How can we explain the sharp increase in residential property prices per m² in Paris? Are we at the height of a Parisian property bubble?

How can we explain the sharp increase in residential property prices per m² in Paris? Are we at the height of a Parisian property bubble?

September 2018 | 8 MINUTE READ

For years now, we have witnessed complex and often senseless debates on property prices in Paris. The clichés are legion, and expert debates most often end in chaos. Ultimately, estate agents, notaries and bankers are hauled over the coals.

Essentially, there are two conflicting viewpoints:

1. Paris is unique; Paris is a niche market; Paris is magical.

2. This growth cannot last forever; the bubble is at bursting point and the aftermath will be tough.



According to Parisian notaries, the price per m2 grew from €3,463/m2 to €8,782/m2 between 1991 and 2017. This is indeed a significant increase, averaging 3.6% per annum. At the same time, Insee recognises inflation between 1991 and 2017 to be an average of 1.4% per annum.

In short, the value of the Parisian square metre grew on average 2.6 times more quickly than consumer prices as estimated by Insee.

In the graph below, we can observe the curve representing the actual increase in property price per m2 in Paris versus a curve representing the 1991 price per m2 subsequently inflated each year using Insee’s inflation rate.


Average price per m² in Paris [in €/m²]


We can identify two phases:

In the period from 1991 to 2004, the actual price per m2 remained below the 1991 Insee-inflated price. Property prices grew significantly in the period prior to 1991 then underwent a major correction of around 35% between 1991 and 1997. Only in 2004 did the actual Parisian property price curve exceed the Insee-inflated curve1.

In the period from 2004 to 2017, the actual price per m2 grew massively, much more quickly than economic inflation: +5.5% per annum on average for the actual price versus only +1.2% for inflation.

In this later period, the disparity between property price and the average increase in standard of living warrants further investigation.



Many defend the following theory: demand for Paris is growing, whilst supply is limited.

The demographic reality is not quite so categorical. The number of inhabitants in Paris was 2.15 million in 1990 and 2.17 million in 2018, with a peak of 2.24 million in 2010.

Demographic pressure, however, is much more prevalent in the rest of the Ile-de-France region (excluding Paris), with the population growing from 8.4 million to 10.1 million inhabitants between 1990 and 2018 (+0.6% per annum).

We can conclude that demographic pressure in Paris has remained constant or has grown, if we include the inner suburbs. This enables us to rationalise prices per m2 remaining stable at a high market level2. However, it in no way rationalises the major growth dynamics witnessed for actual property prices in Paris.



In Paris, transaction volumes are higher during periods of increased prices (between 35,000 and 40,000 transactions per annum in periods of increasing prices), whilst these volumes fall significantly during periods of decreased prices (25,000 to 30,000 transactions).

It seems high time to put an end to the common misconception: a fall in the volume of properties for sale does not automatically increase prices, as some property professionals would have you believe.

The economic reality is different: when prices are high, property owners are more inclined to sell their property, either to crystallise a capital gain or to undertake a buy–sell transaction (incidentally, often in the reverse order) because they have confidence in the market.

Conversely, when prices are shrinking, the market seizes up and property owners delay their potential sales waiting for better days.

The conclusion is striking: the classic economic mechanisms of supply and demand simply cannot explain the historical growth in residential property prices per m2 in Paris.

These price dynamics should really be considered as ‘contra-economic’: supply grows in volume when prices increase; supply falls in volume when prices decrease.


Does demand for Paris explain the rise in prices? [FALSE]

Does a fall in the number of transactions explain the rise in prices? [FALSE]



As we cannot use demography and standard economics to explain the increase in prices observed, what are the variables that really can explain this sharp increase?

To answer this question, we have built a multi-variable regression model using a long historical series (1990–2017) which comes to the following conclusion.

Since 1990, the development of residential property prices per m2 in Paris can be explained ‘entirely’ and ‘mathematically’ by two financial variables.

To put it simply, this means that it is possible to explain – and potentially predict – the price per m2 of residential property in Paris with an extremely high degree of accuracy using only two financial variables.

– For those familiar with such techniques, our multi-variable regression model reached a correlation index (R2) level of 93%3.

The first variable involved is the following:

Variable 1: spread (French 10-year OAT rate – Insee inflation rate).

– Taken in isolation, this variable explains the price per m2 with an R2 of 73%.


In simple terms, it represents a borrower’s interest rate adjusted for economic inflation. It verges on the notion of the actual net interest rate for the borrower.

