Based on our research, HDB flats depreciate, on average, 0.67% per year across the first 4 decades of their lease term, all else being equal*.

During the recent National Day Rally speech, PM Lee devoted substantial airtime to speak about the new HIP2 and VERS initiatives. These policies specifically target the expiring lease issue with public housing. Since then, debates on this topic have dominated news headlines in Singapore.

While there were plenty of commentary around the potential depreciation effects of expiring leases, data scientists have been silent on this subject.

A key challenge of pinpointing the precise effects of expiring leases is that property values are driven by a multitude of other factors.

The public generally assumes that the older the property, the larger the depreciation effect from a shortened lease. However, this effect has been largely negated by a multi-decade rising market, a testament to Singapore’s wildly successful economic policies.

Therefore, while leasehold properties (be it private or public for that matter) may be theoretically depreciating assets if one only considers the expiring lease, the observable trend in home prices may actually be the opposite – a steady rise through the years along with economic growth and a well-balanced housing market.

In addition to the temporal effect are other confounding factors like location, flat size, floor difference, etc; all of which conspire to make any quantitative analysis tricky. So the key question here is:

What is the specific price impact of expiring leases on HDB flats, all else being equal?

In other words, how much value would say a 40-year old HDB flat recover if its lease is restored to 99 years?

Controlling the effects of time and location

 

In truth, it’s impossible to completely isolate the effect of expiring lease from all other effects in order to get to an “all else being equal” scenario. Nonetheless, if we can control the 2 largest factors, namely time and location, then we can get to some useful, albeit imperfect, insights. Fortunately, the real world offers us opportunities to do just that (or in data geek speak, a “quasi-experiment”).

A Quasi-Experiment is an empirical interventional study used to estimate the causal impact of an intervention on target population without random assignment. – Wikipedia

The map below shows a particular area in Kallang along Upper Boon Keng Road.

In this 100m radius circle, there are 4 HDB blocks with past resale transactions. Incidentally, they were built in different decades (70s, 80s, 90s & 00s). Searching through past data, we can group together transactions in these 4 HDB blocks that also occurred in the same quarter of a given year. For example:

Given that these HDB blocks are adjacent to each other, and that the selected transactions happen within the same quarter, we can make the broad assumption that any difference in the “price per square foot” is due primarily to the difference in their remaining lease term, and not location or timing.

For the purpose of this blog post, we have omitted handling of factors like floor level, flat size, interior condition, etc. While we recognize that these factors can have an impact on individual transaction prices, we find that when aggregating transactions across long enough periods of time and across the whole of Singapore, such effects are effectively spread out across both older and newer flats*.

To that end, we combed through all completed HDB blocks to identify pairs of adjacent blocks in whole of Singapore within 100m of each other, before segmenting them into buckets based on the difference in lease commencement year between the corresponding pair.

For instance, Blk 13 Upper Boon Keng Road’s lease commenced in 1999 while Blk 12 started in 1980. Therefore, this pair of adjacent HDB blocks will be captured into the “18-19 years” bucket.

Armed with this database of adjacent HDB block-pairs and their past transaction records, we can now explore the impact of expiring leases.

Results discussions

The chart below shows the percentage difference in price per square foot between transactions recorded in adjacent blocks in the same quarter, starting from 1Q13 to 2Q18 over a 5-year period.

Despite a few caveats (listed at end of this article), we think it’s still instructive to look at how the median percentage difference in prices become increasingly negative as HDB flats get older compared to their newer neighbors.

In fact, if we plot a straight trend line across these median points, we can infer the following:

HDB flats depreciate, on average, 0.67% per year across the first 4 decades of their lease term, all else being equal*.

Another way to look at this is to break up the chart into specific time frames. Here, we’ve divided the chart into 4 different “regimes” (basically longer time horizons) and look at the median percentage difference in prices across all observation pairs in these regimes:

0-9 years: In the first decade, there’s negligible downward pressure for older flats.

10-25 years: In the subsequent 15 years, older HDB flats start to display a non-trivial amount of depreciation, with a median discount of 12% compared to the newer neighbors.

26-37 years: In the next 13 years, older HDB flats face further downward pressure, with a median discount of 21%.

