News, analysis and opinions since 1998
Search the site
Get our free newsletter!
Hong Kong Land Lease Reform, Part 2
|Lease term||Sold from||Ending in||Ground rent||Location|
|999 years||1843-1898||2842-2897||Fixed at start||HK Island and a few former wharves in Old Kowloon|
|75+75 years||1899-1985||2049-2134 (renewable once in 1974-2060)||Fixed at start and re-fixed on renewal date to 3% of then rateable value||HK Island and Old Kowloon|
|99+99 years (rare)||2097-2183 (renewable once in 1998-2084)|
|50 years or until 30-Jun-2047 (whichever is later)||27-May-1985 onwards||30-Jun-2047 onwards||Fixed until 30-Jun-1997 then 3% of adjustable rateable value||New Territories (including New Kowloon) for leases granted before 27-May-1985, and everywhere after that.|
|50+50 years||2000||2100 (renewable in 2050)||3% of adjustable rateable value||Hong Kong Disneyland|
There are also 21-year leases granted on petrol filling stations, which generally get re-tendered on expiry for a premium, and 15-year leases granted to recreational clubs, which generally get renewed at 3% of rateable value unless the Government needs the site.
At least until World War II, ground rents were usually a significant part of the deal for new leases, typically around 5% of the premium. For example, Rural Building Lot 441 on Island Road, near Stanley, was auctioned on 3-Feb-1941 with an annual rent of $180 and a reserve price of $3925. Assuming a long-term discount rate of 5% (that is, $1.05 payable in a year's time is worth $1 today), then the fixed ground rents over 75 years had an initial net present value (NPV) of about 20.5x the ground rent, or about the same as the premium. In other words, about half of the land value was payable up front, and the rest over 75 years in fixed instalments. If we take a higher discount rate of 8% (anticipating higher inflation) then the NPV of the ground rents was still about 13.5x the annual ground rent.
However, after World War II, through inflation and the real appreciation of land values from a booming economy, ground rents gradually diminished in significance, and eventually they were just set at token levels, even for large sites. For example, part of Pacific Place, Inland Lot 8571, which was one of the last pieces of land to be leased before the Sino-British Joint Declaration took effect on 27-May-1985, has an annual ground rent of only $1,000 per annum until its first 75 years expire in 2060.
The 50-year leases granted today pay 3% of adjustable rateable value in ground rent, so in effect about 97% of the land value is the up-front premium and 3% is collected over the life of the lease. They have no right of renewal, but the current policy, which may change in future, is that unless "circumstances require in the public interest", the Government will normally grant a new lease upon expiry, at a ground rent of 3% of rateable value, with no premium. Therefore the working but uncertain assumption by many owners is that they have a perpetual lease on those terms, subject only to the usual eminent domain - the right of the state at any time to resume the land for roads, railways and so on upon payment of compensation, which right is contained in the lease and in law.
This high-premium-low-rent structure has the following consequences:
From the foregoing, the solution should be obvious. HK should shift the balance back towards ground rents, so that:
In order to eliminate inflation from the equation, the ground rent should be expressed as a percentage of adjustable rateable value, as it already is for leases granted since 1985. This means that in times of economic slowdown, when rents fall, ground rents would fall in line, as rateable values are reduced.
How far can we shift the balance without risking the land having negative value? Keep in mind that the Government only sells land, not the buildings, so the buyer still has to pay for the building. However, construction costs are only a small fraction of the average property value, perhaps as low as 10% in luxury houses or grade-A office districts, and up to 30% for apartments in distant parts of the New Territories.
As we will show below, ground rents could be set at 30% of rateable value without any material risk that owners would default and hand the land back, because it would always be worth someone's while to buy the property, even if they have to redevelop the building, and collect the rest of the rent as their gross return on investment (before expenses and tax). We propose a name for these new leases, the Ground Rent 30 or GR30 lease. In order to ensure that potential buyers of properties developed on such land are aware of their obligations, the law could be changed to require the type of land lease to be disclosed in all marketing material, including estate agents' adverts in the secondary market. Leases currently carrying a ground rent at 3% of rateable value could be called GR3 leases, and those older sites with fixed rents could be called GRF leases.
