In Microeconomic Theory by Mas-Colell, Whinston, and Green, the definition of monotone preference relations is given as follows:

Definition 3.B.2$\quad$ The preference relation $\succsim$ on $X$ is monotone if $x \in X$ and $y \gg x$ implies $y \succ x$.

So suppose that $\succsim$ is monotone. I was wondering if it is possible for two bundles $x$ and $y$ to be such that $y \geq x$ but $x \succ y$?

I'm not sure about this, but I think that the above statement is not possible; that is, I think:

If $\succsim$ is monotone, then $y \succsim x$ whenever $y \geq x$.

However, I couldn't prove it. Could someone please help me out? Thanks a lot in advance!

Note that $x \geq y$ means $x_n \geq y_n$ for all $n=1,\dots,N$ and that $x \gg y$ means $x_n > y_n$ for all $n=1,\dots,N$.

  • $\begingroup$ I think this works. Why can't my preferences be defined by a preference for less? That is, $U(x) > U(y)$ wherever $x<y$ such that $x\succ y$ whenever $y > x$. I was never under the impression that it was necessarily an increasing relation. Preference relations can be monotonically increasing or decreasing. Not sure how to go about directly proving this though, I would just suppose the preference relation I defined and then prove it by contradiction. $\endgroup$
    – Brennan
    Commented Apr 6 at 1:31
  • $\begingroup$ @Brennan But that's bizarre. If you take a look at my first question in this post "economics.stackexchange.com/questions/58053/… =", it should indicate that "$y \succsim x$ whenever $y \geq x$". $\endgroup$
    – Beerus
    Commented Apr 6 at 2:51
  • $\begingroup$ I don't think it's that bizarre. My preference relation is monotonically increasing in less suffering. Or it's monotonically decreasing in suffering. I dont see how monotonicity is violated if I just say I prefer less of something in a monotonic way. $\endgroup$
    – Brennan
    Commented Apr 6 at 16:00

1 Answer 1


Throughout my answer, I assume that $x \succ y$ is read as "$x$ is strictly preferred to $y$". Refer to footnote (a) if that is not the case.

I was wondering if it is possible for two bundles $x$ and $y$ to be such that $y \geq x$ but $x \succ y$?

  1. Counterexample: $X = \{(1,1), (2,2), (3,1)\}$ such that $(2,2) \succ (1,1) \succ (3,1)$. Completeness, transitivity and monotonicity are satisfied. Now pick $x = (3,1)$ and $y = (1,1)$.

  2. With the assumption that $X \cong \mathbb{R}^m \ (m \in \mathbb{N})$ and continuity of $\succeq_X$, we have $y \geq x \implies y \succeq x$.

    Proof. Suppose $y \geq x$ for some $x,y \in X$ and assume for the sake of contradiction that $x \succ y$. Define $Z := \{z \in X : z >> y\}$. Consider any two open balls $B_x$ and $B_y$ and observe that $z_y \in B_y \cap Z$ must satisfy $x \succ z_y$ (due to continuity and our assumption that $x \succ y$) and $z_y \succ x$ (due to monotonicity) simultaneously. Contradiction!

(a) If $x \succ y$ is read as "$x$ is weakly preferred to $y$", consider $U(X) = 1$ and see that $(y \geq x) \land (x \succ y)$ holds for all $x,y \in X$ satisfying $y \geq x$.

(b) Definition of continuity of preferences: A preference relation $\succeq$ on $X$ is continuous if whenever $a \succ b$, $\exists$ open balls $B_a$ and $B_b$ such that for all $x \in B_a$ and $y \in B_b$, $x \succ y$.

(c) Note: I would appreciate an answer that has a weaker condition on $X$ compared to $X$ being isomorphic to $\mathbb{R}^m$ $(m \in \mathbb{N})$. Having said that, the above proof works for $X = \mathbb{R}^m_{\geq 0}$ $(m \in \mathbb{N})$ which makes more sense from the economic point.

  • $\begingroup$ Correctme if I'm wrong. For your definition of continuity, I feel it should be: $a \succ b$ whenever there are open balls $B_a$ and $B_b$ such that $x \succ y$ for all $x \in B_a$ and $y \in B_b$. $\endgroup$
    – Beerus
    Commented Apr 6 at 14:51
  • $\begingroup$ @Beerus No, your definition is incorrect. Counterexample: Pick $X = \mathbb{R}^2, a = (2,2), b = (1,1)$. For $\epsilon = 0.1$, there are open balls $B(a,\epsilon)$ and $B(b,\epsilon)$ such that $x \succ y$ for all $x \in B(a,\epsilon)$ and $y \in B(y,\epsilon)$, and also $a \succ b$. However, lexicographic preferences not continuous. I took the definition from Ariel Rubinstein's Lecture Notes in Micro Theory, Lecture 2 (Utility), Page 16, Definition C1. $\endgroup$ Commented Apr 6 at 16:58
  • $\begingroup$ Oh, I see. Yes you are right! $\endgroup$
    – Beerus
    Commented Apr 6 at 17:34

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.