Policy Research Working Paper 10998 Reforming Land Valuation and Taxation in Ukraine A Path towards greater Sustainability Fairness, and Transparency Klaus Deininger Daniel Ayalew Ali Eduard Bukin Andrii Martyn Development Economics A verified reproducibility package for this paper is Development Research Group available at http://reproducibility.worldbank.org, December 2024 click here for direct access. Policy Research Working Paper 10998 Abstract The shift from administrative to market-based valuation of prices that are 25 to 33 percent above current valuations assets is a key part of Ukraine’s transition from a planned to hide vast interregional differences. Given their proximity to a market economy. This shift will improve the functioning active conflict, predicted prices in the East and South are of financial markets and the ability of local governments to close to or even below the normative value, but they are 80 obtain revenue effectively for local service provision. This percent above it in the West of the country. A transition to paper describes the rationale and evolution of Ukraine’s reg- market-based valuation is thus a precondition for fair and ulatory framework for market-based valuation. The analysis equitable taxation and incentives for productive land use. uses prices and publicly available parcel attributes for the By aligning with globally accepted standards for banking nearly 200,000 agricultural land sale transactions during regulation and improving credit access in areas where land 2021–24 to estimate a land price model that is then used values have increased, the transition could also affect the to predict prices for the roughly 20 million hectares of com- speed and quality of reconstruction. The paper discusses mercial agricultural land in the country. Mean predicted legislative steps to move in this direction. This paper is a product of the Development Research Group, Development Economics. . It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at kdeininger@worldbank.org; dali1@worldbank.org; ebukin@worldbank.org, and agmartyn@gmail.com. A verified reproducibility package for this paper is available at http://reproducibility.worldbank.org, click here for direct access. RESEA CY LI R CH PO TRANSPARENT ANALYSIS S W R R E O KI P NG PA The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Reforming Land Valuation and Taxation in Ukraine: A Path towards greater Sustainability Fairness, and Transparency Klaus Deininger Daniel Ayalew Ali Eduard Bukin Andrii Martyn Keywords: Mass land valuation, Ukraine, property taxation, geospatial modeling, land cadaster, economic reforms JEL codes: Q01, D23, H20, Q15, R14 The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Funding support from the European Union (ENI/2017/387-093 and ENI/2020/418-654) is gratefully acknowledged. We thank Denis Bashlyk, Florian Blum, Ben Hell, Thea Hilhorst, Roman Hrab, Sergyi Kubakh, Vasyl Kvartiuk, Daria Manzhura, Sevara Melibaeva, Irina Schuman, Karlis Smits, and Taras Vysotskyi for helpful discussions and insightful comments and as well as Tania Khorzovskaya for outstanding coordination and administrative support. Reforming Land Valuation and Taxation in Ukraine: A Path towards greater Sustainability Fairness, and Transparency 1. Introduction The signals provided by market prices for land and property are important for many reasons, including for national accounts and consumer price indices to indicate macro-economic activity, fine-tune monetary policy, measure and regulate the financial sector’s risk exposure or stability, and inform individual decision- making (European Comission 2013). Accurate property valuations are a key precondition for investment and financial market operation, transparency in real estate markets, and taxation of capital gains or losses due to land price changes as well as recurrent property tax valuation as a basis for the social contract and provision of services that, in turn, tend to be capitalized into land values. Ukraine has traditionally valued rural land based on a Soviet-style normative monetary value (NMV) that is based on the productive capacity of land indexed to inflation. Since independence in 1991, it has moved gradually towards building the institutional and legal infrastructure for a land market that could support land valuation methods that are more flexible in terms of reflecting actual conditions. Building on reforms enacted in 2020 data on agricultural land sales are available from July 1, 2021, when the country allowed agricultural land sales and put in place a sophisticated system for public monitoring of land relations. This system leverages central electronic information platforms, including the State Land Cadaster and the State Register of Real Property Rights, to increase transparency and effectively collect and manage data on all market transactions involving land plots and other real estate assets. The paper uses data from this system to demonstrate how price data for more than 175,000 land transactions for commercial agricultural land, together with publicly available information on soil quality, land use and infrastructure access, allows to estimate land prices using a regression approach. As cross validation tests suggest the model consistently explains at least half of the variation in the data, we use it to predict land prices for the approximately 20 million hectares of commercial agricultural land in Ukraine. Comparing results to the NMV shows not only that mean land prices are now between a quarter and a third above the NMV but that the size of the discrepancy varies systematically across regions: while mean predictions are slightly below the NMV in the East, they are well above it in the Center and the West. Use of the NMV to assess tax liabilities or value land posted as collateral would thus risk to constrain the ability of enterprises and local governments in regions characterized by an increase in relative land values to raise credit for investment or taxes for local service delivery in line with new opportunities or needs. While Ukraine's experience offers valuable lessons for other transition countries, two aspects are relevant from a policy perspective. First, although this paper shows digital registry data can be leveraged to 2 implement an objective and replicable approach to mass valuation, applying it routinely in practice requires legal amendments to valuation and tax codes. Such amendments should include provisions for a gradual expansion to non-agricultural urban land by integrating data on land and buildings that are currently managed in separate registries, in line with the roll-out of a unified address registry. Second, the discrepancies between NMV and market-based values identified here imply that a transition from NMV to market values based on mass appraisal as a basis for taxation may change specific local communities’ receipts or taxpayers’ bills. Making provisions for a gradual transition, including the ability to set local tax rates within a wide band, will be important to maintain fiscal stability for local governments and mitigate backlash by taxpayers. In fact, war-induced drops in land prices may make transition towards a market-based land valuation easier and put in place a buoyant system to help recovery. The rest of the paper is organized as follows: Section 2 discusses global background and principles of transparent property valuation as a basis for own source revenue generation by local governments. Section 3 links to the Ukrainian context by documenting the country’s journey towards establishing a land market and improving land valuation and remaining challenges. Section 4 describes the empirical model, discusses results from estimation of land values based on publicly available data as well as predicted prices for the 7 million parcels of commercial agricultural land. Section 5 concludes with policy implications. 