# Risk reduction

This is categorised as a potential [benefit to members](/_P1DHaeIShuTZh6gpShPnA).

In the context of extending [trade credit](/CncoUGlDQU-1ZRS5KpfLsg), there are several 'events' associated with risk; late payment, default and loss of credit value. The first two could be considered 'normal' risks, whilst events that affect the value of the mutual credit units (e.g. club failure) can be categorised as sources of 'systemic' risk. They are discussed separately, although in reality they are correlated as the default of a particularly important club member could plausibly precipitate collapse.

A common definition of 'risk' is 'probability of event mutliplied by impact of event'. In a business context, impacts of late fiat payment or default depend on multiple factors (including the affected party's response), but may include forced reliance on expensive sources of finance or disruption to spending plans (see 'consequential harm' from around page 30 [here](http://www.hblr.org/download/HBLR_1_1/Ricks-Regulating_Money_Creation.pdf)).

In situations where outcomes are predominantly financial by design (a casino or financial markets, for example), it may make sense to measure all impacts in purely financial terms; quantifying risk in fiat units seems appropriate. Even where monetary gain and loss is not the central consideration of success, it may be useful to use such measures of risk in addition to other impacts that cannot be reduced to purely financial terms.

In the case of default the primary financial impact is simply loss of the amount that was due, and the ['expected loss'](https://en.wikipedia.org/wiki/Expected_loss) approach seems reasonable. In the case of late payment, however, the amount due is not lost but merely delayed. Economists have attempted to use the difference between the quantity of an expected payment and the quantity multiplied by a 'discount factor' (given by some time-dependent [discounting curve](https://en.wikipedia.org/wiki/Hyperbolic_discounting) determined by subjective time-preferences for money) to establish a consistent way to quantify the value people place on different payments at different points in time.

In principle, this approach could be applied to analyse financial risks associated with late trade credit settlement; the interval between an invoice being issued and settled could be used to discount payment amounts, resulting in an estimate of the impact (measured in fiat units) for any given realised interval. With sufficient data, late payment risk could be presented in the form of statements such as 'after joining a trade credit club, member x experienced an average late payment risk reduction of £1504.32 per month'.

This makes the quantification of risk in fiat explicit, but it seems wrong-headed (even within the limited scope of purely financial risk). It assumes that time-preferences for money are fixed and context-independent and can therefore be established on the basis of responses to abstract scenarios in a survey, and gives the impression of providing an 'objective' basis for making decisions, whereas a key part of the calculation is based on subjective preferences. Some specific dangers:
* Because has the appearance of being objectively measureable, it could distract from non-financial impacts.
* It gives a misleading impression of precision.
* The common unit of fiat risks mixing up these estimates with actual savings that can be directly observed (e.g. a reduction in interest paid to financial institutions).

Ultimately, it wouldn’t convey anything meaningful to anyone who wasn’t intimately familiar with the business, even if contextualised with other numbers like turnover. The most appropriate response would be ‘what does that mean for me?’ to which there is no simple answer. It therefore seems best to confine late payment analysis to comparing only payment interval probability distributions (the first, objective component of the usual risk formula) and letting people decide for themselves what trade credit club membership would mean for them with respect to late payment risk.

This is actually consistent with a [more recent definition of risk](https://en.wikipedia.org/wiki/ISO_31000#Definitions) as 'effect of uncertainty on objectives', which suggests a much more open-ended perspective. Choosing to limit quantitative analysis does not, therefore, mean that 'late payment risk' is never referred to - data showing that members experience late payment half as often (or the associated delays are much reduced) as they did before joining can meaningfully be interpreted as demonstrating a reduction in late payment risk, but it does mean that the analysis does not presume to collapse multi-faceted and context-dependent outcomes into a one-dimensional quantitative measure as a supposedly objective way of evaluating benefits.

To summarise, whilst the financial risk of default might be quantified in fiat units without making potentially misleading assumptions (although [non-ergodicity](/JL-hFK7BRjuxx6zbz3O6Qg) is a concern when dealing with expected values - how important this is in practice needs further research), it seems best to confine the analysis of late payment risk to the probability distributions of different payment intervals being realised without attempting to express the result in fiat terms. The same probably applies to systemic risk, although it does seem intuitively obvious that a club member's exposure to this risk is directly proportional to the amount of mutual credit they are holding at any given time, which does suggest one plausible quantitative approach beyond simple probabilities of systemically-relevant events occurring.

(Philosophical sidenote: I think there’s a broader point here - some aspects of economics are inherently qualitative and there are things that cannot/should not be quantified even in principle. It would seem a shame (not to say ironic) if we were to reinvent the means of exchange and then interpret the outcomes through modes of thinking that have done much to prop up the old system. If we can build an alternative (that works!) on a basis that is in large part explicitly qualitative (whilst of course applying quantitative methods where appropriate) this could be an even more profound challenge to the status quo.)

In any case, the condition that must be met for [trade credit club](/Q93p0MdKRr-D07uK3iAVHg) participation to make sense for each potential member is that the risks are less than those they are exposed to in the course of their usual bilateral trading relations. This will vary on a case-by-case basis and compelling evidence can only come from real-world data and experience, but the entries listed here outline one possible way of operating a club, consider the general conditions that would make membership attractive, establish some plausible mechanisms for achieving these and specify the necessary variables and data to draw appropriate quantitative conclusions.

* [Late payment](/iIVkdpz_S76a_6FexHZWAw) (also includes a description of possible trade credit club operation)
* [Default](/nO2IbqcMTbiNSY0AG0PCnA)
* [Systemic risk](/H3VYE7o3QwWqkw3dw4l3GQ).

## Further work

Only one arbitrary trading activity scenario has been explored for late payment and default, and in a fairly limited way. It would be worth building an agent-based model to easily generate and systematically explore various trading activity patterns and club configurations. This could potentially include [parasitic strategies](/fNhqlAK8TNqmDiXtEuLUWg) by which members gain from imposing risks on others.

Federated clubs have not been considered.