Capitalization of Urban Rail Investments in Land Values:
The Case of Athens
1. INTRODUCTION
It is widely recognized that transportation accessibility influences land prices and location choices (e.g. ANAS, 1982; DAMM et. al., 1980; McMILLEN and McDONALD, 2004). The practice of transportation planning, however, has mostly ignored the feedback effects of major transportation investments on urban development (e.g. WADDELL et. al., 2007). This overlooking hinders a proper coordination of land use and transportation policies. A possible reason for this omission is the observation of a fading impact of an already developed ubiquitous highway system. It is true that, in the scale of urban subareas, relative differences in road accessibility have declined over time. But still new, especially peripheral, highways complementing radial transit systems significantly affect road accessibility.
In view of the increasing congestion of radial arterials, urban rail investments improve the transit accessibility of the areas served, providing travel time savings to the affected population. The value of such time savings may be capitalized into land value gains in areas close to urban rail stations. However, the relationship between proximity to stations and land values is not consistent among empirical contexts. The context of the present study refers to new Athens metro stations opened just before the Olympics and the residential values around them. The methodology followed compares residential values in areas around stations as well as further away, thus isolating the unique contribution of rail accessibility to property values.
The aim of the study is threefold. First, to develop a conceptual framework of the study (section 2) and a typology of revenue generating mechanisms based on land value gains due to urban rail investments (section 3). Second, to investigate the existence and magnitude of the distinct effect of the improved urban rail accessibility on residential values in the Athens context. Third, to estimate the revenue potential of land value capture schemes in the surroundings of metro stations. Value re-capture would internalize some of the external benefits arising to private parties due to the said public investments.
In section 4 a single market approach is described, alternative data sources are discussed, the spatial and temporal frame of the data collection are selected and issues of data processing are addressed. Section 5 presents the empirical findings of a land value surplus and demonstrates the revenue potential of a particular value capture scheme. The paper ends with conclusions, policy implications and some suggestions for further research.
​
2. CONCEPTUAL FRAMEWORK
The value of a property equals the value of land and, if the site is improved, the value of the building. Buildings and other site improvements are reflected not on the land value but on the property value. Price-relevant structural characteristics refer i.a. to the size, age, floors, structural condition, ancillary spaces and outfit of the building. The land value depends on natural characteristics of the site (e.g. intrinsic usefulness for productive land use) as well as on locational characteristics. The rail accessibility, as a locational characteristic per excellence, impacts solely the value of land, not of the structure. Travel time savings may be capitalized into land values.
Residential development needs, apart from land, capital input for housing production (ANAS, 1982; ANAS, 2003). The elasticity of supply of land in a delimited area [1] is considered as inelastic, whereby the response of supply of capital (including bank loans) with respect to price is considered as highly elastic. The land share of the property value depend i.a. on locational characteristics, land availability or capital cost (e.g. interest rate). Planning constraints in a particular area (e.g. building regulations such as the maximum floor area ratio) alter the land share of the property value too. All other thing being equal for given locational characteristics, land prices increase with floor area ratio.
However, real world hand-overs of housing units relate to property rather than land. It is difficult in practice to disentangle the value of land covered by structures. An accurate value estimation of improved land implies high transaction costs. Therefore, even if land values matter, virtually considered in relevant studies are property values.
McMILLEN and McDONALD (2004) associate in an elegant way the change in value of a house due to a transit line with the anticipated change in house rent after the opening of the station. Their conceptual frame considers, however, the change in value only in the period before the opening, assuming perfect prior information about travel time savings, no lagged response and a truncated value adjustment. The authors apply a Fourier analysis of the temporal changes in house prices and transit station gradients. The 95% confidence interval for the Fourier estimate begins to reveal a price impact 4 years before the line opening. The full price impact of the transit station gradient takes place one and a half year after the line opening. They found also that homes located immediately adjacent to the rail line exhibit a small price discount rather than a premium, due to negative extemalities such as crowding, traffic, temporary disruptions and noise. In other words, the transit impact on housing prices does not seem to be a monotonically decreasing function of distance.
Following DIPASQUALE and WHEATON (1996), land prices consist of two parts. The first part is a mean price level representing a general price trend driven by short-term shifts of interest rates and imbalances between supply and demand for housing as well as by long-term cycles of economic growth and decline, demographic trends, changes in tax regulations and income. Housing supply and demand are mostly in disequilibrium because the former evolves as a longer-term and the latter as a shorter-term process. The second part is the relative price of land across locations in the city. Relative prices are based on comparative advantages of sites with differing natural and locational characteristics. Comparative advantages explain the intra-urban land price variation. The intrinsic (relative) price of each individual characteristic is of a permanent nature when reaching an equilibrium state, as long the respective comparative advantage sustains.
