NSO Malta | Valuation of Land underlying dwellings in Malta: A residual approach

Article: Valuation of land underlying dwellings in Malta: A residual approach

Author: National Accounts Methods, Standards and Sector Accounts

Published on: 1 September 2025

Abstract

Land, as the earliest and most historically recognised form of non-produced capital, plays a fundamental role in economic systems, providing a continuous flow of capital services to production. Given its significance, accurately estimating the value of land underlying dwellings is critical for both economic analysis and policy formulation. This paper aims to determine land valuation associated with residential properties, employing the residual method over the period from 1995 to 2024. The analysis draws on a comprehensive set of data, including national census records and dwelling prices. Furthermore, this paper also examines the trends of these valuations, offering insights into the evolution of residential land values over nearly three decades.

1. Introduction

Assets “are entities that must be owned by some unit, or units, and from which economic benefits are derived by their owner(s) by holding or using them over a period of time” (System of National Accounts (SNA) 2008, ¶ 1.46). Economic assets may be classified as either financial or non-financial assets. Within the frameworks of the SNA 2008 and the European System of Accounts (ESA) 2010, non-financial assets are grouped into two broad categories: produced and non-produced assets. Produced assets are non-financial assets that have been created as outputs from production processes falling within the production boundary defined by the SNA 2008. Conversely, non-produced assets are non-financial assets that have come into existence through means other than production processes.

“The first and oldest recognised form of non‑produced capital is land. Land is special in that, under good management, the value is assumed to remain constant from year to year except for the effects of inflation in land prices. That is to say, there is no depreciation of land and all the contribution to production can be regarded as income” (SNA 2008, ¶ 20.41).

Non-financial assets, or capital, play a dual role in the economy by providing capital services in production and serving as a store of wealth. Although land is classified as a non-produced asset, it is well established in the economic literature as a factor of production and therefore recognised as an asset that generates a flow of capital services into production. Land can be considered both an environmental and an economic asset, as it typically possesses economic value. Consequently, the valuation of land is important not only for compiling a comprehensive non-financial balance sheet but also for its application within the System of Environmental‑Economic Accounting.

Historically, there has been considerable debate, both within official statistics and among academics, regarding the best approach to obtaining reliable estimates of land. However, no consensus has been reached among National Statistical Institutes (NSIs) on a common or best practice methodology. A major challenge in land valuation stems from the frequent combination of land values with those of dwellings and other buildings and structures situated on it. The approach adopted in this statistical research paper is the use of an indirect estimation method, the residual approach, which will be discussed in detail in the following chapter. Additional details on the residual approach are available in the Eurostat-OECD compilation guide on land estimation (Chapter 6).

2. Methodology

The residual approach is widely employed by many countries as a method for estimating the value of land underlying dwellings, primarily due to the high accessibility and availability of data derived from real estate transactions. This method offers a practical solution where direct land valuations are not separately recorded or are otherwise difficult to obtain. In essence, the residual approach involves determining the total market value of a property, the combined value of the land and the dwelling constructed (\(CV_t^i\) ) upon it, and subsequently deducting the estimated value of the dwelling itself, based on the net capital stock of constructions, which excludes land at current prices, adjusted for depreciation and other relevant factors (\(C_t^i\)). The remaining value, after this deduction, is attributed to the underlying land (\(LV_t^i\)), at time t.

\(LV_t^i = CV_t^i – c_t^i\)

This approach is particularly favoured in official statistics and valuation practices because market data on the sale prices of residential properties are more readily available and dependable than isolated land values. Furthermore, by utilising information on net capital stock of constructions, it allows for a systematic and replicable method of disaggregating property values. However, the residual approach also requires vigilant observation to ensure that construction costs are accurately estimated and appropriately reflect the quality, age, and condition of the dwelling as otherwise, inaccuracies can significantly affect the precision of the derived land values. Despite its limitations, the residual approach remains as one of the most practical and commonly adopted methods for land valuation in the context of national accounts and property statistics.

3. Data sources and treatment

3.1. Dwelling prices

Dwelling prices from 2015 onwards were based on property transactions established on final deeds of sale and promise of sale agreements registered with the tax authority. Dwelling price data for the period from 1995 to 2014 were extrapolated using the Property Prices Index published by the Central Bank of Malta, which was based on advertised property prices. Such data was disaggregated by dwelling type, including apartments, maisonettes, terraced houses, and others, were available from 2000 to 2014. For the years 1995 to 1999, only the aggregate index was available, thus, the breakdown by dwelling type for this period was extrapolated accordingly.

