Tag Archives: credit risk

How Insurtech Changes Credit Risk

Risk management activities of insurance companies are mainly based on three risk types: whole portfolio, supplementary and others. In “others,” two risk types — operational risk and credit risk — stand out with their financial impacts and frequencies.

Credit risk is defined as “the potential that insurance company’s borrowers or counterparties will fail to meet their obligations in accordance with agreed terms.” The main goal in credit risk management is maximizing insurance company’s risk-adjusted rate of return by maintaining credit-risk exposure within acceptable parameters. Credit risk has six sub-types:

  1. Credit default risk
  2. Concentration risk,
  3. Counterparty risk,
  4. Country risk,
  5. Sovereign risk and
  6. Settlement risk.

Furthermore, traditional credit risk management is based on manual or semi-manual assessment of these domains:

  • Detailed assessment of counterparties,
  • Financial strength,
  • Industry position,
  • Qualitative factors and
  • Underlying credit exposures.

The first trigger of change in credit risk management was Solvency II. After implementation of the capital regime in Euro Zone, insurance and reinsurance companies integrated further credit risk assessment tools into their internal models, because the credit risk management approach was found very weak in standard model of EIOPA. The second evolution in credit risk management comes not with another capital regime but with technology: insurtech. Insurtech is converting credit risk management into a new form like many other components in insurance business.

See also: A ‘Credit Score’ for Your Cyber Risk?  

For bringing into the complex structure of risk management with basic inputs, we can classify the insurtech effect on credit risk management mainly on two points. The first point defines the philosophy behind risk management activities, and the second point defines actions:

  1. Maximizing a company’s risk-adjusted rate of return by maintaining correct credit risk exposure within the risk appetite of the company and maintaining sufficient risk-return discipline in credit risk management process.
  2. Covering all insurance/reinsurance transactions and identification, measurement and monitoring of transactions with embedded credit risk.

The risk-adjusted return is generally defined as a concept that measures real value of risk and enables a company to make comparisons between risk taking and risk aversion. This variable shows real value of business and aims at maximizing efficiency on capital management. In business today, correct allocation of limited capital should be the main object behind all activities of a company, and risk-adjusted rate of return is the pointer that makes this objective visible. Insurtech also converts the calculation methodology of risk-adjusted return. With a more sophisticated methodology, risk managers can cover thousands of variables and calculate a value very close to real, risk-adjusted return exposure.

The second point, covering all transactions where credit risk arises, is the inception point of actions. The definition covers not just financial transactions but also all insurance/reinsurance transactions performed during daily business cycles. Furthermore, because of the complex structure of finance, not just loans, the most obvious source of credit risk, but also other structured financial instruments, like trade financing, foreign exchange transactions, financial futures, swaps, bonds, equities, options, etc., should be assessed in an effective credit risk management function.

Naturally, the variety of sources brings a huge amount of data, which could not be managed manually, especially by a function like risk management, which should be always preventive and pioneer. One of insurtech’s dimensions, big data management, helps risk management professionals especially on this point. With the organization, administration and governance functions of big data management, not just structured data but also unstructured data coming out from mentioned transactions will be measured, analyzed, grouped and monitored according to their likelihood and magnitude within seconds.

See also: How to Adapt to the Growing ‘Risk Shift’  

Credit risk management is a crucial tool among other risk management functions. Effective credit risk management and efficient capital management make companies ready and solid for their next step on investment, acquisitions and every step they take for their existence.

How to Create Risk Transparency

There was a time not long ago when a bank originated a loan and kept that loan on the balance sheet until it was repaid. The amount banks could lend was limited to the deposits they had on hand and the banks’ own ability to borrow. Today, credit risk is traded regularly, with specialized data and analytical services giving investors confidence they understand the risks they are assuming. But there has been limited opportunity for investors to deploy capital against specific pools of insurance risks, because of a lack of that sort of transparency. With the vehicles that do exist, it has been difficult to structure the transfer of risk to meet investors’ respective objectives and risk tolerances.

