1
$\begingroup$

I was looking into convexity while reading Varian's intermediate micro textbook, and an article by Richter and Rubinstein said, "The canonical definition of convex preferences requires that if a is preferred to b, then any convex combination of a and b is also preferred to b." (I interpret "preferred" in the strong sense in this statement, although that might be wrong.) I looked at Kreps' textbook, and he defined it slightly differently: a preference relation is convex if for all $x$ and $y$, if $x\succeq y$ and $z$ is a convex combination of $x$ and $y$, then $z\succeq y$. I am trying to prove that Kreps' definition implies Richter and Rubinstein's. (As I think about this, though, I'm not sure whether this is true given my understanding of the latter.)

For the forwards direction, let $x\succ y$ and $z$ be a convex combination of $x$ and $y$. Because $x\succeq y$ and $\succeq$ is convex, $z\succeq y$. Now, I need to prove that the statement $y\succeq z$ is false, which is where I am stuck. I tried a proof by contradiction by trying to get $x$ as a convex combination of $y$ and $z$, but that didn't get me anywhere.

$\endgroup$

1 Answer 1

1
$\begingroup$

Let $X$ be a convex subset of $\mathbb{R}^n$. We define

C1: $\forall x, y, z\in X$, if $x\succeq y$ and $z$ is a convex combination of $x$ and $y$, then $z\succeq y$.

C2: $\forall x, y, z\in X$, if $x\succ y$ and $z\neq y$ is a convex combination of $x$ and $y$, then $z\succ y$.

Proposition. C1 does not imply C2.

Proof. Consider $X=[0,2)$, define $\succeq$ on $X$ as follows $x\succeq y$ if and only if $\lfloor x\rfloor\geq \lfloor y\rfloor$. In other words, $\succeq$ is represented by the utility function $u:[0,2)\rightarrow\mathbb{R}$ defined as $u(x)=\lfloor x\rfloor = \begin{cases} 0 & \text{if } x \in [0,1) \\ 1 & \text{if } x \in [1,2) \end{cases}$. Clearly, $\succeq$ satisfy C1 but not C2. To see that it does not satisfy C2, consider $x=1.5$ and $y=0$ such that $x\succ y$, now $z=0.5$ is a convex combination of $x$ and $y$, but it is not the case that $z \succ y$. Therefore, C2 is not satisfied.

$\endgroup$
1
  • $\begingroup$ Thank you so much! $\endgroup$
    – RHyp
    Commented Aug 22 at 21:30

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.