From 2015 onwards, the same dataset utilised by the National Statistics Office (NSO) Malta to produce the Residential Property Price Index (RPPI) was used. Data was grouped according to price segments as provided by the Price Statistics Unit within the NSO. Segment groupings were determined following an analysis of median prices for several years. This level of disaggregation facilitated a more detailed analysis and enabled the identification of missing data and outliers. Missing data was catered for via interpolations in between the years, whereas outliers were identified manually through visual interpretations. This method was used due to the small number of observations available for some classes which prevented the use of more elaborate techniques. Extrapolations were based on observations with characteristics that were mostly similar to those of the missing data or outliers.

3.2. Dwelling stock

The valuation of the dwelling stock and Gross Fixed Capital Formation (GFCF) is intrinsically linked to the size, type, and condition of dwellings. Information on the number of dwellings, including their distribution by tenancy, number of rooms, dwelling type, age, and state of repair, was sourced from the Census of Population and Housing (CPH) conducted in 2005, 2011, and 2021.

Due to the unavailability of the full household-level database for 1995, published census results were used to estimate the required information on the dwelling stock. For occupied dwellings, the publication provided data disaggregated by tenancy, dwelling type, locality, and age. However, information on the number of rooms was only available by age group, without cross-classification by dwelling type or locality. As a result, the average dwelling size calculated for 2005 was applied. Likewise, since the state of repair was not collected for occupied dwellings in 1995, data from the 2005 CPH was used as a proxy. In contrast, data on vacant dwellings was available by dwelling type, state of repair, and locality, while age distribution was estimated using proxies derived from the 2005 CPH. Data in between censuses was interpolated. The total dwelling stock figures for each year were consistent with those used in the compilation of Real Estate Activities from the National Accounts Unit within the NSO. Distributions by age, quality, and price segment were interpolated based on the patterns observed in census years.

Since the CPH reports the number of rooms rather than the floor area in square metres (SQM), estimates of dwelling size for the benchmark year 2005 were derived by combining data from the 2002 Living Space Survey. This survey provided average dwelling sizes in SQMs by number of rooms. These averages were used alongside CPH data to convert the number of dwellings by tenancy, type, and room count into total SQMs for the relevant categories (Gross National Income Inventory Malta, 2023, ¶ 3.18.22). For 2011 and 2021, the total stock of dwellings in SQMs was estimated using a top-down approach. The calculation began with the stock in 2005, expressed in SQMs, which served as the benchmark. To this, the estimated SQMs of new dwellings constructed between 2005, and each reference year, were added, and classified by dwelling type: houses, apartments, and maisonettes. The resulting total was then distributed across the number of rooms and tenancies using the average SQMs per room derived from the total stock by dwelling type.

The RPPI dataset did not offer the same level of detail as the census in terms of dwelling quality, age, and type. To address this, a bridging methodology, as shown in Table 1, was developed to align census information with the available RPPI data. The bridge may vary across census years, reflecting differences in the scope and structure of census data over time. The total dwelling stock was considered in full, encompassing both occupied and vacant dwellings, including those that were owner-occupied and rented. The combined value of the dwelling stock was derived by applying the prices expressed SQM, to ensure that values are relative to the size of the dwelling rather than representing absolute prices.

The net capital stock at current prices or as sometimes referred to in this paper, the construction cost of the dwelling stock, was derived from the National Accounts Production Unit within the NSO (Gross National Income Inventory Malta, 2023, ¶ 3.18.22). Valuation of dwellings depends on their size, type, and condition (state of repair). While the CPH provided data on tenancy, number of rooms, dwelling type, and repair status, it lacked information on floor area in SQMs, which more accurately reflects size. To address this, data from the 2002 Living Space Survey, which includes average SQMs by number of rooms and tenancy type, was combined with CPH data to estimate total floor area. These floor areas were then converted into construction costs using cost per SQM estimates gathered from architects, differentiated by dwelling type and size. This information was used to derive the gross capital stock at current prices. The net capital stock at current prices was obtained after subtracting the accumulated Consumption of Fixed Capital (CFC) or depreciation. CFC was computed by using the straight-line depreciation model, which assumes that a dwelling depreciates by the same amount each year over the service life of that dwelling, which is assumed to be equal to 70 years¹

¹ This differs from what is stated in the GNI Inventory (Malta) ¶ 3.18.21 which was published in 2023. During the benchmark revision of 2024, the service life was revised to 70 years.

