THE INVESTIGATION of contingency payments that started last year in one way or another will affect even the smallest independent insurance agency-and probably every insurer and general agency using independent agents and brokers. These investigations focus on whether these various incentives have prompted agents and brokers to improperly steer clients to carriers that provide the best contingency bonus to them, rather than offer the best coverage or price to their clients.
Some in the insurance industry, myself included, have contended that true performance-based contingency contracts benefit all parties involved. The most important factor affecting bonuses in performance-based contingency contracts is loss ratio. (The bonus also can be affected by volume, growth and retention, however.) When an agency can earn a bonus based on performance, the agency encourages insureds to practice loss control, which, if successful, reduces the insureds' insurance cost (plus they avoid all the uninsurable costs associated with claims-e.g., time spent on defense). When clients experience fewer, smaller claims, the agent's loss ratio drops, which generates a bigger profit-sharing bonus for the agent. With lower loss ratios, insurance companies pay less for claims, so their profits increase. Everyone wins.
(For purposes of contingency contracts, the loss ratio generally is defined as losses divided by earned premium. The definition of losses varies considerably from company to company, though. Some include one or more of the following: allocated loss adjustment expenses, incurred but not reported losses and unallocated loss adjustment expenses. These elements constitute a significant bone of contention between agencies and companies, as will be noted later. Other factors affecting the loss ratio include whether the losses are calendar-year or accident-year, and whether a stop loss adjusts the applicable loss ratio. Some companies calculate loss ratios using one year of losses, while others use two, three and even four years of losses.)
Other types of "contingency" contracts are available to some large agents and brokers. These contracts, which include placement service agreements (PSAs) and market service agreements (MSAs), are volume-based, not performance-based. Some people argue that insurance companies do not gain cost efficiencies from size. In other words, a company's expense ratio doesn't decrease as it grows larger. Therefore, such people ask, why would a company offer bonuses for more business?
I agree that growth does not guarantee a company a lower expense ratio. However, companies do have reasons to pay bonuses for volume. Furthermore, it makes no sense to pay bonuses to agencies with little volume or to those that obviously are minimizing their relationships with the insurer. Also, depending on the line of business and reserving methodology, a growing company can show better loss ratios, because revenue is realized much earlier than losses. If investment opportunities are adequate, the upfront premiums also can generate additional income.
In time, though, losses catch up with premiums-sometimes with a vengeance. Therefore "cash flow" underwriting has few, if any, long-term benefits. Most carriers probably understand this but are tempted by the short-term gains, the desire to become a big player or the chance to capture market share. Competitive pressures also can force insurance companies to play the game just to maintain their current market share.
Testing an assumption
If performance, rather than volume, is the basis for most contingency contracts, then the amount insurance companies pay in contingency expense should vary with loss ratio. To test this assumption, Burand & Associates, with the help of Edward Fuller, Ph.D., of Pfeiffer University, undertook an extensive statistical analysis of contingency expenses and loss ratios reported by insurance companies. (Dr. Fuller's areas of expertise are computer information systems and quantitative methodologies.) The study covered all P&C companies' 2002 results, as reported in a customized database purchased from A.M. Best. We consulted the 2003 edition of Best's Aggregates & Averages (reporting 2002 industry results) to obtain total industry data, later cited in this report, that was not available in the database we bought.
We analyzed loss ratios relative to contingent commissions, both in dollars and as a percentage of written premiums. Surprisingly, we found NO relationship between companies' loss ratios and their contingency expenses. In other words, the bonuses paid were random, relative to loss ratios.
The product of our calculations was a series of "R square" values. Any R-square less than .15 indicates no relationship between two variables. The higher the value above .15, the greater the relationship between the two variables. (Complete information about the methodologies used in this study can be found on our Web site, www.burand-associates.com.)
Our analysis of the relationship between contingencies and loss ratios of direct writers showed an R square of .0019. The lack of a relationship is to be expected, because direct writers generally do not pay contingencies. More interesting though, our analysis of agency insurers' contingencies and loss ratios showed an R square of 0.0001-even less than that of the direct writers.
There are several possible reasons for the randomness of contingencies relative to loss ratios. Among them is the strong probability that the loss ratios insurers experience are materially different from those upon which contingency bonuses are paid. For example, asbestos and environmental reserve strengthening will increase an insurer's loss ratio, but it will not always affect individual agency and broker loss ratios upon which bonuses are paid. Contingencies earned by the companies themselves (via reinsurance) might also affect the overall results. (We did attempt to minimize this factor by using direct data, rather than net.) Another probable factor is that costs that are not really contingencies are reported as contingency expenses, although the impact of this factor probably is small.
Another factor worth considering is the effect that a few brokers with extremely large market shares could have on the results. If these brokers were earning volume-based bonuses (PSAs or MSAs) rather than loss-ratio-based contingencies, their volume could be large enough to skew the results. (According to one report, Marsh McLennan, Aon and Willis together account for 61% of brokerage revenue worldwide.)
