34 O Hasu and T-K Lehtonen
design of insurance products concerns the relation between the complexity
of the models used and their intended purpose. An increase in complexity
tends to lead to higher resource costs and less predictable performance of
the instrument.
RM includes a didactic section in which the authors lay out a theoretical
framework for justifying how a system is objectified in the design of index
insurance. They explain the thinking behind probabilistic modelling choices
and detail how models helpfully simplify reality and serve as tools that fulfil
context-specific goals. The guide concentrates on examining systems operationally; that is, as defined on the basis of how they work rather than what
they are. The emphasis on operationality is elaborated further in defining the
hierarchy of different models that comprise the totality of an index insurance
product. Index insurance development uses several submodels for processing
economic and ecological data, each of which has additional models defining
parameter values. In the formal hierarchy of index insurance design, payout ratio modelling is at the top, while indices for environmental risk data,
such as drought frequency and drought severity, are situated as submodels (RM: 102). Importantly, this multi-layered apparatus is too complex for
calculating definitive values. Instead, index insurance relies on probability
simulations that generate value approximations with 10,000 simulation repetitions recommended for each variable (RM: 106). The contrast with traditional forms of insurance is marked, as risk modelling for index insurance, as
developed by RM, is not founded on historical variation. Instead, simulation
constructs a system that is predetermined in terms of its variation (on the difference between the ‘archival-statistical’ mode of traditional insurance and
‘enactment-based’ knowledge provided by simulations, see Collier 2008).
The use of probabilistic simulations underscores that the reductive objectification of the soil is a process where financial theory is constitutive of the
categories used in mapping ecological uncertainties. Here, it is noticeably
difficult to separate empirical data from the theoretical models that condition how data is instrumentalised into a tool of weather-related risk prediction (e.g. Edwards 2010, p. 282). The role of the submodels is heavily reduced
in the final product. Instead of taking into account environmental factors, it
focuses on modelling payout ratios:
[T]he model is not actually simulating the weather (such as rainfall), nor
is it simulating the weather index (for example, drawing from a distribution of index trigger values). Instead, the model directly simulates the
uncertainty around the actual payout amounts. An important advantage of this approach is its simplicity and the relative ease of explaining
and understanding its results.
(RM: 269)
The choice is elaborated by detailing the assumptions and conditions
behind successful modelling practices. The suggested strategy presupposes
design of insurance products concerns the relation between the complexity
of the models used and their intended purpose. An increase in complexity
tends to lead to higher resource costs and less predictable performance of
the instrument.
RM includes a didactic section in which the authors lay out a theoretical
framework for justifying how a system is objectified in the design of index
insurance. They explain the thinking behind probabilistic modelling choices
and detail how models helpfully simplify reality and serve as tools that fulfil
context-specific goals. The guide concentrates on examining systems operationally; that is, as defined on the basis of how they work rather than what
they are. The emphasis on operationality is elaborated further in defining the
hierarchy of different models that comprise the totality of an index insurance
product. Index insurance development uses several submodels for processing
economic and ecological data, each of which has additional models defining
parameter values. In the formal hierarchy of index insurance design, payout ratio modelling is at the top, while indices for environmental risk data,
such as drought frequency and drought severity, are situated as submodels (RM: 102). Importantly, this multi-layered apparatus is too complex for
calculating definitive values. Instead, index insurance relies on probability
simulations that generate value approximations with 10,000 simulation repetitions recommended for each variable (RM: 106). The contrast with traditional forms of insurance is marked, as risk modelling for index insurance, as
developed by RM, is not founded on historical variation. Instead, simulation
constructs a system that is predetermined in terms of its variation (on the difference between the ‘archival-statistical’ mode of traditional insurance and
‘enactment-based’ knowledge provided by simulations, see Collier 2008).
The use of probabilistic simulations underscores that the reductive objectification of the soil is a process where financial theory is constitutive of the
categories used in mapping ecological uncertainties. Here, it is noticeably
difficult to separate empirical data from the theoretical models that condition how data is instrumentalised into a tool of weather-related risk prediction (e.g. Edwards 2010, p. 282). The role of the submodels is heavily reduced
in the final product. Instead of taking into account environmental factors, it
focuses on modelling payout ratios:
[T]he model is not actually simulating the weather (such as rainfall), nor
is it simulating the weather index (for example, drawing from a distribution of index trigger values). Instead, the model directly simulates the
uncertainty around the actual payout amounts. An important advantage of this approach is its simplicity and the relative ease of explaining
and understanding its results.
(RM: 269)
The choice is elaborated by detailing the assumptions and conditions
behind successful modelling practices. The suggested strategy presupposes