Expenditure Minimisation Problem: Non-Convex Preferences

Consider the utility function u(x, y) = \sqrt{x^2+y^2}.
For a target utility \mu \geq 0, find the Hicksian demand correspondence (X^h,Y^h)(p_x, p_y, \mu), and the expenditure function e(p_x, p_y, \mu).

1 Like

The Hicksian problem is:

\min_{x,y \in \mathbb{R}^2_+}\; p_xx + p_yy \quad \text{s.t.} \quad \sqrt{x^2 + y^2} \ge \mu

The slope of the budget line is \frac{P_x}{P_y}

Expenditure is minimized by buying only the cheaper good. Therefore,

(X^h, Y^h)(p_x, p_y, \mu) = \begin{cases} (\mu, 0), & \frac{P_x}{P_y} < 1 \\ (0, \mu), & \frac{P_x}{P_y} > 1 \\ \{(\mu, 0), (0, \mu)\}, & \frac{P_x}{P_y} = 1 \end{cases}

And the expenditure function is

e(p_x, p_y, \mu) = \mu min\{p_x, p_y\}
1 Like
\min_{x, y} \quad p_x x + p_y y \quad

subject to:

\sqrt{x^2 + y^2} \geq \mu

now :

\sqrt{x^2 + y^2} = \mu
{x^2 + y^2} = \mu^2

As this is an equation of circle

We are taking MRS

MU_x = \frac{x}{\sqrt{x^2 + y^2}}, \quad MU_y = \frac{y}{\sqrt{x^2 + y^2}}

Hence,

MRS = \frac{MU_y}{MU_x} = \frac{y}{x}

At an interior optimum,

\frac{y}{x} = \frac{p_y}{p_x}

So,

x = \left(\frac{p_y}{p_x}\right) y

Substituting into the utility constraint equation:

x^2 + y^2 = \mu^2

Therefore

\left( \frac{p_y}{p_x} y \right)^2 + y^2 = \mu^2
y^2 \left( \frac{p_y^2}{p_x^2} + 1 \right) = \mu^2
y^2 = \frac{\mu^2 p_y^2}{p_x^2 + p_y^2}

Therefore,

\boxed{y = \frac{\mu p_y}{\sqrt{p_x^2 + p_y^2}}}

Similarly,

\boxed{x = \frac{\mu p_x}{\sqrt{p_x^2 + p_y^2}}}

Checking Corner Points
Corner 1

(x, y) = (\mu, 0) \implies U = \sqrt{\mu^2} = \mu
E = \mu p_x

Corner 2

(x, y) = (0,\mu) \implies U = \sqrt{\mu^2} = \mu
E = \mu p_y

So by this
Hicksian Demand Function :-

\boxed{ (X^h, Y^h)(p_x, p_y, \mu) = \left\{ \begin{array}{ll} \{(\mu, 0)\} & ,\text{if } p_x < p_y \\ \{(0, \mu)\} & ,\text{if } p_x > p_y \\ \{(\mu, 0), (0, \mu)\} & ,\text{if } p_x = p_y \end{array} \right. }

Expenditure Function
Since

\sqrt{p_x^2 + p_y^2} > \min\{p_x, p_y\}

the interior solution is always more expensive than buying only the cheaper good.

Therefore ,

\boxed{e(p_x, p_y, \mu) = \mu \min\{p_x, p_y\}}
1 Like