Jonathan Portes makes a vitally important point - that statistics are reality. Apparently arid numbers summarize the lived experience of real people. As Raymond Wolfinger said, "the plural of anecdote is data". Statistics tell us about facts, whereas anecdotes are subject to sampling error.

Sadly, however, all this misses something - that it takes a theory to kill a theory. When Niall McCrae denies that migrants haven't depressed wages by saying "Try telling that to my landscape gardener friend, who has been severely undercut by waves of east Europeans" he's appealing for something more than just statistics. He needs a story.

Here are some I'd tell.

First, not all migrants are competing against native workers. Some are complements, and so help to raise wages. Polish roofers for example allow electricians to get more work done and so earn more. And Filipina maids allow women to work longer hours and earn more.

Secondly, even where migrants tend to reduce wages, there are offsetting mechanisms. For example:

- The fall in wages should increase employment: if prices fall demand rises.

- Immigrants spend their wages. When the Polish landscape gardener spends his money in Wetherspoons, he's creating work and wages for bar staff.

- The man who employs a cheap Polish landscape gardener has more money left over to spend on other things. This too should create jobs and raise wages.

Now, I'll concede that these mechanisms are imperfect offsets. The price elasticity of demand for labour is low: we know this because higher minimum wages haven't destroyed very many jobs. Migrants send some of their money home and so don't wholly recycle it into the UK economy; and the savings made from employing migrants might be saved or spent overseas.

But then, other mechanisms come into play. If migrants do tend to cause weaker wages and aggregate demand, this should be offset by looser monetary policy. And it has been, at least in part: real interest rates have fallen as migrants have arrived. Nobody credits immigrants with cutting their mortgage payments, but perhaps they should.

And even if all this fails, there are still offsets - most obviously, rising minimum wages have mitigated any tendency for low wages to get lower.

All I'm doing here is spelling out mainstream economics (explaining why the lump of labour fallacy is a fallacy) and telling stories to explain a fact - that immigration (pdf), mostly, hasn't (pdf) greatly depressed wages.

Yes, "migrants arrived and now my wages are lower" is a powerful story. But so too was "my son got the MMR jab and has since been diagnosed with autism. " Both stories confuse correlation with causality. Which is not surprising. People are terrible at identifying causality in economics, and they pay insufficient attention to evidence that is important but not obvious.

Of course, it is the case that real wages have been under pressure in recent years. But why blame immigrants for this? There are many other culprits: weaker trades unions; fiscal austerity; financialization (pdf); the financial crisis; power-biased technical change; and all the things that have depressed productivity.

And let's remember something else. It's not just migrants who have entered the labour force in recent years. So too have older people: there are 2.2 million more over-50s in work now than ten years ago. Economically speaking, entering the labour force from retirement is the same as entering it from Poland. So why blame one for lower wages when you're not blaming the other?

My point here is simple. Facts matter. But we must back them up with stories, with mechanisms. And there are plenty of these that largely exonerate migrants from blame for falling wages.