2. Putting property taxes in context Immovable property taxes are rather simple to administer; target users of local government services with limited spillover across jurisdictions; are less likely to distort decisions by economic actors than taxation of production factors with a higher elasticity due to its immobile base; and as land and real estate is owned by more affluent individuals, are progressive. To capture this potential, it is important to have a complete tax base, an automated process for valuation that is responsive to market conditions, credible enforcement with minimal exemptions, and appropriate assignment of revenue. 2.1 Conceptual evidence in global context As land is immobile, taxing it allows to raise revenue to finance local public goods in a way that is less distortive (Schwerhoff et al. 2020) and, if the rich hold more of their wealth in land, more progressive (Bonnet et al. 2021) than taxes on mobile factors such as capital and labor. Property taxes increase owners’ incentives to use land effectively (Oates and Schwab 1997) and discourage speculation (Norregaard 2013), an issue that is important as in all countries for which data are available, prices for land and property grew strongly in real terms since the 1950s (Knoll et al. 2017). Property taxes also encourage compact urban development (Ermini and Santolini 2017; Song and Zenou 2006) rather than sprawl. 3 Although fiscal cadasters have long been a key and growth-enhancing element of state capacity (D’Arcy et al. 2024), 1 land and property taxation also provides the basis for the social contract between taxpayers and the state (Besley and Persson 2014). Even in fragile, low-income and conflict-affected settings, increased property tax collection triggered a large “participation dividend” by raising citizens’ political participation and improving their perception of government responsiveness to their concerns (Weigel 2020). Provision of public goods increased property tax payments even by those not affected in Argentina (Carrillo et al. 2017) with willingness to pay property taxes enhanced by better service delivery (Kresch et al. 2023). Public goods, including transport infrastructure, are generally capitalized in land prices: For example, in mid-19th century Chicago, public provision of water and sewer access doubled property values, generating a total benefit about 60 times construction costs (Coury et al. 2021). In New York, infrastructure extension increased land prices by 8 percent, creating US$ 5.5 billion in property value, though less than one-third of the private value created was captured via higher property tax revenue, pointing to considerable unexploited potential from property taxes to allow governments to capture a greater share of the added value from such investment (Gupta et al. 2022). Experimental evidence suggests that infrastructure improvements, in this case first-time asphalting of residential streets, will result in appreciation of local land values in an amount that is approximately equal to construction costs (Gonzalez-Navarro and Quintana-Domeque 2016). In Colombia, higher own tax revenue had a bigger impact on public good delivery than an equally sized increase in central government transfers while increasing accountability by local officials (Martínez 2023). Land taxes are not only an effective way to finance local public goods, but also one of the least distortive ways of raising public revenue as land is immobile and land rents are a true scarcity rent (Schwerhoff, Edenhofer and Fleurbaey 2020). 2 In the standard urban model, a recurring local tax on land can finance the optimum amount of local public goods (Arnott and Stiglitz 1981). Property taxes can reduce wealth inequality without requiring distortive taxes on capital (Bonnet et al. 2021). If labor productivity in manufacturing is greater than that in agriculture and agricultural and manufactured goods are substitutes (or the economy is open to world trade), land taxes increase aggregate economic output (Kalkuhl and Edenhofer 2017). Urban property taxes increase incentives to use land effectively (Oates and Schwab 1997) and discourage land speculation (Norregaard 2013), creating scope to reduce sprawl (Ermini and Santolini 2017; Song and 1 In Europe, increased capacity for tax collection was historically a key determinant of long-run economic growth (Dincecco and Katz 2016). 2 US evidence also shows that if housing supply is inelastic, land and property taxes are capitalized in land (Lutz 2015) or property values (Livy 2018) so that future owners are not affected (Hoj et al. 2018)—a finding supported for high-value properties in Sweden (Elinder and Persson 2017). 4 Zenou 2006) and vacancy rates (Segú 2020). 3 Although high property taxes may result in wealth effects, 4 revenue from land taxes is less volatile than revenue from income or business taxes that vary with the business cycle, making projected public revenue from land taxes ideal for financing long-term local investment, as recognized by the fact that municipal bonds are often backed by property tax revenues. Instead of a recurrent property tax, many countries levy high transaction taxes, partly as such taxes can be collected easily from landowners who come to register their transfers, thereby avoiding the fixed cost of an apparatus for property tax collection, including maintenance of land records and a tax roll and property valuation. In a digital environment, the cost of establishing the infrastructure for property tax collection is much lower while the benefits from having this infrastructure in place increase dramatically. Studies show that transfer taxes’ modest contribution to public revenue comes at very high cost in terms of governance risks and economic distortions as transfer taxes (i) push property transactions into informality, undermining the registry’s quality and currency; (ii) encourage corruption via underreporting of prices, often in return for a side payment to officials who turn a blind eye reducing the informational content of registry price data; and (iii) impose high deadweight losses. For example, in the United Kingdom, transaction taxes caused large distortions in the volume, timing, and price of property transactions (Best and Kleven 2018) and negatively affected land markets and labor mobility (Eerola et al. 2021; Hilber and Lyytikainen 2017). A 1.1 percent land transfer tax in Toronto led to significant welfare losses (Dachis et al. 2012) and, if decisions on whether to own or rent are accounted for, it was estimated to incur a deadweight loss of 79 percent of revenue (Han et al. 2022). Replacing transfer taxes with an equivalent recurrent land tax on a much larger base is easy technically and can increase revenue, welfare, efficiency, and transparency. If combined with the abolition of transfer taxes, establishment of a recurrent property tax could thus be attractive politically and yield large welfare gains. 