Section X A (KZ1621DOC;2)

By Johnny Miller,2014-05-14 13:15
8 views 0
Section X A (KZ1621DOC;2)

Section X - Classification Pricing Section X-A

    Subsection A - Summary of Methodology Page 1






     In this section we calculate the rate change for each Industry Group as it varies around the statewide overall indicated rate change (Section X-E, Exhibit 1). We have also

    updated the indicated class relativities within each Industry Group (Section X-N). In doing

    so we have revised and updated the manner of calculating credibilities (Section X-F). We

    have also incorporated updated classification groupings within the Office and Clerical,

    Goods and Services, and Miscellaneous Industry Groups.

     The indicated relative rates by class within each Industry Group are determined in

    Section X-N. Then proposed average rates by class are calculated such that the proposed

    rate change by Industry Group is achieved (Section X-O). Proposed average rates are

    capped in the traditional manner (Section X-G and X-P).

     Finally, proposed manual rates are determined by applying offsets to the proposed

    average capped rates. Offsets for various programs, including the Bureau’s proposed

    $100 loss constant, are calculated based on latest available data and are shown in Section

    X-R. Section X-S displays the proposed manual rates and rating values for this filing.

     Section X-B gives an example showing each of the steps in classification pricing.


     The basic data used is the Unit Statistical Plan data

    1 from Schedule Z which shows

    detailed composite policy year payroll, premium and loss data for each classification in

    Massachusetts. The data used excludes the experience of both large deductible policies

    and of Self-Insurance Groups. In this review, the latest available five composite policy

    years of data are 7/1/90 - 6/30/91, 7/1/91 - 6/30/92, 7/1/92 - 6/30/93, 7/1/93 - 6/30/94, and

    7/1/94 - 6/30/95.

1 The Bureau receives, edits and compiles Unit Statistical Plan Data on an ongoing basis. A typical

    composite policy year of data consists of about 600,000 records of exposure data and about 100,000

    records of loss data.


Section X - Classification Pricing Section X-A

    Subsection A - Summary of Methodology Page 2


    Adjustment of Massachusetts Data

     Before reported data can be utilized to calculate classification rates it must be adjusted to reflect current conditions. First, reported Schedule Z data is adjusted for large losses by capping each claim; this year the cap is $200,000. Capping ensures stability in the calculation of rates by eliminating the impact of infrequent occurrences of large losses. However, through the application of excess loss factors (discussed below) an average excess loss portion is added back to the rate to account for the possibility of large losses.

     Excess losses are subtracted from reported losses to determine primary losses. These primary losses are multiplied by on-level factors (Section X-D) in order to adjust each composite policy year's experience to a comparable basis. On-level factors include benefit level adjustments, loss development factors, loss adjustment expense loadings, escalation, and the effect of Chapter 398. On-level primary losses are the product of the primary losses and the on-level factors.

     An excess loss factor calculation more equitably distributes expected excess losses between classifications. Excess losses are loaded back through the application of factors which vary by hazard group and injury type, as shown in Section X-D, Exhibit 6. These factors are calculated by adding unity to the quotient of on-level excess losses and on-level primary losses. Adjusted on-level losses are the product of the on-level primary losses and the excess loss factors.

     The adjusted on-level losses, which vary by composite policy year and injury kind, are summed by composite policy year and injury type, are displayed by classification or class combination in Section X-N and are denoted Loss and Loss Adjustment Expense. Indicated pure premiums for Massachusetts are calculated by dividing those losses by payroll in hundreds.

    Statewide Credibility

     A credibility procedure is incorporated to account for the varying levels of payroll and loss dollars by class. Credibility is determined based on the volume of expected losses by injury type (Section X-F). Expected losses are calculated by multiplying the pure premiums underlying the present rate by payroll in hundreds. Pure premiums underlying present rates are the product of the adopted pure premiums from the previous rate revision 4279

Section X - Classification Pricing Section X-A

    Subsection A - Summary of Methodology Page 3


    and pure premium underlying present rate factors. The factors are comprised of the rate

    level factor implicit in the current rate and an adjustment for law level changes since the

    previous revision (Section X-C).

