Author: Chris Boggs
Some say work comp experience mods will eventually go the way of the dinosaur because of analytics and other underwriting methodologies. Some still focus heavily on experience mods and predict their continued importance in workers' compensation underwriting and pricing.
Regardless who is correct about the future – experience mods cause agents and insureds much angst - TODAY!
The National Council on Compensation Insurance's (NCCI's) Workers' Compensation Experience Rating worksheet looks daunting at first. But once all the acronyms, shorthand and calculations are understood – it can be appreciated for the thing of beauty that it really is (OK, maybe that's a bit of hyperbole).
Grab a copy of an experience mod worksheet, go ahead – I'll wait. Now let's start from the top; though we are going to skip the self-explanatory terms such as Risk Name, Risk ID and State.
Rating Effective Date: As the name suggests, this is the date the experience mod is or becomes effective. Historically the rating effective date related to the Anniversary Rating Date (ARD), but NCCI filed to cease using the ARD as of May 1, 2017. The ARD was introduced in 1923 (per NCCI) to prevent insureds from cancelling policies mid-term to take advantage of lower premiums if lower rates are filed in the state. Many have opined that eliminating the ARD will create the problem it was originally designed to prevent. However, this will likely not be the case due to short-rate cancellation penalties and some carrier's underwriting guidelines against cancelling and re-writing policies mid-term except for certain specific reasons.
Production Date: This is the date the experience mod is calculated. But in preparation for this date, agents need to be aware of a much more important date, the Unit Statistical Report Date (the Stat Date) sometimes referred to as the "valuation date." This is the date all loss information is sent to NCCI or state rating bureau – this includes paid and reserved loss information (all incurred losses). The Stat date is approximately 6 months (180 days) prior to the Rating Effective Date. To best serve clients, request and gather from the insurance carrier all the loss information before the stat date. Reviewing the loss information before the Stat date gives the agent the opportunity to check and question reserves BEFORE they become part of the experience mod. Randy Sieberg provides great information on this topic, here is a link to his article.
Experience Period: No, this term is not on the worksheet. This term describes the policy years used to develop the experience mod. Loss experience from three policy periods are included on the majority of experience mod worksheets. The years chosen are referred to as the "Experience Period." Calculating the experience period, and thus the policy periods used to develop the mod, is easy. The experience period comprises the three policy periods ending 12 months before the Rating Effective Date. For example:
|Rating Effective Date:||June 1, 2017|
(The policy terms used):
June 1, 2015 – June 1, 2016
June 1, 2014 – June 1, 2015
June 1, 2013 – June 1, 2014
Basically, the experience period comprises four years, but only the losses occurring during three oldest years count in the mod calculation. Injuries occurring before June 1, 2013, and after June 1, 2016, in this example do not count towards the work comp mod with a Rating Effective Date of June 1, 2017.
The worksheet presents the experience period from oldest (top of the form) to newest (bottom of the form) audited policy years.
Information Recorded Within Each Policy Period of the Experience Period
Each of the three policy periods used as the experience period are presented separately. Several pieces of information are provided for each policy term.
Code: This is the class code assignable to the insured. The number of class codes is a function of two factors: 1) the governing class code, and 2) exceptions to the governing class code rules.
ELR: The "Expected Loss Rate" is the amount of expected losses per $100 of payroll for the specific class code. Each state develops a state-specific ELR for each class code. How the ELR is presented may differ depending on which bureau is promulgating the worksheet (NCCI or state rate bureau). NCCI uses a decimal basis whereas some states present the ELR without a decimal. NCCI may present the ELR as 2.35 while a state rate bureau may use 235. If there is no decimal, remember to input it following the second number from the right (235 = 2.35 and 076 = 0.76). The ELR is applied to the audited payroll to develop expected losses.
D-Ratio: This ratio is used to develop the expected primary losses (Exp Prim Losses). Basically, it's the percentage of total expected losses (calculated using the ELR) that the bureau expects to fall below the relevant split point and thus be considered primary losses. Like the ELR, the D-Ratio is state-specific. The D-Ratio is similar to the ELR in that it may also be presented with or without a decimal; just remember to apply the same "decimal" rule presented under ELR when there is no decimal shown.
(Note: The ELR and D-Ratio factors used in the worksheet are those in effect when the experience mod is calculated, NOT those in effect during the policy years of the experience period. All policy years are assigned the same ELR and D-Ratios. However – and there is always a however – there are at least two states that do not apply this rule – Pennsylvania and Delaware. ELR and D-Ratios track by year in those states, not the factors in effect when the mod is calculated.)
Payroll: As the name suggests, this is the audited payroll assignable to the specified class code. If the payrolls are incorrect, expected losses are skewed and the final mod may be incorrect. Because the audited payrolls are used to develop expected losses, the experience mod is skewed if the audit is incorrect, this is one key reason a close review of the audits is required.
