Why have inventories increased so much since 2008? Traditionally, housing starts, housing under construction, completions and sales have all broadly grown in line with each other. From around 2002 until 2007 housing sales accelerated, reducing the stock of unsold housing. However, since 2008 housing starts and construction were ramped up, initiated by the stimulus package in November 2008, when liquidity quickly spread through China's economy. This created an oversupply of housing.

Oversupply is not ubiquitous. In particular, it is not really an issue in first-tier cities－Beijing, Shanghai, Shenzhen and Guangzhou. There the demand-supply situation has remained fairly tight because of healthy demand trends, the maintenance of relatively strict restrictions by local governments on house purchases in their areas and the relatively disciplined behavior of property developers. In particular, housing prices have surged in Shenzhen in the last two years.

While it is hard to generalize, the average property developer in the large cities tends to be relatively sophisticated and disciplined because of the demand prospects. In contrast, developers in smaller cities have often been less astute, undertaking projects if the financing was available without necessarily doing serious research into the underlying demand prospects. With ill-disciplined developers, strong supply and weaker demand trends (people tend to prefer the large cities as places to work, live and invest), inventories of unsold housing have not fallen enough in the smaller cities in recent years, even as the market has remained tight in first-tier cities. Indeed, while housing prices have been rising strongly in first-tier cities since early 2015, they have only recently started to increase again on a month-on-month basis in third-tier cities.

This matters because smaller cities are where most of the construction is taking place these days. The share of third- and fourth-tier cities in nationwide real estate sales and construction has risen steadily from 44 percent in 2004 to 64 percent in 2014, and nowadays 95 percent of nationwide real estate sales are outside of first-tier cities. Thus, while real estate developments in high profile first-tier cities tend to receive a lot of attention in the media, and probably affect sentiment economy-wide, second- to fourth-tier cities are the key drivers of nationwide property construction trends and their macroeconomic implications.

In the short term the pick-up in property construction is for real. However, we do not think that the pick-up will translate into a sustained recovery of construction and investment until housing sales grow healthily for a prolonged period and inventories of unsold housing come down more significantly. And even then the recovery in construction is unlikely to turn into the kind of boom that China has seen in the past. This is because the speed of urbanization has slowed down somewhat and the demand for housing as an investment vehicle is less strong than it used to be.

The real estate construction pick-up supported overall growth in the first quarter of this year and should do so in (at least the start of) the second quarter. However, we think it is likely that the property construction cycle will ease again and that the government will need to continue to rely on other drivers, including infrastructure investment, to meet its ambitious growth target of 6.5 to 7 percent for 2016.

The author is head of Asia economics at Oxford Economics.