    Inclusion of Countrywide Data

     Countrywide data is incorporated into the filing through the use of data supplied by

    the NCCI or other independent rating bureaus.2 Relativities based on countrywide data

    supplement relativities based on Massachusetts data to the extent they are not credible.

    The remainder is assigned to the relativities underlying present rates. Classes with little or

    no Massachusetts experience will reflect to some extent the experience of these classes

    elsewhere in the country. The countrywide relativities are based upon the latest available

    three composite policy years, although the specific years vary by state (see Section X-E,

    Exhibit 3). Before the data can be used for ratemaking purposes, it must be adjusted to

    reflect the Massachusetts distribution of payroll and loss level.

     Pitch factors are used to adjust the countrywide data to a corresponding

    Massachusetts level (see Section X-B, Exhibit 5). These pitch factors are calculated as


     1. Multiply each state's classification pure premiums by the

    corresponding Massachusetts payroll to estimate the loss level

    that would occur if the employers in that state were located in


     2. Sum these losses by state, industry group, and injury type to

    determine the aggregate adjusted losses for each state.

     3. Calculate aggregate adjusted losses for Massachusetts in a

    similar fashion.

     4. Divide the aggregate adjusted losses for Massachusetts by the

    aggregate adjusted losses for each state to determine the pitch


2 The New York and New Jersey data is "A-Sheet" data comparable to that in Section X-N of this filing.

    The data is limited on a per-claim basis and developed in a manner similar to Massachusetts data. The

    data from NCCI is Unit Statistical Plan Data similar to Schedule Z, and then developed by NCCI. The

    Bureau has capped the average claim by injury kind by class by state at $200,000 (Indemnity + Medical).


Section X - Classification Pricing Section X-A

    Subsection A - Summary of Methodology Page 4


     The pitch factors represent the amount by which the other states' data must be adjusted to be on the same payroll distribution and loss level as the Massachusetts Schedule Z data (Section X-E, Exhibit 2). Each state's pure premiums are multiplied by their corresponding pitch factor to derive adjusted pure premiums. The adjusted pure premiums by state and class are multiplied by each state's payroll for the class, summed, and divided by total payroll to derive countrywide pure premiums by class.

     Finally, the countrywide adjusted pure premiums are further adjusted to the Massachusetts "indicated" loss level by application of the on-level and excess loss factors in a manner similar to the adjustment of Massachusetts data. These countrywide pure premiums are now on a comparable basis with the Massachusetts indicated pure premiums and can be used for ratemaking.

     For each Industry Group, for each injury type, the countrywide indicated pure premiums are converted to relativities, so that they balance to unity on Massachusetts exposures.

    Countrywide Credibility

     Credibility of adjusted countrywide pure premiums is determined based on claim counts by injury type (Section X-F). Payroll can not be used as a basis of credibility since wage levels, and hence payroll, vary significantly between states. Countrywide credibility is limited to one-half.

    Formula Relativities

     Next, the credibility formula is applied to determine the formula relativity by class. The indicated Massachusetts relativities for each of five years are multiplied by their credibilities, the indicated countrywide relativity is multiplied by its (limited) credibility and the relativity underlying the present rate is multiplied by the remaining credibility. The formula relativity is the weighted average of the relativities by injury type, using the Industry Group pure premiums by injury type as weights.

    Proposed Rates

     These formula relativities determine the indicated relative rates by class within Industry Group. Proposed average rates by class are calculated such that the proposed 4279

Section X - Classification Pricing Section X-A

    Subsection A - Summary of Methodology Page 5


    rate change by Industry Group is achieved. Proposed average rates are capped in the traditional approved manner (Section X-G).

     Finally, proposed manual rates are determined by applying offsets to the proposed

    average capped rates. Offsets for various programs are calculated based on latest

    available data and are shown in Section X-R. Section X-S displays the proposed manual rates and rating values for this filing. The derivation of the proposed manual rate for one classification is displayed in Section X-B for illustrative purposes. The remaining sections display the factors and calculations used to determine the proposed manual rates.