Locational site characteristics are decomposed to neighbourhood and accessibility characteristics. Expressed in positive terms, neighbourhood amenities are public goods such as green, open spaces, even an adequate garbage collection level. To put with are environmental goods such as quietude and clean air, socioeconomic attributes such as local income level or employment rate, as well as densities of various land uses.
Leaving aside the discussed road accessibility, rail accessibility contains a local as well as a regional dimension. Speaking of local rail accessibility, the range of influence of an urban rail station on land values is approximated by its catchment area within walking distance. In the case of the Athens metro Lines 2&3, when considering Line 1 as part of the metro system 57% of the riders access or egress L2&3 on foot (AMEL, 2005). People located closer to an urban rail station have, all other things being equal, a greater propensity to ride than people located further away. DILL (2003) finds evidence in the San Francisco Bay Area that for every 10% increase in walking distance, urban rail use declines about 10%. Note that in the Athens context, the walking time is consistently valued almost twice as onerous as in-vehicle time (SPANOS et. al., 1998; OASA, 2006). In other words, the local rail accessibility is perceived as more important in relative terms than the regional rail accessibility. The planning implication is the need for an improved walking environment, i.a. direct sidewalks to the stations, pedestrianizations, unbroken tiles and adequate lighting. In order to increase the rail use, it is more (cost) efficient to strive for a more pleasant and shorter access time for walk-in riders, than to invest on higher rail frequency or shorter station spacing (GIVONI and RIETVELD, 2007). In absolute terms, travel time savings achieved by using urban rail as compared to road modes are a consequence of improved regional rail accessibility. This may be especially the case when using peripheral stations with longer line haul distance to the city centre.
3. VALUE CAPTURE AND REVENUE GENERATING MECHANISMS
It is possible to distinguish two broad, non-traditional mechanisms to support financially urban rail investments. Both instruments rely on land value gains due to improved rail accessibility. The first pertains to transit-oriented joint development projects undertaken in proximity and in concert with urban rail investments. The second mechanism refers to revenues from charging land value gains around stations (DELOUKAS, 2007).
3.1 Transit-oriented Joint Development
A transit-oriented joint development is based on the idea of a mutually supportive land use and urban rail planning. In one of the variants which have been set up, local, regional or national authorities purchase with public funds low-priced land in the proximity of a planned urban rail line. They will then become beneficiaries of the land value gains generated by the rail project. The land acquired is used to support a transit-oriented joint development. Aimed are increased densities of residential and/or office development, while applying parking constraints and avoiding car-dependent land uses (e.g. regional shopping centers). The endeavour, typically run by a special purpose vehicle (SPV), has a long-term horizon, similar to that of public private partnerships in the rail sector (30 to 40 years). The revenue generated through the sale of high-priced landed property over time, is used i.a. for the repayment of the rail project debt. The concentration of development close to the transit system ensures multiple synergies. Higher densities enable increasing levels of ridership and an improved modal split. On the other side, a compact growth justifies a higher level of rail service, expressed in frequency and/or network connectivity. Two recent examples of this approach will be further on discussed, namely the Copenhagen metro and the tramway of Parla, south of Madrid. Both concern a greenfields development, but there exist cases of brownfields re-development and urban regeneration too.
The city of Copenhagen purchased during the nineties undeveloped land which has been transferred to a SPV. The latter has taken up a state-guaranteed loan to construct the transit system and related infrastructure. The SPV will sell the land in the mid/long-run to much higher market prices to private developers, and pay off the debt in the course of time. Targeted is the development of 3 mi. sq.m. floor space for 15.000 inhabitants and 50.000 job positions within a 40-year horizon. A basic aim is the dense concentration of mixed activities within walking distance of the stations, while expanding public space and green amenities. The scheme displays two capital advantages. First, the tax payers do not assume the burden of the metro investment. Second, the land value gain induced by a public investment goes back to the public wealth. Financing, land use and urban rail planning are integrated herewith in an innovative way.
