Catastrophe Risk Insurance and its Pricing Issues for Emerging Markets P R O F. D R. A. S E V T A P K E S T E L M I D D L E E A S T T E C H N I C A L U N I V E R S I T Y T H E I N S T I T U T E O F A P P L I E D M A T H E M A T I C S D E P T. O F A C T U A R I A L S C I E N C E S
Basics Hurricane Andrew (1992) estimated losses $13 billion actual losses $15.5 billion, nine insurance company bankrupt Natural disasters involve potentially high losses that are extremely uncertain and high endangers the financial stability of the company. 25.02.2019, SKOPJE, MACEDONIA 2
Number of catastrophic events between 1970-2017 Insured losses (million USD) in 2017: Maritime disasters 20.3 Aviation: 40.7 Other: 50.5 Total: 101 Source: Sigma (2018) Source: Sigma (2018) 25.02.2019, SKOPJE, MACEDONIA 3
Economic Losses as % of GDP North America: 1.17 Latin America and Caribbean: 0.59 Europe: 0.12 Africa: 0.14 Asia: 0.11 Oceania/Australia: 0.22 Source: Sigma (2018) 25.02.2019, SKOPJE, MACEDONIA 4
Total losses by region in 2017 Source: Sigma (2018) 25.02.2019, SKOPJE, MACEDONIA 5
Components Hazard: Characteristics of the disaster Earthquakes: magnitude, the location of its epicenter and other relevant parameters. Hurricane: projected path and wind speed. Portfolio of properties at risk: The location of each property facing risk, construction type, the number of stories in the structure, the building s age 25.02.2019, SKOPJE, MACEDONIA 6
Components Vulnerability or susceptibility to damage the physical impact of the natural hazard which may differ from model to model. A model might construct damage curves and relate the damage to the structure to a severity parameter, such as spectral acceleration or peak gust wind speed. Other models, (i.e. HAZUS) classifies a structure as being in a Slight, Moderate, Extensive, or Complete damage state. 25.02.2019, SKOPJE, MACEDONIA 7
Components Loss: direct or indirect in nature. Direct losses the cost to repair and/or replace a structure, Indirect losses business interruption and relocation costs of residents The key challenge: how to allocate catastrophe risk, which is less frequent but more extreme in a similar manner as other insurable risks? The limited availability of data makes determining the probabilities of the occurrence of catastrophic events and their likely outcomes difficult 25.02.2019, SKOPJE, MACEDONIA 8
Specification of a premium rate with sufficient demand producing a consistent cash inflow of revenues to cover the expenses on a given catastrophe still yield a net positive profit over a pre-specified time horizon is the main challenge for insurance companies 25.02.2019, SKOPJE, MACEDONIA 9
Estimation of frequency of catastrophic event the most probably losses based on catastrophic modeling The insurer is able to formulate what is known as an Exceedance Probability Curve which depicts the probability that a certain level of loss will be exceeded on an annual basis. 25.02.2019, SKOPJE, MACEDONIA 10
Pure Premium: EQ case (Yucemen (2005) Expected Annual Damage Ratio (EADR): For the kth type of structure EADR SH: seismic hazard with intensity I. The mean damage ratio (MDR) DR: Damage State (none, light, moderate, severe, collapse) CDR: Central damage Ratio P(DS,I): Probability transition matrix 25.02.2019, SKOPJE, MACEDONIA 11
Pure risk premium (PRP) for the kth structure having insured value (INSV) 25.02.2019, SKOPJE, MACEDONIA 12
Risk Management The losses from a natural disaster not independent, as they do not follow the law of large numbers. involve spatially correlated losses, or the simultaneous occurrence of many losses from one single event. Lack of risk diversification for many insurance firms in the realm of catastrophe insurance is precisely the purpose of reinsurance firms. The existence of reinsurance firms allows individual insurers to match their retained risks with their capital resources. 25.02.2019, SKOPJE, MACEDONIA 13
Reinsurance Increase underwriting capacity Stabilize profits Reduce the unearned premium reserve Prov ide protection against catastrophe loss Pro-rata Method: the ceding company and reinsurer agree to share losses and premiums based on a certain proportion Excess of Loss Method: Reinsurer pays only when covered losses exceed certain level 25.02.2019, SKOPJE, MACEDONIA 14