There is no reason to limit the terms of leases to 50 years if in practice the land will only be taken back for public interest reasons. We already have compulsory purchase systems for that, so the Government could start granting much longer or perpetual leases. Of course, there is the risk that the entire "capitalist system and way of life" will be scrapped by the Central Government after 2047 (Article 5 of the Basic Law only promises 50 years from 1997), but we are all in this together and proceeding on the basis that it won't. So the GR30 leases could be granted either for 999 years or in perpetuity - take your pick, because it makes almost no difference to theoretical value for the first 900 years or so.
The beauty of reforming the land revenue system through lease reforms rather than say, imposing a development tax, is that once signed, the terms of leases cannot be touched by future SAR governments, because they are binding contracts, whereas taxes could change by legislation or policy at any time. So as long as the new lease policy was in place, it would build up a stream of revenue, and a market of lower-premium properties, which would exist after any future policy change, and leaseholders would have certainty.
So what would be the effect GR30 leases on the value of developed properties and on current land premiums? Well, the easiest way to prove that in a pilot phase would be to take a pair of sites in say Kai Tak, West Kowloon, Aberdeen or Tseung Kwan O, of equal size and development potential (but not immediately adjacent, so there is no combination value), and give one a GR3 lease and the other a GR30 lease, then put them up for auction at the same time. The pilot phase should do that a few times, to eliminate other factors, so that people have confidence that this is not just a one-off result. No legislation would be needed to do this - the Government could start tomorrow. However, we can tell you now what this pilot phase would probably show, with some simple desktop analysis.
Theoretically, the value of any developed property is equal to the net present value of the future after-tax income the owner can earn from it. If "R" is the gross market rent per year, then the rateable value is typically about 80% of R, because the Rating & Valuation Department routinely underestimates market rentals when assessing rates, to avoid challenges. Based on the rateable value, the owner of a GR3 lease pays 3% in ground rent and 5% in government rates. So that 8% cost is equal to 6.4% of the gross market rent. Next, keep in mind that a property owner does not in reality receive 100% of the gross rent, because there will be periods when the property is vacant, between tenants or while being renovated, and every 30-50 years or so, while the building is being redeveloped. So let's deduct 10% from the gross rent for that, and another 10% for the cost of repairs and maintenance (including saving up for rebuilding). That leaves the owner with a pre-tax income of about 73.6% of R. Next, deduct profits tax at an assumed rate of 16.5%, then we have post-tax income of about 61.5% of R. Of course, there may be ways to enhance the return with tax structuring including bank loans and interest deduction, but you get the general idea.
Now, if we increase ground rents by 1% of Rateable Value, this will reduce post-tax income by (1-0.165)*0.01*0.8*R = 0.00668*R. That is a decrease by 0.00668/0.615, or about 1.09%. Since this affects all future net income by the same factor, the value of the developed property would come down by the same factor. So in summary, a 1% increment in ground rent equates to about a 1.09% decrease in property value.
However, remember that the value of the property comprises the value of the land plus the value of the building. The building value includes construction costs, plus a profit for the developer, plus financing costs during the construction period. Looking at the residential sector, we'll take the mid-points in the 2010 Q2 report by Davis Langdon & Seah for construction costs. We'll allow 6% for professional fees and add financing costs at say 5% p.a. for an average of 1.5 years (in a 3-year construction period). Let's assume selling prices 3 years from now of $5,000/gsf in outlying areas, $12,000/gsf for mid-range apartments in urban areas and $18,000/gsf for luxury apartments. Keeping it simple, we'll estimate $24,000/gsf for luxury terraced houses and $30,000/gsf for detached houses in luxury areas. Allow a profit margin of 20% for the developer, then discount the residue by 3 years at 5% p.a. to estimate the original land premium. These are consistent with the auction prices in 2010:
Now, based on our estimate of 1.09% reduction in developed property value for every 1% additional ground rent (as a percentage of rateable value), a GR30 lease, involving a 27% increment in ground rents compared with a GR3 lease, would involve a 29.4% reduction in sale prices for new property. Backing that out into land premiums, here is what would happen:
As the table shows, we estimate that land premiums in current auctions would be between 32% and 40% lower for GR30 leases than for the current GR3 leases, depending on the type of property. Note that because of lower premiums, the developer's dollar profit per square foot would be reduced, without any change in the percentage profit margin. This is because part of the land value would go directly to government as ground rents, by-passing the developer's mark-up. Furthermore, one of the effects of lower premiums ought to be a bit more competition as the barrier to entry will be lower, so margins should decline somewhat. While our example has focused on the residential sector, the office and retail (commercial) sector would work in the same way.
