Vigorous debate about reinsurance market cycles and the viability of traditional business models has increased since HIM made landfall last year. After all, large-loss years in the past brought about significant and sustained reinsurance rate increases, particularly in the property-catastrophe market. This did not happen through 2018’s major renewal season, however, confounding predictions of double-digit rate increases in response to record insured catastrophe losses of more than USD 140 billion.
Figure 1 shows that propertycatastrophe pricing, in historical terms, barely moved at 1 January and 1 June in 2018. There was no repeat of the spikes that followed Hurricane Andrew in 1992, the 9/11 terrorist attacks in 2001 or hurricanes Katrina, Rita and Wilma (KRW) in 2005. In fact, Figure 1 underlines how post-loss
pricing volatility has reduced over the last decade, particularly in Florida, given the more measured pricing reaction in 2011/12 and now 2018, even after successive years of pricing declines between both years.
JLT Re’s Global Risk-Adjusted Property-Catastrophe Reinsurance Rate-on-Line (ROL) Index rose by 4.8% on 1 January 2018, but that did not even return the market to where it was in 2016.
Figure 1: JLT Re’s Risk-Adjusted Global and Florida Property-Catastrophe ROL Indices – 1992-2018 (Source JLT Re)
Perhaps more tellingly, only negligible rate increases were recorded during mid-year renewals as rate momentum slowed through the course of the year.
For Florida property-catastrophe renewals at 1 June, prices increased by an average of only 1.2%, failing to match the rate increases recorded for US property-catastrophe business earlier in the year despite a greater number of loss-affected programmes renewing at mid-year after Hurricane Irma’s landfall in Florida last year. As a result, the cost of property-catastrophe protection both globally and in Florida remains competitive, with pricing still down on 2012 levels by 30% and 40%, respectively.
This is, of course, not to say that more pronounced rate increases have not occurred in 2018. Classes significantly impacted by last year’s losses, such as retrocession and direct and facultative (D&F), have typically seen higher rate increases for both loss-free and loss-affected programmes, although these again moderated through the course of the year. Whereas loss-free retrocession programmes saw price rises of up to 10% at 1 January, flat to up 5% were more typical by mid-year. A similar decelerating trend was recorded for loss-free D&F programmes, although rates here remained positive at midyear at up 5% on average. Pricing in non-property lines has been equally as mixed, with continued moderate reductions for better performing classes and pockets of firming for lines with more challenged loss ratios.
Some now believe that 2017/18 represents a watershed moment for the reinsurance sector, marking the end of post-loss pricing paybacks for reinsurers and the point where rate movements became more muted, for property-catastrophe in particular. Using previous large-loss years as a point of reference, evidence seems to back up this theory, although there are a number of interacting factors at play.
1. QUANTUM OF LOSSES
Figure 2 provides an overview of how major insured catastrophe losses since 2000 impacted property-catastrophe reinsurance pricing during the subsequent 12-month period. It shows that the magnitude of catastrophe losses alone does not necessarily correlate strongly to subsequent rate changes. The most meaningful rate increases recorded since the turn of the century (of more than 25%) followed the 9/11 attacks in 2001, when industry insured losses for the entire year reached just USD 50 billion in constant dollars, the lowest of all six years included in the analysis. 2005, with its USD 137 billion in inflation-adjusted insured losses, was the only other year that resulted in double-digit property-catastrophe rate increases at both 1 January and 1 June. Conversely, single-digit rate increases were recorded after insured catastrophe losses reached similar levels in 2011 and 2017. Furthermore, significant losses from Superstorm Sandy in 2012 did little to prevent (often substantial) pricing declines the following year.
Figure 2: Changes to Property-Catastrophe Pricing After Major Catastrophe Losses (Source: JLT Re, Swiss Re)
1 ROL changes shown for 1 June are for the following year, apart for 2011 where the same year hasbeen used to account for major losses occurring before the renewal took place (e.g. Tohoku and
Arguments that the magnitude of insured catastrophe losses in 2017 alone would be sufficient to trigger double-digit reinsurance rate increases now seem to have been misplaced. Comparisons to 2001 and 2005 must also account for key differences in loss characteristics: 2017 saw an accumulation of several mid-sized losses rather than one or two large events on the scale of 9/11 or Katrina, for example, when insured losses totalled approximately USD 45 billion and USD 60 billion, respectively. As a result, the percentage of losses ceded to the reinsurance market for both these events was significantly higher compared to HIM. Indeed, most US nationwide catastrophe limits were totalled after Katrina (as they were for London and retrocession placements) whereas 2017 was much more sporadic and specific. History shows, therefore, that the quantum of insured catastrophe losses alone has not been a strong catalyst for reinsurance market firming. Meaningful pricing corrections have occurred only when losses penetrate deep into reinsurance layers and bring a shock element. Neither happened in 2017, as HIM losses were mostly retained by the primary market and fell within modelled expectations.