Reinsurance Pricing Experience rating Data on written premiums and incurred claims Exclude extreme losses Estimate future claims and premiums Exposure rating Historical experience- risk profiles Claim distributions Total Exposure (Sum insured, MPL,EML) 25.02.2019, SKOPJE, MACEDONIA 15
The Basis of Exposure Curve Source: Akarsu,Centendo, Kestel (2018) 25.02.2019, SKOPJE, MACEDONIA 16
EXAMPLE: Univariate Exposure Curves Source: Bernegger, 1997; Akarsu (2018) 25.02.2019, SKOPJE, MACEDONIA 17
Reinsurance Pricing 25.02.2019, SKOPJE, MACEDONIA 18
Reinsurance Pricing The graphs imply that high loss ratio portfolios have more diagonal curves. The concavity depends on differrent type of risk, class of risk, type of peril and size of risk. 25.02.2019, SKOPJE, MACEDONIA 19
Pricing Example: Swiss Re curve of a policy with limit Source: Akarsu (2018) 25.02.2019, SKOPJE, MACEDONIA 20
Cat Bonds available through a Special Purpose Reinsurance Vehicle (SPRV). The insurer purchases from SPRV and pays premiums to the SPRV SPRV holds the premiums and proceeds from the bond sales in a trust, invests the funds in treasury bonds or high quality assets Source: Edesess, 2015. 25.02.2019, SKOPJE, MACEDONIA 21
Cat Bonds The insurance firm establishes a special purpose vehicle (SPV) The SPV sells bonds to investors Sale proceeds are invested in a collateral account If a specified event occurs, funds are withdrawn to cover losses If no event occurs, principal+coupon is paid to investors Cat Bonds pay relatively high interest rates, help diversification of the portfolio Have less moral hazard and counterparty risk Reinsurance can not be traded or transferred but, Cat Bonds can be in secondary markets 25.02.2019, SKOPJE, MACEDONIA 22
Cat Bonds Pricing Three-parameter Lane Model: i.e. full premium is expected loss plus the risk premium. PFL: Probability of first loss CEL: Conditional expected loss S: Spread is the additional coupon rate for taking on additional cat risk. Interest provided by investors LIBOR rate. The price of a cat bond : Total coupon rate % to investors = LIBOR % + spread % 25.02.2019, SKOPJE, MACEDONIA 23
Cat Bond Pricing: Bayesian approach Ahrens, Fuess, Kestel, 2014 Cobb-Douglas model applied to Cat Bond Data (2001-2008) The annual yield spread (risk premium), S, over the six-month LIBOR is adjusted for the reinsurance price cycle. The annual expected loss (EL), probability of first loss (PFL), conditional expected (CEL), risk load, R, (load) are determined. Additionally, based on the analyses on the data, a significant structural break after Hurricane Katrina is observed (Dummy variable). Rating whether a CAT bond is rated BBB or above being set to 1 and zero otherwise of CAT bonds higher. 25.02.2019, SKOPJE, MACEDONIA 24
Cat Bond Pricing: Findings The increase in CAT bond supply after the hurricane season 2005 lead to a reduction in the spreads. CEL increased significantly after Katrina Investors put a higher weight on the expected severity of the CAT bond. Also, during the catastrophe model adjustment process, the implied correlation structure was changed. Additionally, an investment-grade rating is found to increase the impact of the conditional expected loss. The linearized version of the LFC model does perform better than the non-linear Cobb- Douglas form. 25.02.2019, SKOPJE, MACEDONIA 25
Emerging Markets Characteristics i. Lower than average per capita, ii. iii. iv. Which provides rapid growth, Leading to social change which results in high volatility due to natural disasters, external price schoks and domestic policy instability, Requirement on investment capital capital markets are less mature, v. Successful growth leads to higher than-average-return for investors costing more to cover for additional risk of companies in the market 25.02.2019, SKOPJE, MACEDONIA 26
Government Role: Catastrophe Pools Since catastrophic events are so unpredictable and risky for insurance companies, most of them do not offer catastrophe insurance anymore. Governments need to intervene and create various plans to mitigate the problems found in catastrophe insurance. Example: Turkish Catastrophe Insurance Pool (TCIP) 25.02.2019, SKOPJE, MACEDONIA 27
