The Cobb-Douglas calculator above evaluates the production function Y = A Kα Lβ, the most widely used functional form in applied economics. It returns output, the marginal product of each input, the returns to scale implied by the exponents, and the output you would get if both inputs were scaled by a percentage you choose.
Arb Digital publishes it as a teaching tool. It is worth saying plainly at the outset: the Cobb-Douglas function is a model, not a measurement. It is a compact algebraic form that happens to fit aggregate data reasonably well and has convenient mathematical properties. It does not describe how any particular factory turns inputs into outputs, and no number it produces is an observation about the world. Our economic profit calculator and marginal cost calculator deal with the cost side of the same firm.
What This Cobb-Douglas Calculator Does
It computes output from four inputs — total factor productivity, capital, labour and the two output elasticities — and then derives the quantities that make the function useful for teaching. The marginal product of capital is the extra output from one more unit of capital with labour held constant, and it equals alpha times output divided by capital. The marginal product of labour is the mirror image. Both fall as more of the corresponding input is added, which is diminishing marginal returns and is a property of the algebra whenever the exponent is between zero and one.
The returns to scale figure is simply the sum of the two exponents. Below one, doubling both inputs less than doubles output; at exactly one, doubling both inputs doubles output exactly; above one, output more than doubles. The scaled-output field demonstrates this concretely rather than asserting it, which is far more convincing than the algebraic statement.
The bar panel holds labour fixed and steps capital upward, showing the resulting output at each level. That is the clearest possible display of diminishing returns: the bars keep rising but by ever smaller amounts, which is the single most important qualitative feature of the function.
How to Use It
- Enter capital and labour in whatever units your problem uses, keeping them consistent throughout.
- Set the two elasticities. The textbook default is 0.3 for capital and 0.7 for labour, giving constant returns to scale.
- Set total factor productivity, which is usually normalised to 1 in exercises and estimated as a residual in empirical work.
- Choose a scaling percentage and compare the scaled output with the original to see the returns to scale in numbers.
- Read the marginal products to see what one more unit of each input would add at the current levels.
The Formula and How It Is Calculated
The function is Y = A Kα Lβ. The exponents are elasticities: alpha is the percentage change in output caused by a one percent change in capital with labour held constant, and beta is the same for labour. That interpretation is exact rather than approximate, and it is the reason the form is so convenient — taking logarithms turns it into the linear equation ln Y = ln A + α ln K + β ln L, which can be estimated by ordinary least squares. Our logarithm calculator and exponent calculator handle the two operations involved.
Work the default through. A = 1, K = 100, L = 50, alpha = 0.3, beta = 0.7. Then Y = 1000.3 × 500.7. Taking logs, 0.3 × ln 100 = 0.3 × 4.60517 = 1.38155, and 0.7 × ln 50 = 0.7 × 3.91202 = 2.73842. The sum is 4.11997, and e raised to that is about 61.56. So output is roughly 61.56 units.
The marginal product of capital is αY/K = 0.3 × 61.56 ÷ 100 = 0.1847, and the marginal product of labour is βY/L = 0.7 × 61.56 ÷ 50 = 0.8618. Since alpha and beta sum to exactly one, doubling both inputs doubles output: with K = 200 and L = 100 the answer is 123.11, exactly twice as much. That is constant returns to scale, and it is a consequence of the exponents summing to one and nothing else.
The Assumptions Built Into the Form
Choosing this functional form commits you to several strong claims, and they are worth stating because they are usually left implicit. First, the elasticity of substitution between capital and labour is exactly one, always, at every input combination. That means a one percent rise in the relative price of labour produces exactly a one percent shift in the capital-to-labour ratio. It is a specific empirical claim, and it is the reason economists reach for the CES family when they want to test it rather than assume it.
Second, the factor shares are constant. Under competitive markets and profit maximisation, capital earns exactly the fraction alpha of output and labour exactly beta, whatever the levels of K and L. Historically this looked like a good approximation, which is much of why the form was adopted. The observed decline in the labour share across many economies since around 1980 is one of the reasons that assumption is now actively debated.
Third, output is zero if either input is zero. Both factors are strictly essential, with no possibility of producing anything with capital alone or labour alone. Fourth, the form is smooth and continuous everywhere, so it cannot represent lumpy investment, indivisible machines, or a minimum viable scale. Each of those assumptions is a modelling decision, and each is violated by some real production process.
Where Total Factor Productivity Actually Comes From
A is the term that carries the most weight and is the least directly observed. In empirical work it is not measured; it is the residual left over once the contributions of measured capital and measured labour have been accounted for. Robert Solow's 1957 calculation attributed the majority of US output growth per worker to this residual rather than to capital accumulation, which is why it is sometimes called the measure of our ignorance.