Table 1: Bridging methodology between Prices data and Censuses

Housing Table

	1995	2005	2011	2021
Age:
Pre WW2	Up to 1945	Up to 1945	Up to 1945	Up to 1945
Over 2 decades	1946 – 1970	1946 – 1990	1946 – 1990	1946 – 2000
Last 2 decades	1981 – 1995	1991 – 2005	1991 – 2011	2001 – 2021
Type:
Apartment/Penthouse	Flat/Apartment/Penthouse (with and without access to lift)
Maisonette	Ground floor tenement/Maisonette (with and without own airspace)
Townhouse	Townhouse/Terraced house/Semi‑detached house
Villa/Farmhouse/Bungalow	Semi/Fully detached farmhouse (unconverted), Semi/Fully detached house (including villa, bungalow), Fully detached house/Palace
Quality:
Good	In good state of repair/Needs minor repairs
Adequate	Needs moderate repairs
Poor	Dilapidated/Needs serious repairs

4. Results

Table 2 presents the detailed results for the combined value of the dwelling stock, the net capital stock at current prices of the dwellings stock, and the corresponding underlying land value. The ratios presented in Table 2 provide context on the shifting balance between land, construction, and total property value. The data captures market reactions to economic events, including the 2008 global financial crisis and the COVID-19 pandemic, while revealing longer-term structural trends in Malta’s property sector.

As the breakdown clearly shows in Figure 1, townhouses consistently contributed the highest total dwelling value from 1995 up to 2019. This reflects their deep-rooted presence in the Maltese housing stock, particularly in central and urban localities, and their strong demand among both local buyers and foreign investors seeking traditional or character properties. From 1995 to 2008, townhouses saw steady growth, briefly plateauing during the global financial crisis, from 2009 to 2012. Following this, values resumed an upward trend, albeit more modestly, until around 2019, when the rate of growth began to be outpaced by apartments and penthouses. This suggests that while townhouses retained high value, their relative market share began to diminish in the face of expanding apartment development and shifting demand toward more readily available housing formats.

Apartments and penthouses have shown a remarkably strong and accelerating growth, especially since 2013, overtaking townhouses in total value around 2019, reflecting a change in the composition of the dwelling stock. The continued increase through the COVID-19 period (2020–2021) demonstrates the resilience of this sector, which benefited from sustained investment activity, housing incentives, and the evolution of the short-term rental market. By 2024, apartments and penthouses dominate with the highest total value, reflecting both high development volume and increased price per SQM.

Maisonettes have maintained a steady upward trajectory throughout the period. While their total contribution to dwelling value is lower compared to townhouses and apartments, they have experienced consistent growth, particularly post-2014. On the other hand, villas, farmhouses, and bungalows are the smallest contributor in terms of total value throughout the period. Although there are spikes, particularly around 2021 – likely due to pandemic-driven preferences for properties with larger open spaces – the overall value remains modest in comparison (Micallef, 2021).

Chart 1. Combined value by dwelling type

in € millions

The net capital stock increased steadily from approximately €3 billion in 1995 to over €14 billion in 2024. The increase in net capital stock is the result of the increase in the stock of dwellings and inflationary pressure related to rising labour and material costs, regulatory changes, and intensified construction activity. Notably, there is an increase in total net capital stock from 2017 onwards, coinciding with a combination of broader economic developments, evolving policy frameworks, increased immigration, and the rise of short-term rental markets, as well as Malta’s post-EU accession development boom and an increase in both private and public sector building projects (European Commission, 2025). A further significant jump in net capital stock arose due to the COVID-19 pandemic, which inflated material costs globally, as reflected in sustained construction cost growth from 2020 to 2022, as shown in Figure 2 (The Construction Association, 2021; Kapadia, 2024).

By subtracting net capital stock at current prices from the total combined value of the dwelling, the land value underlying dwellings is derived. This figure expands sharply over the decades, from €4.7 billion in 1995 to €88.3 billion by 2024. This growth has been driven by land scarcity, high demand for central locations, and speculative investment. Figure 2 shows rapid growth in land value in the early 2000s, peaking at 26% in 2004, followed by a marked slowdown during the global financial crisis, from 2009 to 2012. These years reflect a market correction as investor confidence diminished, and global credit tightened. A second significant growth phase occurred between 2016 and 2019, marked by strong demand, increased development activity, and favourable lending conditions. Post-COVID, land value growth initially slowed, dropping to 5% in 2022, before gradually increasing to 7% in 2023, and then surging to 16% in 2024, reflecting renewed market confidence and investment activity.

Chart 2. Land value underlying dwellings

in € millions

Figure 3 provides insights on the growth in the dwelling stock in SQM which is relatively smooth across the 30-year period under review. In contrast, the increase in combined value per SQM is volatile and tends to follow the business cycle with a significant increase prior EU accession, a drop during the economic crisis of 2008 and an upward trajectory from 2012 onwards. As expected, the growth in the combined value largely follows the same path. It is noteworthy that the growth in net capital stock generally mirrors the trend of stock in SQM; however, from 2017 onward, net capital stock begins to increase at a significantly faster rate. Until 2021, data on construction costs was obtained directly through Bills of Quantities (BoQs) provided by architects; thus, no estimates were necessary for that period. In the absence of BoQs from 2022 onwards, the Harmonised Index of Consumer Prices (HICP) category 04.3, Maintenance and Repair of the Dwelling, has been used to estimate constructions costs across different dwelling types. It should be noted that HICP 04.3 has experienced a significant increase starting in 2022. This, combined with the inclusion of new dwellings and their associated total floor area, has further contributed to rising construction cost figures, which in turn have led to increases in the net capital stock at replacement costs. Data from 2022 onwards will be subject to revision upon the integration of new BoQs which are updated regularly by the NSO.

