Annual Insurance Loss Report

Transcription

1 +44 (0) Annual Insurance Loss Report Operational risk loss data for insurers submitted between 2012 and 2017 June 2018 Managing risk together

2 2 Annual Insurance Loss Report Contents Foreword 3 Introduction 4 Trends 5 Event frequency by business line and event type 6 Event severity by business line and event type 7 Annual summary of loss data reported to ORX between 2012 to Quarterly summary of loss data reported to ORX between 2012 to Loss data by size reported to ORX between 2012 to Geographic distribution of loss reported to ORX between 2012 to Loss by business type as a proportion of gross written premium exposure 14 Life insurance 14 Non-life insurance 15 Asset management 15 Disclaimer: ORX has prepared this document with care and attention. ORX does not accept responsibility for any errors or omissions. ORX does not warrant the accuracy of the advice, statement or recommendations in this document. ORX shall not be liable for any loss, expense, damage or claim arising from this document. The content of this document does not itself constitute a contractual agreement, and ORX accepts no obligation associated with this document except as expressly agreed in writing. Operational Riskdata exchange Association (ORX) 2018

3 Annual Insurance Loss Report 3 Foreword The ORX Insurance Service was established in 2015 with six original members. By June 2018, we had grown to 17 members, including a high proportion of the world s largest global insurance and re-insurance firms. This growth has enabled ORX to present the second annual report on the insurance operational risk loss database. It provides an overview of the data reported since inception up to the end of 2017 and illustrates the industry operational risk trends that can be observed. This report will continue to be produced annually and will be enhanced as the database size increases. ORX has a strategic commitment to continue investing in and growing the service to support enhancement in both operational risk management and measurement in the insurance sector. In addition to the anonymous sharing of loss data, becoming a member of the ORX Insurance Service enables you to gain access to a leading edge risk management resource and to an extensive peer network of insurance operational risk professionals. Benefits include: The opportunity to participate in leading edge operational risk research and the development of risk management practices and tools to help enhance risk management at your firm. Access to a range of online resources including benchmarking against your peers Access to a global network of operational risk professionals Free access to ORX forums, including dedicated insurance forums and industry leading Head of Operational Risk forums ORX has leveraged its extensive experience to provide the most comprehensive and professional operational risk management service for the insurance industry. In addition to the sharing of loss data, insurers benefit from leading edge, insurance focused, risk management research and access to a global network of operational risk professionals. ¹ Claudia Meyer, Group Head of Operational and Reputational Risk Management, Allianz SE Subscribing to the Global Insurance Service One-off joining fee of 47,000 The joining fee covers becoming an equal owner of the assets of the Association. It also allows a new member to have access to all our intellectual property and an extensive library of research, data and information that ORX has built up since it was established in Annual membership fee of 34,000 This fee is paid every year and covers the costs of running ORX and delivering our core activities. If you would like further details on this report or you are interested in the Insurance Service, please visit our website at A dedicated insurance community resource on the members website Steve Bishop Head of Insurance & Relationship Management ORX ¹ This quote is a personal view and does not represent the opinion of Allianz.

4 4 Annual Insurance Loss Report Introduction This is the second annual report on the ORX insurance operational risk loss database. It summarises our database and illustrates trends in operational risk losses within the insurance and reinsurance industries. This edition is based on submitted loss and income data from 16 insurance firms, up to the end of ORX members reported a total of 5,438 loss events totalling approximately 3,703m in gross loss for the six years from 2012 to Each loss event submitted has a gross value greater or equal to our minimum reporting threshold of 20,000. Members also submit their quarterly gross income, gross written premium exposures and assets under management. Banking This year, we published our eighth annual global banking report. The report provides a summary of recent operational risk losses within our global banking data. It also highlights trends in frequency and severity by year, business line, event type and region. You can view this year s banking report here: This edition is based on submitted loss events and exposure indicators for life, non-life, asset management and corporate item business lines that all have recognition dates between 1 January 2012 and 31 December This report provides a summary of the data for 2012 to It also highlights trends in frequency and severity by year, business line, event type and region. ORX reporting standards All ORX members are required to submit data in line with a common standard and in a common format. The standards are developed and maintained by the ORX Definitions Working Group. They are published, in detail, as the ORX Insurance Operational Risk Reporting Standards (I-ORRS). The reporting standards will continue to be updated as industry and supervisory practice evolves. Visit here to request a copy: standards