As part of our study, we also analyzed the relationship between contingencies and direct-written premiums. We found a much higher correlation between these two variables than we did between contingencies and loss ratios. For seven different groups of companies, as defined by A.M. Best aggregation codes, the R square ranged from .0149 to .5351. For five of the seven A.M. Best aggregation codes, the value was greater than .15, indicating at least a minimal relationship between contingencies and premium volume.
Contracts: the real McCoy
To look at this issue from another perspective, we analyzed the performance-based contingency contracts of 12 national and regional insurance companies. These contracts are typical of those possessed by more than 90% of all independent agencies.
These carriers accounted for 14.87% of 2003 premiums and 28% of the market share available to independent agents within the top 90 carriers. (The contracts actually represent a significantly higher market share because many of the carriers in the top 90 are not available to retail independent insurance agents on a direct basis, or only part of the company is available to independent agents. If the data were available to adjust for these factors, the market-share percentage of these companies would increase significantly.) Some of these companies had small market shares, and we included them to see if their contracts were significantly different from those of larger carriers. We found they were not.
The results of this second analysis were very different from those of the first. Using the actual contracts and simple regression analysis, we found that the correlation of contingencies and loss ratio for a hypothetical agency with $3 million in written premium, 10% growth and a 45% loss ratio is extremely high (R square of .76). The correlation between bonuses paid and volume is even higher (R square value of .87.) Despite the higher R square for volume, our analysis showed that changes in loss ratio had a much greater effect on expected bonuses than did changes in volume, demonstrating that it is the more important variable.
The results also showed a significant variance between what one would expect to be paid under the majority of contracts and the actual amounts paid by companies. According to our calculations, the insurers A.M. Best identifies as distributing products mainly through agents (as opposed to brokers, employees or other distributors) had an average loss ratio (including loss adjustment expense) of 70% to 74% in 2002 (depending on the segmentation code) on both a weighted and non-weighted basis.
(By "non-weighted," we mean we simply added up the reported loss ratios and divided the sum by the number of companies. By "weighted," we mean the sum of loss ratios times premiums per company divided by total premium for these companies. The premiums in this case, however, are written premiums. We could not use earned premiums, as one normally would when calculating loss ratios, because they were not included in this particular A.M. Best database).
Assume an agency has a 70% loss ratio with a company with which it places $2 million in premium volume. Assume it also has a 90% retention rate and a 5% growth rate. Under the contracts we analyzed, such an agency would earn, on average, a contingency bonus of just .02% of written premium. (See charts 1 and 2 below to see how the bonuses paid under these contracts vary with loss ratio.) According to the A.M. Best data, however, the average contingency paid (at the aforementioned 70% to 74% loss ratio) by all insurers primarily distributing through independent agents equaled 1% to 1.3% of written premium, depending on the segmentation code. (The average in this case is the "non-weighted" contingency percentages, calculated in the same manner as the "non-weighted" loss ratio previously described.) That is 4,900% to 6,400% higher than one would expect.
Overall, the results of our study suggest a lack of any relationship between what companies are paying in contingencies and the amount one would expect them to pay under the contingency contracts they have with the vast majority of their agents. As mentioned earlier, part of this is probably due to the carriers' loss ratios being affected by reserve changes while the agents' loss ratios are not always similarly impacted. However, these results also suggest that carriers probably are paying a lot of money simply for volume-as is the case with MSAs and PSAs. Both pay strictly for volume or growth, and both are available primarily to the largest brokers. This is not good for the insurance companies because they are paying these bonuses even when the business is not profitable. Conversely, with almost all true performance-driven contingency contracts, it does not matter how much volume an agency has with a given company; if the agency's loss ratio is inadequate, it will not get a bonus. These contracts are much better for insurance companies because they compel a carrier to pay a bonus only if it profits, too. The results strongly suggest true contingency contracts, which fall far short of accounting for the sums insurers are paying out as contingencies, are not the problem some critics suggest they are.
Many companies state their goals in terms of profitability and have said their contracts are designed to reward good loss ratios. The companies, however, are paying for volume, too. This suggests companies do not adequately understand their actions, or they are saying one thing and doing another. Our experience with helping companies develop contingency contracts supports the former view. When we begin working with companies, they have ambiguous goals for their contracts. They often do not understand how their agents view the contracts and how they affect the bottom line. By providing guidance and education, we have helped these insurers define their goals and then align their contingency contracts with their goals.
CLICK ON THE IMAGES ABOVE TO VIEW LARGER
Can contingencies affect placement?
Are bonuses paid under performance-based contingency contracts potent enough to cause agents to steer business to specific companies, regardless of customers' needs? Our analysis of the contracts indicated that for the vast majority of agencies, the motivation to place any given account with a particular company just for the extra profit-sharing bonus related to volume is extremely minimal. Assume an agency has a 50% loss ratio, a 90% retention rate and a 5% growth rate. Under a typical contingency contract, such an agency earns a contingency bonus of .9% of written premium for $500,000 of business. For $5.5 million in premium, the bonus increases to 1.65%. (Chart 3 and 4 above.) In other words, for a 1,000% increase in volume, the agency potentially earns .75 percentage points more in contingency bonuses, all else being equal. This amounts to an increase of .00000015 percentage points for every $1 increase in premium (.75 / 5,000,000). For an agency to consider placing a $1,000 (or even a $10,000) premium account with a certain carrier just to potentially increase its contingency bonus by .000015 (or .00015) percentage points is ridiculous-especially when you consider that the agency risks losing the account if it is moved to an insurer offering an inferior value and that the bonus itself is not even guaranteed.
