Ian Mulheirn, the Director of Consulting at Oxford Economics and a former HM Treasury economist, has published an article arguing that high property prices are not the result of an excess of demand over supply of housing.

This surprising claim for an economist is justified primarily by a comparison between the rate of change in the stock of dwellings, and the rate of change in the number of households.

In most of England (all but the East and West Midlands), the number of dwellings is rising faster than the number of households. QED, there cannot be a shortage of dwellings in which households can be established.

But how can you tell from the statistics how many households have not been formed because houses are neither available in sufficient numbers nor affordable on typical household incomes?

Mulheirn compares only these two data sets. He does not consider a third relevant figure – the population level. Population tells a different story. From 2004 to 2016, the population of London increased by around 113,000 per year. The number of households increased by around 25,000 per year. And the number of dwellings increased by around 27,000 per year.

These figures would square if the average household size were just over 4. Unfortunately for Mulheirn’s theory, although London has the highest household size in England, and that figure is rising, it is still only around 2.5.

At that average household size, London needs over 45,000 new dwellings per year to keep pace with its increasing population. The small discrepancy between the numbers of dwellings and households is dwarfed by the large discrepancy between the number of dwellings required and the number added to the stock.

The small increases (2,000 per year) in the number of unoccupied dwellings should be compared with the total of 3,200,000 households in London. It is a drop in the ocean, and can easily be accounted for by the well-noted phenomenon of unoccupied properties bought as financial investments in a rapidly-inflating market. The 2,000 discrepancy on this account cannot account for the 20,000 discrepancy between the numbers of dwellings or households added each year and the numbers required to match the rising population at typical occupancy levels.

The 2011 census from which the average household size is drawn notes that the average occupancy levels are increasing in London, faster than in the rest of the country. That provides the easiest, indeed the only credible explanation for the discrepancy between the increases in population and dwellings.

Are we to assume that the increasing household sizes reflect changing preferences in favour of larger households, unrelated to the shortage and high cost of property? People have simply decided that they like living together in larger numbers? Or is it more likely that property prices that are unaffordable to many are making larger household sizes inevitable?

For instance, another well-noted phenomenon is the increasing age at which children are moving out of their parents’ homes. Is this simply because today’s twenty- and thirty-somethings don’t want to leave the nest? Or is it because it takes them longer to save for the deposit and/or reach the necessary earnings level to be approved for mortgages that will cover the cost of inflated London prices, even with low interest rates and rising multiples? Mulheirn’s theory requires us to believe that it is the former. Perhaps he needs to talk to more young(ish) people.

Following Mulheirn’s approach, demand can never exceed supply. The amount of goods bought can never exceed the amount of goods offered on the market. If demand is quantified by the amount bought and supply by the amount offered, demand cannot exceed supply. In fact, as there will always be some spoilage and transactional friction, supply must always exceed demand, by this definition. Rising prices can never be a function of excess demand over supply, because excess demand cannot exist. If an earthquake destroyed half of London, we would probably still not have a housing shortage under Mulheirn’s definition.

“Ce qu’on voit et ce qu’on ne voit pas”. What Mulheirn does not see is the transactions that would have happened had prices or availability been different. The fact that 25,000 households are formed each year does not tell us that only 25,000 families want to form households, but that only 25,000 families can afford to form households given the availability and cost of housing. That it is just over half the household-formation rate that one would expect for the rate of population growth tells us that there is indeed a shortage of housing. Anyone familiar with aspiring homeowners’ real-world experiences rather than abstract and partial data sets would not think to dispute that reality.

This is not to deny Mulheirn’s point that monetary policy has played an important role in property-price inflation. But it reasserts the fundamental role of supply and demand in pricing. The excess of demand over supply created the upwards pressure on pricing. The artificially-low interest rates elevated the level to which that pressure could drive prices. In a world of more-normal interest rates, but facing the same disparity between properties available and properties required by the rising population, housing would have been equally unaffordable. The composition of those costs would change: you would pay less to the previous owner of the house, and more to the bank, in the form of higher interest payments. But there is no reason why the overall level of housing costs, either in absolute terms or as multiples of income, would have been any lower.