5 Annual Insurance Loss Report 5 Trends Insurance members reported a total of 5,438 loss events with a gross loss of approximately 3.7bn between the six years from 2012 to Following a post-crisis period where operational losses were volatile within the insurance industry, the past two years have brought relative stability and typically lower losses. As the ORX Insurance service continues to grow, we have seen a steady increase in the annual number of loss events submitted to ORX from 724 in 2012 to 1,038 in 2017 (page 8). By Spring 2018, the database contained over 5,400 loss events. Average loss size is decreasing The typical severity of losses, however, appears to be decreasing. The average size of a loss event in 2017 was around 430K less than half the average mean loss in 2012 of 1.1 million. As a proportion of gross income, this equates to 0.96 lost per 1k in 2017, which is less than half of 2012 s figure of 2.06 lost per 1k (page 9). The large reduction in average operational loss size appears to be due to a fall in the size of the largest industry losses (page 12). Indeed, in 2012 the ten largest loss events accounted for 77 per cent of total loss, while in 2017, this figure is 42 per cent. or process management errors. Losses of this nature account for 49 per cent of total operational loss between 2012 and 2017 (page 7). Majority of losses in North America and Western Europe When examining regional operational loss, it is evident that North America and Western Europe dominate. Loss within Western Europe accounts for 72 per cent of total loss and 14 per cent can be attributed to North America (page 13). This is largely a reflection of the ORX membership. Loss rates decreasing Considering loss as a proportion of gross written premiums in the life and non-life business lines highlights that both areas have seen a decrease in loss in recent years. The life insurance business line experienced losses of 1.67 per 1k of gross written premium exposure in 2012, which fell to 0.63 per 1k in The non-life business line has witnessed a more significant change where loss was as high as 2.53 per 1k of gross written premium exposure in 2012, though in 2017 this fell to just 0.88 per 1k (page 15). EDPM dominates losses ORX members report the event type of each operational loss exceeding or equalling our reporting threshold as well as the line of business in which the event occurred. A breakdown of loss by type of event shows that the most significant losses are classified as Execution, Delivery and Process Management (EDPM). Losses classified as EDPM include those due to failed transaction processing This was the global industry loss between 2012 and 2017, from a reported total of 5,438 loss events.

6 6 Annual Insurance Loss Report Event frequency by business line and event type ORX members report their operational losses by business line (life, non-life, asset management and corporate items) and by the type of event, which resulted in the loss. Losses are reported as corporate items if the loss is experienced by the entire firm rather than a single revenue-generating line of business. The event type categorisation used by ORX insurance members is similar to the standard risk types used in our banking services. Figure 1 illustrates the events submitted to ORX for the reporting period 2012 to 2017 for each business line and event type. The size of the square is directly proportional to the number of events reported within each business line and event type combination. The most common event type Execution, Delivery and Process Management (EDPM) accounted for approximately 48 per cent of total loss events. In contrast, the least common event type was Damage to Physical Assets, which only accounted for less than 1 per cent of total events. Event frequencies for life and non-life businesses were alike 41 per cent of events originated from life insurance and 43 per cent of events were due to the non-life business. Additionally, the distribution of events between the seven event types for these two key business lines were very similar. Figure 1. The number of events submitted in each business line and event type between 2012 and 2017 Internal Fraud External Fraud Employment Practices & Workplace Safety Clients, Products & Business Practices Damage to Physical Assets System Failure Execution, Delivery & Process Management Asset Management Life ,245 Non-Life Corporate Items

7 Annual Insurance Loss Report 7 Event severity by business line and event type Figure 2 depicts the total loss due to events submitted to ORX for the reporting period 2012 to 2017 for each business line and event type. The size of the square is directly proportional to the total loss reported within each business line and event type combination. The event type Execution, Delivery and Process Management (EDPM) was also the costliest, it accounted for 49 per cent of total loss between Losses of this nature pertain to failed transaction processing or process management. Loss severity for the life and non-life business was also comparable with 42 per cent of loss originating from the life business and 39 per cent due to the non-life business. Moreover, the distribution of loss between the seven event types for these two key business lines was very similar. Figure 2. The total loss submitted in each business line and event type between 2012 and 2017 Internal Fraud External Fraud Employment Practices & Workplace Safety Clients, Products & Business Practices Damage to Physical Assets System Failure Execution, Delivery & Process Management Asset Management 4M 0.6M 4M 48M 0.0M 3M 73M Life 85M 65M 28M 659M 10M 51M 654M Non-Life 63M 88M 25M 554M 5M 43M 673M Corporate Items 37M 1.4M 23M 67M 6M 37M 399M