70% of the population lives in the 1st and 2nd degree earthquake zones. Marmara Earthquake (17 August 1999) is a milestone. Established in September 27, 2000. Disaster Insurance Law (Nr.6503) has been in force since August 2012. All of the residential buildings that fall within municipality boundaries The Scheme reduces the burden on government budget. 25.02.2019, SKOPJE, MACEDONIA 28
Aims of the Scheme EQ insurance at affordable prices Creating insurance capacity to limit government exposure Ensure risk sharing by homeowners Improvement of risk & insurance awareness Building up catastrophe reserves over time 25.02.2019, SKOPJE, MACEDONIA 29
Main Characteristics Insurance coverage in return for an affordable premium Use of alternative risk transfer instruments at an attractive cost Realiable distribution channels through insurance industry Enforcement of compliance (checkpoints) Real estate registration offices Mortgage credits Electricity and water subscriptions 25.02.2019, SKOPJE, MACEDONIA 30
Structure of TCIP THE TREASURY REINSURANCE CONTRACT TCIP BOARD FUND MANAGMNT IT MANAGEMENT OPERATIONAL MANAGER PUBLIC RELATIONS INSURANCE COMPANIES LOSS ADJUSTMENT INS AGENTS INS BROKERS BANKS DIRECT SALE LOSS ASSESSMENT HOMEOWNERS 25.02.2019, SKOPJE, MACEDONIA 31
Basic Figures Total no. of policies : 7.5 million Penetration : 42% Annual premium($) : 5.9 million Avg. sum insured ($) : 24.6 million Avg. premium ($) : 37.9 Total claim files : 21.852 Total claims paid ($) : 56 million Max. Coverage ($) : 55,172 25.02.2019, SKOPJE, MACEDONIA 32
Tariffs and Premiums Rates ( ) 1st Zone 2nd Zone 3rd Zone 4th Zone 5th Zone Steel, concrete Masonry buildings Other buildings 2.20 1.55 0.83 0.55 0.44 3.85 2.75 1.43 0.60 0.50 5.50 3.53 1.76 0.78 0.58 Premium amount = sum insured x the tariff rate. 25.02.2019, SKOPJE, MACEDONIA 33
Premium Calculator 25.02.2019, SKOPJE, MACEDONIA 34
Reinsurance Claim capacity of TCIP $4,3 b. (12.5 b.tl) Reinsurance Cat Bond (Bosphorus Re) Funds 2.8 m. Excess of Loss Reinsurance Bosphorus Re: 3-year Cat bond upsized $400m (by April 2013) Funds: 82% Deposit, Govn.Bond 18% $1,29 b. (3.75 b. TL) 25.02.2019, SKOPJE, MACEDONIA 35
TCIP- Cat Bond Issued: 2013 2015 Issuer/SPV: Bosphorus 1 Re Ltd Cedent/Sponsor: TCIP Structuring Agents: Munich Re and GC Securities Size: 400 mi. USD Ratings: BB+ (S & P) Coupon and Maturity: 2.5% and 3 years Issuer/SPV: Bosphorus Ltd Cedent/Sponsor: TCIP Structuring Agents: Munich Re, Swiss Re and GC Securities Size: 100 mi. USD Ratings: BB+ (S & P) Coupon and Maturity: 2.5% and 3 years 25.02.2019, SKOPJE, MACEDONIA 36
Achievements Sustained low prices Less than US$ 40 Improved penetration From 4% to 42% Strong reserves Over US$ 1.25 billion Increased public awarenes Over %90 brand and product recognition Release on national budget US$ 5 billion claim payment capacity Effective PPP structure Become a global role model 25.02.2019, SKOPJE, MACEDONIA 37
Thank you! Благодарам! 25.02.2019, SKOPJE, MACEDONIA 38
References Sigma (2018). Swiss Re, Natural catastrophes and man-made disasters in 2017, No:1/2018 http://www.artemis.bm/deal_directory/bosphorus-ltd-series-2015-1/. TCIP, www.dask.gov.tr Ahrens, F., Fuess, R., Selcuk-Kestel, A.S., 2014. A Bayesian Pricing Model for CAT Bonds Springer Proceedings in Mathematics & Statistics 73. DOI 10.1007/978-3-319-04849-9 4. Ünal, B., Askan, A., Selcuk-Kestel, A.S., 2017. Simulation of large earthquakes and its implications on earthquake insurance rates: a case study in Bursa region (Turkey), Natural Hazards, 85(1), 215-236. Morton N. L.2000. Pricing Risk Transfer Transactions, ASTIN BULLETIN, 30(2) 259-293. Bernegger, S., 1997. The Swiss Re exposure curves and the MBBEFD distribution class 1, ASTIN Bulletin, 27(1), 99-111. 25.02.2019, SKOPJE, MACEDONIA 39
Thank you very much for your attention! Contact details: A. Sevtap Kestel address: ODTU, UME 06800 Ankara, Turkey phone: +90 (0)312/210-5614 mail: skestel@metu.edu.tr web: www.iam.metu.edu.tr 25.02.2019, SKOPJE, MACEDONIA 40