A cost elasticity describes proportional scaling, not dollars per extra unit

A financial model may assume that a 1% activity increase produces approximately a 0.7% cost increase. That is an elasticity assumption. It does not mean the next activity unit costs 70 cents, nor does it mean 70% of total cost is variable under every operating condition.

One way to represent a constant elasticity is a power relationship: cost equals a scale parameter multiplied by activity raised to an exponent. NIST describes the mathematical form and its logarithmic transformation in its modeling guidance. Whether that form fits a business's costs must be established separately from the algebra.

Use the model when a proportional scaling relationship has evidence over the relevant activity range. It should compete with alternatives that represent fixed commitments, linear variable cost or capacity steps, rather than being selected because its curve looks smooth.

Assume an invented cost model is $10,000 multiplied by activity divided by 1,000, raised to the power 0.7. At 1,000 units, the activity ratio is one and modeled cost is $10,000. The exponent controls how cost changes as activity moves away from that point.

At 1,100 units, cost is $10,000 times 1.1 raised to 0.7, or approximately $10,689.93. Activity rises 10%, while modeled cost rises about 6.90%. The small-change approximation of 0.7 times 10%, or 7%, is close but not exact.

At 2,000 units, cost is approximately $16,245.05, an increase of 62.45%. Multiplying the 100% activity increase by 0.7 would predict a 70% cost increase and would overstate the exact power-model result. Use the formula for material changes rather than extending the local approximation indefinitely.

In the example, average cost falls as activity rises because the exponent is below one. At the 1,000-unit operating point, the local slope is 0.7 times $10,000 divided by 1,000, or $7 per additional unit for a very small activity change.

That local slope differs from the $10 average cost. It also differs from the average incremental cost of moving all the way from 1,000 to 2,000 units, which is $6,245.05 divided by 1,000, or about $6.25 per additional unit.

Keep the decision size visible. Pricing a small extra order, budgeting a major expansion and allocating historical total cost are different applications. A coefficient described only as unit cost can blur those differences and lead to the wrong amount being used.

Taking logarithms of a positive power relationship makes log cost linear in log activity. The activity coefficient is the exponent. For two hypothetical points of $10,000 at 1,000 units and $16,000 at 2,000 units, the connecting exponent is log 1.6 divided by log 2, approximately 0.6781.

Two points determine that connecting curve but do not establish a stable cost law. More observations are needed to examine variation, other drivers and changes over time. A pair chosen from different price regimes could attribute supplier inflation to volume scaling.

Specify the error model when fitting real data. A relationship estimated on logarithms has implications for the distribution on the original dollar scale. Converting a fitted log value back to dollars does not automatically produce the expected dollar cost without an appropriate treatment of the error distribution.

A business with fixed cost of $5,000 and variable cost of $5 per unit also costs $10,000 at 1,000 units. Its cost elasticity at that point is 0.5 because an extra unit adds $5 against an average cost of $10. At other activity levels, that elasticity changes.

A constant-elasticity power curve and a fixed-plus-variable model therefore describe different structures. The power curve can imply that all cost tends toward zero as activity approaches zero, while real rent, salaries or minimum supplier commitments may remain payable.

Use contractual and operating evidence when deciding which costs disappear after a volume reduction. A historical exponent is not permission to remove the same percentage of payroll or rent from a downside cash forecast. The time needed to change those commitments also matters.

An exponent below one can make average cost decline continuously as volume grows. A real business may eventually need another team, facility or system tier. That capacity step can interrupt or reverse the smooth scaling assumed by the power function.

Identify the range represented by the historical observations and the capacity available within it. Flag forecasts outside that range and obtain an operating estimate for the extra resources. A fitted curve does not establish that the organization can deliver arbitrary volume with progressively lower unit cost.

At the other end, logarithms cannot directly represent zero activity or nonpositive cost observations. Investigate whether such records indicate a shutdown, credit adjustment or data problem. Adding an arbitrary constant to force them through the calculation changes the model and its interpretation.

A declining average cost can reflect supplier discounts, simpler work, a changed product mix or spreading a standing resource over more activity. Those mechanisms have different prospects for continuing. Separate them where the data and decision justify it.

Use a consistent cost perimeter and activity unit across periods. If support work moved to another department, the apparent scale benefit in the original cost center may simply be a transfer. Reconcile the total business cost before treating the fitted exponent as an efficiency improvement.

Present the operating anchor, exponent, evidence range and implied costs at the proposed volumes. Include an alternative where capacity or commitments change the curve.

The decision maker should be able to ask why doubling activity is expected to raise cost by 62.45% in the example and which operating facts support that relationship. A useful elasticity model makes that proportional assumption visible while preserving the separate questions of capacity, avoidability and cash timing.