Currency hedging advisor
Premise: currency hedging only. A South African gold producer earns US dollars and pays its costs in rand. The purpose of currency hedging is to strengthen the business through risk mitigation: more predictable rand cash flow for wages, costs, debt and investment, whatever the rand does next. This exercise asks one question: how much of the next 12 months' dollar revenue should be converted to rand at a fixed rate now? Gold price hedging is deliberately left out. It has fallen out of favour with many producers, because those who locked in their gold price missed out on large bull runs. It can be incorporated later, but not at this stage.
Move the sliders and watch the reasoning change. The advisor weighs margin, balance sheet and where the rand sits against its 3-year average.
Illustrative only. This is a demonstration of decision logic, not financial or treasury advice, and the figures are sample values, not live gold prices or exchange rates. It runs entirely in your browser; nothing you enter is sent anywhere.
Inputs
Type a value in any box, or open Range under a lever to widen its slider. The fuzzy points sit with their charts further down.
Policy limits
Hard rules around the fuzzy result: the board's maximum and the lender's minimum.
Market corrections
The rand tends to weaken over time because South African inflation usually runs above US inflation. Tick the box to compare today's rate with the 3-year average adjusted for that drift, instead of the plain average.
of the next 12 months' US-dollar revenue, hedged into rand,
Scenario log
The default inputs are already locked in as your first scenario. Change a lever or pick a scenario, then lock it in to compare side by side. The last five are kept in this browser; click a row to restore it.
The 12 rules
The advisors' rule set gives one outcome for every combination of the three readings.
- Fires: how true the rule is right now, the weakest of its three readings.
- Share: its part of the decision; all firing rules together make 100%.
- Points: share × its hedge level. The points add up to the recommendation.
Hedge levels
The hedge % each rule outcome stands for.
By degrees, not cut-offs.
The five levers become three readings: margin (gold price, cost and the rand), solvency, and the relative forex rate (today's ZAR/USD rate against its 3-year average). Margin is Low, OK or High; solvency and relative forex are Low or High; each to a degree, never a hard cut-off.
The amber line is where the business sits now; the numbers show how true each reading is, from 0 to 1. Readings are sinusoidal: between two points, one fades out along a half-sine while the next fades in, always adding up to 1. A rand at its 3-year average (0%) is half Low, half High, so the strong-rand and weak-rand rules each count for half; the result lands halfway between the two. The exception is margin from OK down to Low, which drops off exponentially: it falls away fast below the OK point and is almost fully Low well before 0%. The rand works twice. A weaker rand lifts the margin (more rand per dollar), which leans towards hedging less, and it also lifts the relative forex rate, which leans towards hedging more. Both effects are real; the rules weigh them together. Response charts show the recommendation as that one reading moves, with the other two readings held where they are (the dashed line is before policy limits).
Where the model goes next.
These are not in the model yet. Each would be added as a new reading with its own Low, OK and High, and its own rules, so the reasoning stays visible.
Hedge or postpone in rash markets
When the currency market overreacts to short-term volatility, decide whether to act on it or wait. A sudden rand slump could be a chance to lock in part of the rate; a disorderly market, or one where forward cover is poorly priced, is a reason to postpone. Adds a timing signal (act now, stage in, or wait) on top of the hedge %.
Composite indicators
A multi-variable parameter we develop to capture related issues in one reading, for example regional stability: political and regulatory risk, labour action, power supply reliability, and logistics and port disruption, each scored and weighted into one index from 0 to 100.
Expert sentiment
The view of a recognised thought leader or analyst, scored from bearish to bullish with a confidence level. Kept to a small, capped weight and logged with its source, so it informs the decision without driving it, and a person still approves.
Forward premium
Rand forward rates usually sit above the spot rate, because South African interest rates are higher than US rates. That makes hedging a little more attractive than the spot rate suggests. It is linked to inflation, since central banks use interest rates to control inflation, so it would sit with the market corrections.
Forex volatility
How unsettled the rand is against the dollar. A volatile rand makes rand revenue less predictable and strengthens the case for cover.
Gold price hedging
Locking in the US-dollar gold price as well as the exchange rate. Left out on purpose: it has fallen out of favour because hedged producers missed out on large bull runs. The same approach can incorporate it later, as a separate recommendation with its own readings (for example gold price against its 3-year average).
Keeping the rules manageable
Each new reading multiplies the rule count: 12 today, and two more three-level readings would make 108. Two ways to keep it workable: give new readings only Low and High (no OK), so two more readings make 48 rules instead of 108; or group readings into small sub-models, for example market (forex, forex volatility, and later gold price) and business (margin, solvency), whose outputs feed a short final rule set.
Want this logic around your own decisions?
The same approach works for stock levels, maintenance timing, production planning and pricing: rules your people already use, written down, weighed consistently, and explained every time. Our founder built fuzzy-logic controllers for minerals processing plants long before AI was a buzzword.
A real currency hedging policy also weighs the revenue profile, the instruments (forwards, options, collars) and their cost, counterparty limits, and tax and accounting treatment.