That is not a criticism so much as a description. The residual absorbs technological change, but also improvements in management, better matching of workers to jobs, institutional change, mismeasurement of capital quality, and any specification error in the functional form itself. Attributing all of it to technology is a step beyond what the arithmetic supports. OpenStax's chapter on labor productivity and economic growth in Principles of Macroeconomics sets out the aggregate production function with capital, labour and technology in an introductory treatment.
The empirical history is worth a mention because it is regularly misremembered. Charles Cobb and Paul Douglas published their regression in 1928 on US manufacturing data, and their estimated exponents were close to 0.25 for capital and 0.75 for labour. Jeff Biddle's article "Retrospectives: The Introduction of the Cobb-Douglas Regression" in the Journal of Economic Perspectives (2012) documents how the form was originally received, including the substantial criticism it drew at the time, and how it nonetheless became standard.
Reading the Marginal Products Correctly
The marginal product of capital is a derivative, which means it is the rate of change at the current point and not the actual gain from adding a discrete unit. Because the function is concave in each input, the true gain from one more whole unit is always slightly less than the marginal product suggests. At large input levels the difference is negligible; at small ones it is not, and the tool's working panel reports both figures so the gap is visible.
The ratio of the two marginal products is the marginal rate of technical substitution — how much capital you can give up for one more unit of labour while holding output constant. Under cost minimisation a firm sets that ratio equal to the ratio of input prices, which is where the familiar tangency condition comes from. Note that the ratio depends only on alpha, beta and the capital-to-labour ratio, not on A or on the absolute scale of production, which is a direct consequence of the constant elasticity of substitution.
A word of caution on interpretation. None of these quantities is a wage or a rental rate. They become factor prices only under competitive markets, profit maximisation and the assumption that firms are price takers. Where those conditions fail — market power in output markets, monopsony in labour markets, adjustment costs — the marginal product and the observed payment diverge, sometimes substantially. Our price elasticity calculator covers the demand-side elasticity concept, and our comparative advantage calculator works through the trade model that uses similar production reasoning.
Arb Digital builds free tools that state what a formula assumes rather than presenting its output as a fact.
Browse All Free Tools Talk To Our TeamCommon Mistakes to Avoid
- Assuming the exponents must sum to one — constant returns to scale is an assumption commonly imposed, not a property the function has automatically.
- Confusing returns to scale with diminishing returns — the first is about scaling both inputs together, the second about adding one while the other is held fixed. A function can have constant returns to scale and diminishing returns to each input at the same time.
- Treating A as measured technology — it is a residual that absorbs specification error and mismeasurement alongside anything genuinely technological.
- Reading marginal products as wages — that equivalence holds only under competitive markets and profit maximisation, which are assumptions rather than observations.
- Mixing units between A and the inputs — A is scale-dependent, so an A estimated for annual hours means nothing when the labour input is headcount.
Related Free Tools From Arb Digital
Work out the cost of one more unit with the marginal cost calculator, take account of implicit costs with the economic profit calculator, measure demand responsiveness with the price elasticity calculator, value a forgone alternative with the opportunity cost calculator, or evaluate a power with the exponent calculator. The full free online tools hub lists every economics tool we publish.
Frequently Asked Questions
Output equals total factor productivity multiplied by capital raised to one exponent and labour raised to another. The two exponents are the output elasticities of the respective inputs.
Each is an elasticity: the percentage change in output caused by a one percent change in that input while the other is held constant. The interpretation is exact rather than approximate.
When the two exponents sum to exactly one. Below one it has decreasing returns to scale and above one increasing returns, and the sum is the only thing that determines which.
Diminishing returns means each extra unit of one input adds less output while the other is fixed. Returns to scale describes what happens when both inputs are scaled together. A function can have both at once.
The scale term A. In empirical work it is not measured directly but calculated as the residual after the contributions of capital and labour are accounted for, so it absorbs measurement error as well as technology.
Only under competitive markets, profit maximisation and price-taking firms. Where market power, monopsony or adjustment costs exist, the marginal product and the actual payment can differ substantially.
Output is zero. The multiplicative form makes both inputs strictly essential, which is one of its more restrictive assumptions and rules out producing anything with only one factor.
It is a model chosen for its tractability, not a measurement of any real process. It imposes an elasticity of substitution of exactly one and constant factor shares, both of which are testable claims rather than facts.
This page explains an economic model for educational purposes. The Cobb-Douglas function is a simplified representation rather than a measurement of any real production process, and nothing here is business, investment or policy advice.
