Real estate bubbles: supply elasticity and expectations

Summary:

– Housing supply is inelastic in the short term and becomes elastic in the long term.

– If agents do not anticipate future changes in supply, prices will rise excessively.

– A recent analysis of the history of real estate bubbles in the United States tends to confirm this hypothesis.

In a February 2013 working paper, Edward Glaeser, professor of economics at Harvard, examines various bubble phenomena that have occurred in the United States from 1790 to the present. He attributes most of these phenomena not to irrationality but to the cognitive limitations of agents. In particular, they do not take into account the elasticity of supply. The purpose of this note is to present, in a simple manner, the effect that supply elasticity and the errors in anticipation by economic agents can have on real estate price variations.

The sensitivity of housing supply to price only becomes significant in the long term.

In the short term, the supply of real estate is relatively fixed: it takes time to set up a real estate project (acquisition of patents, definition of the type of buildings) and to implement its construction. The supply of real estate is therefore relatively inelastic to short-term prices [1]. Consequently, when demand increases in the short term (accommodative credit conditions and/or strong growth in household income), an upward trend in prices is observed.

However, while supply is inelastic in the short term, it becomes elastic in the longer term. Higher prices attract more real estate developers, which tends to increase the volume of construction and therefore the supply of real estate, contributing to a slowdown in the rise or a contraction in housing prices [2].

This long-term elasticity of real estate supply is limited due to the availability of land in densely populated areas and/or regulatory constraints relating to urban planning policies.

However, on the one hand, the availability of land in densely populated areas is not always a constraint, as price increases have been observed in areas where, a priori, the lack of land was not a constraint [3].

On the other hand, above a certain price level, measures to expand supply (such as increasing the height of buildings) will necessarily be implemented.

It therefore seems important to study the effect of this variation in supply on real estate prices.

In the rest of the analysis, we assume that demand for real estate is demand for financial assets, which implies that demand for real estate must be strong:

– if real estate prices are low

– if credit conditions are accommodating

– if taxes on this property are low

– if income is high (increase in available savings)

– if the anticipated appreciation of real estate prices is high (high expected return on capital).

We examine two situations: the first where supply is perfectly inelastic and the second where supply is perfectly elastic in the long term.

First situation: changes in real estate prices with perfectly inelastic supply

Let us first consider a situation of perfectly inelastic supply, i.e., a case where supply is assumed to be fixed over time.

We can represent this situation in the following figure (Figure 1). The downward sloping line is the demand function: the quantities demanded decrease with prices given future price expectations. The vertical line is the supply curve.

The initial situation is the intersection of the two curves at point A, with demand increasing permanently (e.g., credit conditions becoming more accommodative, income increasing, or property taxes decreasing). In this case, if agents do not anticipate future price increases, the equilibrium is at point B. Prices increase once and for all and remain at this high level. There is no sign of a bubble.

However, there are two phenomena that are reminiscent of a « bubble »:

– The case where the increase in demand is temporary, i.e., where the demand curve returns to its initial position. This is the case, for example, if credit conditions have been lax for a certain period of time and return to their « normal » level. We then see an increase in prices followed by a decline as demand returns to its « normal » level.

– If the increase in prices following an increase in demand leads agents to anticipate a similar increase in prices in the following period. This tends to increase demand because the expected capital gain (expected return on real estate assets increases) increases. We then find ourselves at point C: the anticipated appreciation is realized. The phenomenon can continue indefinitely as long as agents do not change their expectations. In this case, bubbles tend to be more durable in theory than in the case of supply elasticity, which we will now discuss.

Figure 1

Second situation: the evolution of real estate prices with perfectly elastic supply

Let us now imagine a case where supply is perfectly elastic in the long term. We are now in the situation shown in Figure 2, where the shape of the demand curve is unchanged from the previous case, but the supply curve is now a horizontal straight line.

The assumption here is that in the long term, prices must return to their initial level through an adjustment in the quantity of housing.

In the short term, if agents do not anticipate the change in supply and assume that prices will remain constant in each period, then there is a movement from point A to point B, with supply being inelastic in the short term.

Then, as new construction progresses, the equilibrium shifts along the demand curve until it reaches point C. In effect, this is equivalent to a gradual shift of the vertical line in Figure 1 to the right.

If, on the other hand, agents perfectly anticipate the change in supply, prices should increase less. This is because, in this case, they anticipate a future decrease in prices, which automatically reduces their short-term demand as they anticipate a loss in capital. In this case, we move from point A to point B’, and then the equilibrium shifts along the curve linking B’ to C (we have the same shift in the supply curve as in the previous case).

Figure 2

This is the fundamental mechanism behind Glaeser’s (2013) thesis: if agents do not anticipate future changes in supply, prices tend to rise « too » much. Agents do not anticipate future price declines, and real estate demand is too high, even though we have not assumed that agents anticipate any price increases. If we consider that these agents anticipate an increase, then the demand curve tends to shift upward once again. However, at the same time, as supply increases, expectations tend to cease to be self-fulfilling, and the bubble is very likely to burst because the increase in supply mechanically lowers prices [4].

This is certainly theoretical, but it is extremely important for understanding bubble phenomena. If agents base their decisions on a « simplistic » model, such that they base their expectations of future price growth on price growth observed in the recent past, then they will tend to overestimate the possible appreciation of prices following an initial price increase [5].

The cases of Spain and the United States come to mind as examples of how plausible this intuition is. We have seen sharp price increases in certain regions, but as the supply of housing has increased, prices have tended to fall sharply. Over long periods and across different countries, inflation-adjusted real estate prices appear to be relatively stable. The recent decline in real estate transactions in France could support this observation.