    Classification Combinations for Ratemaking

     Classifications having operations with similar exposures are combined for

    ratemaking when they develop a relatively small amount of payroll with resultant low credibility, over a period of years. The current combinations have been used since the filing for 1/1/89 rates.


     The classification combinations presently used for ratemaking for classes in the former All Other Industry Group have been reviewed. The review was performed jointly by the Bureau’s Actuarial and Underwriting staffs. The result reflects a combination of

    underwriting judgment and an analysis of data.

     The current review was performed in a manner similar to that presented in the 1989 filing. Similar judgments were made. However, a number of things have changed over the last decade.

    Since the last review during 1988, the former All Other Industry Group has been

    split into three: Office and Clerical, Goods and Services, and Miscellaneous. This has produced more homogeneous Industry Groups. Each class within the new Industry Groups is now more similar to its Industry Group. Now classes are more able to stand on their own, since the complement of credibility is applied to the indicated rate change for the new more homogeneous Industry Group.

    3 The relevant portions of the 1/1/89 rate filing are included in the Technical Appendix, Section XIII, of this filing.


Section X - Classification Pricing Section X-A

    Subsection A - Summary of Methodology Page 6


     Also since the 1988 review, we introduced the current practice of supplementing

    Massachusetts' data with “countrywide” data. This makes it more likely that a smaller

    class can stand on its own.

     Finally, since the last review of class groupings we have about a decade of additional

    data. Over time the size of classes and their relative experience changes. Unless

    combinations are reviewed and updated from time to time, they will not serve any useful


     The relevant question is whether the proposed combinations are preferable to the

    current combinations.4 It is a comparative question, which at a minimum requires that one

    5be aware of the source of the current combinations. The information provided allows a

    reasonable person to decide that issue.

     The purpose of this analysis was to determine if:

    1. A classification has similar exposure to the other classifications with which it is

    grouped, so that it should continue to be combined for ratemaking purposes.

    2. A classification has very low credibility so that it should be merged with another

    classification and thus eliminated as a separate classification.

    3. A classification should be combined with a different classification(s) with which it is

    more appropriately grouped for purposes of setting classification rates.

    4. A classification should be newly combined with other classification(s) for purposes

    of setting classification rates.

    5. A classification has sufficiently high credibility so that it should no longer be

    combined with other classifications for ratemaking.

    6. A classification has had no exposure for the last five years and should therefore be

    considered for elimination as a classification.

     The attached exhibits summarize the results of our analysis of the class codes

    comprising the former All Other Industry Group:

    1) Exhibit 1 provides a list of the Industry Groups class codes reviewed by staff

    along with the description of each class and current manual rate.

4 Or to some other set of combinations proposed by a party to the rate hearing.

    5 They were not handed down on Mt. Sinai.


Section X - Classification Pricing Section X-A

    Subsection A - Summary of Methodology Page 7


    2) Exhibit 2 shows the current and proposed grouping status of each class code

    in the Industry Groups.

    3) Exhibit 3 displays the estimated long-term effect of the change in grouping

    status for the class codes that are proposed to change. Included on the

    exhibit are relevant statistics involved in the analysis including the latest three

    composite policy years Schedule Z payrolls, the estimated credibility and the

    average uncapped rate.6

    There were no classifications which are proposed to be eliminated or merged

    into other classifications. Twenty-two classifications with sufficiently high

    credibility are proposed to stand on their own rather than be combined. Nine

    classifications with low credibility are proposed to be combined with other

    classes rather than stand on their own.

     The proposed new combinations are used in the calculation of classification rates.

    It should be noted that the rate change of every class was separately capped so that no

    class received more than the maximum allowed increase. It should also be noted that this

    proposal has no effect on the overall statewide rate level or on the average rate level for the

    classifications within an Industry Group.

6 It should be noted that due to the capping of classification rate changes, the full impact of some of the

    revisions in groupings may take several rate changes to be fully felt.


Report this document

For any questions or suggestions please email