Of course, there is nothing magical about the 30% ground rent figure; the Government could easily opt for a higher number. New GR40 properties (with ground rent at 40% of Rateable Value) would sell for about 40.3% less than GR3 properties, and here is what the premiums would look like:
The land value is still substantial, even for apartments in outlying areas, so there would still be a very low probability of owners handing back their land or any of the units in a multi-owner building. In the event that an individual owner, for whatever reason, defaults on his ground rent, there is an existing process for dealing with that breach of the lease. In practice, the assets of a bankrupt property owner are sold off and the new owner still has the obligation to settle the overdue rent and remedy the breach.
Note that we are not expressing any view on future property prices in the above assumptions. For what it is worth, Webb-site has been saying for some time that we are in a bubble which will eventually burst, but that is a different story. All that matters for this article is that, in the event of a major correction, the GR30 model would still be robust.
One good thing about higher ground rents is that owners would have a stronger incentive to rent the units out rather than just hold them empty in the hope of capital gain, because they would still have to pay the ground rent. This would reduce complaints (justified or not) that speculators are "hoarding property", although we think that is their right to do so.
It is important to underline that no existing lease terms would be violated. All existing leases would be honoured. However, that still allows the Government to make the following changes to policy:
On 30-Jun-2047, all leases granted in the New Territories (including New Kowloon) before 1-Jul-1997 will expire. So will any lease in HK and old Kowloon granted between 27-May-1985 and the Handover. So this key date represents a mass expiry of 50-year leases, and after that date, some more 50-year leases (granted since the Handover) will expire each year. Most leaseholders are working on the assumption that their lease will be regranted on some unknown terms, possibly even the extremely generous GR3 terms without a premium, but there is no legal certainty of this - it is a policy at the whim of future HKSAR governments. As we get closer to 2047, for many leaseholders this uncertainty could otherwise become a major issue, particularly if a more socialist government was minded to exercise its right to resume the land or to levy a premium or higher rent for regrant. Remember that in 1982-3, 15 years before the expiry of the British lease on the New Territories, that there was a crisis in HK due to the uncertainty over what would happen, including for property lease terms. We should take early steps to avoid such uncertainty.
There is also the issue of existing 75+75 year leases. They fall into two categories - those which are in their first term, and those which are in their second term. For those which are already in their second term (call them 75-T2 leases), they have no right of renewal when their lease expires. Again, they are at the whim of policy, which is currently the extremely generous GR3 regrant without premium. As almost all such leases were granted after 1899, there are very few which expire before 2049, so this isn't such a big problem, but our research has uncovered a few 75-year leases on Rural Building Lots on the Peak and South side of HK Island which do expire sooner, probably because they were granted in exchange for surrender of older leases.
For those who are in their first term, the 75-T1 leases, they have a right of renewal for another 75 years at "such rent as shall be fairly and impartially fixed...as the fair and reasonable rental value of the ground at the date of such renewal", according to their standard lease terms. The legislation on what those words should mean, now known as the Government Leases Ordinance, was passed in 1973, and as we explained in Part 1, there was heated debate resulting in something far more generous to landlords than the natural meaning of the lease wording, namely that leases would be renewed at a ground rent of just 3% of rateable value at the time, fixed for the duration of the second term. There is nothing to prevent the Legislative Council from passing an amendment to this law. However, holders of renewable leases may raise questions of legitimate expectation if this legislative windfall is undone without reasonable notice, so it would be wise not to make dramatic changes.
