In the long-running game show of the same name, contestants vie with one another to figure out the right price to win valuable prizes. Insurance pricing isn’t all that different. Managing the ups and downs of pricing cycles and related profit ratios can pose a big challenge. Much has been written and discussed about the numerator of those ratios — which reflects losses.
However, sometimes forgotten in the discussion is the denominator — the earned premiums — and the price movements that drive those premiums. To maintain long-term profitability and viability, companies that write long-duration business must manage the “information emergence” cycle and establish “early warning” mechanisms triggered by abrupt price or loss changes.
Accurate internal price monitors are imperative to help a company keep on top of the pricing cycle and the effects of price competition — especially during prolonged soft markets.
In monitoring the effects of dynamic competitive forces, it is also critical to attain accurate and credible external market information. Not surprisingly, in a highly competitive marketplace such as insurance or reinsurance, the pricing actions and results for an individual company are strongly linked to the pricing actions and results of its competitors.
In the past decade, subsequent to the devastating soft market cycle of 1997 to 2001, companies have invested many resources to create robust internal price monitoring systems. Companies invest heavily in building and expanding their data information structures. They then tap into those huge data warehouses to extract whatever knowledge they can about specific markets to turn their data into actionable insights.
As a way of supplementing their internal price monitoring information, either for comparison or credibility enhancement, companies require accurate external pricing indications that look both back in time and toward the future. Underwriters and actuaries have access to numerous rate-level survey sources to help measure historical external pricing changes. And to estimate future expected changes, there are many learned opinions from various industry segments and leaders.
For both past and future time frames, survey procedures, motivations and results are as varied as the sources. Given the extreme diversification of results and the difficulty in getting accurate price movements, using the information poses many challenges.
By comparing the survey results to an insurer’s actual data, the insurer can gain deeper insight into those differences. This comparison involves testing the “skill” of not only the historical survey results but of future expectations as well. Both involve comparing survey results to actual data, either internal or from various external sources. Ultimately, for a better understanding of its position in the pricing cycle, an insurer needs a direct comparison of its results with the industry aggregate.
Many independent organizations create price monitors based on surveys. Insurance companies frequently use these survey results to help benchmark against their own results. These external price monitors are constructed in a variety of ways. All have their strengths and weaknesses.
There are at least four common, often-quoted, survey sources that have produced results at various levels of detail for many years. The first of these has generated results since the late 1970s, while the most recent source started in 2005. Most use some version of survey data and send out many requests for survey completion, with varying levels of responses. Some sources try to supplement the survey information with various external data.
Some have a large-account bent, such as those taken from risk managers whose clients tend to be large corporations. However, others will focus on agents and brokers or clients that make up a particular segment of the marketplace. Even when administrators take the utmost care, the nature of surveys is such that biases and sampling errors can be significant and prone to human error.
By their very nature, future expectations — even from learned experts — can still be more problematic. When an exhaustive search for exact pricing changes is impractical, and especially when trying to guesstimate future ones, heuristic methods (experience-based techniques) speed up the process of finding a satisfactory solution.
The behavioral sciences are rich in research of heuristics and biases approach. But experts can be excessively overconfident in the ability to replicate test results from small samples. This overconfidence rings true not just for management and actuaries, but in almost any professional field, such as meteorology, catastrophe modeling or golfing.
An accurate “skill” assessment against real data for both historical survey indications and forward-looking expectations is necessary in all fields. Only then can insurers begin to discern how much weight to give both the various backward- and forward-looking indications at the disposal of company management.
Each of these assessments is critical to determine a carrier’s position in the pricing cycle. And, like the game show contestants vying for cash and prizes, companies that can accurately assess their own competitive pricing strategy will win their own reward — long-term marketplace viability.
Accurate Price Movements
As an example of the importance of accurate price movements, assume an insurer is trying to determine the overall profitability of a line of business. The chief actuary determines through public information that 2004’s loss ratio is 60 percent (that is to say, an insurance company pays out $60 in claims for every $100 in collected premiums). The actuary will use that ratio along with results for all other years until 2012, with varying degrees of credibility, to estimate the expected loss ratio for 2013.
The impact of pricing changes is one adjustment among all types of adjustments for loss levels and exposure changes. If the only (therefore, best) source of rate changes is the values in Figure 1, then the results would be very different depending upon which survey was selected. Under survey 4, the price-level adjustment factor over this eight-year period would yield a 2012 indicated loss ratio of about 70 percent. Using survey 1, a comparable estimation would yield a loss ratio of more than 110 percent. Surveys 2 and 3 yield indications between those extremes. Clearly, this example offers a broad range of conclusions, leading to much different decisions by top management on the viability of a particular market segment.
Premium Size Makes a Difference
Premium size is one of the most important areas to differentiate pricing among groups. Actual pricing statistics verify that larger accounts experience much lower lows and higher highs, depending on a carrier’s position in the underwriting cycle. For example, Figure 2 shows how significant the impact can be. In this example, accounts less than $10,000 in average premium behave very differently than those larger than $100,000.
Therefore, survey sources that include a substantial portion of small premium accounts may significantly understate the pricing swings for insurers interested primarily in monitoring larger accounts.