2.2 Key aspects for effective administration of property taxes Tax map completeness. Lack of a complete cadastral map is the most fundamental impediment for effective revenue generation from property tax. Historically, creating the fiscal cadaster that is needed to underpin land taxation required manual mapping and data collection, a process that took developed countries years 3 In developed countries, the visibility and salience of property taxes (Cabral and Hoxby 2012), while increasing accountability (Presbitero et al. 2014), often make getting political support for them challenging (Bahl and Wallace 2008). Fiscal policy and land use regulation interact strongly, and governments must align those policies carefully if they are to achieve their land use objectives (Blochliger et al. 2017). 4 With imperfect credit markets and limited liquidity of the underlying asset, rising property values will result in higher property tax bills without increasing incomes, reducing owners’ liquidity, an issue that can be addressed through deferment of tax liability, which is more pronounced in rural areas (Hoff 1991). The wealth effects generated by house price movement affect labor market participation (Liu and Yang 2020), consumption, mortgage defaults, and attitudes toward property taxation (Wong 2020). While deferment of bills by putting a lien on the property is a response widely practiced by developed economies (Slack and Bird 2014), provision of liquidity to constrained taxpayers has been raised as an alternative policy instrument (Brockmeyer et al. 2020). 5 or decades to complete (D'Arcy and Nistotskaya 2018). 5 Building footprints are now available freely or can be generated from high-resolution satellite imagery or, at city scale, drones (Gevaert et al. 2020) at minimal cost, 6 eliminating the need for costly and complex aerial photography (Carolini et al. 2020). The use of such imagery to update property tax rolls and improve fairness of taxation by minimizing tax evasion in Italy led to large increases in collection of local property taxes. Beyond providing economic benefits, this also pushed up rates of incumbent re-election. Political returns were higher where public goods were provided efficiently and where tolerance for tax evasion had been lower at the start (Casaburi and Troiano 2016). Since 2004, Bogotá has been a pioneer in introducing policies of land taxation and land value capture, to very positive effect (Garza and Gonzalez 2021). Automated valuation based on market prices. Once the universe of taxable properties is identified, a transparent and fair process for determining each property’s tax liability needs to be established, ideally distinguishing land and buildings. Automated computer-assisted mass valuation (CAMA) involves regressing price data on property characteristics and using the coefficients thus obtained to generate property-specific predicted values. CAMA can overcome the disadvantage of flat rates and is thus a preferred option (Hill 2013). Beyond data on property characteristics and coordinates (Diewert and Shimizu 2022), it requires price data that are ideally obtained from registry records or, if these are not available or distorted due to high transaction taxes, sample enumeration that can be checked against other data sources, such as listings of high-end properties. 7 The potential of CAMA is illustrated by the mass valuation of properties in Rwanda, which suggests that a 1 percent ad valorem tax could increase revenue up to tenfold and be more equitable (Ali et al. 2020). Since the 1990s, institutional mechanisms to allow low-cost implementation of CAMA models have been adopted (Gouriéroux and Laferrère 2009) and are now routinely used in most developed countries, with positive impacts on revenue and the perceived fairness and transparency of property taxation. In the Netherlands, valuations of some 8 million properties via CAMA are conducted annually, yielding 2.8 percent of GDP in revenue from real estate taxes. Assessed property values are publicly available, improving market functioning and preventing fraud (for example, related to mortgages), owners can update their data online, and interoperability provides additional benefits (Kuiiper and Kaathman 2015). 8 A 5 The Swedish cadaster was launched in 1530 and achieved complete national coverage in 1628. Greece still lacks a functioning national cadaster. This variation is associated with vast differences in tax discipline and trust. Even a decade ago, establishing tax maps for individual cities in World Bank–supported projects took years, as the example of Tanzania shows. 6 As the use of drones is not always well-regulated (Stöcker et al. 2017), obstacles to their use may be more regulatory than technical. 7 Registry records can be used if land markets are sufficiently representative and liquid and if transfer taxes do not discourage truthful land value declaration. Ben-Shahar et al. (2020) propose a method to identify transactions suspected of underreporting. They apply this to all transactions in Israel over 1998–2015 and conclude that some 8 percent of transactions are underreported, with an average price that is 30 percent below the projected true price. 8 In Europe, owners’ estimated property values are about 9 percent above assessed values, ranging between 3.2 percent in Germany and 22 percent in Italy, a discrepancy that can partly be explained by household characteristics, in particular the level of indebtedness (Le Roux and Roma 2019). 6 programming algorithm to predict an adjusted house price index for arbitrary spatial units from repeated cross-sections of geocoded micro data has been used to generate a panel of German property prices and rents that is unprecedented in its spatial coverage and detail. It provides price and rent indices for residential and commercial properties while also raising questions for follow-up research (Ahlfeldt et al. 2023). Based on such experiences, the methodology for property valuation has been standardized (Diewert et al. 2020), allowing quick transfers to other contexts. In Lithuania, assessed property values more than quadrupled as a result of adopting and gradually refining the mass valuation system (Almy 2015) that is now at the core of a private sector–driven registry system that provides many ancillary benefits (Grover et al. 2017). Most African countries still require valuation of all property by a licensed valuer (Franzsen and McCluskey 2017), despite the high cost and potential arbitrariness. Using CAMA for residential properties and focusing scarce valuation capacity on commercial property has proven to be a more viable way to ensure buoyancy by allowing frequent revaluation. Credible enforcement and minimal exemptions. Reducing compliance costs, for example, via e-billing and mobile payment options (McCluskey et al. 2018), can help increase property tax collection. 9 To improve enforcement, treating property tax arrears as a high priority lien on the property or requiring tax clearance before any transfer, mortgage, or other encumbrance can be registered on a property seems to be a low-cost but feasible option. By implicitly deferring tax payment until the point when the asset is liquidated, this can also address concerns about negative liquidity effects of property taxes, especially on seniors who, if they are unwilling to move, might have to reduce consumption to pay increasing property taxes levied on the assessed value of their home (Brockmeyer et al. 2020). 