Only an overview can be used to estimate whether the measured price development represents a real trend or simply a fortuitous fluctuation within the framework of statistical uncertainty.
No one is indifferent to the state of the real estate market. Everyone is affected by developments in this market, whether as a tenant or owner. Real estate accounts for by far the largest share of Swiss household wealth and housing costs are the biggest expense item in the average household budget. Therefore, it is not surprising that everyone pays special attention to this topic when it arises.
The public eagerly absorbs all new information about the latest developments in the real estate market. The media is always rushing to almost every news post by just about anyone and thus is virtually certain to generate the coveted clicks – the main currency of the digital age. Information quality? It is not subject to any verification or critical review. Information content? It doesn’t matter as long as the story is succinct, engaging and, if possible, easy to understand.
Measuring the evolution of property prices is undoubtedly the most revealing example to illustrate the struggle to draw attention to the analysis of the property market. In this sense, consulting firms, banks, start-ups, real estate portals, realtors and universities are fighting a fierce but fleeting struggle each time to attract the public’s attention. Measuring the evolution of prices of residential real estate transactions is considered the main discipline in the field. Because, unlike bid prices, transaction prices measure the property prices actually paid and are therefore a valuable aggregate of supply and demand.
For years, a large number of indices have provided quarterly information on the evolution of transaction prices for single-family homes and condominiums. The best known and most established of these award series are provided by leading real estate consultancies Wüest Partner, CIFI and Fahrländer. But Raiffeisen has also published its own recognized transaction price index since 2016. The ZKB regionalizes these indices in the Zurich market and now the Federal Statistical Office (FSO) is also making its voice heard in this area.
In many media, the publication of each new price change rate by each index publisher gives rise to an article dedicated to each quarter. What is the problem? The numbers published by these reputable organizations sometimes differ greatly, even though they measure more or less the same thing. Thus, it sometimes happens that a media announces three different rates of price increase for the same quarter within the space of a few days. Neither editors nor readers seem to notice these contradictions. Yesterday’s flash is long forgotten today, given the flood of information that falls on us. Even the various arithmetic signs fall prey to this frenzy and do not raise any questions.
There are so many clues and possible explanations for these differences in the price dynamics they measure are also many. Each index provider has its own pool of OTC trading data that it incubates like a treasure trove. Due to the lack of a central database, no one has access to all home purchases. Each vendor therefore always only measures the price development of a fraction of all transactions.
Other properties are also negotiated quarter by quarter. And since a price index is not intended to measure differences in the quality of accommodation itself (e.g. different locations, areas and levels of development) but is intended to capture the underlying price trend, the quality must be rectified each time using complex statistical models and a large number of variables. Each provider uses its own model and, of course, is convinced that it is better than the competition. There is also no uniform practice regarding the modalities of creating an index from the data and correcting its quality. Paasche, Laspeyres or Törnqvist, which is the best track? Each favors a different method for building an index. And that’s why the indices don’t all measure exactly the same thing, but at most approximately the same thing.
Added to this is always a certain statistical uncertainty arising from the limited size of the sample of property sales made available each quarter. Despite these uncertainties, all vendors still only publish point estimates. And that voluntarily with digits after the comma. This suggests to the public a precision and certainty that does not exist in the measurement of real estate prices. The confidence intervals for estimating index values, a measure of statistical uncertainty, are often so wide that even the change in the arithmetic sign of the price change is within the usual range of statistical uncertainty. But these relativizations are obviously ignored when communicating price dynamics. They contradict today’s public need for short, punchy, and explicit messages. But I think it’s also important, given our own index construction, to make such relativizations. You also have to see the building limits and not just the raw value.
The Confederation, which had long relied on these indices established by private entities due to the lack of its own data and models, undoubtedly had another explanation for the differences between the price indices. Taking into account the often very different price series, it ended up assuming that private organizations were overloaded with measuring prices or simply incapable. Kind of a market failure, in a way. The State therefore had to take matters into its own hands to correct them. After several years of intensive and costly development and the complex establishment of a database to which the top 25 banks must provide transaction data for their mortgages, the Federal Statistical Office (FSO) has for the first time published its home price index housing in 2020. Since then, this official index has been added to the multitude of private price series mentioned. It cannot be said that this official index has done much to shed light on the question of real price developments. Now there’s just one more data point the media can report on. However, real developments in housing prices remain at least as vague as before, even after state intervention.
Some examples? While the OFS announced a significant drop in condo prices of -1.2% in Q1 2020, all other index providers mentioned saw a price increase. Wüest Partner recorded an increase of 2.2%. A solid 3 percentage points difference in a single quarter. For an average home, that’s almost 30,000 francs given the current price level, so it’s far from insignificant. Obviously, such quarterly variations should not be overinterpreted, given the aforementioned statistical uncertainty. Sometimes values level off over time and indices move closer together over an extended horizon. But even if we consider longer periods, some questions still remain. Thus, the Fahrländer price index for multi-storey properties has increased by almost 23% since the beginning of 2019. The corresponding CIFI index, on the other hand, shows only a 14% price change. After all, who is right? Obviously, no one has the answer and never will. This is how all index providers will continue to publish their indexes, consider theirs the best and be sure to attract media attention.
The indicator represented by the transaction price index is not, however, completely useless, despite all the differences and uncertainties. Because all indices agree on at least one thing, namely the general development of price dynamics in recent years. Knowing that residential property prices have risen a lot in recent years is valuable information for the market, which can be taken for granted given the unanimity of opinions. On the other hand, it is important not to overdo the market analysis of the quarterly announcements on property price developments from the various index providers. The anomalies relating to these quarterly figures are, at best, an indication of the interest in scrutinizing the market a little more closely. As with all economic indicators, property prices can only be used to assess the market situation in combination with many other data. Only an overview can be used to estimate whether the measured price development represents a real trend or simply a fortuitous fluctuation within the framework of statistical uncertainty. Therefore, caution is needed in case a price drop is announced again…
In that spirit, I wish you and your loved ones a happy holiday season and a happy new year, which I personally hope means a little more boredom. The last few years have just been hectic! We will meet again on January 4, 2023.