DECISION AND ORDER

LYNN ADELMAN, District Judge.

Mai Nhia Thao seeks to represent a class comprising owners of certain life insurance policies issued by Midland National Life Insurance Company ("Midland"). Before me now is Thao's motion to certify this case as a class action pursuant to Federal Rule of Civil Procedure 23.1

Thao asks that I certify the following class and designate her as the class representative:

The base policy forms identified in the class definition are components of universal life insurance policies. Universal life insurance can be contrasted with term life insurance and whole life insurance. With term life insurance, the insured pays a premium in exchange for a death benefit that the insurer pays to the beneficiary only if the insured dies during the term of the policy. See 1 Jeffrey E. Thomas & Francis J. Mootz III, New Appleman on Insurance Law Library Edition § 1.08[2][b][ii] (2011). With whole life insurance, the insured pays a premium in exchange for both a death benefit (the insurance component) and a savings plan (the savings component). The savings component can be described as an investment that gets bigger over the term of the policy. Should the insured choose to do so, she can "surrender" the policy in exchange for whatever cash value has accumulated in the savings component at the time of the surrender. Id.

Universal life insurance, like whole life insurance, has both an insurance component and a savings component. Universal life is different than whole life in essentially two ways—flexibility and transparency. Universal life is said to be more flexible than whole life because it allows the insured to make various decisions over the life of the policy. With whole life, the insured pays a fixed premium over the life of the policy, and the death benefit is fixed at the policy's face value. With universal life, the insured can pay premiums in almost any amount at almost any time and may increase or decrease the amount of the death benefit at any time (subject to certain limits). See Richard G. Schectman, New Concepts in Life Insurance Planning: Universal Life, 13 Cumb. L. Rev. 219, 222 (1982). Relatedly, the insured can use the savings component to pay for the death benefit. That is, instead of making premium payments, the insured can allow the cost of the insurance component to be deducted from the savings component. So long as the policy's cash value is sufficient to pay the cost of the death benefit, skipping a premium does not result in the policy's lapsing. Id. at 224. Universal life is said to be more transparent than whole life because the policyholder receives periodic statements detailing how her premiums are being used. The statements itemize the various costs, fees and expenses that the insurer deducts from her premiums. With whole life, the policyholder typically does not know how her premiums are being used. Id. at 224-25; Douglas I. Friedman, Universal Life: Product Development and Tax Aspects, 13 Cumb. L. Rev. 499, 503-04 (1982).

The mechanics of the Midland universal life policies at issue in this lawsuit are as follows: When a policy is purchased, the policyholder pays an initial premium and specifies the amount of the death benefit. Midland deducts a "premium load" from this initial payment and then applies the rest to the "policy fund," which is the savings component. If the policyholder makes additional premium payments, Midland adds that payment to the policy fund. Each month, Midland calculates the policy fund by taking the existing amount in the fund, adding any premium payments (and any interest earned on the savings component), and subtracting various charges. If after the relevant additions and subtractions are made there is still money left in the policy fund, then the policy will continue in force. If there is no money left in the policy fund, then (subject to certain exceptions), the policy will terminate.

The focus of this lawsuit is on the cost-of-insurance charge, which is one of the charges that Midland deducts from a policyholder's policy fund each month. In general, "cost of insurance" refers to the amount that an insurance company charges to cover its risk—i.e., the cost of paying the death benefit upon the death of the insured. Midland's policies specify that Midland will calculate cost-of-insurance charges by multiplying a "cost-of-insurance rate" by the difference between the amount of the death benefit and the amount of the policy fund.2 The cost-of-insurance rate is taken from a set of tables produced by Midland's actuaries. Each policy type—that is, each Midland "product"—will have its own set of tables. The rates on the tables are organized by certain characteristics: the insured's age when the policy was issued ("issue age"), the number of years the policy has been in force ("policy years"), the insured's sex, the amount of the death benefit (the "specified amount"), and the insured's "premium class" (which is determined by certain characteristics that affect the insured's mortality risk, such as whether the insured uses tobacco or participates in hazardous activities). These characteristics determine which rate Midland will apply to a given policyholder at a given time. For example, suppose that the insured is a female nonsmoker (as Thao is) who purchased a Century Universal Life-G policy with a $100,000 death benefit (as Thao did). To calculate her cost-of-insurance rate for a particular month, Midland will first find the set of tables designed for the Century Universal Life-G product. It will then find the table in that set that contains rates for female non-smokers who selected a death benefit of between $100,000 and $1 million.3 The table will have cost-of-insurance rates organized by issue age and policy years—the rows are issue ages and the columns are policy years. Thus, assuming that the insured was twenty-six when the policy was issued and the policy has been in force for ten years, Midland will select the rate in the cell of the table located at the row corresponding to age twenty-six and the column corresponding to ten years' duration. Midland will then plug that rate into the cost-of-insurance formula to determine the cost-of-insurance charge.