To sweep all this up, we propose that the GR30 policy for renewals and regrants be phased in over 27 years at an increment of 1% p.a., both for non-renewable leases, and for those renewable leases which have not yet arrived at their single renewal date. This would be done by amending the Government Leases Ordinance to cover all the types of leases discussed above. If the amended law was enacted in 2012, this would mean that leases expiring in 2013 would get a GR4 lease, those expiring in 2014 would get a GR5 lease, and so on until those expiring in or after 2039 would get a GR30 lease. The law would provide certainty of renewal terms, and it would be open to leaseholders to make an early exchange of their lease (without accelerating the ground rent) so that they could lock in the new terms and remove any risk of future changes to the law. As noted above, one of the incentives for this early exchange could be that the GR30 lease would have either a 999-year or perpetual tenure, removing any renewal risk.
Because 2047 is still over 36 years away, switching from GR3 to GR30 in 2047 would not have a large impact on current valuations, if any. We are talking about around 30% of what the property will be worth in 2047, discounted back to the present day. Given a risk premium of say 4% above the risk-free rate (because returns on property are volatile), this computes to around 1/(1.04)^36 * 0.3 or around 7.3% of current value, but keep in mind that this could be offset by the removal of uncertainty by locking in the terms beyond 2047 now, for 999 years or in perpetuity.
Another option the Government could allow is what we will call the Premium Release Scheme, or PRS. Upon tender from an owner of a GR3 lease, the Government could pay a premium to convert the lease from GR3 to GR30. This would, as indicated above, allow the release of 30% or more of the value to the owner, while still giving her full leasehold tenure. The Government might want to limit the amount of capital made available for this, perhaps on a quarterly basis. Adopting a fair and market-driven approach, each quarter, the Government would announce how much it was willing to spend, and invite applications from owners stating how much, per dollar of rateable value, they would accept to exchange their GR3 lease for a GR30 lease. The Government would then accept the bids in order of ascending price until it had filled the quota.
For example, let's suppose that a unit in a building on GR3 land has a market rental of $30,000 per month ($360k p.a.), a rateable value of $288k p.a., and an $8m market value (implying a gross yield of 4.5% p.a.). Then the owner might be willing to switch from a GR3 lease to a GR30 lease, in exchange for (using the estimate above) about 29.4% of $8m, or about $2.4m. If her tender is successful, then in exchange for $2.4m, her ground rent will increase by 27% of $288,000, or $77,760 p.a., or $6,480 per month, adjustable in line with rateable value. The Government has swapped $2.4m of capital for an income of $77,760 p.a., or an initial rate of return of 3.24%. This may seem low, but keep in mind that the income is linked to rateable value, which should rise in line with inflation, so it is more like investing in a perpetual inflation-linked bond.
At 30-Sep-2010, the Government had about HK$1,113bn of liquid reserves - the fiscal reserves and the accumulated surplus in the Exchange Fund, HK's own quasi-sovereign wealth fund. To be clear, that is the amount in excess of what is needed in the Exchange Fund to back the HK$ peg. It also excludes the money sloshing around in various Government funds and statutory bodies such as the SFC, MPFA, URA, Post Office, Housing Authority, Innovation and Technology Fund and so on. Some of them place their funds with the Exchange Fund, currently at $68bn, while others place their funds without external managers under an investment committee.
The Premium Release Scheme could also apply to those who hold leases which are currently on fixed ground rents, including even 999-year leases. Again on a voluntary basis, these GRF leases could be exchanged for GR30 leases. The tender for these would be on the basis of 30% of adjustable rateable value, with a tabulated schedule provided by Government to deduct the net present value of the fixed rent stream (plus any renewal obligation). So the GRF leaseholder would be swapping his obligation to pay fixed ground rent for an obligation to pay GR30 rent plus receipt of a premium.
Apart from returning hoarded capital to the market in the Premium Release Scheme, the Government could also finance the scheme by issuing inflation index-linked bonds, or by securitising the ground rents (wrapping them into a bond) and selling the bonds back to the market as PRS Bonds. The ground rents, being linked to rateable value, are a natural hedge against inflation. For long-term investors, such as retirement funds, there is latent demand for a product which will provide inflation-linked income.
