returns to scale

Suppose you scale up everything at once — double the workers, double the machines, double the floor space — not just one input but all of them together. What happens to output? Does it double too, more than double, or less than double? Returns to scale is the answer to exactly that question, and it is a purely long-run idea, because only in the long run can every input be changed.

There are three cases. Constant returns to scale: double all inputs and output exactly doubles (a tidy benchmark — clone the whole factory and you get two factories' worth). Increasing returns to scale: double the inputs and output more than doubles, perhaps from specialisation or bigger, better machines — this is the engine behind economies of scale, where average cost falls. Decreasing returns to scale: double the inputs and output less than doubles, usually from the management strain of size — this sits behind diseconomies of scale, where average cost rises. A firm may show different returns at different sizes: increasing while small, constant in a middle stretch, decreasing when it gets too big.

Returns to scale is the production-side mirror of the cost-side economies-of-scale story: the two describe the same phenomenon from opposite ends. Increasing returns means each unit needs proportionally fewer inputs, so it costs less — that is an economy of scale. The concept matters because it shapes whole industries: where increasing returns are strong (software, networks, utilities) markets tend toward a few large firms; where returns are roughly constant (farming, restaurants) firms of many sizes coexist. The key contrast to keep straight: returns to scale changes all inputs together, whereas diminishing marginal returns changes just one while holding the rest fixed.