Thao's contention in this lawsuit is that Midland has been setting its cost-of-insurance rates in a way that is inconsistent with the following provision in her policy: "Cost of Insurance rates are based on the Issue Age, completed Policy Years, Sex, Specified Amount, and Premium Class of the Insured." Thao Policy § 7.7, ECF No. 57-1. Now, Midland's cost-of-insurance rates are in some sense "based on" issue age, policy years, sex, specified amount, and premium class in that, as explained above, Midland's cost-of-insurance rate tables are organized by these five factors. However, Thao contends that the policy language requires more than using tables organized by the five factors. According to her, the policy language imposes a constraint on how Midland sets the rates that appear in the cells of the tables. She contends that the policy language required Midland to consider nothing other than the five factors listed in the policy—which are all factors related to mortality expectations—when setting those rates. Thao contends that, contrary to this language, Midland considered factors unrelated to mortality expectations when setting its rates.

Thao contends that this case is appropriate for class treatment. She points out that all of the proposed class members' policies incorporate one of nine base policy forms and that each of the nine forms contains the same or virtually the same operative language as her policy. She also points out that there are no significant differences in state contract law or individualized issues concerning extrinsic evidence that might result in the policies' having different meanings for different class members. Thus, argues Thao, whether the policies allowed Midland to consider factors unrelated to mortality expectations when setting cost-of-insurance rates is a common question of law or fact. See Fed. R. Civ. P. 23(a)(2).

It is true that whether Midland acted contrary to language that appears in all of the proposed class members' policies is a common question of law or fact. The operative language in all of the base policy forms is the same or virtually the same, and differences in state law and extrinsic evidence will likely not result in different meanings for different class members. Thus, there is likely a single answer for all policyholders: either the policies allowed Midland to consider factors unrelated to mortality expectations when setting its cost-of-insurance rates, or they did not. However, it does not follow that this case is appropriate for class treatment. This is so because, as we will see, the class members may not agree on what the answer to the common question should be. Some class members might prefer rates that are not based exclusively on mortality expectations, while other class members might, like Thao, prefer rates that are based exclusively on mortality expectations. Put differently, there may not be a common claim that Thao can litigate on behalf of a class of policyholders. Thus, in the remainder of this opinion, I examine whether Thao's individual claim is representative of claims that other Midland policyholders might have.

Thao's individual claim is that Midland is overcharging her. More specifically, she claims that Midland is deducting too much from her policy fund each month because the cost-of-insurance charge is based on rates that are higher than they would be if Midland had set its cost-of-insurance rates using only the five factors listed in the policy. Essentially, Thao claims that if Midland had set its rates using a methodology based exclusively on the five factors, all of the numerical values in the table containing her cost-of-insurance rates would be either lower than they are now or unchanged.4 According to Thao, all of the other proposed class members have the same claim against Midland. She contends that if Midland had set its cost-of-insurance rates using only the five factors listed in the policy, then all of the numerical values in all of Midland's tables would be either lower than they are now or unchanged. Thus, argues Thao, all members of the proposed class have a claim against Midland that arises out of Midland's use of improper cost-of-insurance rates. She believes that each proposed class member would want (1) declaratory and injunctive relief requiring Midland to adjust its rates so that they are based exclusively on the five factors and (2) a refund of the excess cost-of-insurance charges paid in the past on account of Midland's using rates that were based on other factors.