Chart 3. Growth in prices compared to the dwelling stock

The accompanying ratios provided in Table 2 give further insights into the evolving dynamics of Malta’s housing market. The land-to-construction cost ratio rose from 155% in 1995 to a peak of 627% in 2020, underscoring how the value of land increasingly outpaced building costs. This is especially critical in an island nation where developable land is limited and tightly regulated. Similarly, the land-to-combined value ratio increased from 61% in 1995 to a high 86% in 2021. This shift reflects the growing dominance of land as the primary driver of property value.

In contrast, there is a decline in construction costs to combined value ratio, from 39% in 1995 to 17% in 2024, as shown in Figure 4. This may primarily be due to Malta’s post-EU accession development boom and property prices rising faster than construction costs, rather than gains in construction efficiency. While building costs have gradually increased, land values and market prices have surged, especially post-2013. Additionally, smaller average dwelling sizes, particularly in apartment developments, may have contributed to lower construction costs per unit.

Chart 4. Ratio analysis

Table 2: Results

5. Data limitations

This statistical research is subject to several data limitations that may affect the accuracy and reliability of the findings. Firstly, any biases in the derivation of the net capital stock at current prices or the broader methodology used to calculate combined asset values can significantly impact the estimated value of the underlying land. This is because errors in the valuation of structures directly affect the residual land valuation. Additionally, if the combined value is based on data with systematic distortions, such as self-reported property values for tax purposes that are not independently verified, these values may be consistently underestimated due to possible tax evasion. Furthermore, inaccuracies can also arise from incorrect assumptions about the service lives and depreciation rates of dwellings and structures, which influence net stock estimates. Errors in the time series used for GFCF, or incorrect initial values, can further distort the estimated land values (Eurostat-OECD Compilation guide on land estimations, ¶ 6.108 – 6.110).

Another significant constraint arises from missing data, which necessitated the use of interpolation methods to estimate values where information was absent. Additionally, certain outliers within the dataset required extrapolation based on the most closely related category available, which may introduce a degree of uncertainty. A further limitation involves the absence of comprehensive dwelling price data for the period between 1995 and 2014. To address this gap, advertised property prices published by the Central Bank of Malta were used as a proxy. While useful, such data may not fully reflect actual transaction values and could lead to an under- or overestimation of market trends. Furthermore, the 1995 CPH did not collect data on SQMs per room or on the quality of dwellings. As a result, estimates for 1995 were derived using data from the 2005 census, which may not have accurately captured conditions a decade earlier. Moreover, there is a potential undervaluation of dwelling units, which could skew results, particularly in the context of land value underlying dwellings.

Nevertheless, the data required to apply the residual method is often readily available or measurable. The values of real estate properties are frequently observable or can be estimated for dwellings. (Eurostat-OECD Compilation guide on land estimations, ¶ 6.104).

6. Conclusion

This paper applies the residual method to estimate the value of land underlying dwellings in Malta from 1995 to 2024. Over this 30-year period, the total value of land increased substantially from €4.7 billion in 1995 to €88.3 billion in 2024. Townhouses initially dominated the housing stock in value terms, but apartments and penthouses overtook them by 2019 due to rapid development and market shifts. Construction costs also grew steadily, rising from €3 billion to over €14 billion, largely influenced by inflation, regulatory changes, and increased building activity. Key ratios underscore the growing prominence of land in property value. The land-to-construction cost ratio rose from 155% to 504%, while the land-to-total value ratio peaked at 86% in 2021. Conversely, the share of construction costs in total dwelling value declined from 39% to 17%. These trends highlight land’s increasing dominance as the main driver of property value in Malta’s constrained geographical context. Despite data limitations and the need for interpolations, the residual approach provided a replicable and insightful framework for land valuation, supporting both statistical analysis and policy planning.

In the ESA 2010 transmission program, Member States are obliged to compile Land as part of the Balance sheets for non-financial assets for the Household and Non-Profit Institutions Serving Households. The derivation of land valuation underlying dwellings was the first step. The office will focus on the valuation of land related to agricultural holdings and other land owned by households in their function as employers and own-account workers.

Valuation of Land underlying dwellings in Malta: A residual approach