The city of Parla (100.000 inhabitants) and the Madrid region realize through a PPP concession an 8,2 km-long tramway line costing 120 mi. € incl. rolling stock. A SPV shared by both authorities has purchased 2,9 mi. sq.m. land close to the tramway and will develop in stages up to 1 mi. sq.m. residential space for 35.000 inhabitants. The private concessionaire who constructed and operates the tramway will get availability payments during the 40-year concession. The business plan foresees a 100% operations cost recovery of the tramway operation. A funding source of the availability payments are the surplus revenues of the SPV generated from the sale of the housing units developed.
Modal choice exhibits a significant behavioural inertia. A behavioural change is difficult, even when the constraints of an initially car-dependent environment loosen through the provision of a transit service. Therefore, the behaviourally adequate sequence of planning actions to achieve a transit-oriented joint development is the following:
​
Urban rail infrastructure development at future growth poles
Land value gains effect
Development of activity centres with dense mixed uses
Relocation effect of households and firms
Attraction of newly induced traffic demand to urban rail
Modal split improvement effect
Figure 1 : (Chrono)logical sequence of planning actions
The particular sequencing of planning actions maximizes the desired impacts. A concentrated new development without prior rail provision and parking control management would enhance car dependency.
​
3.2 Value Capture Taxes
The taxation of land value gains due to public infrastructure investments is an alternative mechanism to support financially urban rail development. Three instruments of value re-capture are distinguished.
Planning gain
The planning gain is a one-off charge paid up-front by landowners or developers in case of a hand-over or of granting planning permissions around transit stations. A scaled fee may be set in relation to the value of the transaction or of the project to be developed. The fee encourages public-private collaboration and deters land speculation. It addresses sales transaction as well as building and zoning permissions. However, the limited circle of addressees makes the amount of revenues to be generated uncertain. Section 106 of the Town & Country Planning Act in the U.K. (1990) provides for agreements that developers will pay a charge as part of planning conditions. An example is the currently under construction 25km-long guided busway promoted by the Cambridgeshire County Council (500.000 population). Developers of almost 20.000 planned new housing units close to the guideway will contribute 25% of 265 mi. € investment costs in form of Section 106 charges.
Business charge
Urban rail proximity means a fast, frequent and reliable service for employees. Consequential are productivity gains for the respective firms. Firms located close to stations increase the pool of potential labour, customers, visitors or shoppers. An urban rail investment enlarges the market area and the turnover of adjacent retailers. DAMM et. al. (1980) report that the effect of distance to the station on property values seems to decline quite rapidly and, moreover, the impact on retail is stronger than on residential properties. The catchment area for residential uses is more stretched than for commercial uses, but the price gradient exhibits more peakiness in the latter case (DEBREZION et. al., 2004).
A business charge may be paid on a recurring basis by business occupiers of properties within walking distance of stations. The charging takes place after the station opening. It may be calculated as an increment on municipal taxes or business income taxes or even salary costs of employees. The latter way is followed in Paris where a ‘Versement de Transport’ is imposed on business occupiers. The charge is in fact justifiable but works as a disincentive for business activities.
Land value tax (LVT)
Land value is not the result of actions undertaken by the individual owner but of the community at large, e.g. in the form of urban rail investments. A LVT is, therefore, not a disincentive for economic activities (as are taxes on labour or capital) but rather an incentive for productive use of vacant land. As such encourages a compact urban growth, thus favours a better utilized public transport service provision. A LVT is justifiable because the public investment led to capitalized accessibility benefits, thus to ‘unearned’ land value gains. Note that the use of LVT revenues is a distinct, more contentious matter. However, an earmarking of the LVT revenues for further urban rail investments makes the LVT politically more acceptable. ANAS (2003) mentions that ideally land and building values should be taxed separately. This way is currently applied i.a. in Pittsburgh and other Pennsylvania cities. It is difficult to disentangle in practice the value of land covered by buildings, so that property taxes rather than land taxes prevail. However, land value is incorporated in property value, so an adequate property tax is an indirect way to capture land value gains. Los Angeles applies a differentiated property tax rate within ‘Special Assessment Districts’ close to new rail stations and uses the revenues to fund the transit system.
This kind of a location benefit tax may be paid on a recurring basis by property owners within walking distance of stations. The charging takes place after the station opening. It may be calculated as a differentiated tax rate and/or an increased tax base. A periodic re-appraisal of property values is necessary in this respect. This tax instrument exhibits stable revenue streams and stands in the focus of the present approach.