While the increase for any given company's bonus for extra volume under a performance-based contingency contract is minimal, the differences among what various companies will pay for any given scenario is significant. Some companies will pay double or triple what other companies pay. (Chart 3 above) However, again, on any given account, even triple is a minimal gain. For example, one company in this study averages a .34% of premium bonus at $1 million dollars of premium, while a direct competitor in the study averaged .82% of premium. So on a $10,000 account, the difference is a $34 bonus versus an $82 bonus. But I have not met an agent who would risk an E&O claim or risk losing a $10,000 account to a competitor by placing it with an inferior company or product for an extra $48.
It is also worth noting again that this $48 is a potential gain. It is not guaranteed because the agency still must achieve certain other results (such as an adequate loss ratio) to make that $48-and agencies often do not attain these additional goals. In fact, one of the most significant concerns agencies have about profit-sharing bonuses is that their companies will manipulate the results so the agencies won't earn contingencies. As noted earlier, in calculating contingency bonuses, some companies include allocated loss adjustment expense, unallocated loss adjustment expense and incurred but not reported claims, and almost all companies include standard reserves on known claims. Many agents believe companies often manipulate these reserve charges to increase applicable loss ratios and thereby decrease contingency bonus payments. I must emphasize that many agency owners strongly believe companies do this. So again, why would an agent risk losing an account for $48 that is not even guaranteed, especially when the agent believes the company will manipulate the agency's results to reduce its bonus?
Forty-eight dollars might not be enough to sway one account, but is it enough to sway the agency to place a large number of accounts-say, 100-with a company offering an inferior product? First, keep in mind the resulting $4,800 bonus still would not be guaranteed. Additionally, we have not yet considered the cost of trying to gain that extra $48 per account. According to Business Management Group's 2003-2004 Non-producer Compensation & Benefits Survey, the average commercial lines CSR makes at least $18.50 per hour. Moving a client will take at a minimum an extra hour of work, reducing the potential bonus by 38.5%. When the other headaches and costs are included, I estimate the potential extra bonus will be reduced by at least 50%.
Several years ago, I completed a peer-reviewed study of the cost of moving a book of business in a year's time. I presented the results to several audiences of agency owners in the early 1990s in an effort to help them understand the importance of representing only stable insurers, so they could avoid the cost of moving books of business from distressed carriers to those in good financial health. The study showed the cost of moving a book was 48% of the book's annual commissions.
This is a huge cost. Suppose the aforementioned agency moved a $1 million book (100 of those $10,000 accounts) from one carrier to another. Assume the agency's commission on this book is 12%, or $120,000. Forty-eight percent of $120,000 is $57,600. If all went well and the agency earned all of the potentially higher contingency bonus, using the example above, the extra revenue would come to only $4,800.
Why would any sane person spend $57,600 to make $4,800? The upshot is that even when you consider accounts en masse, rather than individually, it seems implausible that an agency would select an insurance company not in their clients' best interests just to potentially earn a larger performance-based contingency bonus.
Producer compensation arrangements also argue against steering accounts to maximize contingencies. Agencies' producers typically are paid only on the commission generated; they rarely share in contingencies. Therefore, producers would be unlikely to comply with management's desires to place accounts with specific companies, if doing so risked the producer's ability to keep the account-and its renewal commission.
Final thoughts
The findings of our analysis suggest the bonus payments one would expect to be paid under the applicable contracts held by more than 90% of independent agencies do not correspond with the bonus payments reported by companies. More investigation and better data collection is needed for both regulatory purposes and so insurers can better manage their expenses and combined ratios. For now, it appears companies are not getting any extra profit for the bonuses they are paying their brokers and agents. This makes no fiscal sense. Despite this incongruity, however, most independent agents have little or no financial incentive to place business with companies not offering the best overall price or product for their clients.
(Disclaimer: Burand & Associates, LLC, advocates that agencies constructively manage and improve their contingency contracts by learning how to negotiate and use them more effectively. We maintain that agents can achieve considerably better results without ever taking actions that are detrimental or disadvantageous to their insureds. We have never, and would not ever, recommend or otherwise advocate that an agent or agency implement a policy of putting an increase in contingency income ahead of insureds' interests.
Further details are necessary for a complete understanding of the subjects covered by this article. None of the materials in this article should be construed as offering legal advice, and the specific advice of legal counsel is recommended before acting on any matter discussed in this article. Regulated individuals/entities should also ensure that they comply with all applicable laws, rules and regulations.)
© Arc, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.