10 Tax codes often contain ample exemptions, including for owner-occupied properties or pensioners that are difficult to justify from an economic perspective (Englund 2003) as all of these consume the local services supplied using property tax proceeds. These can significantly reduce tax revenue (Balan et al. 2020) and yield (Ali, Deininger and Wild 2020) and may negatively affect governance, efficiency of land use, and perceived fairness of the system. While public disclosure of tax information or social recognition have often increased compliance (Slemrod et al. 2022), public disclosure in Uganda reduced compliance, consistent with the notion that individuals overestimate compliance levels and, once they learn the truth, stop complying not to be taken advantage of (Regan and Manwaring 2023). 11 9 In a randomized experiment in Zambia, text message reminders significantly increased payment, especially for those on a network that allowed mobile payment. Effects were particularly large for payers outside the capital (Ali and Deininger 2021). 10 As housing wealth is illiquid, changes in wealth, for example, via housing booms or property tax liabilities, can trigger large wealth effects and, through these, affect consumption or economic activity. In Italy, tax increases on the main dwelling led to large expenditure cuts among mortgagors who hold low liquid wealth despite owning sizable illiquid assets (Surico and Trezzi 2019). In the United States, property tax limits reduced female labor force participation by 0.7 to 1.4 percentage points during the 2005–06 housing boom (Liu and Yang 2020). 11 Reminders to pay taxes were effective if they were credible (Castro and Scartascini 2015) or described the effects of nonpayment (Chirico et al. 2017). Beliefs about the ability of the authorities to penalize defaulters also affected truthful declaration of other taxes (Lopez-Luzuriaga and 7 Revenue assignment, competition, and capacity building. Efforts to make property taxation more transparent and increase revenue may encounter opposition from stakeholders, including politicians, whose time horizon is short and accountability weak. In Brazil, local politicians failed to take advantage of a free national program to update the cadaster and increase property tax revenue by some 10 percent on average if they feared that doing so would jeopardize their chances of re-election (Christensen and Garfias 2020). Revenue from property taxes was particularly low in municipalities with high levels of inequality where affluent individuals lowered their tax liability by blocking efforts to increase capacity or bribing officials (Hollenbach and Silva 2019). In Colombia, property was systematically undervalued in municipalities where large landlords were influential even though this was associated with lower provision of local public goods and ultimately growth (Sanchez-Talanquer 2020). As a local tax, revenue from property taxes should accrue to local bodies. If local governments lack capacity to administer such taxes transparently, better tax administration can possibly improve tax collection (Basri et al. 2019) as demonstrated in France (Breuillé et al. 2018), Canada (Feir et al. 2023), India (Awasthi et al. 2020), Mexico (Larreguy et al. 2018), and Africa (Tourek et al. 2020). Beyond improving the scope for local revenue collection, generation of a comprehensive tax map and land valuation could have political benefits by (i) providing the basis for a reduction of transaction taxes that would remove a key obstacle to keeping property registries up to date; (ii) allowing for a more realistic land use plan land and increasing the effectiveness with which public services will be provided (including by improving the informational basis for subcontracting and monitoring private sector delivery); (iii) reducing the transaction cost in land markets by publishing valuations and using them as an objective benchmark for payment of compensation in case of compulsory land acquisition, for example, for infrastructure construction; and (iv) using the information generated by tax maps and evidence of property tax payment to open up a low-cost path toward recognition of landownership rights. 12 3. Property taxation in the Ukrainian context As market-based valuation is impossible without a market for land and real estate, this section first describes Ukraine’s journey towards a market for agricultural land. Valuation methods have not yet caught up with these advances, resulting in a disconnect that risks distorting of private and public land use and services in ways that impede effective decentralization and reduce efficiency, equity, and resilience. By mandating recording of purchase prices, land market reforms have, however, provided the basis for a more flexible approach to property valuation and the use of such values in multiple contexts. Scartascini 2019). Appeals to civic duty had limited effect (delCarpio 2015) even if they included information about neighbors’ compliance levels, and making delinquency more visible increased compliance only for small debts (Perez-Truglia and Troiano 2018). 12 For example, in Peru, expected land value increments from titling even informal slum land, if adopted jointly with land taxation, are argued to allow implementation of a titling program that would be fully self-financing (Hawley et al. 2018). 8 3.1 Reforms initiated by Ukraine moving towards a market-oriented economy Since gaining independence in 1991, Ukraine has undertaken a sweeping set of land reforms to move from a centrally planned economy where all land was state-owned to a market-oriented system characterized by predominantly private ownership. This change was pivotal in repositioning Ukraine as a significant global supplier of agricultural products, leveraging its vast arable land resources to contribute to global food security. The trajectory of land reform in Ukraine comprises several key phases, each marked by legislative milestones and institutional changes that together facilitated a shift to market-based land relations. 1990–1992: Initiation of Land Reform and Legal Pluralism in Land Ownership: The early 1990s marked the inception of land reform in Ukraine, coinciding with the dissolution of the Soviet Union. The adoption of laws recognizing state, communal, and private land ownership was revolutionary. This period established the foundational legal framework necessary for a market economy, challenging the system of collective farms (kolkhozes and sovkhozes) and exclusively state ownership of land. 1992–2001: Adoption of the Land Code and Large-Scale Privatization: In 1992, Ukraine enacted its first Land Code, codifying principles that facilitated the transition to a market economy. This phase was characterized by extensive privatization campaigns, resulting in the redistribution of approximately 27 million hectares of agricultural land to private individuals and entities by 2001. The privatization process aimed to dismantle collective farming structures and empower individual landowners, thereby increasing efficiency and productivity in the agricultural sector. 2001–2013: Legislative Development and Institutional Strengthening: The adoption of a new Land Code in 2001 marked a significant advancement in land legislation. This period saw the introduction of specialized laws concerning land management, protection, valuation, and compulsory acquisition for public needs. Notably, the Law on Land Valuation (2004) sought to standardize valuation practices; however, the reliance on "normative monetary valuation" methods persisted, often leading to discrepancies between assessed values and market realities (OECD, 2018). These normative valuations failed to reflect actual market conditions, resulting in inefficiencies and inequities in property taxation. 