However, insurance may be reaching a point in its evolution where the information gap will begin to narrow. Up until today, insurance risk had often been opaquely and highly subjectively valued. Today, actuaries set reserves based on highly summarized data, and underwriters set premiums based on claims experience that is extrapolated forward using historical loss development patterns and subjective future “trend” projections (or ad hoc substitute measures for risk), neither of which may represent future risk of loss. Outside of property catastrophe risk, where the data elements are generally available in some detail, granular risk data simply has not existed. However, rapid change could now be approaching. Vehicle telematics, wearable sensors, connected machines and other components of the Internet of Things (and Beings) are producing real-time data that allow us to look at risk in real time, rather than relying on current industry practices.

See Also: A Better Way to Assess Cyber Risks?

Credible, data-driven risk indices may create a variety of opportunities, including:

  • Capital Providers: Investment in specific index-based structured insurance pools that are aligned with respective objectives and relative risk tolerances could improve on the alternatives available today, where those who want to invest in insurance risk are often restricted to investing in insurance companies or risk pools that involve assuming underlying exposure to the operational, asset and credit risk, as well as the insurance risk of the originating insurer’s business.
  • Insurance Clients: Clients are likely to observe premium and associated underwriting decisions more transparently and could thus anticipate the cost/benefit implications of decisions taken to reduce risk.
  • Regulators: Regulators could gain greater confidence in the balance sheet valuation disclosed by insurance companies, which has the potential to decrease the regulator’s view of risk capital necessary to support risk.

One could argue that straightforward consumer and commercial loans are much simpler than the risks underwritten by insurers. However, when taking a critical look at the complexity of the financial projects currently being traded by investors, that notion is hard to support. In fact, many of the underlying risks facing lenders are very closely related to the risks facing insurers. Perhaps the biggest differentiating factor is the lack of standardization of contracts, which creates a degree of complexity.

From a contractual perspective, however, complex derivatives, other hard-to-value instruments and non-transparent assets can be at least as opaque and complex. Yet the core elements for assessing risk are available, and credible calculations exist within the valid range of assumptions.

The insurance industry could benefit from the increasing availability of relevant data. That data could be the byproduct of other applications, such as route data from fleet management software; vehicle data from predictive maintenance applications; traffic density data from road management applications; or environmental data from various sources. Or, it could be data that has been custom-generated for insurance applications, such as the data from telematics devices used by personal auto insurers to capture driving behavior. I see the biggest promise in using the data exhaust from other applications. I suspect clients would be averse (in many cases) to additional data capture specifically for insurance but would be open to sharing data already captured—as long as there are appropriate safeguards to ensure that it does not disadvantage them as clients.

The industry will need to invest in new analytical techniques to leverage these new data sources. In many other sectors of the economy, “big data” is having a real impact. This has required new tools and algorithms that might be unfamiliar to most analytical professionals within insurance. David Mordecai and Samantha Kappagoda, co-founders of the RiskEcon Lab at Courant Institute of Mathematical Sciences, which is among the world’s leading applied mathematics and computer science research institutions, explained the necessary evolution:

“The increasingly pervasive proliferation of remote sensing and distributed computing (e.g. wearable tech, automotive telematics), and the resulting deluge of ‘data exhaust’ should both necessitate and enable the emergence of digitized risk management programs. Ubiquitous peer-to-peer interactions between human ‘crowds’ and machine ‘swarms’ promise to dominate commercial and consumer activity, as already observed within omni-channel advertising exchanges and high-frequency algorithmic stock trading platforms. Financing and insurance functions involving risk-transfer, risk-sharing and risk-pooling will increasingly be facilitated by and executed seamlessly within code. Among others, Bayesian statistical and adaptive process control methods (e.g. neural networks, hidden Markov models)—originally employed within the telecommunication, electricity, chemical industry and aviation during the mid-20th century, and more recently adapted for voice, visual and text recognition, along with other supervised and unsupervised data mining and pattern recognition methods—will need to be widely adopted to identify, monitor, measure and value underlying risk factors.”

In my opinion, new data and new techniques are likely to create a degree of transparency in insurance risks that has never existed. That transparency could benefit capital providers (both insurance company investors and direct investors in insurance risk), clients and regulators. A new era is quickly approaching where information and analysis have the potential to remove the cloud engulfing insurance risk. There are likely to be substantial benefits for those forward-thinking companies that exploit the opportunity.