2. EXCESS CAPITAL
The amount of excess capital in the reinsurance sector has traditionally had more of a bearing on pricing than any other factor. Figure 3 shows the strong historical (inverse) correlation between the sector’s excess capital ratio2 and property-catastrophe pricing. Back in the early-to-mid- 2000s, the sector’s capital position was in a very different place to where it is today. Much of this was down to large property losses early in the decade, unrealised investment losses caused by the dotcom equity market crash (in an era when reinsurers were more equities geared) and, most importantly, the liability crisis of 2001-2005, which ultimately cost carriers hundreds of billions of dollars worldwide. This meant that the sector’s capital position was under considerable strain during this time, resulting in persistent, negative excess capital ratios. Widespread balance sheet impairments, compounded by additional devastating catastrophes in 2004 and 2005, led to significant (mostly double-digit) rate rises as demand for reinsurance exceeded supply.
The situation has reversed since the financial crisis. After years of strong capital growth and lacklustre premiums, the reinsurance market currently has more capital relative to risk than at any time in living memory. The marked increase in dedicated reinsurance capital has been driven by several factors, including higher asset values, net reserve releases since 2006/07 and, crucially, the entry of tens of billions of dollars of alternative capital into the sector. Relatively low loss activity has also contributed to the build-up of sector capital: insured catastrophe losses have only exceeded the 10-year average on three separate occasions since 2008 (see Figure 4). And in each of these years (2011, 2012 and 2017), any increase to reinsurance pricing has been short lived and capped to single digits due to the supply glut that dominates the market.
Figure 3: Property-Catastrophe Pricing Versus Reinsurance Excess Capital
2 Calculated by JLT Re using its proprietary dedicated reinsurance capital and premiums data. dominates the market.
3. PERCEPTION OF RISK
Events last year confirmed that mid-sized losses emanating from well-modelled risks are currently unlikely to result in a sustained and broad hard market. Any future market turn is likely to require the withdrawal of capital. This did not happen in the months following HIM, as a sophisticated investor base with a deep understanding of the sector had the confidence to increase allocations almost immediately (more on this shortly). The reaction was supported by the reasonable performance of the latest generation of commercial hurricane models for HIM (notwithstanding the divergence in estimates for Maria), and the fact that no immediate major model revisions have been announced by the leading modelling firms.
But a ‘shock’ event that causes risk perceptions to change could still transform market conditions. This is not without precedent. After 9/11, given the event’s unforeseen nature and multi-line impact, risk assessments were overhauled immediately and coverage was maintained only through the formation of several public-private terrorism backstops and pools. Challenged underwriting assumptions also materialised after Hurricane Katrina, prompting rating agencies to enforce stronger capitalisation requirements for catastrophe risks. Even events in 2011 brought an element of surprise as diversification strategies were called into question by unexpectedly expensive ‘cold spot’ losses.
Figure 4: Global Insured Catastrophe Losses – 2008 to 2017 (Source: JLT Re, Swiss Re)
Although the recalibration of risk played a secondary role in driving up rates in each of these instances, an unforeseen loss that changes underwriting risk assumptions, or causes a spike in global risk premia which depletes asset values, could still turn today’s supply and demand imbalance on its head – a pertinent point given the current diverse and complex risk landscape.
BACK TO THE FUTURE
Put in historical context, the muted pricing reaction to HIM is perhaps not as surprising as it may first appear. Whilst many market participants expected prices to rise significantly after the events of 2017, especially following five years of decreases, our analysis shows that, rather than the cost of claims, it is the nature of the losses, and the capital buffer available to absorb them, that typically determines the market response. Indeed, the last prolonged hard market was brought about only by a confluence of losses and events that included an unforeseen terrorist attack, an equity market crash, a reserving crisis and a succession of US landfalling hurricanes.
That said, events over the last year or so have represented a break with the past to the extent that rate firming post-HIM has been less pronounced and sustained than in previous large-loss years. Clearly, this has huge implications for reinsurance pricing, and underwriting more generally, and it merits further analysis.