2013–2020: Digitalization, Delineation, and Decentralization: Between 2013 and 2020, Ukraine focused on modernizing its land administration system through digitalization and the delineation of state and communal land ownership. The launch of the digital State Land Cadastre in 2013 and the State Register of Real Property Rights enhanced transparency and accessibility of land information. By 2020, over 75% of land parcels were registered electronically, streamlining transactions and reducing opportunities for corruption (Ministry of Agrarian Policy and Food of Ukraine, 2020). Decentralization of land management 9 transferred significant authority to local governments, aiming to promote more responsive and efficient land use planning. 2020–2021: Liberalization of the Agricultural Land Market and Spatial Planning Reforms: The period from 2020 to 2021 was one of the most active phases of land reform. In March 2020, Parliament passed legislation lifting the long-standing moratorium on the sale of agricultural land in two stages: Ukrainian individuals could buy agricultural land up to a maximum of 100 ha each from July 1, 2021. On January 1, 2024, the limit was increased to 10,000 ha and fully Ukrainian owned legal entities were also allowed to purchase land. Laws were also put in place to mandate public reporting of land prices and establish a public land market monitoring system; restructure institutions to increase transparency via digital registry-cadaster interoperability; improve public land management by transferring ownership of such land to local governments subject to central land use controls; require transfer of lease rights to public land via electronic auction; 13 and establish of a partial credit guarantee facility to ease small producers’ access to financial markets. 14 Beyond deregulation of land management procedures, reforms initiated a national geospatial data infrastructure, and involved the adoption of modern legislation on spatial planning, aligning Ukraine's practices with European Union standards. 2022 and Beyond: Adapting to Wartime Challenges and Continuing Reforms: Intensified conflict due to Russian aggression in early 2022 posed unprecedented challenges and prompted government to enact emergency legislation to simplify procedures to allocate land for urgent public needs, particularly defense and reconstruction. Spatial planning processes were adjusted to account for security considerations, and measures were implemented to ensure the resilience of public registers and the protection of land records under wartime conditions. 3.2 Advances and remaining challenges for more transparent land valuation and property taxation The fact that laws on land valuation were among the first tax-related Acts by independent Ukraine highlights the importance of the topic. 15 Yet, mechanisms established remained rudimentary with tax rates stipulated by law for each region, irrespective of land plots’ characteristics or location. To establish prerequisites for more differentiated agricultural land valuation, a comprehensive assessment of soil quality for crop agriculture referred to as soil bonitation was conducted in 1993. Scores between 1 and 100 provided a measure of soil fertility based on soil composition (clay, sand, humus), depth of the horizon, acidity, and 13 The mandatory shift from centralized, in-person to fully electronic auctions run by local authorities for any transfer of use rights to public land instantaneously increased lease prices by 175% (Deininger et al. 2023). 14 The numbers of relevant laws are 552 (turnover law opening agricultural land sales market), 340 (anti-raider law, including price reporting), 711 (local planning), 554 (spatial data interoperability), 985 (state support), 1423 (institutional reform of SGC and decentralization), 1444 (mandatory e-auctions for public land ), and 3205 (partial credit guarantee facility). 15 Law of Ukraine dated July 3, 1992, No. 2535-XII "On Payment for Land". 10 nutrient availability. While of great relevance for agricultural production, this focused only on one of several elements of a market-aligned land valuation system. Subsequent regulation put in place ‘normative monetary valuation’ using data of the 1988 Soviet economic evaluation of land. 16 Valuation of agricultural land was based on rental income from grain crop production whereas zoning and capital investments in infrastructure and public utilities were relevant for urban land. Albeit intended as a temporary measure pending the establishment of a market, this valuation, indexed to which is annually updated based on the index of consumer prices, became entrenched, forming the basis for various taxes and fees, including land tax, rent for state and communal lands, state duties, and simplified taxation systems for farmers. Attempts by government to refine this methodology in 2011, 2016, and 2021 did not correct its fundamental flaws. The discrepancy between normative and market-based valuation leads to economic distortions, violates the principle of tax fairness, and may encourage inefficient land use. As local revenue is not linked to increases in land values, local governments’ ability to benefit from and their incentive to undertake actions such as planning and infrastructure or public service provision to promote economic development, is limited. Ukraine’s local governments also hold a large amount of land as state or communal property and transparent valuation would ensure that lease revenues reflect the true economic value of the land, maximizing public income and discouraging undervaluation or corrupt practices. Land valuation is also key for valuing land pledged as collateral to compute reserve adequacy ratios for banks and pivotal for the effective functioning of land and credit markets. Not having a way to identify the market value of land also makes it impossible to tax capital gains from land appreciation, an issue that is particularly important in Ukraine where both agricultural and non-agricultural land is likely to appreciate after the war, creating a potential revenue source that could provide an important and incentive-compatible source of funding for reconstruction. Land valuation would also help identify areas where public and private investments will yield the highest payoff, facilitate strategic planning to rebuild infrastructure, housing, and economic activity centers, and make provisions for enhanced resilience via spatial analysis. Recognizing the need for more efficient valuation methods, legislation introduced centralized registration in the State Register of Property Rights (SRPR) of information on the price (value) of land plots, property rights, and the amount of payment for land use during real estate transactions. While initial data quality was sub-standard, with prices missing for more than 50% of transactions, significant improvements have been 16 Resolution No. 213 on the "Methodology for Monetary Valuation of Agricultural Land and Settlements" by the Cabinet of Ministers of Ukraine (March 23, 1995). 