Mike Smith argues that the time is coming soon when the growing number and extent of poor earnings reports will force the management of insurance operations to address the fundamental weaknesses of their operations.
Wall Street investors and insurance company chief executives alike have been wearing a crease across their posteriors, sitting on the edge of their chairs waiting for pricing to improve in the commercial lines property/casualty markets. They may not have to wait too much longer (in the context that pricing first began to erode 12 years ago!), but between now and then, earnings reports are likely to turn truly ugly. But this will be good, because the poor earnings results will eventually force managements to address the fundamental weaknesses of their operations, beginning with pricing.
As far as investors are concerned, insurance is a business where bad news is good, and terrible news is even better, unless you are the news. What other industry aside from housing will see the stocks move higher after a major hurricane has racked up $16 billion of insured losses?
The quandary faced by investors is that if history repeats, the insurance stocks will likely begin to advance well before the up-turn in pricing becomes visible, to the extent that it becomes imperative to be fully invested in the group well before the turn. But, which companies will be able to avoid being “the news” that helps precipitate the upturn? One view is to focus on the quality players in the industry, those that have consistently out-performed within the industry and where investors can have the most confidence in balance sheet strength.
Investors and underwriters have always looked to natural events to trigger stronger pricing and bring on better earnings. But in reality, no broad upturn in domestic pricing has been triggered by a natural catastrophe since the early 1960s. The catastrophe that has turned cycles has simply been poor underwriting results, the cumulative trauma of poor risk selection and deficient pricing, finally hitting the earnings statement. Business failures among underwriters have been a better signpost for an upturn, especially since 1976.
Many observers believe the industry is so over-capitalised, as evidenced by the premium-to-surplus ratio, that pricing cannot soon improve.1 Two points can be made on this subject:
• If prices were higher, the premium-to-surplus ratio would be higher;
• A significant amount of surplus is held by one company, Berkshire Hathaway, to the extent that, when adjusted for this one position, the ratio is materially higher.
However, it is fair to note that property/casualty industry surplus is currently about double the level where it once stood, in comparison to the economy as measured by GDP.
By the same token, the common wisdom today is that the industry is so big, so global and that capital is so abundant, that the industry will never again see a classical upturn in pricing. But such an environment, one of relative stability, demands a state of homeostasis that the industry has never seen! The property/casualty insurance industry has always been fragmented, and the reality is that its capital account may not be significantly out of line in comparison to other measures of risk exposure. But the fact is, this industry does not have the ability to maintain the degree of stability required to support a no-turn thesis. It always runs to extremes, and it is in the process of achieving one of those extremes now.
Measured against GDP
For the past 12 years, property/casualty industry premiums written have grown at about half the rate of the economy, as measured by nominal GDP (neither variable, GDP or premiums written, has been adjusted for inflation). This relatively slow growth of premiums is indicative of price discounting. In the simple model, insurance is a safety net for the economy. If something has value, it will be insured and, in providing the economic protection, the underwriter will seek to charge a premium that reflects the economic value of the insured property or service.
Adding some validity to our simple model, industry-wide written premiums have grown at a compound growth rate approximately 1% faster than the economy since the end of World War II. Perhaps speaking volumes about the industry, claims have grown 2% faster than the economy! So we see that insurance claims have been a true growth business, while the underwriters have struggled to keep up.
Premium growth from year to year is not so smooth or perfect in its relationship to the economy as suggested by the compound growth charts. In fact, the cyclicality of the industry is plainly visible in a chart that compares the annual percentage change of premium growth with that of the economy. During up-cycles in the mid-1970s and from 1984-87, written premiums grew sharply faster than GDP, while down-cycles are characterised by relatively slow growth. Occasionally, there may be a false signal of an upturn, when the economy slows to meet the growth rate of the insurance industry. What actually happens in these instances is that underwriters, using hindsight, base their estimated premium needs on the previous year's business activity of the insured which, if this activity slows markedly, means that the underwriter collects more premium than is economically necessary.
Three year cycles
One of the great challenges is to call the turn, to accurately predict when the cycle will turn. Over the past 20 years, attempts at this have occasionally produced disastrous financial results. In late 1981, any number of underwriters dived into the market, intentionally under-pricing significant amounts of business. The rationale was that mid-year 1982 marked the end of the third year of downturn, and everyone knew underwriting cycles usually lasted three years on the downside. Thus, these underwriters wanted to be well positioned for the turn at mid-year. The cycle did turn at mid-year - 1984. But by that time, it was too late for many of the early entrants, who subsequently found themselves being acquired by the insurance commissioner in a different sort of consolidation than most investors normally think about.
Very clearly, the industry no longer runs in the three year up, three year down pattern of historical cycles. Currently, the industry is enduring the second consecutive longest downturn in history, leading many people, investors and underwriters alike, to surrender, having determined that there will never again be an upturn.
