Have mainland home prices peaked? How reliable is official price information?

Since the start of the bull market in early 2015, the magnitude of average price rises of the 70-city basket in China has reached 30%; the trend accelerated further in early 2016 (eg. Shenzhen prices were up 16% in Q1 alone), triggering a wave of government crack down policies in Q4 2016 in an attempt to rein in the run-away prices. To date, over 20 cities have launched their versions of policies, but the impressive performance of the domestic property stocks in recent months could herald further home price increases ahead. This article attempts to look into the nature of the official statistics that stand behind these incredible price increases.

Official statistics suggest further price rises ahead

Many market commentators and analysts follow the property market in China by tracking the ‘Price Indices of Second-Hand Residential Buildings in 70 Large and Medium-Sized Cities’ published by the National Bureau of Statistics (NBS) – we will refer to this below as the official indices.

Under normal circumstances, the property market should be close to, or at a peak when the bulk of the basket cities are showing price rises (green bars in Chart 1). To illustrate, when the green bars begin to grow from the bottom (e.g. in April 2012 and April 2015), home prices usually rise, as proxied by Guangzhou (red line); on the other hand, when the green bars begin to shrink from high levels, a down-trend in property prices should also ensue (as shown in July 2011 and April 2014).

Chart 1: Official data suggest that the bull market has more to run

If one applies this relationship to the above chart, one may reach the conclusion that the current bull market in prices will continue because the green bars are still some distance from reaching the peak yet. However, the reality is not quite as straightforward as the above simple relationship suggests.

Is it the market or market statistics that local governments are trying to manage?

While gathering data for analysis, this writer has noticed a curious pattern – that whenever there is a meangingful surge in home prices, the official indices tend to lag comparable non-official ones; as shown in Chart 2, from a common starting point of 100, when steep rises in the non-official indices (red arrows) take place, the official equivalents (blue arrows) often underperform, with the blue line either remaining flat or even falling. It seems there is a mismatch between the official indicators and reality. In a short period of under eight years, the non-official index is now some 69% higher than the official index.

Chart 2: Official index always lagged behind in bull markets

Chart 3 provides another perspective on this discrepancy. Rebounds in home prices during bull markets (i.e. when the red line rises, represented by the upward sloping red arrows, e.g. from mid-2012 to mid-2013) have been accompanied by big outperformances in non-official indices vs their official equivalents (represented by rising non-red lines, which are collectively annotated by the upward blue arrows). Furthermore, this deviation happens in all the cities, and magnitudes of outperformance are not reverted during the downcycles (e.g. between late 2013 and early 2015), suggesting that the issue is not limited to isolated cases, but is a systematic aberration.

Chart 3: Non-official price index rises tend to outpace official ones 

Another intriguing observation is that how far official indices deviate from non-official indices depends on how far that city is away from the capital: while the official index of Beijing (light blue line in Chart 3), which is right under the nose of the central government, did not drift too far away from the non-official indices in the past few years, that of Tianjin (dark blue line in Chart 3), which is a bit more ‘out of sight’, was able to lag non-official readings by over 100%!

Chart 4 clearly shows the local governments’ practice of understating the substantial rises in home prices in order to avoid drawing attention or policy crackdowns from the central government.

Chart 4: The bigger the price increase (x-axis), the bigger the gap against non-official indices (y-axis)

This is understandable behaviour given the civil service performance system in China – officials’ future prospects are tied closely to the rate of GDP growth in their sphere of influence, providing great incentive for them to maximise their reported economic output. At the end of the day, gross capital formation has been major component of GDP calculations, making up 40%-50% of GDP growth (Chart 5), and in turn the housing sector with its related industries are a significant driver of capital formation. It is no wonder that local officials are tempted to grow their own housing investments as fast as possible while under-reporting this fact, because their political careers are heavily dependent on the continued boom of this sector.

Chart 5: Significant contribution of Capital Formation induces data manipulation

Once the drivers of behaviours are revealed, it is not hard to understand why the gentle rises in official indices are so inconsistent with both non-official figures, and even more at odds with news reports of soaring home prices. Tampering with official statistics may be the easiest way for local officials to produce flashy GDP figures while avoiding being accused of stoking housing bubbles!

Non-official data suggest home prices may have peaked

Given the lower reliability of the official statistics, it is perhaps more appropriate to deduce (using non-official data) the ratio of basket cities experiencing inflating home prices discussed in Chart 1. This exercise results in a picture from which one would draw an opposite conclusion on the state of the market: by as early as Q2 2016, the home prices in all of the basket cities could have already been rising. In other words, the green bars have rocketed to 100% (Chart 6) well before indicated in Chart 1, and at speeds far more rapid than the up-cycles in 2010-11 and 2013-14!

Chart 6: Non-official data suggest that home prices may have already peaked

As the green bars have already reached the top for over a year, the author fears that the housing market may soon correct, if it has not already started to do so. It is regrettable that the unreliability of official data makes it hard to diagnose the symptoms, and yet another opportunity to deflate the housing bubbles could have been missed.

To conclude, the above analysis serves as a good example to support reports that official statistics in the mainland lack robustness, and in this case, confidence in housing data may be further undermined as well. In terms of investments, readers might want to consider selling physical properties in the mainland; however, exposure to housing can be maintained by putting the proceeds in mainland property stocks, some of which still trade at reasonable valuations and have growth factors besides just the residential sub-segment.

With special thanks to Eugene Lai and Stewart Ng for their contribution to this article.