Isoquants: The Uneconomic Region of Production

That weird backward-bending chunk of the isoquant has a name — it's the uneconomic region, and here's exactly why it exists and why no sane firm would ever go there.

Last post,

Equation from the original notes

I drew this thing

and said, OK, let’s pop our eyeballs above it and look straight down~~~

Equation from the original notes

When you look at it from up there,

Equation from the original notes

I said it ends up looking like this.

And right here, you’re gonna squint at me.

Huh??????????????

Not that — shouldn’t it be

Equation from the original notes

THIS?!?!?!?!?!?!??!?!?

Yes yes yes yes yes — that’s correct.

That one over there is the right one @@@@

OK so, instead of putting two of them down, let me work with just one line, and slap names on a few specific points so I can actually talk about what’s going on.

Equation from the original notes

I’ll explain with this picture.

Among the isoquants here, look at the leftmost extreme point of the red one —

Equation from the original notes

this guy.

Right now, our firm is using $L_1$ worth of L and $K_1$ worth of K, and cranking out output $Q = \alpha$.

From here, let’s bump K up past $K_1$, all the way to $K_3$.

With that $K_3$, how much L do we need to throw in to still produce the same $Q = \alpha$ as before?

The answer is: we have to use $L_3$, which is more L than the original $L_1$, just to crank out the same $\alpha$ we had before.

Wait wait wait. We just increased K, and yet to make the same amount of output we also have to increase L….??

On the flip side,

let’s look at L going from $L_1$ up to $L_2$.

When L goes up, K can come down a notch and we still get the same $\alpha$ — right??????????????

But that blue chunk is saying ’nope, you have to crank up both of them just to break even,’ isn’t it?????????????

That’s insanely inefficient, right????? Who in the world would make a choice like that???

Ah @@@@ you say inefficient??@?@?@

Why does this inefficient stretch even exist in the first place. We gotta go back and look at how the isoquant was born.

Equation from the original notes

We pulled the isoquant out of this kind of coordinate space, remember?!?!?

Hold up @@@@

That light green stretch???!?!? What was that thing again??@?!?!?!

Yes yes yes yes — that was the part where humans get ’too’ numerous, which causes a negative efficiency change, and that

cancels out the efficiency boost from adding more humans, so $\Delta Q$ ends up negative —

Equation from the original notes

that’s what we figured out with this picture @@@@

Ah, but —

when $Q = \alpha$, we grabbed all the points that share the same Q value and dragged them out as a line,

and that line is the isoquant,

Equation from the original notes

since we yanked it out like this,

Ohhh, I see~~~~

Equation from the original notes

that weird “inefficient stretch” showing up in blue here

is the same thing as the light green chunk in the picture up above,

Equation from the original notes

so it ends up looking like this @@@@@@

OK so, that blue chunk actually has a name — it’s called the “uneconomic region of production.”

What kind of CEO is gonna pick an L, K combo from the uneconomic region.

Simple version:

Equation from the original notes

past a certain amount of L,

if you throw even more labor in, output actually drops,

and yet — is there any CEO out there who’d hire even more L??????????????

Raise your hand once heh heh heh heh

Yeah, nobody heh heh heh heh heh

There’s nobody who’d voluntarily go do something inefficient.

So,

Equation from the original notes

that blue-colored chunk is called the “uneconomic region of production” :D

So~~~

when we draw an isoquant,

we just erase the uneconomic region of production and only draw the economic region.

Because nobody’s ever gonna pick that area,

and that stretch isn’t something economists care about anyway @@@@@@@

So they say you just need to draw it like this~~

Equation from the original notes

Yay hoo hoo

Comments

Discussion happens via GitHub Discussions. You'll need a GitHub account to comment.