​
4. DATA COLLECTION
The effective sales price of a housing unit is the point where the willingness-to-pay of the buyer and the asking price of the seller intersect. In Greece, however, the effective sales price in most of the cases is not known publicly, for reasons of tax evasion. The tax assessed ‘transaction’ price is recorded and is in the typical case lower than the market price, even up to -50%. Unfortunately, the tax assessment is not lower than the market price in a systematic way. Appraisers commissioned by banks, assess property values in the process of provision of mortgage loans too. DAMM et. al. (1980) modelled with the same set of explanatory property characteristics, once the effective transaction prices and once the assessors’ values of multi-family housing units close to Washington metro stations. The respective hedonic regressions show that assessors understated by factor 9 the effect of the transit system on the values of the housing units. The authors conclude that market prices are subject to a wide range of influences that local assessors may understate or ignore. An alternative source of published information on property prices are asking prices of the sellers. Selling price is equal or exceeds the effective sales price. Published reports citing local estate agents state a discount varying between -5% and -10% as well as a tight correlation between sales price and selling price. The above-mentioned range of a negotiated rebate seems plausible in case of a successful transaction. A higher discount may reflect a longer period of the property remaining unsold. In business cycles where demand exceeds supply, a lower discount may be more acceptable.
The housing market is multifaceted because housing is a heterogeneous good. It is advisable to increase the data homogeneity and the accuracy of findings by focusing on a single but dominant market. In the Greater Athens Area apartments - as opposed to single-family homes - form to 71% the prevailing type of housing (MDS, 1998). Pertaining to housing hand-overs, local reports citing estate agents mention a similar share of used over newly built housing units (KATHIMERINI, 19/1/2008). Market power is wider distributed in the market of used apartments than of newly built apartments. In the latter case, developers influence the prices to a strong degree. Regarding housing occupation, market makers are to 73% owner-occupiers rather than tenants, proportion which remained stable over the last decade (MDS, 1998; OASA, 2006). Tenants tend to overweight accessibility characteristics of housing and live closer to work (ANAS, 1982). Owners trade-off in a more balanced way structural, neighbourhood and accessibility characteristics. This may reflect the fact that renter occupation compared to owner occupation is typically a shorter-term decision, less lasting and involving less transaction costs. Besides that, values of properties for sale may rise in anticipation of the station opening, whereby housing rents may rise only after the opening. The high owner-occupation rate, the fact that the bulk of the housing production and consumption in Athens till 2000 was virtually independent of bank credits as well as the low interest rates after the year 2000 are reasons that the local housing prices have not decreased in real terms over time.
The current terminal stations of the Athens metro Line 2 Agios Antonios and Agios Dimitrios have been opened by mid 2004, just before the Olympic Games. Their planning goes back to 1994, when the northwestern and the southeastern extensions of Line 2 have been first announced. Due to the fact that the base metro system was set in operation not until year 2000, uncertainty prevailed however. The financing of Ag. Dimitrios was secured by 2000, that of Ag. Antonios earlier due to an internal reallocation of funds [2]. By 2001 the preliminary design was finalized, the station sites were exactly known and the construction works have been tendered out. It may be assumed that the housing market began to react to a noticeable degree by then. The construction starts by 2001, the technical risks are better identified then, and due to the Olympics requirements a gradual certainty evolves about a timely completion. In the pre-opening phase, people do anticipate the travel time savings to come and this reflects on the land and property values. In the post-opening phase, the built-up effect takes place. People collect step by step more information about scheduled frequency, service reliability, fares or other dimensions of generalized time savings, and a consolidation process unfolds. In view of the exceptional local owner-occupation rate, there are good reasons to assume that the speed of adjustment of the housing market around the stations to the long-term equilibrium is high. In other words, the price adjustment takes place in the early operations phase already.
The northwestern section of L2 is 6km long and contains 7 stations, whereas the southeastern section is 5km long and contains 6 stations. The running time by metro from the center to Ag. Antonios accounts 10 min. and to Ag. Dimitrios 9 min. respectively. Due to the better road-based CBD accessibility of Ag. Dimitrios, Ag. Antonios achieves almost 20% higher time savings using metro to the center than Ag. Dimitrios. The parking availability and the metro service frequency for the wider districts Ag. Antonios-Peristeri and Ag. Dimitrios-Ilioupoli are comparable. The daily station boardings at Ag. Antonios amount to 33.000 and at Ag. Dimitrios to 28.000 respectively (2005). The within walking distance catchment area of a station is a function i.a. of the topography, traffic of the streets to be crossed, walking environment, land use mix and local demographics (e.g. age distribution, vehicle availability). In the local context, a conventional distance threshold of 500m from the station is quite adequate for the vast majority of the walk-in riders. The said threshold defines the impact area (=station area) as well as the control area (=nonstation area) of the study. The control area lies further away from the station but within the respective district, so that all but station proximity characteristics are comparable with the impact area’s ones. The study period covers the years 2003 to 2005. The start of operations coincides with the mid of the study period.
