Problem
The freehold premium is usually discussed as a single number, while lease decay in 99-year leasehold property is commonly explained through a theoretical curve such as Bala’s Curve.
Neither framing fully explains where returns are actually concentrated within the lease cycle, why the measured freehold premium varies across locations and periods, or why the premium appears to have compressed over time.
Approach
Using a repeat-sales methodology applied to URA REALIS transaction records, this study analysed 155,301 completed buy-sell pairs to compare realised returns across tenure types and remaining-lease bands.
It also constructed an empirical lease-decay curve directly from transaction PSF data rather than relying solely on the theoretical curve.
The study additionally:
- decomposed returns using a 2x2 entry-type-by-tenure matrix
- examined district-level variation in the freehold premium
- compared realised performance across remaining-lease bands
- investigated the structural supply conditions underlying the freehold premium
The supply analysis highlighted that government land sales have not released freehold residential sites for some time. New freehold supply therefore now comes primarily from the en-bloc redevelopment of existing freehold land.
Why It Mattered
Freehold and leasehold properties produced almost identical median annualised returns:
- Freehold: 2.96%
- Leasehold: 2.93%
However, this aggregate result concealed a more useful segment-level finding.
Properties purchased with 70-80 years of remaining lease delivered a median annualised return of 4.89% among viable completed exits, the highest of any tenure and remaining-lease combination measured.
The empirical decay curve also showed that the market priced properties with approximately 70-75 years of remaining lease materially below what the theoretical curve would predict.
This suggests that the market may historically have over-discounted mid-cycle lease decay relative to the CPF usage and financing rules that actually apply at that stage of the lease.
The paper frames this as a historical segment-level finding rather than a recommendation concerning any individual property or project.