8 8 Annual Insurance Loss Report Annual summary of loss data submitted to ORX The following results show annual trends in the frequency and severity of operational risk losses, which our members report. These statistics are based on the reporting date of loss events exceeding our threshold. The average loss size in 2012 was more than twice the average loss in 2017 (Figure 3). In 2012, the top ten largest losses accounted for 77 per cent of total loss. The corresponding percentage for 2017 was nearly half at just 42 per cent. Hence, the notable decrease in average loss size in 2017 was most likely due to the significant reduction in large losses within This is also evident when examining the changes in total loss within the largest loss size bins (page 12). When observing total loss as a proportion of total income, we also see significant reductions in Our member firms lost 0.96 per 1k of gross income in 2017, which was less than half of the loss in 2012 of 2.06 per 1k (Figure 5). In contrast, the frequency of reported loss events is steadily increasing as the ORX insurance service continues to grow. In 2012, 724 loss events were reported by our members, whereas, this figure rose by nearly 45 per cent to 1,038 events in 2017 (Figure 4). Figure 3. Average gross loss size of events reported per year in Year Average loss size 1, ,036, , , , ,374 Mean gross loss reported ( k) ,888 0 Figure 4. Number of loss events reported per year in Year Total number of reported events 1, Total number of reported loss events

9 Annual Insurance Loss Report 9 Figure 5. Total loss per 1k of gross income per year in Year Total gross loss per 1k of gross income Total gross loss reported per 1k of gross income

10 10 Annual Insurance Loss Report Quarterly summary of loss data submitted to ORX in Data is submitted to ORX on a quarterly basis and analysis of the quarterly frequency and severity of our member firms reveals the volatility of operational losses. In contrast to the variation shown between 2012 and 2016, loss for 2017 was relatively stable. On a quarterly basis, average loss size varies between around 200,000 in quarter three of 2016 to approximately 1.7 million in quarter one of 2012 (Figure 6). As with the annual results, this change was driven by the occurrence of large loss events predominantly in 2012 and Similar levels of volatility can be seen when we consider total gross loss as a proportion of income. This varies between 0.46 in quarter three of 2016 to over 4 per 1k of gross income in quarter four of 2013 (Figure 8). Although reporting levels remained relatively stable ranging between 150 and 350 events reported each quarter, we have seen a steady increase in event frequency as the ORX Insurance service continues to expand (Figure 7). Figure 6. Average gross loss size of events reported by quarter in 2012 to Mean gross loss reported ( m) Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Average gross loss size Q1 1,736,751 1,080, , , , ,408 Q2 229, ,261 1,048, , , ,122 Q3 484, , , , , ,735 Q4 1,492,557 1,493, , , , ,454

11 Annual Insurance Loss Report 11 Figure 7. Number of loss events reported per quarter in Total number of reported loss events Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Total number of reported events Q Q Q Q Figure 8. Total loss per 1k of gross income per quarter in Total gross loss reported per 1k of gross income Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Total gross loss per 1K gross income Q Q Q Q

12 12 Annual Insurance Loss Report Loss data by size reported to ORX in 2012 to 2017 Observing the frequency and severity of operational losses in size bins highlights the skewed distributions, which are inherent in operational risk loss data. Analysis of the frequencies shows that most operational losses are relatively small loss events with just over 1 per cent of loss events exceeding 10 million between the years of and in 2017 (Figure 9). In contrast, the severity of the large loss events is such that as a proportion of total loss, events over 10 million accounted for over 55 per cent of losses between , though this fell to just above 40 per cent in 2017 (Figure 10). This decrease in the proportion of large loss events drives the significant reduction in total loss within 2017 in comparison to previous years. Figure 9. Percentage of loss events falling in each loss size bin Loss size K 100K 40.1% 40.7% K 250K 31.2% 32.7% 250K 500K 12.4% 12.3% 500K 1m 7.4% 7.3% 1m 5m 6.6% 5.2% 5m 10m 1.1% 0.6% 10m+ 1.3% 1.1% Proportion of loss events falling within loss size boundaries 40% 30% 20% 10% 0% 50K - 100K 100K - 250k 250k - 500K - 1M - 500k 1M 5M 5M - 10M 10M+ Figure 10. Percentage of total loss falling in each loss size bin Loss size K 100K 3.2% 5.1% K 250K 5.6% 9.0% 250K 500K 4.9% 7.6% 500K 1m 6% 9.4% 1m 5m 15.2% 20% 5m 10m 9.2% 8.1% 10m+ 55.8% 40.8% Proportion of gross loss falling within loss size boundaries 60% 50% 40% 30% 20% 10% 0% 50K - 100K 100K - 250k 250k - 500K - 1M - 500k 1M 5M 5M - 10M 10M+