The key premise in Thao's argument for class certification is that if Midland had set its cost-of-insurance rates using only the five factors listed in the policy, then all of the numerical values in all of Midland's tables would be either lower than they are now or unchanged. If this premise is false—if, for example, a set of rates based exclusively on the five factors would result in higher rates for certain policyholders—then the members of the proposed class would not have a common claim against Midland.5 Only those members who would pay lower rates would have a claim, because only those members would have been injured by Midland's conduct. See Wal-Mart Stores, Inc. v. Dukes, 131 S.Ct. 2541, 2551 (2011) (holding that to show commonality plaintiff must show that the class members have suffered the same injury); Bieneman v. City of Chicago, 864 F.2d 463, 465-66 (7th Cir. 1988) (individual opposed to noise created by O'Hare airport could not represent a class of property owners in area around airport because some property owners benefitted from proximity to airport). Thus, before the class proposed by Thao may be certified, Thao must prove—not merely assert—that if Midland had set its cost-of-insurance rates using only the five factors listed in the policy, then all of the numerical values in all of Midland's tables would be either lower than they are now or unchanged. See Wal-Mart, 131 S. Ct. at 2551-52 (satisfaction of all Rule 23 requirements must be proved, not merely asserted, even if proving those requirements requires resolving an issue relevant to the merits of the underlying claim); Szabo v. Bridgeport Mach., Inc., 249 F.3d 672, 675-76 (7th Cir. 2001) (same).

In an attempt to prove that under her interpretation of the policies all policyholders would pay lower cost-of-insurance rates, Thao points to a collection of spreadsheets known as the "COI Solver" workbooks. See Pl.'s Br. in Supp. at 7-8, ECF No. 57. In the next few pages of this opinion, I will explain what these workbooks are and why they are important to plaintiff's argument for class certification.

The COI Solver workbooks are a set of Excel spreadsheets that Midland's actuaries developed and used in the course of designing the policies at issue in this case. See Bill Decl. ¶¶ 24-27, 37, 39, ECF No. 64. As Midland seems to be using the term, "designing" refers to the overall process of setting the rates that will determine the various credits to and deductions from a policy that Midland will make over the life of that policy. Id. ¶ 24. The cost-of-insurance rates are set during this process, but other rates, such as interest-crediting rates, are also set during this process. In designing a policy, Midland's actuaries attempt to set rates that will both further the policyholder's objectives in purchasing the policy and earn Midland a reasonable profit. Id. ¶¶ 24, 28. Although Midland's objective—making money—is the same in every policy, the policyholder's objectives may be different. For example, some policyholders might want a policy that is designed to have a high cash surrender value at a certain point in time, while other policyholders might want a policy that is designed to have a high death benefit and lower cash value. Id. ¶ 24.

Part of Midland's design process is known as "pricing." Id. ¶¶ 31-38. During pricing, Midland's actuaries feed a proposed set of rates into a computer program to examine various financial metrics relating to Midland's profitability (such as cash flow and distributable earnings). The actuaries assign probabilities to uncertain future events and feed those probabilities into the program along with the proposed rates. After running the program, the actuaries use the results to get a sense of what is likely to happen in the future if Midland adopts the proposed rates. The goal of this process is to determine whether the rates are viable—not so low that Midland loses money, but not so high that Midland loses business to its competitors. If after testing the proposed rates Midland's actuaries determine that the rates are problematic, they will make adjustments. The actuaries might revise a single rate scale or multiple rate scales, and the cost-of-insurance rate scale is one of the scales that might be adjusted.

