URA revises private residential property price index
The new method tracks prices of all private home transactions, and factors in finer attributes such as floor level and nearness to MRT stations.
A KEY property price index widely tracked by market players and the man-in-the-street has undergone a major enhancement.
The Urban Redevelopment Authority (URA) has revised the methodology behind its private residential property price index (PPI), and expanded the data it covers as well. The revised approach addresses a major grouse with the earlier index by including sales data from stamp duty submissions to the tax authority. This means that the index now covers all private home transactions.
URA had previously relied on caveats lodged with the Singapore Land Authority - non-mandatory legal documents that typically represented only 80 per cent of all transactions - and a monthly survey of developers' sales. By switching its index methodology to "stratified hedonic regression", the index is also now able to distil finer property attributes so that purer changes in price are captured.
Using the new approach, URA's first-quarter flash estimate of its PPI showed a 1.1 per cent quarter-on-quarter fall in overall private home prices, following a similar drop in the fourth quarter of 2014. Back-testing of the current and revised index approach shows no change in the general direction of property price movements in the past.
URA's revision of its index followed similar moves by other government agencies to revise their property indices. Its method is similar to that used by the Housing & Development Board (HDB) for the HDB resale price index. Both agencies consulted NUS real estate professor Lum Sau Kim, who earlier led the launch of the NUS Singapore Residential Price Index, which also uses hedonic regression.
Dr Lum said URA's new methodology is more relevant for tracking market-wide price movements in Singapore, given the wide spectrum in housing attributes here. "It is also less susceptible to low volumes and can be more flexibly adjusted should market conditions change," she said.
Under the new method, transacted units are grouped based on property types before price change over time is computed using hedonic regression. These groups are assigned fixed index weights, based on five quarters of transacted values, which will be updated every three years.
The hedonic regression works like a fine-tuning dial, accounting for fine differences between the attributes of units such as their size, age, floor level and even micro-location factors such as proximity to MRT stations. URA said the revision of its index is timely, given that housing attributes in Singapore have widened with the ageing of properties and the emergence of "shoebox units" since the index was last reviewed in 2000.
The private residential rental indices, which will be out in the URA real estate statistics for the first quarter, will also adopt stratified hedonic regression.
Sin Lye Chong, URA group director for land sales and administration, noting industry players' observation that prices of a few high-end projects may have fallen by about 20 per cent, said some projects in the Core Central Region (CCR) have recorded single-digit price falls.
A micro-study by URA on projects in the CCR found that the median prices of eight projects fell 8.9 per cent between H2 2013 and the six-month period to the end of the first quarter of this year. Meanwhile, CCR PPI has dropped 7.2 per cent from its peak in Q1 2013.
Mr Sin said: "The property price index reflects broad price trends for the entire private housing market, rather than small pockets of the market. URA has been publishing unit-level property price data in specific developments on its website for free, and the public should refer to these transactions for price movements at a local or development level."
Most consultants agreed that looking at transactions data already available on the URA website is more relevant in pricing a property.
Chua Yang Liang, JLL head of research for Singapore and South-east Asia, said: "Indices are meant to show an average trend line. Of course, properties are so diverse, it is inevitable that there are outliers and unique properties out there."
ERA Realty key executive officer Eugene Lim noted that the improved index is like "water has been made purer", while SLP International executive director Nicholas Mak likened it to a car model upgraded for better performance.
He said: "Certain parties may have their own interests. Property sellers may want to show that prices have dropped a lot and would like the government to relax the cooling measures. They may give the example of St Regis or a bungalow at Sentosa Cove being transacted at certain prices, and as a result, prices dropped by 25 per cent. But one single transaction does not make the whole market."
Savills research head Alan Cheong, however, pointed out that there was "still a big abyss between what developers and us are seeing on the ground and what the indices are showing", especially for developers who find themselves slashing prices by 20 per cent to push sales.
Since the fourth quarter of 2013, URA has asked developers to disclose their selling prices, net of direct and indirect discounts.
When the amendments to the Housing Developers (Control & Licensing) Act come into effect in the first half of this year, it will be mandatory for licensed developers to submit to URA the net prices of units sold. URA will publish these net prices on its website.
The Business Times understands that the number of units transacted by developers with indirect discounts comprises only 5 to 10 per cent of all transactions; non-arms length transactions will still be sifted out of the data as they are not reflective of open-market transactions. Bulk purchases are typically excluded from the index coverage, but mortgagee sales are included.
URA called for an open tender last month to appoint a consultant to review the methodologies for its office and retail property price and rental indices, but said it may not necessarily lead to a change in methodology for these indices.