Expected Losses: These are developed by applying the ELR to the payroll of each individual class code.
- (Class Code Payroll / 100) X ELR = Expected Losses
This calculation establishes the total of expected losses for that class code for that given policy year. Ultimately all expected losses are added together and the total is input in box (D) in the calculation section.
Exp Prim Losses: Once Expected Losses are calculated, Expected Primary Losses are developed. Again, this is the amount the bureau expects to fall within the classification of "primary losses" for the specific class code in a given year. All expected primary losses are added together and the total is placed in box (E) for mod calculation.
(Primary vs. Excess Losses: To lessen the effect of a single large loss on the "mod," and enhance the fact that the loss occurred (frequence), losses are broken into two parts – "Primary" and "Excess." Primary losses are given more "weight" in the mod calculation as there is no weighting or credibility factor applied to these losses in the final calculation. Although Excess losses are considered and applied as part of the "mod," they are subject to a "credibility" factor lowering the amount of the excess loss that is considered in and applied to the calculation.)
Claim Data: Claims are presented in one of two ways: 1) as an individual claim (with the claim number provided); or 2) as a group (using the number of claims in the group). Whether the claim is presented individually or can be grouped is a function of the claim amount. Small claims (generally below $2,000) can be grouped together into one claim amount. Claims above $2,000 must be presented individually. One important note regarding the information contained in the claims data column – it is NOT code specific. Just because claims data is in the same row as a specific class code does not mean the loss was to a person in that class code. In fact, the class code applicable to a specific claim is irrelevant because the experience mod calculation is ultimately a comparison of actual losses versus expected losses. This will make more sense as we progress.
IJ: The type, classification or severity of the injury(ies) is reported in the Claim Data column. Nine IJ codes are available:
1 – Death
2 – Permanent Total Disability
3 – Major Permanent Partial Disability
4 – Minor Permanent Partial Disability
5 – Temporary Total or Temporary Partial Disability
6 – Medical Only
7 – Contract Medical or Hospital Allowance
8 – Compromised Death – CA only
9 – Permanent Partial Disability
"Medical only" claims (IJ code 6) are often reduced in the experience mod calculation. Many states include only 30 percent of medical only claims in the experience mod calculation. These are called ERA or "Experience Rating Adjustment" states.
The ERA-approved states appear to include (as of the date of this writing): Alabama, Alaska, Arkansas, Arizona, Connecticut, Florida, Georgia, Hawaii, Idaho, Illinois, Indiana, Kansas, Kentucky, Louisiana, Maine, Maryland, Michigan, Minnesota, Mississippi, Montana, Nebraska, Nevada, New Hampshire, New Mexico (?), North Carolina, Oklahoma, Rhode Island, South Carolina, South Dakota, Tennessee, Utah, Vermont, Virginia, West Virginia, Wisconsin, and the District of Columbia.
Small, grouped claims must be grouped based on the IJ code.
OF: Indicates the status of the claim(s) presented. "O" indicates the claim is open. "F" means "final" or closed.
Act Inc Losses: These are the total actual incurred losses taken directly from the insured's loss runs. Many states cap large actual incurred losses and each state develops its own maximum loss for individual and "group" (catastrophic) losses. Remember, incurred losses include the amount paid and the amount reserved. All Actual Incurred Losses are added together and the total is found in box (H) in the mod calculation section.
Act Prim Losses: The insured's Actual Primary Losses are recorded in this column. Grouped losses are carried over in full; but large individual losses are subject to a "split point." Only the incurred loss amount up to the split point is included in this column. Until relatively recently the split point was set at $5,000; meaning that only the first $5,000 of an individual loss was recorded. The split point is now designed to fluctuate and is currently at $16,500. If an individual loss is $20,000, only $16,500 is input in this column. Once all actual primary losses are developed, they are added together and recorded in box (I) in the calculate area of the form.
Compiling the Information to Calculate the Mod
Nearly all the information necessary to calculate the experience mod comes from the insured: class codes, audited payrolls and losses. But to calculate the mod, NCCI or the state rate bureau must provide two factors or values: 1) the weighting factor (W), and 2) the ballast value (B).
Box (A) contains the weighting factor - also known as the credibility factor and box (G) applies the ballast value.
The weighting (W) or credibility factor represents the rating authority's opinion about the credibility of the loss data as it relates to the ability to predict future losses. The higher the number, the more weight or credibility is given to the loss data; and, likewise, the lower the number, the less credible the past losses are considered as a factor for predicting future losses.