So at the same time as releasing equity to property owners, those who needed to plan for the cost of living could buy the PRS Bonds or alternatively just bonds with income linked to the Consumer Price Index. The US already has such index-linked products, known as Treasury Inflation-Protected Securities, or TIPS. The Hong Kong version could, of course, be called HIPS. Importantly, those who are retiring and will need an income stream to pay future ground rents on their GR30 lease could invest in PRS Bonds or HIPS as a hedge, perhaps using some of their MPF money, if there is anything left of it.
Government revenues have fluctuated wildly over the years, partly because of land premiums, and partly because of the increasing fiscal addiction to stamp duty, a frictional tax which puts sand in the wheels of the economy and should be abolished. However, the overall result has been a structural cash surplus at the expense of future income from ground rents. By shifting the balance towards ground rents, the Government would defer income and build a more stable revenue base to match its future expenditure. It would be less tempted to blow lump sums on wasteful projects.
The latest available figures from the Ratings and Valuation Department show that HK rateable premises had a rateable value of HK$368.7bn at 1-Apr-2009:
That's probably up about 10% by now, so call it $400bn. If just 10% of that could be converted into GR30 leases, it would bring in an extra 27% of $40bn, or $10.8bn p.a., adjustable with rateable values. As a rough estimate, the gross rental yield on this property might average 4.5%, so if rateable value is about 80% of gross rent, then the market value of the properties would be about $1,111bn in GR3 lease form. At an estimated cost of about 30%, conversion to GR30 leases would cost about $333bn.
Even if none of the existing base was touched, there are new land sales and lease modifications every year, including vast tracts of former farm land in the New Territories still to be developed. In the first 7 months of fiscal 2010-11, the Government has already raked in $37.9bn in land sales (excluding petrol stations), $11.7bn in private treaty grants to the MTRC and several billion in other land exchanges and lease modifications.
Just taking the first two of those categories, the premiums are $49.6bn. If the leases had been on GR30 rather than GR3 terms, then we estimate the premiums would have been lower by about one third, or about $16bn. Instead, the Government would have received additional ground rents on about 6m gross sq ft of property. Assume an average gross rent of $35/gsf per month. Taking a rateable value per square foot of about $35*0.8*12=$336/gsf, that's a ground rent increment of about 27%*$336*6m=$544m. The year still has 5 months to run, but it's been a good one, so let's just call it a year and assume that $50bn (in current dollars) would be the average in land premiums for the next 10 years under GR3 terms. By shifting to GR30, instead of $160bn of land premiums that it doesn't need, the Government would build up (in 2010 terms) an incremental ground rent of about $5.5bn per year, indexed to rateable values forever.
When the current bubble bursts, yields will go up and premiums will come down. Then the trade-off between land premiums and ground rents will look even more attractive from the Government's point of view, as there has not been as large an increase in rents as there has been in valuations. So they will forego less premium for each recurrent dollar of ground rents.
Hong Kong's high-premium low-rent land lease policy, which has evolved since World War II, and the developer cartel it has fostered, is at the root of so many economic and even social problems. Disregarding the current bubble, the system is still structurally crying out for reform. By 1-Jul-2012 (if not before), Hong Kong will have a new Chief Executive. The candidates are already emerging. Whoever wins may lead HK until 2022, just 25 years before the mass lease expiry in 2047. If any substantial change is to be made to the renewal and regrant policy for leases, it must be done during the next few years to have negligible impact on valuations. So candidates for 2012 should focus on this issue seriously and consider the many merits of rebalancing the land revenue system in their first term, particularly given the possibility that they will face a re-election contest by universal suffrage in 2017 rather than a 1200-member committee. The concerns of property tycoons must give way to the greater good of Hong Kong.
© Webb-site.com, 2010
Important notice: All material on this site, except where otherwise accredited, is copyright to Webb-site.com. Media and researchers are welcome to quote from articles on this site, provided that such quotation is attributed to Webb-site.com. The information in this site should not be relied upon by any person in making any investment decision. No responsibility or liability is accepted by Webb-site.com or any person related to it for any loss arising from or in reliance upon the whole or any part of the contents of this site. Persons who are in any doubt about an investment or potential investment should take professional investment advice. From time to time parties associated with Webb-site.com may own long or short positions in securities issued by or related to companies or governments on which we comment.