When rates fall, the borrowing capacity of an individual borrower rises significantly. A quick back-of-the-envelope calculation shows that if an individual borrower’s rate decreases by one point, his or her borrowing capacity increases by approximately 10%. The borrower can therefore obtain more cash for the acquisition. Unfortunately, the market in Paris immediately takes this factor into account in its prices, which increase proportionately. The fall in rates enables a rise in borrowing capacity for buyers but not in terms of the number of m2 that they can buy. The market absorbs any increase in borrowing capacity in the prices per m2.

In 2017–2018, it is interesting to note the scissor effect between the French 10-year OAT rate and inflation. In this period, interest rates remained at their floor. Based only on these gross rates, property prices should not have increased because of this variable (cf. appendix 1).

However, in parallel, inflation picked up significantly in 2018 thus reducing the spread. In concrete terms, the actual net borrowing rate reduced further, relative to inflation. For the first time, the spread (10-year OAT – inflation) became negative, meaning that for the first time, individual borrowers could borrow at actual negative net rates: their mortgage rate would be lower than inflation.

This scissor effect explains the continued increase in residential property prices per m2 in Paris in 2018 (2018 forecasts: growth in price per m2 of approximately 8% in Paris).


Difference noted in prices per m² [actual price – 1991 inflated price]

This fall in spread, however, is not the only or even the best driver explaining the increase in prices per m2 in Paris.


The second variable is the following:

Variable 2: the size of the European Central Bank (ECB)’s balance sheet.

– Taken in isolation, this variable explains the price per m2 with an R2 of approximately 92%. It is itself significantly correlated with the first variable, owing to the coordination of decisions on the development of the ECB’s monetary policy for these two variables.


The impact of this variable highlights the major impact on the valuation of financial asset classes – including Parisian property – of the quantitative easing policy fervently implemented by the ECB for several years.

Currently, the main driver of the increase in residential property prices in Paris is the ease of obtaining credit, that is, the ECB’s provision of a growing money supply to banks – coupled with falling and subsequently low rates.

Since the 2009 crisis, the size of the ECB’s balance sheet has developed as follows:


Development of the size of the ECB’s balance sheet [1999-2017 – in €m]

The impact on prices per m2 has been direct. Since 1999, it is possible to obtain a 92% correlation between prices per m2 in Paris and this second variable alone (cf. appendix 2).

Note that this correlation also implicitly integrates the fact that the rate spread has fallen.

Indeed, the ECB’s monetary policy can be summarised as adjusting the level of reference rates and the size of its balance sheet jointly with the aim of relaunching economic growth all while maintaining inflation below the fateful threshold of 2%.


Development of residential property prices per m² in Paris versus the ECB’s money supply
[1999-2017 – Base 100 in 2006]


We have said it all!

The ECB’s monetary policy explains ‘entirely’ and ‘mathematically’ the development of prices per m2 in Paris (with three strings constituting our two variables: rates, inflation and money supply). The price per m2 of residential property in Paris is no different from the index of a purely financial asset driven by these variables.

It is possible to predict future developments in prices per m2 in Paris using a mathematical formula and an extremely high relevance. To do this, we have to listen to the European Central Bank and anticipate and model its decisions.

This model will be solid and valid for as long as the ECB controls the banking environment in Europe and macroeconomic indicators are ‘conventional’.

In the short term, the latest releases from the ECB provide the following guidance: stabilisation of its balance sheet and the first increase in reference rates forecast for the beginning of H2 2019. An adjustment to be anticipated.

In the case of a major financial shock, the story will be very different but also able to be modelled with a high degree of precision.



The in-depth combination of economic and financial expertise allows us to obtain better conclusions than market consensus and enables us to build powerful, differentiating tools for forecasting and modelling.

To be limited to consensus and go without such precise tools is not something worth considering for a sophisticated investor.



Appendix 1:
Development of interest rates (10-year OAT), inflation and their spread in France since 1991


Appendix 2:
Development of the size of the ECB’s balance sheet versus prices per m² in Paris

1 Note that 1991 was, at the time, a peak year in the cycle closing out the upswing period in Paris property speculation.
2 Or, at best, exhibiting a level of intermediate growth.
3 The correlation index is a statistical indicator representing the link between relative variations in share prices or different indices. A correlation index of zero indicates a total absence of link; a correlation index of one indicates an extremely high degree of proportionality.

Download the article

More accuracy Perspectives