Focus on Terminology

We understand that terminology can be a hurdle for actuaries looking to incorporate climate risk, as an example, into their assessments. Terms that have been used for decades within specific specialties may be confusing to those outside the field, and terms and definitions can change over time.

This article, the first in a “Terminology” series, addresses inconsistent, evolving and complex terms that may require further explanation or examples to help risk practitioners. These terms are either new to actuaries or may be used differently by actuaries, climatologists, sustainability experts and other stakeholders. By increasing vocabulary awareness, we hope to facilitate better communication among these professionals.

We believe this series, developed in collaboration with the SOA Research Institute’s Catastrophe and Climate Research Steering Committee, will be a valuable resource for actuaries and other professionals working with climate risk.

For this first in the series, Max Rudolph, Matti Goldberg and Steve Bowen examine the use and understanding of the term “models.”

“Models”

Models play a key role in the world of actuaries and scientists, but do we all mean the same thing when we use this term?

In Actuarial Standard of Practice (ASOP) Number 56 Modeling, a “model” is defined as:
“A simplified representation of relationships among real-world variables, entities, or events using statistical, financial, economic, mathematical, non-quantitative, or scientific concepts and equations. A model consists of three components: an information input component, which delivers data and assumptions to the mode; a processing component, which transforms input into output; and a results component, which translates the output into useful business information.”

Useful business information is not defined but is likely to be primarily financial in scope—driving insurance or pension cash flow models forward, resulting in income statements and balance sheets. Although models can be quantitative or qualitative, little thought goes into indirect outcomes like future demographics or greenhouse gas emissions. Scenarios typically change one variable at a time and show the impact of that change, although additional complexity is added at times (e.g., interest rate increases may also increase default rates for a non-investment-grade floating rate bond).

Deterministic scenarios are predefined while stochastic scenarios randomly choose changes in assumptions for each point in time from a distribution. These results are then used to manage risks, either accepting, hedging or avoiding them for current or in-force blocks of business.

Models, further defined

The Intergovernmental Panel on Climate Change (IPCC) defines “models” as:

“Structured imitations of a system’s attributes and mechanisms to mimic the appearance or functioning of systems, for example, the climate, the economy of a country, or a crop. Mathematical models assemble (many) variables and relations (often in a computer code) to simulate system functioning and performance for variations in parameters and inputs.”

A climate scientist models climatic changes and at times the social aspects of that change.1 Climate models simulate, based on mathematical equations, the physical processes in the atmosphere, ocean, ice and land. At times, these are complemented with models of human social and economic activities, forming “integrated assessment models” to estimate, for example, the social consequences of climate change or the climate effects of specific development pathways.

For climate modeling to support socioeconomic narratives, a broad set of financial and governmental land-use, agricultural, infrastructure, education, health care, and other inputs are integrated, and these interactively move from the current time to future points and environmental or social conditions globally.

One way this is accomplished is for modeling teams from around the globe to agree on methodologies in so-called “coupled model intercomparison projects” (CMIPs) using supercomputers and operating under the IPCC seven-year assessment cycle. Combining the outputs from all the modeling teams results in a multi-model ensemble set of scenarios that is then reviewed by the IPCC and converted from climate-forcing scenarios of greenhouse gases into a limited number of narrative scenarios that project impacts on gross domestic product and other domains. These share conditions likely to be present if that scenario plays out, like trade policy and health expenditures.

In Closing

Climate scientists use models to project future scenarios—both for environmental and social change—given specific greenhouse gas emission levels and development pathways to inform international climate policy debates. Actuaries interpret the inputs to do what-if analysis and test the outcomes of various economic inputs. The layperson is unlikely to understand the difference between these, but it is useful for the experts to understand the differences and how tools used by their counterparts could help their own analysis.