The purpose of the COI Solver workbooks is to expedite the process of developing scales of cost-of-insurance rates for testing in the pricing program. Id. ¶¶ 37 & 39. As noted, the workbooks are Excel spreadsheets. Numerical values are entered into the cells of the spreadsheets in a manner that allows arithmetical operations to be performed in accordance with specified formulas. A workbook for a particular policy will contain many spreadsheets—one for every permutation of policyholder characteristics (issue age, sex, premium class, etc.). Here is an excerpt from a COI Solver spreadsheet for a particular Midland product (labeled "UL-CV") and permutation of policyholder characteristics (male nonsmoker "MN", issue age 35, and band 2):

When Midland's actuaries create a spreadsheet for a given permutation of policyholder characteristics, they start by entering a base scale of numbers into one of the columns. Each row of the spreadsheet represents a policy year. Any set of numbers having some reasonable pattern could be used as a base scale. Id. ¶ 40. However, Midland's actuaries have chosen to use numbers that resemble (but do not match) Midland's "pricing mortality" rates.6 Id. The reason for this is that mortality rates generally increase with age, and in most policies the design objectives will call for cost-of-insurance rates that increase with age. Thus, using an approximation of Midland's pricing-mortality rates as a base scale gives the actuaries a convenient starting point. Id. However, there is nothing special or significant about the actuaries' decision to use an approximation of Midland's pricing-mortality rates as a base scale. The actuaries could just as easily have used any one of a number of different mortality tables published for use in actuarial science. Id. ¶¶ 40-41. Or the actuaries could have used something other than a mortality table. For example, in one Midland policy (which plaintiff has excluded from the class definition), the actuaries used reinsurance premium rates as the base scale. Id. ¶ 41.

The column of the COI Solver spreadsheet containing the base scale for a particular permutation of policyholder characteristics is labeled "Pricing Mortality."7 To turn this base scale into usable cost-of-insurance rates, the COI Solver spreadsheet relates the numerical values in the Pricing Mortality column to values in other columns in accordance with a specified formula. One of the other columns is the "Select Factor" column. The first step in turning the base scale into cost-of-insurance rates is to multiply the values in the Pricing Mortality column by the values in the Select Factor column. The numerical values that result from this operation appear in the column labeled "Product."8 The value in the Select Factor column is usually one or below for non-tobacco users and one or above for tobacco users, and thus the Select Factor usually decreases the numbers in the base scale for non-tobacco users and increases them for tobacco users. Id. ¶ 45. However, the value of the Select Factor generally decreases as the number of policy years increases—that is, the value gets smaller as it moves down the Select Factor column and into the lower rows, which represent later policy years. Thus, use of the Select Factor has the effect of producing lower rates in later policy years. Id. ¶ 49.

Three columns in the COI Solver spreadsheet involve something known as the "Add-On." As we will see, the Add-On is the centerpiece of Thao's theory for class certification. The three columns are labeled "Add-On," "Mult of Add-On," and "Tot Add-On." The values in the Add-On column are taken from a separate spreadsheet in the workbook labeled "Expense Add-On." Id. ¶ 48. This spreadsheet is organized by issue age and policy year. Id. ¶ 47. According to Midland, the values that appear in the cells of the Expense Add-On spreadsheet do not represent any particular cost, expense, or mortality characteristic. Instead, those numbers are nothing more than a scale that Midland adjusts as needed to meet design objectives—i.e., producing cost-of-insurance rates that meet policyholder objectives and Midland's objective of earning a reasonable profit. Id.

Once the values in the Expense Add-On spreadsheet are imported into the COI Solver spreadsheet, the values in the Add-On column are multiplied by the values in the Multiple of Add-On column (which are determined by the sex, premium class and specified amount of the policyholder), and the result of this operation appears in the Total Add-On column. Id. The values in the Total Add-On column are then added to the values in the Product column. The result of this operation appears in the "Product + Add-On" column.

The Add-On is generally structured so that its value declines towards zero over some initial number of policy years. Id. ¶ 49. That is, the greatest numerical values in the Add-On and Total Add-On columns generally appear at the top of the columns (i.e., in the rows representing early policy years), and then the values gradually decline towards zero, and eventually become zero, as they move down the columns. Thus, the purpose of the Add-On is to adjust the base rate upwards during the initial years in which the policy is in force. This is contrary to what Midland would do if the Add-On were related to mortality characteristics, since the risk of dying generally increases with age rather than decreases. However, for various reasons, including producing favorable tax consequences, some policyholders prefer to pay higher rates in earlier policy years in exchange for lower rates in later policy years. Id. ¶ 50. Using the Add-On, which generally produces higher rates in earlier years, in conjunction with the Select Factor, which generally produces lower rates in later years, helps Midland's actuaries produce a rate scale that achieves this particular policyholder objective. Id. ¶¶ 49-50.

