When you phrase a problem in terms of variables x and y and not real world quantities it allows you to solve it objectively without bringing in preconceived notions about the variables. For this same reason it is also helpful to look analogies. So here is one:
Lets say there is a pond into which water is poured from x different sources at various rates and leaves from y different sources at various rates. How does the water level change over time? To solve the problem you need to know the rates at which water enters and leaves the pond at different places relative to one another and the initial water level. Maybe valve two shuts off when valve one opens etc.
Now instead of a pond, lets say its a village. And instead of water, lets say its money coming in as loans, salaries and business revenues and leaving as loan repayments and purchases. How does the spending power of the village change over time? When you structure the problem in this way, you can play with these rates and quantities and see what it does.
You might argue that this kind of approach obfuscates the local changes: what happened to the specific person who took *your* money? To this I say that if we care about real change, what matters is the system as a whole and not local changes. Local changes can fool you. It could well be that one person's gain is another's loss. Or consider this, in a wave, each water molecule only moves a little so if you only knew how individual molecules were moving you could well miss the tsunami.
Of course, the pond analogy leaves out one very important thing: innovation. Innovation can be a complete game changer. But more about that in another post...