Menu Search through site content What are you looking for?
20 April 2018

Parametric Insurance: closing the protection gap - Legal considerations

20 April 2018 Global

1. Introduction

Even though parametric insurance has an established track record in many countries, some regulatory and legal uncertainty remains. Parametric insurance products are novel and may not be addressed by statute. Regulatory frameworks that might exist in this area are usually not firmly codified and remain largely untested.

Click here to download the full report on Parametric Insurance

In common law jurisdictions, case law has yet to be established that would inform understanding of how these types of policies will operate, be classified for regulatory purposes and be legally enforced. Under English law, insurance is a contract whereby for consideration one party promises to pay another if a specified event occurs that is adverse to the interests of the insured. There is nothing within this traditional definition to preclude a contract based on a parametric trigger – the agreed parameters would simply be the arbiter of the “specified event”.

Parametric insurance products may cause legal or regulatory uncertainty in jurisdictions where:

  1. the insured must have an ‘insurable interest’ at the time the policy is underwritten and/or at the time the loss occurs.
  2. the size of the insurance pay-out must correspond to the actual loss suffered by the insured. This ‘indemnity principle’ can mean that in certain jurisdictions, an insurer may only restore insureds to their pre-loss financial position, such that losses must be valued or assessed before claims can be paid.

Reconciling insurable interest with payments to interested third parties

Parametric insurance policies are often purchased by sovereigns or sub-sovereigns and are becoming increasingly popular with disaster relief and humanitarian organisations.

Under English law it seems readily arguable that local, regional or even national governments have a direct insurable interest in the effects of a natural disaster on their populations, since without insurance backing they would have to fund the full cost of the disaster recovery from state funds and could be expected to suffer losses in tax revenue, for example.

A non-governmental organisation (NGO) might have plans in place to intervene in a crisis (for example if crops were to fail) even though it is not the NGO itself that is suffering the loss directly. Here the “insurable interest” may be slightly more tenuous, although on balance such a purchaser would likely be able to show a legitimate interest in the cover, rather than it representing a mere speculation. Some jurisdictions may take a wide view and find an insurable interest of a government in its territory and population or of an NGO in the people and places within its scope of operations.

There are other interesting potential work-arounds.

In China, parametric insurance products providing protection against drought have been arranged collectively at a regional government level, and reinsured by international players.

Protection is then sold at a subsidised premium to individual farmers, who effectively become the policyholders and receive pay-outs directly.

The farmers themselves clearly have an insurable interest although it is the government purchaser who has arranged and subsidised the insurance as a means of building local resilience within its populace.

Dealing with valuation issues under the indemnity principle

English law has long recognised the concept of valued policies whereby the insurer agrees to pay a fixed sum once the loss is established, without a need for further adjustment or valuation at the time of the loss. However, the indemnity principle can potentially create regulatory and legal challenges in jurisdictions where codified insurance law does not traditionally permit ‘contingent contracts’, requiring instead that any losses are subject to valuation.

In India there has been relatively widespread take-up of parametric products particularly covering agricultural risks such as crop failure due to drought or flood. For example, a number of the larger multinational insurers including AIG, Sompo Canopius and Tokio Marine – in conjunction with their Indian partners – have come together under the umbrella of the Agriculture Insurance Company of India Limited to offer parametric solutions which are generally purchased by farmers as a requirement of their lenders. However, because of Indian regulation and law around contingent contracts, the insured needs to prove to the insurer what the loss has been. This can, at times, slow down the payment process and thereby somewhat dampens one of the key benefits of parametric cover.

The element of indemnity may also have an impact on how a product is classified, regulated and taxed. The UK Law Commissions note that the main difference between parametric insurance and derivatives seems to be that parametric insurance contracts usually require at least some nominal element of loss before the policy will pay out and that often insurers will require the insured to provide a sworn proof of at least some actual loss.

In South Africa, where parametric insurance is being actively encouraged by a government mandate to provide financial services to the commercial farming sector and to agri-business, an insured must prove they have suffered a loss and that they have an insurable interest in the loss to avoid the cover being classed as a derivative product and regulated as such. Unlike in India, however, insureds in South Africa are not required to quantify the extent of the loss before receiving a payment, merely to prove that some loss has been suffered.

Reducing basis risk via improved modelling

Parametric insurance responds to a trigger (e.g. wind speed exceeds a threshold), and the amount payable depends on the modelled outcome of that trigger. So there is an inherent basis risk: the trigger might not cause loss for the insured; or the loss it incurs might not be what was modelled. Of course, conventional insurance does not always fully indemnify the insured for its loss, but the reason is usually more obvious – for example a retention, limit or exclusion.

Insurance regulators are understandably keen to see that basis risk is minimised, but it can create uncertainty and damage trust in the insurance sector.

Reducing basis risk is particularly important where the policyholder is unsophisticated, or a first-time purchaser of insurance.

Accurate modelling is key to ensuring that any gap between losses experienced and parametric payments is as small as possible. The African Risk Capacity (ARC) is a world leader in parametric trigger modelling and capacity building with its bespoke Africa RiskView tool. ARC’s experience has provided an interesting example of the importance of accurate input data. In 2016, a policy issued to Malawi was not immediately triggered even though there was widespread crop failure. Subsequent investigations by ARC revealed that farmers in Malawi had switched to a different crop with a shorter growing cycle. When this data was put into the model it created a more accurate estimate of the drought-affected population and a pay-out was triggered.

In time, basis risk may be minimised by more detailed and accurate modelling and greater availability of data. Sophisticated parametric triggers for hurricane risks are now designed to respond to a storm’s radius, maximum wind speed, and latitude and longitude at landfall. Another example of innovation in modelling is the Kenya Livestock Insurance Program, supported by the Kenyan government, Swiss Re and the World Bank, which uses satellite images to monitor grazing conditions. When livestock are at risk it delivers parametrically triggered insurance payments for feed, veterinary medicine and water directly to pastoralists via mobile phones.

The costs of developing better parametric models are also likely to significantly decrease as enabling data and software are increasingly shared as an open resource.

The Oasis Loss Modelling Framework is one example of an open source catastrophe loss modelling platform. NASA’s Global Flood Monitoring System provides realtime satellite data and hydrological runoff analysis as an online resource and the OpenQuake Platform allows modelling analysts to share datasets and tools to assess earthquake risk.

Leveraging mixed models

For simplicity and to build trust, parametric insurers might also start off by using simple single datasets as the basis for a parametric trigger instead of more detailed and complex models.

Another possible solution to the basis risk issue is the creation of mixed models.

These are insurance products which have a parametric trigger allowing for immediate emergency funding to be released to a policyholder combined with a normal indemnity function that tops up any additional loss after adjusting.

Such hybrid models are increasingly available and can assist in reducing basis risk and building regulatory and consumer trust for parametric insurance whilst such products are in their infancy. This is particularly the case where there is a strong indemnity principle under local law and losses must eventually be quantified or evaluated.

The CCRIF started with a simple parametric index, which allowed the scheme to be up and in place quickly. After three years, and after more research and model building, it switched to a modelled loss basis that enabled new hazard modules and a variety of exposure database formats to be added.

The FONDEN scheme in Mexico consists of a parametric catastrophe bond to provide immediate post-disaster funding combined with an indemnity-based insurance to cover local governmental assets.

About the report


20 April 2018

About the report


20 April 2018