The final step in the COI Solver process is to determine the final rate for testing. This is done by comparing the value in the Product + Add-On column for a particular policy year to the value in the column labeled "Composite Guar COI," which is the maximum cost-of-insurance rate allowed under the policy.9 Id. ¶ 42. If the value in the Product + Add-On column is less than the value in the Composite Guarantee column, then the value in the Product + Add-On column becomes the cost-of-insurance rate for that policy year. Otherwise, the value in the Composite Guarantee column becomes the cost-of-insurance rate for that policy year. The cost-of-insurance rate scale thus appears in the column labeled "Mo COI." This is the rate scale that Midland's actuaries will feed into their pricing program to determine whether that scale—together with all other rate scales for the policy under consideration—will produce the desired results. If after testing the actuaries determine that adjustments to the cost-of-insurance rate scale are needed, the actuaries will adjust the values in the Add-On and Select Factor columns (and possibly other columns that I have not discussed) as needed to produce a revised cost-of-insurance rate scale for testing. Id. ¶¶ 44 & 50. This process will continue through as many iterations as needed to produce a rate scale that satisfies all of the objectives for the policy under consideration. Id. ¶ 37. Once Midland's actuaries find that rate scale, Midland will use it to construct the cost-of-insurance rate tables for that policy.

Having explained the function of the COI Solver workbooks, I can explain why they are important to Thao's theory for class certification. Recall that a class may be certified only if Thao can show that all class members would pay lower cost-of-insurance rates than they do now if Midland's rates were based exclusively on the five factors listed in the policies. In an attempt to make this showing, Thao focuses on the Add-On. The Add-On has the effect of producing higher cost-of-insurance rates during early policy years because the value of the Add-On, which is positive in early policy years, is added to the Product (i.e., the base scale multiplied by the Select Factor). Thus, as a matter of arithmetic, if the Add-On were removed from Midland's cost-of-insurance rates—that is, if the values in the Add-On columns were all set to zero—all class members would pay lower cost-of-insurance rates than they do now in early policy years, but their rates in later years (in which the Add-On is already zero) would not change. Thao contends that all class members thus have a claim against Midland for removal of the Add-On.

The problem for Thao is that her theory for class certification does not match her theory for the merits of this case. In order to obtain class certification, Thao argues that Midland must recalculate all of its cost-of-insurance rates after removing the Add-On from the COI Solver workbooks. However, that remedy is not consistent with Thao's interpretation of the policy language. Under Thao's interpretation of the policy language, cost-of-insurance rates must be based exclusively on Midland's mortality expectations—i.e., Midland's estimates of future death rates for groups of policyholders as determined by age, sex, tobacco use, etc. However, no column in the COI Solver workbooks contains Midland's mortality expectations. Although the base scale is loosely related to mortality expectations in that its values are larger in later policy years than in earlier policy years, the base scale does not itself constitute Midland's mortality expectations for any set of policyholder characteristics. See Bill Decl. ¶¶ 40-41, 63. Likewise, although the value of the Select Factor is generally larger for tobacco users than for non-tobacco users, the Select Factor does not reflect Midland's expectations regarding the effect of tobacco use on mortality. Id. ¶ 47 (stating that the Select Factor has no independent meaning). Removing the Add-On would leave the base scale multiplied by the Select Factor, but since neither the base scale nor the Select Factor (nor the product of the two) is based on Midland's mortality expectations, removing the Add-On from the rates produced by the COI Solver workbooks would not transform a rate scale that is not based exclusively on mortality expectations into one that is. Thus, even if Thao is correct on the merits and the policy language requires Midland to base its cost-of-insurance rates exclusively on mortality expectations, it would not follow that removing the Add-On would bring Midland's existing rates into line with that language.

