I know it doesn't actually matter which axis is which, but it seems less intuitive me. When I see a supply and/or demand graph, I always have to flip it in my head to feel like I really get what's going on.
It just seems like if you have one variable (price) on which two other variables (supply and demand) depend, then the single variable should be on the horizontal axis. So why isn't this the convention?
This objection never made too much sense to me. In the standard model of perfect competition, firms take the price as given and respond by choosing their quantity. So you have a model in which a bunch of actors choose quantity and the market price emerges as a consequence of all of those decisions. This makes it sound awfully like price is the "dependent" variable, which by convention is always placed on the vertical access.
Indeed, this seems to be how Alfred Marshall (who originated the modern form of the Demand-Supply diagram) thought about things. Here's a quote from An Introduction to Postitive Economics, Seventh ed. by Richard G. Lipsey (as quoted here):
"Readers trained in other disciplines often wonder why economists plot demand curves with price on the vertical axis. The normal convention is to put the independent variable on the X axis and the dependent variable on the Y axis. This convention calls for price to be plotted on the horizontal axis and quantity on the vertical axis.
"The axis reversal - now enshrined by nearly a century of usage - arose as follows. The analysis of the competitive market that we use today stems from Leon Walras, in whose theory quantity was the dependent variable. Graphical analysis in economics, however, was popularized by Alfred Marshall, in whose theory price was the dependent variable. Economists continue to use Walras' theory and Marshall's graphical representation and thus draw the diagram with the independent and dependent variables reversed - to the everlasting confusion of readers trained in other disciplines. In virtually every other graph in economics the axes are labelled conventionally, with the dependent variable on the vertical axis."
See also this post on Greg Mankiw's blog.
As @KennyLJ points out I think the reason is partly historical. However it is a useful convention for several reasons.
In most charts that show a relationship the input goes at the bottom.
In this case p = f(q)
The market is what quickly reacts to q (supply). If there is an interruption to supply price will quickly go up. The clearing price is a function of q.
q = f(p) is also valid as if the price is high more producers will enter the market but that is not an immediate effect. If the price is low and stays low some produces will exit the market but again not an immediate effect.
But if you switch x and y it still works if that is what makes sense to you.
A convention is the only real justification.
It does not match the current ways of teaching economics as being the quantity supplied and demanded at given price points.
But also, at no point would it ever have been sensible to consider price the dependant variable (y-axis) and quantity the independant variable (x-axis) like the conventional supply-demand graph would falsely suggest.
Based on what the justifications I've seen, it seems like a misunderstanding of what makes a variable (in)dependant.
People argue that price is dependant because the supply and demand curves are used to find the market value, but this is a resulting intersection of two curves.
The dependant and independant variable show the relations between the variables in a single curve:
Surely, the 1st is sensible whereas the 2nd is not.
The same applies to supply.
It is only when we plot both supply and demand and find their intersection that we find market quantity and market price.
But some still seem to think that curves shifting are always a change in quantity which results in a changed price at the new point of intersection compared to the old point of intersection, and therefore wrongly consider quantity the independant value and price the dependant value.
Curves can shift for all sorts of reasons, but as the curves are often fairly straight lines, it can sometimes be difficult to determine in which direction it shifts:
both look the same on a straight downwards line, and it is easy to wrongly assume that it is all changes in quantity in that case.
So let's consider a non-linear curve.
For example the demand for food in general.
People need a certain amount of food, and as long as their basic needs aren't met, they'll be very willing to spend their money on food (and it will be one of the first things they'll spend their money on).
But once their basic needs are met, they don't really demand much more (maybe a little more, and maybe shift towards more luxurious food products, but the total amount of food demanded should be fairly inelastic at this point).
This should result in a graph which is nearly perpendicular to each of the axes with a bit of a transition in between:
When prices are too high people simply don't have the money to pay for it, but if the price drops even a little, people will want to buy as much as they can, up to a point where their needs for food are satisfied, at which point the demanded quantity stagnates and people won't buy much more at further decreases in price.
Now let's consider two situations:
This means the quantity needed to meet the basic food needs of all people increases.
We should see a shift along the quantity axis (the intersection of the curve with the quantity axis is at a higher quantity).
But people's available funds are unchanged so we wouldn't see a significant shift along the price axis (the intersection point between the curve and the price axis remains nearly unchanged).
As food is a basic need, people will spend their money on food as one of the first things if their needs aren't met.
This should result in a shift of the curve along the price axis (the intersection between the curve and the price axis is at a higher price).
But people don't suddenly buy a significant additional amount of food if their needs are met, so we shouldn't see any significant shift of the curve along the quantity axis (the intersection between the curve and the quantity axis remains nearly unchanged).
With these examples, we can more clearly see that the curve can shift along either the quantity or price axis, so it would be incorrect to think that only changes in quantity collectively demanded at a given price are the cause of shifting curves and the resulting changes in the intersection between supply and demand curves (which changes both the market price and market quantity).
