and the likelihood of any outcome occurring.
[0023] The above application of Monte Carlo simulation to project
the costs of self-insurance is an important aspect of the present
invention. By utilizing Monte Carlo simulation, all costs can be
assessed to predict the total costs of self-insurance and the probability
of such costs. These costs can then be compared to the annual costs
of full insurance to determine which type of insurance is most appropriate.
Currently, this information is not provided to employers that are
in the process of selecting insurance.
[0024] FIG. 2 is a flow chart indicating steps of a method of the
present invention in optimizing insurance estimates utilizing Monte
Carlo simulation. As noted at 10, the initial step is obtaining
a quote for an annual premium for full insurance from a carrier.
As will be appreciated, the quote may be obtained from a broker
or directly from the carrier. To further facilitate selecting the
most appropriate type of insurance, several quotes may be obtained
from different carriers. The annual costs of full insurance provide
a baseline against which the projected total costs of self-insurance
are compared.
[0025] After the employer has obtained quotes for full insurance,
a model of the total costs of self-insurance must be created as
noted at 12. This step involves determining the specific variables
to be included in calculating the costs. As mentioned previously,
these variables may include administrative expenses, the cost of
stop-loss insurance at specific cap levels, broker commissions,
demographics of the group of potential insureds, location of employer
and expected losses. Once the variables have been determined, they
may be expressed as a model such as y=f(x.sup.1, x.sup.2, X.sup.3,
x.sup.4) where y is the total cost of self-insurance and the variables
are x.sup.1-x.sup.4.
[0026] After the model of the total costs has been determined,
data distributions for all of the variables in the model must be
obtained as noted at 14. If the distributions are not known, a maximum
and minimum value for a variable will suffice. As will be appreciated,
certain variables are likely to have existing data, such as historical
industry loss data used to calculate expected losses. Depending
on the software used, existing data may be fit into distributions
before the simulation is implemented.
[0027] As noted at 16, once the distributions and/or maximum and
minimum values have been obtained, the Monte Carlo simulation is
implemented. This is accomplished by randomly selecting a value
from the distribution of values for each of the variables in the
model. This may be accomplished by randomly selecting a value from
an existing distribution or, if the distribution does not preexist,
using a uniform distribution created through the use of a random
number generator. To generate a random number for such variables,
the maximum and minimum values of the variable must be obtained.
The generator will then randomly select a number within the maximum
and minimum values. As will be appreciated, this process may utilize
a uniform distribution or other distribution such as a Gaussian
distribution.
[0028] The random selection of values from the variables in the
model is repeated a preselected number of iterations. The number
of iterations is preferably 10,000, although other values are possible.
[0029] As noted at 18, after the preselected iterations have been
run, the range of possible outcomes and the probabilities of such
outcomes are obtained. In other words, a range of possible total
annual costs of self-insurance is obtained along with the probabilities
of the possible costs.
[0030] As recorded at 20, once the range of possible total costs
and probabilities has been obtained, this data is compared to the
fixed annual costs of full insurance obtained from the quotes. That
is, the projected costs of self-insurance at the probability levels
obtained from the simulation are subtracted from the cost of full
insurance to determine possible savings.
[0031] After the possible savings have been determined, the employer
can use this information to select between self-insurance or full
insurance, as noted at 22.
[0032] An example of the results of the above process is shown
in FIG. 3. For a self-insured plan, also referred to as self-funded
plan, the possible total costs 30 are $1,288,128; $1,397,199 and
$1,517,400 at probabilities 32 of 50, 60 and 70% respectively. Therefore,
in this example, it is more likely that the costs of self-insurance
will be higher rather than lower. The possible costs of the self-funded
plan are then subtracted from the fixed costs for full insurance
34 obtained through the quotes to determine possible savings 36.
[0033] FIG. 4 illustrates an implementation of a method of the
present invention utilizing Microsoft Excel.RTM.. In the implementation
of FIG. 4, the model contains multiple variables 40. These include
the number of potential insured units, referred to as "counts"
and the claims expected per 1000 insured units.
[0034] Although the present invention has been described with reference
to preferred embodiments, it will be appreciated by those of ordinary
skill in the art, that various modifications to this invention may
be made without departing from the spirit and scope of the invention. |