The single market investigated refers to used apartments offered for sale within the said period in the districts Ag. Antonios-Peristeri and Ag. Dimitrios-Ilioupoli. Relevant ads, published within the said period, contained information on district-level location, apartment size, floor level and phone contact. Many observations comprised the selling price and, if newly built, the latter fact (not further considered). Few observations contained age class and exact location of the structure. Any missing information on price, age and address was gained through follow up contacts. Selling prices have been normalized to a standard floor level and a standard age class. Price adjustment coefficients for tax assessment purposes variate in a curvilinear form for floor levels (with 2nd floor as standard level) and in a linear decreasing form for units older than one year -depending on the age class. However, the tax assessment coefficients are of long date (1987) and, in the case of floor level, refer to newly built structures too. Ground floor unit prices are idiosyncratic due to commercial uses and have been excluded from further consideration. The updated study coefficients contain a reduced variability. Table 1 shows the price adjustments by floor level.
Table 1. Floor level adjustment coefficients
​
​
The price adjustments by age class are shown in Table 2.
​
​
Table 2. Age class adjustment coefficients
The age distribution of housing units in Athens reveals a mean age of 28,6 years (OASA, 2006). The weighted adjustment coefficient of the said distribution is close to 0,70, so that the 3rd interval may define the standard age class.
The estimation sample comprises 385 observations decomposed to 96 and 73 in the impact areas of Ag. Antonios – Peristeri and Ag. Dimitrios – Ilioupoli respectively as well as 127 and 89 in the control areas of each district respectively. For the pooled impact areas particularly, the observations of the pre-opening and post-opening phase amount to 86 and 93 respectively. A substantial variation in property prices is observed. The normalized, inflation-adjusted selling prices (in 2004 values) range from 79.318 € to 242.919 € with a mean of 122.680 € and a S.D. of 18.832 €. The average size of the sampled apartments offered for sale amounts to 83 sq.m.
​
5. EMPIRICAL FINDINGS
When comparing both districts, housing units within 500m– buffer zone display a mean premium per sq.m. of 31% in Ag. Antonios – Peristeri and 25% in Ag. Dimitrios – Ilioupoli. The t-statistics are 2,55 and 2,13 for Ag. Antonios – Peristeri and Ag. Dimitrios – Ilioupoli respectively, so that the null hypothesis of equal impact and control area’s price means by district can be rejected at the α=0,05 significance level. The higher premium in Ag. Antonios – Peristeri is attributable to the greater (capitalized) time savings to CBD achieved by its residents compared to Ag. Dimitrios – Ilioupoli.
As already mentioned, the price adjustment due to improved rail accessibility is not instantaneous but stretches over time. The question arises if the capitalization rate within 500m– buffer zones is faster before than after the stations’ opening, reflecting then the power of expectations. Housing units within 500m– buffer zones display in the post-opening phase a mean 2% discount per sq.m. compared to the pre-opening phase. The t-statistic is 1,69, so that the null of equal pre- and post-opening price means for the pooled impact areas cannot be rejected at the α=0,05 significance level. The before-after comparison in the given context is inconclusive.
The aggregate increase of housing prices per sq.m. within 500m– buffer zones induced by the new metro connections, amounts to 28% when compared to the control areas over the pooled districts. The t-statistic is 2,91, so that the null of equal impact and control areas’ price mean can be rejected at the α=0,01 significance level. In view of the dominance of the housing segment investigated, the results represent a trend in the housing market around both terminal stations. Whereas the results are transferable to new stations with similar locational characteristics, one should be careful to generalize them over the full set of metro stations because of their dissimilar accessibility and neighbourhood characteristics. Further analysis has a merely illustrative character in this respect.
