Basis risk, or the gap between the index and yourself

Basis risk is the price paid for speed and simplicity, and it is the central concept of the whole of alternative risk transfer. It denotes the gap between the compensation received under an instrument with a non-indemnity trigger and the loss actually suffered. As soon as the trigger stops following the sponsor's own loss, it follows something else, and that something else never coincides perfectly with what happens to it.

The gap runs both ways and the two are not equivalent. Positive basis risk means receiving more than you lost: financially comfortable, it raises accounting and sometimes regulatory questions about how much risk was really transferred, and above all it announces itself, since nobody forgets to bank a payment. Negative basis risk means receiving less than your loss, or nothing at all despite a heavy one. That is the structural problem of non-indemnity triggers, and it is discovered at the worst possible moment, that is, during a loss.

Three causes widen it, and they are dealt with differently. The first is geographic: the area where the parameter is measured, or the one the index aggregates, does not overlap the area where the portfolio is concentrated. It is the most frequent and the most measurable cause, since it is worked by comparing two maps. The second is the nature of the peril: a physical parameter describes a cause, and actual damage depends on the vulnerability of the building stock, on construction rules, on insured density, all of which the parameter ignores. The third is measurement lag: an index publishes an estimate that is revised over months, and the threshold is judged on a value that is not yet the final one.

Only the first of those three causes is genuinely corrected by the choice of trigger, and that is an important nuance. Moving from a parametric trigger to a modeled loss one, or from a national index to a regional one, narrows the geographic gap. It does nothing against vulnerability, which depends on the portfolio, nor against revision of the estimate, which depends on the index provider. Believing that the choice of trigger settles basis risk means treating one cause out of three.

The experience of Hurricane Irma in 2017 served the market as a shared lesson. Several parametric cat bonds on the Caribbean were indeed triggered by the wind parameters measured, and several sponsors nevertheless found their actual losses diverging noticeably from the compensation received, owing to unusual tracks and geographic disparities in their exposure. The episode fed a lasting debate on trigger design, and it made plain that the gap is observed after the fact while it is decided at the drafting stage.

Hence a working method with nothing sophisticated about it, applied before buying. You take past events for which you have your own loss experience, compute what the contemplated instrument would have paid on each, and compare the two series. The exercise predicts nothing, and makes no such claim: it shows the amplitude of the gaps over what has already been lived through, which is already far more than intuition. An instrument that would have paid nothing on the two worst events of the last ten years says something no price says.

Basis risk is not a defect to be eliminated but a trade-off to be chosen, and that is how practitioners treat it. An indemnity trigger removes it and costs in delay, in the trust demanded of the investor, therefore in price. A parametric trigger pays in a few days, which can be worth more than an exact cover paid eighteen months later when cash decides survival. The right question is not which instrument removes the gap, but which gap is accepted for which speed, and that choice is made once, in writing, before the event.