11 made. 17 A remaining data challenge relates to transactions of apartments that tend to lack cadastral numbers for underlying land. Resolving it requires acceleration of ongoing efforts to fully implement an address register that is interoperable with the urban cadaster through an API and an expansion of existing data reporting requirements and exchange protocols for land parcels to include real estate. While accurate property valuation will be invaluable to guide activities by the private sector, translating it into higher local revenue requires that market values change the tax base taxes on land or real estate from NMV or floor area, respectively, to value. In this context, exemptions and caps on local governments’ ability to adjust taxes may also need to be reviewed, ideally based on evidence of their impact on revenue. 18 4. Implementing a mass valuation model for commercial agricultural land in Ukraine This section presents the valuation model, the public data used to estimate it, and the prices predicted by the model. Comparing predicted prices to normative values for 1.2 million of the 7.2 million parcels for which such values are available shows that, while mean estimated prices are 25% above the normative monetary value, a market-based valuation increases horizontal equity with limited increases in the country’s East and South but land values that are more than 30% above the NMV in the Center, North, and West. 4.1 Regression model and variable construction To provide a market-based valuation for agricultural land for commercial farming, we focus on some 160,000 transactions completed between June 2021 and July 2024 to estimate a hedonic valuation model (Rosen 1974). The underlying assumption is that transacted parcels’ prices are the result of an equilibrium between demand and supply functions that are differentiable in land characteristics relevant to the seller and buyer respectively (Palmquist 1989). If land markets are homogenous (Huang et al. 2006), knowledge of the price and characteristics of a transacted parcel then allows to recover the marginal contribution of each characteristic (Palmquist 2005). Applying these regression coefficients to parcels not transacted then provides the basis for predicting the prices of all parcels. Formally, the hedonic model to be estimates takes the form log = + + + , (1) where is the sale price of a parcel in region in constant 2024 USD per ha, is a vector of parcel- specific characteristics including parcel size, land use at the time of transfer (crops, forest, built up, grass 17 Law No. 340-IX from December 5, 2019, "On Amendments to Certain Legislative Acts of Ukraine Regarding the Counteraction of Raiding" makes price reporting mandatory. Subsequent adjustments to the registry’s software made this mandatory in the registration process with land price data now transmitted automatically to the State Land Cadastre (SLC) through an automated information exchange protocol. 18 The tax -free allowance is 60 m2 for apartments, 120 m2 for residential buildings, and 180 m2 for other buildings in addition to a luxury tax of UAH 25,000/a for apartments and houses above 200 and 500 m2,respectively. Local governments are allowed to impose land taxes of up to 1 percent of NMV per year on land and 1.5 minimum wages. Families with three or more children, disabled individuals, retirees, and war veterans are exempt from land tax irrespective of income or wealth. 12 or uncultivated), distance to key infrastructure, and soil quality; is a vector of transaction characteristics; is a region fixed effect and is an error term. Per hectare land prices are and expressed in January 2024 USD by using the monthly Consumer Price Index (CPI) to adjust for inflation and then convert to USD at the official exchange rate. 19 We eliminate duplicates, defined as the same parcel being transacted multiple times on the same date.20 Extreme outliers are removed by trimming at the 1st and 99th percentiles. 21 Parcel characteristics include the Euclidean distance from parcels’ centroid to the nearest road, railway access point, port, grain elevator, or city from Open Street Map (OSM) data provides data on infrastructure access. 22 We also compute the share of parcel area covered by built area, crops, forest or grassland using crop classification data for the year prior to the transaction. 23 Finally, as we did not have access to nation-wide data on the soil quality index discussed earlier, we use publicly available SoilGrids data (Poggio et al. 2021) to obtain measures of soil quality, in particular pH, soil density, nitrogen and organic matter in the 0-60 cm horizon at the centroid of each parcel. 24 The vector includes variables to capture transaction timing and adherence to legal requirements, and other characteristics. Specifically, we include (i) a war dummy that takes the value of one if the transaction happened after February 2022 when war conditions are likely to have significantly altered agricultural land supply and demand; (ii) the number of buyers if a parcel was acquired in joint ownership; (iii) a variable capturing the order of transactions (from 1 to 5) in case a parcel has been transacted multiple times; and (iv) dummy variables for prices below NMV and for missing NMV to capture non-compliance with legal provision prohibiting transactions below the NMV. Use of fixed effects at the level of 355,663 cadastral blocks, delineated to contain homogenous patterns of land use, is the most appropriate way to control for unobserved heterogeneity in local conditions. However, as agricultural land sales started only in 2021, there may be no or only a limited number of sales transactions at this level, especially in agriculturally marginal areas. Appendix figures A4, A5, A6, A7, A8, and A9, illustrate this graphically by showing that, there were no or only few land sales transactions for many cadastral blocks, zones, villages (KOATUU), communities (hromadas), or districts (rayons). If there are only few non-representative observations, estimation of fixed effects at any given level may result in biased and misleading prediction of land values. 19 Both the CPI and UAH to USD exchange rates are taken from the data of the National Bank of Ukraine https://bank.gov.ua/ua/statistic. 20 As such repetitions are due to a parcel being purchased by multiple owners and the value of each owner’s share is recorded separately, we aggregate the land value over every owner, while keeping other parcel characteristics constant. By contrast, cases where the same parcel is sold multiple times during the period under concern are retained as different observations. 21 Extreme values due to very small plot size that result in very high per-ha prices are likely due to data entry error and removed via trimming. 22 OSM data was downloaded from https://www.openstreetmap.org. Equivalent and possibly richer data on access to infrastructure could in principle be obtained from official topographic maps that were not accessible to us. 23 Downloaded from https://ukraine-cropmaps.com. 24 Downloaded from https://files.isric.org/soilgrids/latest/data. Systematically comparing bonitet to these publicly available data would be one area for improving the data base used for estimation. 