​
Land value gain around stations
The land share in the production of new multi-family buildings may vary from 3 to 5 %-age points within the same municipality and time period, depending on a multitude of factors described in section 2. Following estate agents asked just after the Olympics, the land share in both districts was about 37% on the average. The estimated 28% increase of property value around the stations, would increase then the subordinated land value for newly built structures by (1/0,37=) 2,7 times, i.e. up to 76%. For the existing stock of aged buildings and the standard age class coefficient 0,7, the land share raises up to 53% on the average, and the subordinated land value increases to 53%. The magnitude of the land value gain due to improved rail accessibility indicates the pressure for site re-developments or new developments on vacant land around stations. Site re-development may take the form of land use conversion (KNIGHT and TRYGG, 1977) and/or the form of demolition of the existing structure and replacement with a high-quality one. CLARK (1989) turned the attention to the consequences of an emerging rent gap, while emphasizing the commodity character of housing. A site re-development is the likely the lower the difference of the building value and the land value (‘highest and best use’ of the site).
Land rent in Athens
Following a published journal supplement (KATHIMERINI, 15/9/2004), the tax base of the existing housing stock in the Greater Athens Area (GAA) amounts to 233 bi. € by 2004, i.e. three times the Greek stock market capitalization by that time. Assuming a 53% land value share, the respective subordinated land value amounts to 123 bi. €. The minimum nominal land rent set for tax assessment purposes amounts to 3,5% p.a. In other words, the minimum annual wealth for owner-occupiers and real rent for lessors amounts about 4,3 bi. € p.a. This is equivalent with 6,1% of the revised regional GDP of Attica by 2004. Note that the effective rent of the market valued stock is considerably higher.
Fiscal effect of land value gains around stations
The state budget expects for 2008 390 mi. € property tax revenues from GAA (out of 890 mi. € nationwide). Following Metro Development Study, the GAA covers 50.600 ha. The 500m– buffer zones of the current L1&2&3 metro stations cover 2.500 ha., i.e. 5% of GAA. Making simplifying assumptions about spatially uniform tax revenues collected over GAA and uniform 28% property value increases due to improved rail accessibility over the full set of station areas, the change of tax revenue collected (TRC) may be calculated as the product of the differentiated tax rate (TR) and the re-assessed tax base (TB):
ΔTRC = ΔTR * ΔTB (1)
Selecting a reasonable tax rate 0,4% [3] for properties within 500m – buffer zones around metro stations, Eq. (1) comes to 57 mi. € p.a. This amount matches about 50% of the annual surplus land rent around stations due to improved rail accessibility. If actually recouped, then a fair rule-of-half would apply, leaving 50% of the external benefits (‘windfall gains’) to the benefited landowners. The tax revenue generated over time under this value re-capture scheme, may be earmarked and used i.a. for the repayment of a metro extensions debt. The above estimated surplus tax revenue collected could repay within 35 to 40 years 20kms of metro extensions.
​
6. CONCLUSION
A basic research contribution of the paper is the distinction of local and regional rail accessibility. Local proximity to the station is a necessary condition for land value gains. A policy implication is the need for a more walkable environment and a pedestrian-friendly urban design in the station areas. The sufficient condition for land value gains are travel time savings achieved by using metro, i.e. an improved regional rail accessibility.
The impact of rail accessibility on residential values around stations revealed to be positive and significant in the Athens context. The distance of a station to the CBD has a significant positive impact on residential values when greater travel time savings are achieved, all other things being equal. The latter provision is important because neighbourhood modifiers such as differing land use densities and building regulations do impact land value share and value gains.
A typology of revenue generating mechanisms relying on land value gains due to improved urban rail accessibility has been developed. Relevant case studies pertain to transit-oriented joint development as well as to distinct charging forms of land value gains.
Another contribution of the paper is the interconnection of land and property values in the context of capitalized accessibility benefits. A practical way has been proposed to deduce land value gains from property value changes. Property taxes equivalent to hypothetical land taxes have been suggested in this respect. Taxation of land value gains is fair, efficient and, moreover, welfare-neutral (as opposed to the welfare-reducing taxation of labour and capital). The potential of a land value re-capture scheme through regular property taxation around stations has been estimated, taxing away about half of the surplus land rent effectuated by improved rail accessibility.
Further research would extend the range and the variability of the station areas as well as of the site characteristics surveyed. Determinants of residential values are, apart of accessibility, neighbourhood characteristics and urban amenities. A hedonic price approach could explain then the land value of sites in terms of all contributing site characteristics (ROSEN, 1974).