38+ years: Admittedly, we don’t have a huge amount of observable transactions beyond the 38 years range. However, we do see a marked jump in this median discount, pushing close to the 30% range. This could be due to the regime change in terms of both borrowing limit and CPF usage limit that kick in once the HDB flats cross over the 40 years mark.

Final Thoughts

One crucial point we need to highlight is that all transactions used in this study is based on resale prices. In contrast, Singaporeans who opt for BTO flats stand to enjoy substantial discounts from market rates in addition to grants they receive. Likewise, Singaporeans who opt for resale flats also enjoy various grants that effectively give them significant discounts as well.

In this study, we attempted to design our research in order to isolate the price effect of expiring leases, all else being equal. In reality, the real world is never as straightforward. Instead of finding a leasehold property depreciating from day 1, home owners may actually find that their home values continue to rise for years, as long as Singapore’s economic policies continue to work its magic. And in the later years, the recently announced HIP2 and VERS should also further mitigate the effects of expiring leases.

The hard truth is that Singapore is land-scarce and therefore, in our view, cannot afford a policy of “all land freehold” which allow the “property-owning class” to carry its asset-owning advantage across generations and severely impacts upward social mobility.

Having said that, HDB home owners (in fact, private leasehold home owners too) should be aware of the depreciation effect and take that into consideration when making their home purchase/sale decisions. In that light, we hope that this study provides some useful quantitative insights for the current debate.

[Added note on 6 Sept, 2018:

A common question since we first published this blog post 5 days ago is that why do we try to study the price effect of expiring leases in this “all else being equal” manner; after all, what everyone cares about is whether home prices appreciate or depreciate  when “all things considered”.

Here’s the way we think about this. Imagine a car; what everyone ultimately cares about is the speed of the car. But to the car engineer, he’ll certainly dive deeper to study the individual components – push factors (eg. power of the engine) and the pull factors (eg. frictional force on the tyres) – in order to fully understand how all these opposing forces interact to create the final speed of the car.

Likewise for this study. It’s true ultimately everyone cares about whether home prices appreciate or depreciate. But this shouldn’t stop us from studying the individual components – push factors (eg. economic growth) and the pull factors (eg. expiring lease). Our study is one where we drill down on a single pull factor (expiring lease). Yes it’s not the full picture but it’s certainly still a worthwhile endeavour, in our view.]


About UrbanZoom Research

UrbanZoom Research publishes data-driven insights on the Singapore property market.

We don’t publish regularly. But when we do, we endeavor to contribute meaningful and unbiased insights.

In short, quality over quantity.

Don’t want to miss our next “big thought piece”? Sign up for our newsletter HERE!

[mc4wp_form id=”98″]

 


*Some Caveats

As is often the case in real life scenarios, there’re some “gotchas” along the overall depreciation trend. For example:

  • There is wide distribution of results even within the same bucket. In the “10-11 year” bucket, older flats’ prices are, on average (or more precisely on median), 10% cheaper than the adjacent newer flats. But a closer look reveals that the top tip of the light blue bar actually cross above the 0% line into positive territory, which means a non-trivial number of older flats actually transacted at a higher price (psf) than neighbors which are 10-11 years their junior.
  • Ideally, we want a significant number of observations within every bucket. But in the real world, we have much fewer instances of adjacent blocks with large difference in lease commencement year. For example, we observed 560 pairs of adjacent HDB blocks in the “2-3 year” bucket while we could only find 5 pairs of HDB blocks in the “38-39 year” bucket.
  • This study draws transaction data for the 1Q2013 – 2Q2018 timeframe, which is generally in a down-market. While we feel this is a reasonable tradeoff between selecting relatively recent data and maintaining a decent observation pool size, further study should be taken to investigate into other time periods too, which could give new insights on how the expiring lease effect change over the years.
  • We opted to ignore other factors like floor level, flat size, etc, in order to not set too stringent a selection threshold which will severely impact the number of observations. We did run more complex studies where we control for these variables and the general result still holds.
  • This study concerns the price effect of expiring leases, which is not limited to public housing. In the realm of private housing, the discussion gets even more complex with freehold properties. We hope to publish another study in the future on this topic.