Ekonomistas guest post by Harry Flam: Bubble in the housing market?

Ekonomistas guest post by Harry Flam (in Swedish). Here is an English translation.

This is a guest post by Harry Flam, professor of international economics at the Institute for International Economic Studies, Stockholm University.

The U.S. and several European countries have experienced large falls in housing prices in the wake of the financial and euro crisis. Housing market indices fell 43 percent in the U.S., nearly 50 percent in Ireland, and – so far – 30 percent in Spain. This has contributed to very negative outcomes in these countries: banks have suffered enormous credit losses and have cut down their lending, tax payers were forced to rescue many of them, households became less wealthy, and households and firms decreased demand, resulting in high unemployment.

With hindsight, it has been argued that housing prices in these countries were driven by expectations of future capital gains. Considering the fact that real housing prices in Sweden have experienced a historically large increase since the mid-1990s – prices of single-family houses have increased by about 300 percent – after having increased at the rate of the consumer price index before that, one may ask if we have a pricing bubble waiting to burst, with serious consequences for households, banks and the economy? 

The Riksbank (2011), Sten Hansen (2013) (in Swedish) and the consulting firm Evidens (2013) (in Swedish) have reached the conclusion that the price increase in Sweden can be explained by fundamental factors, such as people moving to the large cities, little new construction, increased real incomes and wealth, abolition of the wealth tax and the lowering of the property tax, and less severe liquidity constraints due to no amortization, higher leverage and lower loan rates relative to disposable income.

Does this mean that there are no signs of a pricing bubble in the housing market? A first sign is usually that housing is bought as a financial investment in expectation of future capital gains and not for own use. Such speculation took place in the U.S. and Spain, for example. However, one can immediately say that the purchase of a permanent condominium apartment to let is prevented by regulation in Sweden, and is probably not very common in the case of one-family houses.

A second sign is that the user-cost per square meter and year in condominiums is higher than in rental housing with the same location and standard – if rents are set by the market. Owners of condominiums accept a higher user cost because it is compensated by future capital gains. One should expect the user cost to be fairly similar in condominiums and rental housing in the absence of speculation.

Rents in the existing stock of rental housing are not set by the market. Rents in newly constructed rental housing are negotiated by the tenants’ association and the landlord, but they are set to cover the cost of production and a reasonable rate of return. If there is competition between different landlords, rents in new rental housing should be equal to those that would emerge in a free market.

In order to compare rent per square meter and year in new rental housing with the user cost per square meter and year in condominiums, I have chosen to make a comparison with newly produced condominiums in a central location in Stockholm and with the existing stock of condominiums in all of central Stockholm. If a price bubble exists anywhere, it should be here.

Table 1.  SEK per square meter and year

New rental housing

New condominiums Existing stock of condominiums
2 300 2 020

2 052

The calculations are based on the following assumptions: inflation rate of 2 percent, interest rate of 6 percent (normal repo rate of 4 percent and a margin of 2 percent to the borrowing rate), the price of condominiums increases at the rate of inflation (which it did before 1996), mortgage equal to the purchase price (own capital invested is assumed to have an opportunity cost equal to the borrowing rate), and a full tax deduction of 30 percent. The basic parameter values are: SEK 60 000 per square meter in new condominiums in central Stockholm (actual purchase price), SEK 66 000 per square meter in existing condominiums in central Stockholm (average for September-November 2013, Maklarstatistik (in Swedish)), and a yearly fee per square meter to the condominium association of SEK 700 (from annual statements) and SEK 600, respectively (stated by real estate agent).

As can be seen, the user cost per square meter and year in condominiums is below rather than above the rent per square meter and year in new production of rental housing with approximate market rents. There is therefore no sign that the prices of condominiums are driven by speculative motives and expectations of future capital gains. (This is supported by the fact that the average market determined rent for a sublet studio was SEK 6 200 in 2013, Dagens Industri January 13, 2014 (Swedish). If the average studio has an area of 30 square meters, the rent per square meter and year is SEK 2 500.)

It is sometimes argued, for example by the Riksbank, that the prices of condominiums are driven by unrealistic expectations of future low borrowing rates. If the above calculation is made with the assumption of a borrowing rate of 2.5 percent – today’s low rate – and an inflation rate of 1 percent – the average for the past five years – and with the assumption that the user cost in condominiums should be equal to the rent in new rental housing, the price of a square meter of condominiums becomes SEK 227 000. An assumption of higher inflation yields an even higher price. Consequently, households do not seem to have unrealistic expectations of continued low borrowing rates. For other indications that households do not have low expectations, see this Ekonomistas post by Lars EO Svensson (English translation).

A third sign of a housing bubble in a particular area is a strong and rapid increase in the land price per square meter of housing relative to other areas. Figure 1 shows that the increase in the  price of land per square meter of new condominiums in the three largest cities has been very strong since 1998, the price has increased five-fold in real terms. The corresponding price in the rest of the country has at the same time increased three-fold. The relative increase of land prices in the largest cities can be explained by an inflow of people and the low rate of new construction. Even if the relative price increase in the largest cities were a sign of a housing bubble, the potential for a price fall to the level of land prices in other parts of the country is not particularly large in percent of total production costs, about 13 percent, see figure 1.[1]

Figure 1. SEK per square meter of condominiums


Source: Statistics Sweden

In summary, there are hardly any signs of a bubble in the housing market. That does not mean that housing prices cannot fall. If they do, this will be caused by a change in fundamental factors, such as an increased risk of unemployment or a stock market downturn.

[1] One may, of course, ask why construction costs – shown in figure 1 – have more than doubled since 1998. Similar increases have occurred in the other Nordic countries. The increases are evidently not caused by higher relative wages or profits in the construction sector.