Outlook on Empty Property Rates
Rate to climb to almost 25% in 20 years at current pace
Hidetaka Yoneyama, Senior Research Fellow
July 9, 2013 (Tuesday)
All across Japan, the increase in empty properties has become more noticeable. In order to guard against the dangers of dilapidated empty homes collapsing, a movement to establish empty property management regulations has become active in many regions. Furthermore, regions which are troubled by their declining populations have created an “Empty Property Bank” on the Internet and list information on empty properties in order to bring in even a few new residents. Using such measures, examples of dangerous empty homes being torn down and new residents found for empty properties have been on the rise, albeit gradually.
However, the problem is expected to become ever greater as Japan’s population continues to decline. What will happen to the empty property rate in the future? Below, let us perform some hypothetical calculations based on a given set of assumptions.
1. Demand Assumptions
Starting with the most recent “Housing and Land Statistics Survey,” that of 2008, let us estimate the numbers after 20 years’ time. When predicting future empty property rates, the important factors are housing demand (number of households) on the demand side and new houses being built (number of new house groundbreakings) and houses torn down (number of demolitions) on the supply side.
First, on the demand side, it should go without saying that changes in population will affect the number of households. For this variable, we will use the estimated number of households in 2030 given by the National Social Security and Population Research Lab (“Japan’s Estimated Number of Households”, March 2008). According to those numbers, the number of households will continue to increase until 2015, at which point it will begin to decrease. While Japan’s population has already reached its peak in December 2004 and is currently declining, the reason the number of households yet continues to increase is that households are growing smaller and the number of single-person households is increasing.
2. Relationship of Housing Stock and Flow
Before explaining the supply side, let us look at the relationship between housing stock (total number of houses, according to MLIT’s “Housing and Land Statistics Survey”) and new home groundbreakings (flow, according to MLIT’s “Construction Groundbreaking Statistics”), which will be the basis for our calculations (see Figure 1). The house stock number is based on the Housing and Land Statistics Survey, which is carried out every five years, and so we can calculate how much stock has increased over that period. Conversely, the number of new home groundbreakings is published each month in the Construction Groundbreaking Statistics, and so five years of groundbreaking data can be similarly accumulated.
There is a small inconsistency between the number of new groundbreakings and the stock increase of this five-year period, which is due to demolished buildings. While new houses are built during these five years, some old houses are torn down as well, which is why the increase in stock is smaller than the number of new groundbreakings. In the case of rebuilding, a new house is built after the old one is destroyed, and so it is counted as both a demolition and a groundbreaking (1 house destroyed followed by 1 house built equals 0 net demolitions). In a case where the land is left vacant or used for other purposes after a house is torn down, it is counted as a demolition only (1 net demolition). Thus, the difference between new groundbreakings and increase in stock is the number of pure demolitions (net demolitions). From 2003-2008, the net number of demolitions was 2,175,000.
Table 1: Relationship between stock and flow in housing
| Initial Housing Stock | 5-year Increase in Stock | New Groundbreakings | Net Demolitions | Avg Rate of Annual Net Demolitions (%) | Number of Households | Number of Empty Properties | Empty Property Rate (%) | |
| A | B | C | C-B | ((C-B)/5)/A | D | E | E/A | |
| 1973-78 | 31,059 | 4,392 | 7,254 | 2,862 | 1.8 | 29,651 | 1,720 | 5.5 |
|---|---|---|---|---|---|---|---|---|
| 1978-83 | 35,451 | 3,156 | 6,196 | 3,040 | 1.7 | 32,835 | 2,679 | 7.6 |
| 1983-88 | 38,607 | 3,400 | 7,147 | 3,747 | 1.9 | 35,197 | 3,302 | 8.6 |
| 1988-93 | 42,007 | 3,872 | 7,628 | 3,756 | 1.8 | 37,812 | 3,940 | 9.4 |
| 1993-98 | 45,879 | 4,367 | 7,269 | 2,902 | 1.3 | 41,459 | 4,476 | 9.8 |
| 1998-2003 | 50,246 | 3,645 | 5,929 | 2,284 | 0.9 | 44,360 | 5,764 | 11.5 |
| 2003-08 | 53,891 | 3,695 | 5,870 | 2,175 | 0.8 | 47,165 | 6,593 | 12.2 |
Figure 2: Years elapsed since construction and rate of homes survival
Source: MIC “Housing and Land Statistics Survey”
Note: the number of houses built in the 1950s and surviving after 8-17 years is unknown, and so this value was calculated as the number of surviving houses built between 1956-65 and surviving after 8-17 years.
Furthermore, the MLIT’s “Statistical Survey of Demolitions” also tells us the gross number of demolitions: 7,325,000. Net demolitions accounted for only 29.7% of all demolitions, which means that approximately 70% of homes are rebuilt after being torn down and therefore were not counted within net demolitions.
We can also calculate the annual average rate of net demolitions by dividing the 5-year number by 5 to give us the annual number and further dividing by the initial number of stock. This rate was 0.8% from 2003 to 2008. This value is tending to decrease, but the main reason for this is that stock is increasing every year (denominator factor). If the number of pure demolitions remains unchanged but housing stock continues to increase annually, the pure demolition rate will continue its decreasing trend.