In an attempt to show that removal of the Add-On is consistent with her interpretation of the policy language, Thao contends that the Add-On represents a surcharge for Midland's "expenses." However, even if the Add-On were a surcharge for expenses, which it is not,10 that would not change the fact that removing the Add-On from Midland's existing rates would not leave rates that are based exclusively on mortality expectations. No matter what the Add-On is or what it represents, it simply is not the case that removing the Add-On from Midland's current rates would produce rates based exclusively on mortality expectations. Again, since no part of the COI Solver workbooks is based on mortality expectations in the first place, removing the Add-On from those workbooks would not result in the creation of rates based exclusively on mortality expectations.

Thus, Thao's removal-of-the-Add-On remedy is not consistent with her interpretation of the policy language. And when we set that remedy aside, we can see that the class does not have a common claim against Midland. This is best illustrated by showing what would happen if Thao's interpretation of the policy language were adopted, as I do below.

Recall that under Thao's interpretation of the policy language, Midland's cost-of-insurance rates must be based exclusively on mortality expectations. There is one rate scale that Midland uses that is based exclusively on mortality expectations, and that is Midland's pricing-mortality rates. See Bill Decl. ¶ 53 (explaining that pricing-mortality rates are estimates of uncertain future death rates).11 However, many class members would prefer Midland's current cost-of-insurance rates to rates based on Midland's pricing-mortality rates. In many (if not all) policies, Midland's current rates are higher than pricing-mortality rates during early policy years and lower than pricing-mortality rates in later policy years. Id. ¶ 50. That is because the design objectives for those policies called for charging a cost-of-insurance rate that is higher than expected mortality in early policy years in exchange for charging a rate that is lower than expected mortality in later policy years, as illustrated in the following graph:

This graph is for a particular policy design (UL-CV) and a particular permutation of policyholder characteristics (male, non-tobacco premium class, issue age forty-five, and specified amount falling within band two). What it shows is that until about the fifteenth policy year, Midland's current cost-of-insurance rates are higher than they would be if Midland had used its pricing-mortality rates as its cost-of-insurance rates. However, in later policy years, Midland's cost-of-insurance rates are lower than they would be if Midland had used its pricing-mortality rates. Many policyholders will prefer to pay higher rates in early policy years in exchange for lower rates later in life—that is, they will prefer the shape of the cost-of-insurance curve to the shape of the pricing-mortality curve. One reason why they may prefer the shape of the cost-of-insurance curve is that it has certain tax advantages. Id. In any event, whatever their motives may be, the fact is that some policyholders will prefer Midland's current cost-of-insurance rates to rates based on pricing mortality. Thus, many class members would prefer Midland's interpretation of the policy language to Thao's.

In an effort to avoid this problem, Thao devised her removal-of-the-Add-On remedy, which basically results in bringing the cost-of-insurance rates down to pricing-mortality rates in early policy years but leaving the lower-than-pricing-mortality rates in place in later policy years. However, as already discussed, removing the Add-On from Midland's current rates is not a remedy that follows from Thao's interpretation of the policy language.

The problems identified above show that the class that Thao seeks to represent does not comprise policyholders with a common grievance against Midland. Instead, the class contains many policyholders who would prefer Midland's interpretation of the policy to Thao's. Thus, Thao's motion for class certification will be denied. To the extent that any individual policyholders feel that they are being injured by Midland's current cost-of-insurance rates, they must bring their own individual suits.

The remaining issue is Midland's motion to strike, which I can deal with quickly. Midland moves to strike the expert reports that Thao attached to her reply brief along with all arguments in the reply brief that are based on such reports on the ground that those materials should have appeared in Thao's opening brief rather than in her reply brief. However, I conclude that the reports and the arguments employing them were properly included in Thao's reply brief, as they are directly responsive to arguments made in Midland's brief in opposition and supporting materials.

Accordingly, IT IS ORDERED that Thao's motion for class certification is DENIED.

IT IS FURTHER ORDERED that Midland's motion to strike is DENIED.

FINALLY, IT IS ORDERED that a telephonic status conference will be held on June 7, 2012 at 11:00 a.m. for the purpose of determining whether further proceedings are necessary. The parties should confer in advance of the call and be prepared to propose a plan for proceeding. Counsel must call the court at 414/297-1285 to advise of their participation.