13 To avoid this, we estimate a fixed effect at a given administrative level only if the volume of sales transactions at this level exceeds a threshold n of transactions and move to the next higher administrative level, either 77,225 cadastral zones; 18,730 villages; 1,779 amalgamated communities; 680 old rayons; 138 new rayons; or 24 regions or oblasts, if this threshold is not met. For a threshold of n=10, table 3 shows that fixed effects at the level of cadastral block, cadastral zone, village, and community are used in 50.6%, 24.2%, 15.6%, and 9.1% of cases, respectively, leaving less than 0.5% of cases for oblast fixed effects. Descriptive statistics for variables used in the regression are reported in table 1 for the 176,000 cadastral parcels of land for commercial agriculture transacted between June 2021 and July 2024 as well as the remaining close to 7.5 million parcels for commercial farming (land use code 01.01) that are mapped in the cadaster. Results from t-tests of the difference between the regression sample of transacted parcels and parcels not transacted are also reported. We note that the average sales price of US$ 1,097 is about 44% higher than the mean NMV (US$ 760) and the difference between transacted and non-transacted parcels in terms of NMV is marginal (US$ 43.2). Appendix Figure A1 also shows that the relationship between price per ha and parcel size is nonlinear. While differences between the two samples are highly significant statistically, reinforcing the need for a regression-based approach that adjusts for such differences, their magnitude is often small: compared to sold ones, non-transacted parcels are located slightly more distant from road infrastructure, cities, and grain elevators. A larger share of their area is covered with grassland (17.2% vs. 11%), buildings (0.27% vs. 0.2%), or forest (1.9% vs. 1.5 %) rather than crops (86.9% vs. 79.9%), and they have slightly less soil nitrogen, a higher soil density, and are marginally more acidic (pH of 6.84 vs. 6.83). Plots of the density functions for key covariates in appendix figures A2 show that, overall, there is strong overlap between the regression sample and parcels that were not transacted for key covariates but that more valuable and larger parcels with lower soil density, higher level of soil N and organic carbon, and less extreme pH are over-represented in the regression sample relative to the Universe. Figure A3 provides the same plots for distances to key infrastructure objects, documenting that the regression sample is slightly more distant from the next city and port but closer to the next motorway and primary road. Such systematic differences provide a strong rationale for a regression-based approach that differentiates land values based on their covariates instead of valuing land parcels based on averages that will be affected by random variation in the characteristics of land parcels subject to transactions in any period. 4.2 Estimation results and comparison to NMV Table 2 reports results from regression analysis for models with fixed effects at the level of cadastral block, cadastral zone, village, community, old and new rayons, or oblast. Key covariates are significant and have 14 the expected signs even for the most disaggregated specification: Land use indicators are highly significant and positive compared to the omitted category of bare land highlighting that, even within cadastral blocks, parcels that contain more built, crop, grass or forest area are more valuable. The distance to tertiary and primary roads is consistently negative as expected. Soils that are less acidic, contain higher levels of organic matter, or an optimal level of nitrogen are also more valuable and land value increases in parcel area at a decreasing rate per ha. Prices for parcels sold after the war are on average 7% to 16% lower than those sold before, with a further 2.5% reduction in 2024. Finally mean prices for parcels sold below the NMV in violation of legal norms are between 25% and 30% lower than when legal stipulations were adhered to. Appendix table A2 presents the assessment of each independent variables’ importance and its marginal contribution to the variance explained measured using the adjusted within R2 (excluding fixed effects) to model fit. Results indicate that area, distance to infrastructure, land use, soil quality, and the NMV dummy together contribute some 21% of the variation in the data for the pooled model although their contribution decreases by some 50% as more disaggregated fixed effects are added. To test whether the regression can be used for prediction, we perform a 5-fold cross validation where data is randomly split into 5 parts and regressions are re-estimated for an 80% sample and used to predict prices for the remaining 20%. The adjusted R2 is used to measure quality of predictions. Table 4 reports results for models 1-6, highlighting that models with more disaggregated fixed effects attain higher R2 in all folds. With an R2 close to 0.5, models with fixed effects at cadastral block, village and cadastral zone level perform best followed by those with fixed effects at community (0.39), new rayons (0.29), and oblast (0.25). 25 To put the regression approach in perspective, table 4 reports R² for the simple arithmetic average of prices at cadastral block, cadastral zone, village, community, old and new district, or oblast level subject to at least 10 observations at the relevant administrative level. This approach explains 33.9% of the variation in the data while regression-based estimates using cadastral block fixed effects explain 53.2%; an improvement by about 20%. Figure 1 illustrates this graphically by showing actual versus predicted prices for different models. 26 The variables used in our regression model area available for about 7.5 million parcels of land for commercial agriculture that, with over 20 million ha, cover the core of Ukraine’s farmland. We can thus use the above estimator to predict land prices for every commercial agricultural parcel in Ukraine and, by 25 Table 4 also reports the goodness of fit for ensemble models with a minimum number n of 5, 10, or 40 observations per fixed effect. The trade- off between better predictions and more limited coverage (as not all the relevant administrative units will have more than n sales transactions) is clearly visible and we opt to set n to 10. 26 The model with cadastral block fixed effects is the best one, followed by the ensemble models with 5 and 10 observations thresholds while regional averages and oblast fixed effect models are the worst, supporting use of an ensemble models with a10-observations threshold. 15 comparing it to actual and trimmed market prices as well as, subject to data availability, the NMV,27 assess the difference between different that a shift to market-based valuation would make. Table 5 rows 4-7 report, respectively the mean actual price, the predicted price from the regression model, the average prices computed using a 10-observation threshold per administrative unit and the NMV. For ease of interpretation, rows 8-10 provide the difference between the NMV and the regional average, the predicted and the actual price. For the entire sample, the predicted price of US$ 994/ha is 24% above the NMV, below the actual price (32.8%). National figures conceal considerable inter-regional variation: predicted land values using our model are close to the NMV in the East and the South but 24.8%, 22.9%, and 80.2% above it in the North, Center, and West, respectively. Using the NMV as a basis for valuing or taxing land would ignore recent inter- regional shifts in economic opportunity and, for example, impose an unjustified tax burden on regions negatively affected by the war while making it impossible for owners in the West to leverage the full value of their land as collateral for investment. 