Additively Separable Utility and Marginal Utilities

Why Additive Separability Matters

A consumer’s utility may depend on several goods, but the effect of one good can sometimes be represented separately from the effect of another. This structure makes marginal utilities straightforward to calculate: the marginal utility of a good depends on that good’s quantity alone. The marginal rate of substitution, however, compares two marginal utilities, so it can still depend on both quantities.

This tutorial develops that distinction using differentiable utility functions. Its treatment is aligned with the SOA VEE Economics (Validation by Educational Experience) syllabus.

The additive form says that utility is the sum of one function for each good. It does not, by itself, guarantee that more of either good is preferred: that requires positive marginal utility on the domain being studied. Nor does the formula assert that goods are statistically independent. It is a statement about the form of the utility representation.

As with other utility representations, the numerical utility level is not generally a directly measurable amount of satisfaction. Moreover, applying an arbitrary increasing transformation to \(U\) can preserve preference rankings while changing whether the transformed function remains additively separable. Separability is therefore a property of the specified representation, not automatically a property preserved by every equivalent representation.

Marginal Utilities and the MRS

Marginal utility is the partial derivative of utility with respect to the quantity of one good, holding the other quantity fixed. For a differentiable, additively separable utility function, the partial derivatives simplify because each component depends on only one variable.

Thus, changing \(y\) does not directly change \(MU_x\), and changing \(x\) does not directly change \(MU_y\). If both marginal utilities are positive, the marginal rate of substitution \(MRS_{xy}\) is the amount of \(y\) the consumer is willing to give up locally for one additional unit of \(x\), while remaining on the same indifference curve.

This convention reports the positive magnitude of the trade-off. The slope of the indifference curve, when it is well-defined, is negative and equals \(-MRS_{xy}\). If \(MU_y=0\), the ratio is undefined; if a marginal utility is negative, the usual interpretation as a consumer’s willingness to give up one good for another may not apply.

Why the MRS Formula Holds

On an indifference curve, utility is held constant. For a small change in the bundle, the total differential of utility is zero. This gives a direct derivation of both the indifference-curve slope and the MRS formula.

For an additive function, substituting its partial derivatives produces \(MRS_{xy}=f'(x)/g'(y)\). The numerator depends on \(x\) and the denominator on \(y\). Consequently, separability simplifies each marginal utility but does not generally make the MRS constant.

Diminishing Marginal Utility

Marginal utility diminishes when an increase in a good’s quantity lowers that good’s marginal utility, with the other quantity held fixed. For an additive function, this condition can be checked separately for each component. If \(f\) and \(g\) are twice differentiable, their second derivatives describe how the corresponding marginal utilities change.

If \(f''(x)<0\) throughout an interval, \(MU_x\) is strictly decreasing there. If \(f''(x)\le 0\), \(MU_x\) is nonincreasing there. The same conditions apply to \(g''(y)\) and \(MU_y\). These statements concern changes in marginal utility as its own good changes. They do not, by themselves, say how the ratio of the two marginal utilities changes.

For a consumer who prefers more of each good throughout a region, one also checks \(f'(x)>0\) and \(g'(y)>0\) on that region. A function can have a negative second derivative while its first derivative eventually becomes negative; in that part of the domain, utility is not increasing in that good. The domain and the signs of the first derivatives therefore matter alongside curvature.

Diminishing MU Does Not Determine the MRS

Even when both marginal utilities diminish, their ratio can rise or fall depending on how quickly each changes. The following calculation shows why it is not enough to inspect the numerator alone. It also illustrates a frequent distinction in consumer theory: diminishing marginal utility is about one good’s marginal utility, while the MRS compares marginal utilities across goods.

Common Mistakes and Practical Checks

• Using utility instead of marginal utility. The MRS is a ratio of partial derivatives, not a ratio of utility levels.
• Dropping the negative sign from the slope. \(MRS_{xy}\) is conventionally the positive magnitude \(MU_x/MU_y\); the indifference-curve slope is \(-MRS_{xy}\) when the derivatives are positive.
• Assuming additive separability makes the MRS constant. Each marginal utility depends on only one quantity, but their ratio can vary with either quantity.
• Checking curvature without checking the domain. A negative second derivative indicates falling marginal utility, not necessarily positive marginal utility. Check the first derivative before interpreting more of a good as desirable.
• Confusing diminishing MU with diminishing MRS. The former checks how one marginal utility responds to its own good; the latter checks the behavior of a ratio. They are not equivalent conditions.