If you remember the book I'd appreciate a tag, because in my experience (not that I'm an expert or anything) this is more an outside looking in view. It's very possible that perfect competition is sometimes assumed where it shouldn't be, in fact I'm certain that's true, but the extent that it happens I suspect is very exaggerated (most research papers I read it's not even relevant). People talk a lot about how economists are influenced by their politics, and that's true, but exactly because economics is so relevant to politics and people's daily lives you end up with a lot of people having some pretty strong opinions about economics as an academic discipline without necessarily being aware of what's going on. For instance, no economic model or theory assumes that agents have perfect knowledge, intelligence or that they are unbiased. What's likely happening here is that economists famously assume that people are
rational. And people aren't rational, at least not all the time, so that sounds very suspect, but in typical academic fashion economic rationality doesn't mean what anyone else would call rationality. Being rational in an economic sense means that your preferences are complete and transitive, nothing more exiting than that. Meaning that if you're presented with two choices, you're able to tell me which one you prefer (or that they're equally good), and that if you prefer choice A over B, and B over C, then you also prefer A over C. And rationality isn't a set assumption either, necessarily, see behavioural economics but also in a lot of other settings that assumption is relaxed.
Have you heard of the
spherical cow? It's making fun of physicists's tendency to simplify. However, the issue with what the physicist is doing isn't that he's simplifying, it's that he's simplifying something that's not appropriate. When we are talking about the milk production of actual cows, then it's pretty relevant that cows aren't spherical, and neither do they exist in a vacuum. They'd die, actually. But, in other situations that simplification doesn't matter, it just makes it easier, so it's fine.
We talked briefly about minimum wage in the labour market a while ago, so I'll go with some of what I touched upon there. If we assume perfect competition, then a minimum wage set above equilibrium wage will lower employment. If we relax that assumption and instead allow for the possibility of employers having market power, then that's no longer necessarily true. Here perfect competition is a critical assumption, so it matters what the model says about it. But, if we want to talk about the effect of a payroll tax on wages, then it doesn't matter if we assume perfect competition or not, the model will say that wages fall. Perfect competition is not a critical assumption here, and relaxing that assumption will only give us the same answer with more equations and lines.
Another example is general and partial equilibrium, where partial equilibrium is looking at a single market and treating everything else as exogeneous, while with a general equilibrium model you take into account feedback loops etc. If we use a partial equilibrium model to look at how immigration impacts wages then the result is very simple: a positive shift in labour supply increases quantum (employment) and lowers the price (wage). Done. If we use a general equilibrium models then things can change rapibly. The people immigrating share a trait with cows; they don't live in a vacuum, no matter what physicists believe. They take their wage and they spend it. They eat, they live places, they do all sorts of things, so aggregate demand increases. This increases the demand for labour, and here the feedback loop comes in (and this feedback will keep going). We're not just dealing with a positive shift in the supply of labour, we're dealing with a positive shift in demand as well. Does that change anything? Employment will still increase, but we can no longer say anything certain about what will happen to wages. Partial equilibrium analysis is a critical assumption here.
If we go back to the farmer, though, and look at what would happen to his income if we decided to help him out with a little subsidy (or an extremely large subsidy if we're talking real life), then partial or general equilibrium doesn't matter so there's no reason to bother with general equilibrium. Even more phycisist talk: Newton is sadly outdated, but for a lot of stuff that really doesn't matter. I don't need Einstein if I want to calculate the trajectory of a football, Newton will do just fine. And it turns out Einstein was wrong anyway.