Short Run, Long Run: Why the Same Firm Acts Completely Differently Over Time

In the summer of 2021, Taiwan Semiconductor Manufacturing Company announced it would spend $100 billion over three years to expand capacity.¹ The chip shortage driving that announcement had started more than a year earlier. Factories had been running flat out, squeezing every extra wafer from existing equipment, but the decision to actually build more capacity came only after management judged that demand would hold long enough to justify the investment. That gap between "we are at full tilt" and "we are breaking ground" is not delay or indecision. It is the short run behaving exactly as economic theory says it should.

The distinction between the short run and the long run is one of those concepts that looks obvious until you need to apply it, and then suddenly feels slippery. Let me walk through what it actually means.

The question that defines the horizon

The short run and long run are not calendar periods. They are defined by a single question: which of the firm's inputs can actually be changed right now, and which is it stuck with? In the short run, at least one input is fixed, and that fixed input is almost always capital: the physical plant, the machinery, the fabrication lines. In the long run, every input is variable. The firm can build, sell, expand, or exit any part of its operation.

What makes this non-obvious is that the calendar length of the "short run" changes completely depending on the industry. The Federal Reserve tracks this constraint directly through its industrial production and capacity utilization data, which measures how close manufacturers are running to the ceiling set by their existing physical plant.¹ When that ceiling is a taco truck, the "short run" might be a week: buy a second propane tank, hire a helper, and the constraint dissolves. When that ceiling is a chip fabrication facility, the short run is years and the price tag starts in the billions. The constraint is the same in structure; the time scale is wildly different.

Capacity utilization data shows that U.S. manufacturers routinely run at 75 to 80 percent of their installed capacity.² What that number is really capturing is the gap between the short-run ceiling and what firms would produce if every input were fully variable. In the short run, you operate within that ceiling. In the long run, you raise it.

What both horizons share

Before getting to where they differ, it helps to be clear about what stays the same. In both the short run and the long run, the firm is trying to do the same thing: produce its target output at the lowest achievable cost. The underlying production function, the technical relationship between inputs and output, does not change between horizons. What changes is the menu of adjustments available. The Library of Economics and Liberty describes this cost-minimization goal as central to how competition actually allocates resources across an economy.³

So when a firm behaves differently across horizons, it is not because its goals changed or because management got smarter. It is because the binding constraints changed.

Where the two horizons genuinely differ

How a firm meets rising demand

In the short run, with plant and equipment fixed, the only lever a firm has is labor. Hire more workers, add shifts, push overtime. That works up to a point, but piling labor onto a fixed amount of capital runs straight into a fundamental production principle: the law of diminishing marginal returns. Each additional worker has less capital to work with than the last, so each adds a smaller increment to output. You can push production up, but at rising cost per unit. The EconLib treatment of production costs traces this constraint through the cost curves that underlie every firm's short-run supply decision.⁴ Productivity per worker, in other words, falls as the fixed input gets more crowded, a dynamic that the Library of Economics and Liberty covers as foundational to how output and efficiency interact.⁶

In the long run, that wall is not there. The firm can expand capital and labor together, building a second production line or an entirely new facility, and the diminishing-returns squeeze dissolves. That is exactly what TSMC's $100 billion commitment was: a long-run response to a demand signal that had become durable enough to justify moving the ceiling rather than just pressing against it.

What the cost structure looks like

Short-run costs split into two categories. Fixed costs are the lease, the loan payment on equipment, costs that do not change with output and cannot be escaped quickly: they are owed whether the firm produces one unit or a thousand. Variable costs, labor and raw materials mostly, rise and fall with output. In the long run, there are no fixed costs, because every commitment can eventually be unwound. The lease expires, the equipment is sold, the loan is paid off. A firm that feels trapped by its fixed costs right now can be completely free of them given enough time.

The shutdown rule versus the exit rule

This is where the theory produces a result that surprises people the first time they see it, and it is worth taking slowly.

In the short run, a firm losing money should keep operating as long as its revenue covers its variable costs. The reasoning is direct: the fixed costs are owed regardless of whether the factory runs. Shutting down eliminates revenue and variable costs simultaneously, but the fixed bill does not go away. If the firm is at least covering variable costs, every dollar of that contribution is reducing the loss below what it would be at zero production.

In the long run, the logic flips. Because fixed costs become avoidable, the firm should exit if it cannot cover all of its costs: fixed and variable together. The question is no longer "can I reduce my losses by staying open" but "should I renew the lease, buy more equipment, stay in this business at all."

Shutdown and exit are separate decisions, governed by different arithmetic, made on different time horizons.

A worked example: the furniture workshop

Follow one firm through a slow quarter on both horizons. A furniture workshop has a fixed plant lease of $8,000 per month. At current output it brings in $30,000 in revenue against $25,000 in variable costs (wood, labor, finishing materials).

The workshop is losing $3,000 a month. Should it shut down immediately? In the short run, no. Closing would eliminate the $5,000 contribution and leave the $8,000 lease fully unmet, an $8,000 loss instead of $3,000. Running and losing the smaller amount is the rational move.

Now the lease comes up for renewal. The fixed cost is suddenly avoidable. The question is whether the business can cover total costs: $25,000 variable plus $8,000 fixed equals $33,000, against revenue of $30,000. The answer is no. Unless demand recovers or the firm relocates into a cheaper space, the long-run decision is to exit. Same business, same revenue, opposite conclusion, because the applicable horizon changed which costs are escapable.

Which lens to use and when

Let me be direct about the practical read here. Use the short-run lens whenever the question is how a firm or industry responds this quarter to a price change, a cost spike, or a demand swing. The answers run through labor markets, overtime, and the shutdown calculation against fixed costs that cannot be escaped. Use the long-run lens for questions about entry, exit, industry capacity, and where supply is heading over the next several years. The answers run through investment decisions and the freedom to vary everything.

There is a reason industries can look stable for years and then see a wave of closures or expansions all at once. Firms accumulate losses or profits within their short runs, quietly, adjusting labor at the margin. When fixed commitments come up for renewal in the same economic environment, the long-run exit or expansion decision gets made by many firms simultaneously. The wave is not a coincidence. It is the long run arriving at the same time for a cohort that started under the same conditions.

Output data reflects the sum of all these horizon-bound decisions. The Bureau of Economic Analysis measures the quarterly swings driven by short-run adjustments alongside the slower structural shifts in what the economy can actually produce.⁵ Underneath both is a set of firms, each operating on a clock set not by the calendar but by how long their most stubborn input takes to change.

When you see a company keep a money-losing plant open while announcing it will not build another one, you are watching both horizons operate at once. That is not confusion. That is exactly right.