In general terms, the credibility factor is based directly or indirectly on the insured's premium or payroll amounts – specifically expected losses; and expected losses are a function of the payroll and the expected loss ratio (ELR). The smaller the risk, the less weight is given to past and expected losses – resulting in a low weighting factor; conversely, the larger the risk, the greater the weight given to past and expected losses – resulting in a higher credibility factor. In essence, the credibility factor is insured-specific.
Two functions served by the weighting factor are:
- It is applied to excess losses (expected and actual) limiting the amount of each used in the calculation; and
- Its inverse is used to develop the stabilizing value.
The ballast value is based on the size of the risk; the larger the risk, the higher the ballast amount. Like the "W" factor, the ballast value is promulgated by the authority developing the mod or the mod factors. As its name suggests, the ballast value is designed to avoid too great of movement away from the center/base (a mod of 1.00). It is part of the stabilizing value, along with the inverse of the Weighting Factor, applied to both actual and excess values in the calculation.
Putting it ALL Together
Now that we understand all the factors, let's review "boxes" A – K and calculate the experience mod.
Box (A): The weighting factor as supplied by NCCI or other rating bureau.
Box (B): No one knows why this box exists, maybe for symmetry. Some states supposedly use it for the ballast value, but other than that, this box is generally blank.
Box (C): Expected Excess Losses. Developed by subtracting total expected primary losses (EXP PRIM LOSSES) found in box (E) from total expected losses (Expected Losses) found in box (D).
(D) – (E) = Expected Excess Losses (C)
Box (D). Total Expected Losses.
Box (E). Expected Primary Losses.
Box (F). Actual Excess. Developed by subtracting the actual primary losses found in box (I) from actual incurred losses presented in box (H).
Box (G). Ballast Value.
Box (H). Actual Incurred Losses. However, this amount may not be the total of the actual incurred losses (ACT INC LOSSES) presented during the experience period in ERA states. In ERA states, the medical only losses are reduced by 70 percent (only 30 percent of these losses count towards the total).
Box (I). Actual Primary Losses (ACT PRIM LOSSES). Like box (H), this is the total of the actual primary losses developed in the experience period section of the worksheet. And also like box (H), the amount input in box (I) is actually the reduced total of primary losses if the risk is in an ERA state and there are medical only losses.
The only "boxes" not yet addressed are (J) and (K).
- (J) provides is the ratable "Actual Total" losses; and
- (K) presents the ratable "Expected Total" losses.
So, the experience mod is a function of actual compared to expected. The ratable actual losses in (J) are divided by the ratable expected losses in (K) to develop the experience mod.
Now! Let's Calculate the Mod
NCCI's calculation is presented first; but not every state promulgates the experience mod like NCCI. Several state variations are presented as well. This is how the bottom of the worksheet reads:
| ||Primary Losses|| ||Stabilizing Value|| ||Ratable Excess|| ||Totals|
|Actual ||(I)||+||C * (1 – A) + G||+|
- * (F)
|Expected||(E)||+||C * (1 – A) + G||+|
- * (C)
| || || || || || || ||Exp. Mod|
| || || || || || || ||(J) / (K)|
To aid in comparing calculations between NCCI and other states, alternative lettering is used so that NCCI and state differences can be seen more clearly because the states don't always use the same lettering system as NCCI. To help further, NCCI's "box" letters are used with this key to make the correlation easier.
|AI = Actual Incurred Losses (H)||EE = Expected Excess Losses (C)|
|AP = Actual Primary Losses (I)||WV = Weighting Factor (A)|
|AE = Actual Excess Losses (F)||BV = Ballast Value (G)|
|TEL = Total Expected Losses (K)||TAL = Total Actual Losses (J)|
|EP = Expected Primary Losses (E)||WMV = Weighted Maximum Value|
|Experience Mod =||AP + (EE x (1-WV) + BV) + (WV x AE)|
|||EP + (EE x (1-WV) + BV) + (WV x EE)|
Michigan, New York, North Carolina and Wisconsin
|Experience Mod =||AP + BV + (AE x WV) + ((1-WV) x EE|
|||EP + BV + (EE x WV) + ((1 – WV) x EE|
Delaware and Pennsylvania
|Experience Mod =||(AI x WV) + (TEL x WMV) + TEL x (1-WV))|
|Experience Mod =||1+||((AI-TEL) x WV) + ((AP-EP) x (1-WV))|
|||||TEL + BV|
|Experience Mod =||(AE x WV) + (AP x WMV) + (EE x (1 - WV)) + (EP x (1 - WMV))|
|Experience Mod = ||[(AP x WV) + (EP x (1 – WV))] + [(AE x WMV) + (EE x (1 – WMV))]|
Hopefully this takes the uneasiness out of reading the experience mod worksheet. Learn more about the complexities of work comp with our on demand webinar recording that addresses the problems created when employees travel to other states to work and the importance of employers' liability coverage.
Last Updated: June 15, 2017