Another possible reason for a reduction in rate (numerator factors) is that the quality of stock has improved over time and it therefore lasts longer, and the number of houses which are torn down after a short period of use has decreased. If we look at how the number of surviving houses built in each decade decreases over time (survival rate for years elapsed; houses built 3-12 years ago are assumed to be 100%) (see Figure 2), we see that more recently built buildings tend to have a greater survival rate and are demolished at a lower rate. This suggests that the lifespan of houses has extended and they are used for longer.
Alternatively, in the case of rebuilding, the landowner could demolish a single house and subsequently divide the land and build more than one house, thus in fact increasing stock. If such cases multiplied, the net reduction in stock could become less. Finally, if a house is no longer being used and is left empty, and if the number of houses which are rebuilt or made into vacant lots decreases, this value will indeed fall as a result.
It is unknown which of these factors has the greatest effect, but one may infer that the denominator factors are the base for the gradual reduction while the numerator factors are additive in nature.
3. Supply Assumptions
In order to predict the future number of stock it is necessary to assume how new groundbreakings and stock numbers will increase. The increase in stock can be calculated using new groundbreakings as a base and assuming the number of net demolitions. The number of new groundbreakings was 788,000 in 2009, 813,000 in 2010, and 834,000 in 2011 — varying around 800,000 in recent years. Let us assume two cases: the case in which this number remains constant (800,000 new groundbreakings per year), and the case in which the status quo holds until 2013, after which the number decreases by 25,000 each year until it is halved to 425,000 in 2028.
Two possible factors which may contribute to the decrease in new groundbreakings are the decline in number of households and the increased use of pre-owned homes. However, as previously mentioned, because the decrease in the number of households during the time period of our prediction will be only slight, such a value will not be achieved unless use of pre-owned homes increases. If sales of pre-owned homes increases to supplement the 50% reduction of new groundbreakings, the current (2008) pre-owned home sales rate of 13.5% (number of pre-owned homes sold/(number of pre-owned homes sold + number of new groundbreakings)) may climb as high as 50% in the end, based on our assumptions.
For demolitions, let us also assume two cases: the status quo case (2,175,000 net demolitions from 2003-2008) and the case in which 1% of the number of empty properties in 2008 (7,568,000) is added to the net demolitions of the previous five years. This assumes that net demolitions increase due to the removal of empty homes. In this case, net demolitions will increase by 757,000 compared to the status quo over the prediction period (2008-2028). Furthermore, under this assumption, the average number of annual net demolitions during the prediction period is maintained at exactly 0.8%: the 2003-2008 average (given that new groundbreakings remain constant).
Using these assumptions, the number of empty properties is calculated as (housing stock – number of households), which value deviates slightly from the actual number. Therefore the ratio of (stock – households) to empty properties in 2008 is used as an adjustment factor for calculating the number of empty properties after 2013.
4. Empty Property Rate in 20 Years
New groundbreakings and net demolitions each have two assumptions, which results in four combinations of these cases.
Case 1: Both new groundbreakings and net demolitions remain at the status quo and the empty property rate increases to 23.7% in 2028. This is a 10.6% increase over the 13.1% of 2008.
Case 2: Net demolitions remain at the status quo and new groundbreakings are halved. The empty property rate rises to 17.1%, a 6.6% decrease compared to case 1. Reducing new groundbreakings by 50% is very effective.
Case 3: New groundbreakings remain constant while net demolitions are increased due to removal of empty properties. In this case, the empty property rate is 22.5% in 2028, a 1.2% reduction compared to case 1. Removing empty properties will not be a large factor in lowering the empty property rate unless a very large number are removed.
Case 4: New groundbreakings are halved and net demolitions are increased, resulting in an empty property rate of 15.7%, 8% less than in case 1. This case is a combination of the conditions in cases 2 and 3 and so the reduction is approximately the sum of each of those cases’ reductions. According to this calculation, the empty property rate will rise by only 2.6% compared to 2008, which suggests that constraining new groundbreakings (i.e., using pre-owned homes) and removing empty properties are both necessary to stave off an increase in empty properties.
5. Support Needed for Buying Pre-owned Homes
Current countermeasures against empty properties involve tearing down dangerous empty homes or calling emigrants from urban areas to live on empty properties in rural areas; these efforts are centered on especially problematic properties or regions with especially high empty property rates.
However, this alone will have a limited effect, and it will be difficult to arrest the rising empty property rate. Fundamentally, the use of pre-owned homes must be promoted more. Japan should create a mechanism whereby buying pre-owned homes is more advantageous than newly built homes, such as receiving tax benefits when buying and renovating a pre-owned home or subsidizing renovations.
It is expected that long-term grade-A homes will spread and high quality housing stock will continue to increase in the future. And, as previously mentioned, the number of households will continue to decrease due to the low birth rate and ageing society. In such a changing housing market, it is necessary to shift the direction of housing policy such that the next generation can inherit high quality housing stock.