13 Annual Insurance Loss Report 13 Geographic distribution of loss reported to ORX in 2012 to 2017 ORX members report the region in which each loss event occurs. These regions are Africa, Asia-Pacific, Eastern Europe, Latin America and the Caribbean, North America, and Western Europe. The geographic distribution of loss is largely indicative of the ORX membership. Figure 11 shows the global distribution of frequency and severity of losses reported between 2012 and The regions with the largest operational losses were Western Europe with approximately 72 per cent and North America with just over 14 per cent of total loss. In contrast, all other regions experienced relatively low levels of loss. We see a similar distribution of loss frequency with 15 per cent of loss events having occurred in North America and 69 per cent in Western Europe. In comparison, all other regions have similar reporting frequencies. Figure 11. Global distribution of frequency and severity of losses reported between 2012 and 2017 North America Frequency: 833 Percentage of total frequency: 15.3% Total loss: 530,255,006 Percentage of total Loss: 14.3% Western Europe Frequency: 3,766 Percentage of total frequency: 69.3% Total loss: 2,678,874,025 Percentage of total Loss: 72.4% Eastern Europe Frequency: 219 Percentage of total frequency: 4% Total loss: 96,491,571 Percentage of total Loss: 2.6% Latin America and the Caribbean Frequency: 123 Percentage of total frequency: 2.3% Total loss: 118,609,290 Percentage of total Loss: 3.2% Africa Frequency: 234 Percentage of total frequency: 4.3% Total loss: 25,878,463 Percentage of total Loss: 0.7% Asia-Pacific Frequency: 263 Percentage of total frequency: 4.8% Total loss: 252,453,909 Percentage of total Loss: 6.8%

14 14 Annual Insurance Loss Report Loss by business type as a proportion of gross written premium exposure As well as reporting income at firm level, ORX members report quarterly the value of gross written premiums by business line, as well as the value of assets under management. Figures 12 and 13 illustrate losses as a proportion of gross written premiums in the life and non-life insurance business lines, respectively. Meanwhile, Figure 14 depicts loss as a proportion of assets under management. Both the life and non-life business lines have seen considerable reductions in operational loss as a proportion of gross written premiums in 2017 compared to previous years. Whereas, the asset management business line witnessed operational losses that vary significantly across the years of 2012 to The non-life insurance sector sees a significant spike in loss as a proportion of gross written premiums in 2012, though this seems to stabilize in the following years. In 2012, approximately 2.50 was lost per 1k of gross gross written premium exposure, this was significantly higher than any other year across both the life and nonlife business lines. However, by 2017 this had fallen to 0.88 lost per 1k of gross premiums (Figure 13). In contrast, insurers that manage assets have witnessed fairly volatile operational losses. In 2012, around 1.80 was lost per 1m of assets under management, this has increased by almost 20 per cent to a figure of 2.23 lost per 1m in Within the life insurance sector, there appears to be a steady decrease in losses ranging from 1.67 lost per 1k of gross written premiums in 2012 to just 0.63 lost in 2017 (Figure 12). Life insurance Figure 12. Total loss per 1k of gross written premiums in the life insurance business line Year Total gross loss per 1k of gross written premiums Total gross loss reported per 1k of gross written premium

15 Annual Insurance Loss Report 15 Non-life insurance Figure 13. Total loss per 1k of gross written premiums in the non-life insurance business line Year Total gross loss per 1k of gross written premiums Total gross loss reported per 1k of gross written premium Asset management Figure 14. Total loss per 1m of assets under management in the asset management business line Year Total gross loss per 1m of managed assets Total gross loss reported per 1m of assets under management

16 +44 (0) Managing risk together ORX believes many heads are better than one. We re here to bring the best minds of the international operational risk community together. By pooling our resources, sharing ideas, information and experiences, we can learn how best to manage, understand and measure operational risk and become less vulnerable to losses. We work closely with over 90 member firms to develop a deeper understanding of the discipline and practical tools. We set the agenda, maintain industry standards, and garner fresh insights. ORX is owned and controlled on an equal basis by its members. For more information about ORX, visit our website at Contact Steve Bishop Head of Insurance & Relationship Management steve.bishop@orx.org Dr Luke Carrivick Head of Research and Information luke.carrivick@orx.org