Forward curves explained

Unlike a stock, which has one price, a commodity has a different price for each delivery month. Gas for next month, next winter, and next year all trade at different prices simultaneously. The forward curve captures all of them at once:

The shape of the curve carries information, and two terms describe it:

• Contango, the curve slopes upward: further-dated delivery is more expensive than near-dated. Often reflects storage and carrying costs.
• Backwardation, the curve slopes downward: near-dated delivery is more expensive, typically signaling near-term scarcity or strong prompt demand.

The same commodity can move between the two as supply and demand shift, and the shape directly affects the economics of storage, hedging, and roll strategies.

Curves are constructed from the most liquid market instruments available, exchange futures, over-the-counter forwards, and swaps, then interpolated to produce a price for every delivery period, including those that do not trade directly. For power and gas, curves also capture seasonality (winter gas above summer) and granularity down to monthly, daily, or even hourly shapes.

The mechanics are simple in outline: anchor at the liquid points, interpolate between them, then apply a shape profile that preserves the average:

pythondef build_curve(anchors, buckets, shape):
    """anchors: {date: price} liquid quotes; buckets: fine delivery periods;
    shape: relative profile per bucket that averages to 1.0 within each anchor."""
    coarse = interpolate(anchors, buckets)        # smooth between liquid points
    return {b: coarse[b] * shape[b] for b in buckets}  # shape to granularity

def interpolate(anchors, buckets):
    xs = sorted(anchors)                          # known delivery dates
    return {b: _linear(b, xs, anchors) for b in buckets}

Because curves change constantly, they are managed as time-versioned reference data, modeled as slowly-changing dimensions in the analytical marts so any valuation can be reproduced as-of a chosen date.

Every mark-to-market and every risk number depends on the curve. To value a forward position you compare its trade price to the current curve; to compute risk you shock the curve and see how the book responds. A small error in the curve propagates into every downstream number, which is why curve quality and governance matter so much.

This is the direct link between forward curves and the real-time valuation and risk covered elsewhere in the Knowledge Center.