A criticism of the industry has always been that insurers “cheat” on their loss reserves during the downturns, and stuff reserves during the upturn. Perhaps, but that does not explain why companies finally run out of room on their balance sheets. Possibly, what is actually happening is that a quantifiable pricing cushion is built during the upturns, a cushion that is subsequently utilised to support earnings during the downturn. The place where this cushion is hidden is the loss reserve liability on the balance sheet.
If we accept the simple model that compares written premium growth with nominal GDP, then by definition the difference between the two growth rates will approximate the degree of excess pricing during the upturn, or the degree of price discounting during the downturn. By measuring the cumulative difference, we find some interesting comparisons.
Using 1987 as the base year, the last year when written premiums grew faster than the economy, we conclude that pricing has fallen 25% in the subsequent 12 years (including our estimate for 1999 of 1.5% growth).
But how meaningful is this? After all, it is generally acknowledged that pricing over-shot the industry's needs with two consecutive years of 20%-plus growth, in 1985 and 1986. Therefore, we can go back to 1978, the end of the previous up-cycle, as the base, to see how far prices fell in the 1979-84 period, and by how much the rebound exceeded the requirement.
Based on the data portrayed in chart 6, pricing had deteriorated by 15% from the base level as of 1983. But, the rebound that began in late 1984 more than captured the difference. Despite a dozen years of price discounting since, the surplus in pricing has only recently been fully utilised. This would suggest that downward pricing pressure will remain for a bit longer, despite the highly visible efforts of a few companies such as American International Group, Chubb, St. Paul Companies and the Hartford to draw the line on pricing. Late in 1998, those companies announced that they had had enough and that they were willing to walk away from hundreds of millions of dollars of premium each, rather than renew business at deficient prices.
The difficulty with the use of 1978 as the base year is that there may still have been some “noise” in the numbers. In 1976 the industry experienced 20% premium growth, until then the strongest up-cycle seen in the post-World War II period. This was a rebound driven by fear, as one of the industry's major personal lines carriers, GEICO, verged on the edge of insolvency. In addition, many of the major underwriters today are mostly focused on the commercial lines market, and achieve little benefit from the greater long term stability of rates in the personal lines market. While our charts have measured premium growth for all lines, if we measure only the commercial lines, the picture becomes far more grim.
In order to deal with the noise, we can go back even further, and measure the cumulative difference between premium growth and the economy, focusing only on commercial lines and beginning with a point in the late 1950s. Based on this measure, the industry has only recently fully utilised the pricing cushion that it has built up over the past two cycles. However, the cumulative price deficiency appears to be approaching a level where the past two downturns have ended.Perhaps it is only coincidence, but concurrent with the current statistical disappearance of the price cushion as chart 8 depicts, investors are now beginning to see an increase in quarterly earnings disappointments reported by property/casualty insurers. Until about mid-year 1998, insurance companies were reporting relatively strong earnings despite the poor fundamentals of slow premium growth and slow investment income growth, and the steady deterioration of the underlying quality of earnings (exhibited most clearly by a decline in cash flow from operations).
Second quarter earnings were generally short of expectations in 1998. True, catastrophe losses were significantly greater in 1998 than they had been in 1997, but when “cat” losses were stripped out of both sides of the equation, earnings were found to be well short of targets. As a result, the group's performance in a falling stock market was significantly worse than the 10% decline of the S&P 500 in last year's third quarter. Subsequently, the broader market index recovered, but property/casualty stocks did not, as investors ignored even the salutary effects of a strong bond market that normally would have propelled insurance stocks higher, and instead walked away from the shares.
Other indicators buttress the view that the underlying strength of the industry has eroded. Specifically, investment income is in an industry-wide decline, only the third year-to-year decline since the end of World War II, but all three declines have come in the current decade.
One view is that cash flow is the ultimate driver behind the underwriter's pricing decisions. The industry's product is cash. Insurance is not about delivering promises, it is about generating cash to feed the investment machine that is the source of most of the industry's earnings, and when there is less cash on the table at 5:00 pm than there was at 9:00 am, there is a problem. Historically, cycles have bottomed when operating cash flow relative to total revenue, reaches a certain level, a level that the industry has been hovering at for about the past five years.
Quite possibly, the real trigger for an upturn is a modification of the cash flow idea. Perhaps it is not until the firm's chairman realises that his personal cash flow is about to be cut off that he visits his underwriters to tell them to raise prices. So far, most of the industry's restructurings have represented nothing more than clerical shakeups. But when we see management shakeups, when ceos are fired, then we can look forward to the long-awaited upturn.Michael A. Smith, is managing director, Bear, Stearns & Co. Inc.
1. The author believes that significant changes in the level of net risk retention, especially with respect to catastrophe losses that have been brought about by changes in the reinsurance market, have rendered historical premium-to-surplus comparisons irrelevant. Simply, premiums do not serve as an accurate proxy of risk exposure in comparison to the representation prior to 1992, when hurricane Andrew and a number of other major losses decimated reinsurance capacity. The missing element in 1999, and the invisible element in 1989, is the capacity of the international reinsurance market to absorb losses incurred from relatively small catastrophes that prior to 1992 were heavily reinsured.