Probably a daft question but I derived an equation for a demand curve from a general Cobb-Douglas utility function $$U(x,y)=\beta x^{\alpha}y^{1-\alpha}$$ given a budget constraint $$M=xP_x+yP_y$$ and found that the quantity of $x$ demanded would be $$x=\frac{\alpha M}{P_x}$$.

which surprised me in that this function is independent of the price of good $y$. I see some references on these pages which suggests that I didn't make any spectacular errors but, my elementary understanding is that the demand curve is a functions of many thing, not least the price of other related goods. Now surely goods $x$ and $y$ are substitutes to some extent?

Now, my question is this, is my understanding of this essentially correct and, what sort of utility functions would generate demand curves $x=f(P_x, P_y, M, etc)$?

Thank you


1 Answer 1


If you take the general class of CES utility functions, of which Cobb-Douglas is a special case, you do indeed get a demand function that depends on other prices. Specifically, the CES utility function (over $n$ goods, $x_1,\dots,x_n$) takes the form \begin{equation} u(x_1,\dots,x_n)=\bigl[\alpha_1x_1^\rho+\cdots+\alpha_nx_n^\rho\bigr]^{1/\rho}, \end{equation} where $\rho\in(-\infty,1]\setminus\{0\}$, $\alpha_i\in[0,1]$ and $\sum_i\alpha_i=1$. We interpret $\alpha_i$ as the consumption share of good $i$ and $\sigma\equiv\frac{1}{1-\rho}$ as the constant elasticity of substitution. Note also that when $\sigma=1$ (or $\rho\to0$), we get the Cobb-Douglas utility form.

Solving utility maximization subject to the usual budget constraint, we get the demand for good $i$ as \begin{equation} x_i(p_1,\dots,p_n,M)=\frac{M(\alpha_i/p_i)^\sigma}{\sum_{j=1}^n\alpha_j^\sigma p_j^{1-\sigma}},\quad i=1,\dots,n. \end{equation} Again, observe that when $\sigma=1$ we get the demand associated with Cobb-Douglas utility.

The elasticity of substitution governs how relative expenditures on different goods change as relative prices change. Take a two-good example. An increase in the relative price $p_1/p_2$, i.e. good 1 becoming relatively more expensive, causes two effects simultaneously:

  1. per unit expenditure on good 1 rises, as good 1 now costs more in relative terms; and
  2. quantity demanded for good 1 decreases due to law of demand.

These are opposing effects on the expenditure on good 1 relative to that on good 2. It turns out that the elasticity of substitution determines which effect dominates. If $\sigma>1$, the second effect dominates, and if $\sigma<1$, the first effect dominates. When $\sigma=1$, which is the case of Cobb-Douglas, the two effects cancel each other exactly, so that relative expenditure is independent of the relative prices, and depends only on preference parameters (the $\alpha_i$'s).

Of course, CES utility is not the only class of utility functions that generate demands dependent on prices of other goods. Another common form of utility function, the quasi-linear utility function, \begin{equation} u(x_1,\dots,x_n)=x_1+v(x_2,\dots,x_n), \end{equation} where $v(\cdot)$ is strictly increasing and strictly concave, also generates demand functions that depend on other prices. A common example is $u(x_1,x_2)=x_1+2\sqrt{x_2}$. I trust that you can verify that both $x_1$ and $x_2$ depend on the two prices $p_1$ and $p_2$.

  • $\begingroup$ Thank you very much, I'll see if I can do the two case. $\endgroup$ Mar 31, 2020 at 7:28
  • $\begingroup$ If this adequately answers your question, please consider accepting the answer @JonathanAndrews $\endgroup$
    – Brennan
    Mar 31, 2020 at 20:15
  • $\begingroup$ Gosh, I tried the two good case. Horrible, when you try substitution but pretty straight forward using Lagrangians. Thanks again for your help. $\endgroup$ Apr 2, 2020 at 10:57
  • $\begingroup$ @JonathanAndrews: Check out cameronharwick.com/blog/… $\endgroup$
    – Herr K.
    Apr 2, 2020 at 19:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.