The methodology and the study results are considered to be of wider interest for planning agencies, transit authorities and policy makers.
​
Footnotes​
[1] In built-up areas around new Athens metro stations has been observed an increasing shortage of housing offered for sale in the period after the station opening. Published reports citing estate agents mention two years after the station opening by year 2000, a decrease of residential space offered for sale, e.g. for Ampelokipoi to -24%, Neos Kosmos to -44%, Dafni to -61% (KOSMOS TOY EPENDYTH, 15/3/2003).
​
[2] Ag. Antonios was first assigned to the Peristeri extension but was later formally re-assigned to the base system as a result of budget considerations.
​
[3] The property tax rate was set at 1/1/2008 to 0,1% for natural persons and 0,6% for legal entities with an exception for main homes of owner-occupiers and a tax exemption amount of 300.000 €.
​
Author
Alexandros Deloukas
​
Disclaimer
The views expressed in this paper do not represent official views of Attiko Metro A.E., the author being solely responsible for the views expressed herein.
References
AMEL (2005), ‘Development and Handling of O/D Matrix Software and Completion of Scheduling Software’, Attiko Metro Operating Company, Athens
ANAS A. (1982), Residential Location Markets and Urban Transportation, Academic Press, N.Y.
ANAS A. (2003), ‘Taxes on Buildings and Land in a Dynamic Model of Real Estate Markets’, in: D. Netzer (ed.) The Property Tax, Land Use and Land Use Regulation, pp. 6-36, Edward Elgar Publishing, Surrey
CLARK E. (1989), The Rent Gap and Urban Change: Case Studies in Malmoe 1860-1985, Lund University Press, Lund
DAMM D., S. R. LERMAN, E. E. LERNER-LAM and J. JOUNG (1980), ‘Response of Urban Real Estate in Anticipation of the Washington Metro’, Journal of Transport Economics and Policy, Vol.14, pp. 315-335
DEBREZION G., E. PELS and P. RIETVELD (2006), ‘The Impact of Rail Transport on Real Estate Prices: Empirical Study of the Dutch Housing Market’, 85th Transportation Research Board Annual Meeting, Washington D.C.
DELOUKAS A. (2007), ‘Alternative Ways of Funding Urban Transit Systems through Land Value Gains and Joint Development’, TranSUrban-INTERREGIIIC International Symposium, Volos, Magnesia
DILL J. (2003), ‘Transit Use and Proximity to Rail’, Transportation Research Record, Vol. 1835, pp. 19-24
DIPASQUALE D. and W. WHEATON (1996), Urban Economics and Real Estate Markets, Prentice Hall, N. J.
GIVONI M. and P. RIETVELD (2007), ‘Developing the Rail Network Through Better Access to Railway Stations’, paper presented at the European Transport Conference, AET, Noordwijk
KATHIMERINI daily newspaper, Greek edition dated 15/9/2004 and 19/1/2008, Athens
KOSMOS TOU EPENDYTH weekly newspaper, Greek edition dated 15/3/2003, Athens
KNIGHT R. L. and L. TRYGG (1977), ‘Evidence of Land Use Impacts of Rapid Transit Systems’, Transportation, Vol. 6, pp. 231-247
MDS (1998), Survey of Household Travel Characteristics, Technical Report 0036, Attiko Metro S.A., Athens
OASA (2006), Study of Trip Origins/Destinations, Technical Report and Analytical Database, Athens Public Transport Authority, Athens
McMILLEN and J. McDONALD (2004), ‘Reaction of Home Prices to New Rapid Transit Line, Chicago’s Midway Line, 1983-1999’, Real Estate Economics, Vol. 32, pp. 463-486
ROSEN S. (1974), ‘Hedonic Pricing and Implicit Markets: Product Differentiation in Pure Competition’, Journal of Political Economy, Vol. 82, pp. 34-55
SPANOS I., A. DELOUKAS and A. ANASTASAKI (1998), ‘A Stated Choice Experiment: Value of Travel Characteristics in the Context of Attika’, in: Papageorgiou M. and A. Pouliezos (eds.), Proceedings of the 8th IFAC Symposium on Transportation Systems, Vol. 1, Pergamon, Elsevier
WADDELL P., G. F. ULFARSSON, J. FRANKLIN and J. LOBB (2007), ‘Incorporating Land Use in Metropolitan Transportation Planning’, Transportation Research Part A, Vol. 41, pp. 382-410
​