Either the observations are wrong or the theory is nonsense.

The observations

1. The larger the conurbation-the denser the population-the better the links to other places, the higher the average productivity-wages-profits. This is partly because of synergies-specialisation-agglomeration benefits; and partly because higher wages at the top of a pyramid depend on how wide the base of the pyramid is. This pulls up the overall average while widening the difference between top and bottom. The fag packet calculation is double the size of a conurbation and average wages go up by 5%.

2. Within any country, there is free movement of people and workers. Between leaving home and ‘settling down’, plenty of people are willing move from lower wage to higher wage areas. Once all housing in the higher wage area is occupied, this pushes rents up to the point where a new equilibrium is reached and the wage differential is equal to the rent differential. At this stage, there is little financial advantage in moving and net internal migration migration is reduced to a trickle.

Anybody who disputes this can leave the conversation right now.

The theory

1. Housing is a perfectly ordinary good, like coffee beans, cars or carpets. If there is an increase in supply, prices will fall, and vice versa. This is easily observable in the short term, like if there is a coffee glut, or in Germany post-unification when East Germans sick and tired of Trabis, were itching to buy decent second hand VWs and second hand car prices doubled for a year or two. This levels off very quickly, as manufacturers make a certain minimum profit margin. Any less than that and some go out of business, any more than that and they manufacture more, or new players enter the market.

(It must be obvious that manufacturers aren’t going to commit hari kiri by increasing output in the long term. Stronger players might do this to force weaker competitors out of the market, but once they have gone out of business, output falls and prices rise to the old level. And it must be obvious that home builders/land bankers have chosen their profit maximising level of output, i.e. roughly keeping pace with population growth, and it’s going to be a teensy weensy bit tricky to persuade them to build more.)

2. Therefore, if somehow home builders/land bankers could be persuaded to increase output (despite it being against their interest to do so), prices and rents would fall. Simple, ‘housing crisis’ solved! Duh, why did nobody think of that before?

Either the observations are wrong or the theory is nonsense

OK, let’s measure ‘affordability’ in terms of how much disposable income internal migrants have left over after paying rent (‘the basic minimum’). We know that within any country, this is pretty much the same everywhere and however many homes there are in any area, the basic minimum will still be the same all over the country.

This is where is becomes nigh impossible to reconcile the theory to the observed facts:

Possibility A. We increase supply of homes in high wage-immigration areas, there will be more agglomeration benefits = higher average wages, so the trajectory is that rents there will increase rather than increase. Assuming that in emigration areas there is the opposite effect and wages fall, then the basic minimum must fall. This actually exacerbates the whole thing.

Possibility B. There are no agglomeration benefits, but people move to the new housing in the higher wage areas anyway. Using the simplistic theory, more supply of labour = lower wages = lower rents. But in terms of affordability, this is no improvement, because wages fall by as much as rents. So no real improvement either.

Possibility C. A greater mind than mine can actually reconcile the observations with the theory.