4.3 Potential refinements and extensions Several refinements in terms of adding additional parcel characteristics could potentially be considered. First, we used the admittedly coarse share of built-up area from sentinel imagery as a measure of built area. A digital address or building register or commercially available building footprints could be possible alternatives. Second, geographical characteristics including elevation, slope, and north and east exposure vary little across parcels used for commercial agriculture in Ukraine that are mostly flat and located at low altitudes but will need to be considered if land for personal farming is included. Third, no information was available on access to irrigation facilities as responsible authorities are still in the process of establishing a registry of public irrigation assets. Cadastral blocks’ high level of granularity of (with an average size of less than 2 km2) together with the rather high volume of land market activity in areas suitable for irrigation (not all of which may currently be operational) suggest that cadastral block fixed effects may absorb most of the land value effect of public irrigation infrastructure, though testing this with actual data, once available, is desirable. It will also be important to expand our analysis to include other types of agricultural and non-agricultural land use. 28 Doing so will require to overcome the legal separation of land parcels and buildings or structures on them that is common in post-socialist countries. This not only implies that prices for land and buildings 27 Owing to idiosyncrasies in the way NMVs at parcel level were recorded, NMV values are available for only about 1.2 million land parcels. SGC is currently working on establishing a database with API connectivity that should provide NMVs for all parcels in the country by early 2025. 28 Land in Ukraine is classified in a wide number of land use classes (e.g., as crop land, hay fields, pastures, or unsuitable for agriculture) that may or may not correspond to actual use. If such data were available, overlaying them with actual land use and soil quality would be desirable. 16 on it are recorded in separate registers but, with building registries the responsibility of local rather than central government, makes it impossible to easily link the two and value property even if prices for structures are reported. An integrated address or building register, fully interoperable with the cadastre, is thus needed to extend mass valuation to residential land. While establishment of such a register across the country is envisaged, piloting relevant procedures with local governments that have already established such a register would be desirable. Finally, to the extent that predicted prices are to be used for tax purposes, the transition from NMV to mass appraisal based on market prices can significantly alter individual taxpayers’ bills or local communities’ property tax revenue. A carefully designed strategy for a smooth transition is thus essential when moving to a new tax base, ensure financial stability, and prevent substantial budget losses during the transition. 29 5. Conclusion and policy implications By providing estimated land values for about 7 million agricultural parcels, this paper demonstrates that the preconditions for a market-based valuation system of land and property in Ukraine are in place. Ukraine's advanced digital transformation in public registries, particularly the mandatory state registration of property rights in centralized electronic systems like the State Land Cadastre and the State Register of Real Property Rights, provides a robust foundation for such a shift as it enables efficient collection and rapid processing of transaction information, including prices and property values, which is essential for accurate market assessments and fostering transparency. Our analysis illustrates both the viability and the greater flexibility and equity of a market-based approach to land and property valuation that can increase fairness in taxation; streamline private investment and direct it toward areas with the highest returns; and discourage speculative land holding by allowing capital gains to be reliably measured and taxed fairly. To institutionalize transparent and reproducible market-based valuation, three steps may be considered: Data Standards and Data Sharing Protocols: The exercise presented in this paper would not have been possible without a legal basis mandating documentation of land sales prices by registrars, the establishment of necessary software, and the publication of such data. Routinely valuing agricultural land will require using data from official sources held or acquired by entities like the State Land Cadastre, such as cadastral maps, soil composition, topographic features, crop cover, and administrative boundaries. These may be complemented with information from other sources if based on a replicable methodology, especially to value agricultural land for personal use that may have buildings or structures affecting land value. 29 For example, the normative valuation of industrial land in Ukraine is often significantly higher than its market value, while residential land in cities may have a normative valuation several times lower than its market value. Therefore, a gradual transition to the new tax base is necessary, including a transitional period and adjustments to tax rates to mitigate potential negative effects on both taxpayers and local government revenues. 17 Integration of Land and Building Data: For urban land, which accounts for a much greater share of wealth, mechanisms including digitally accessible address or building registries that allow unambiguous linking of objects such as apartments to associated land parcels are crucial. Ensuring the reporting of prices for such objects if transacted separately through automated data exchange, similar to what has been implemented for land, will provide the basis for applying an approach akin to the one used here. For public and other types of land that are not regularly traded, a cost approach could be adopted, or lease prices, together with an appropriate discount factor, could be used depending on the case at hand. Publication of such values at the parcel level will increase their credibility and enhance trust. Legal Provisions: Beyond addressing data issues, legislative measures are required. Legislation needs to establish a methodology for a market-based approach to mass valuation, mechanisms for gradually replacing normative valuation, and define implementation responsibilities, including the computation and management of necessary information systems and handling of appeals. To use such values routinely for administrative purposes, including as a basis for valuing real estate collateral or taxation, changes in the tax code and prudential banking regulations will be needed. Evidence from applying a mass valuation model to agricultural land suggests that, in addition to aligning with European valuation standards, such steps will facilitate easier access to credit for enterprises and promote horizontal equity and transparency. Ukraine's shift toward a market-based land valuation system, supported by advanced digital infrastructure, represents a significant step toward economic modernization. Anchoring it in law will have implications for Ukraine's economic recovery, development, and investment climate and be of fundamental importance in a post-war context, where the ability to raise substantial domestic and foreign investment quickly and at low cost and allocate them to their best use will be critical to restore economic stability, recover from the war's devastating effects, and put in place a decentralized system of governance with the means and incentives to promote economic development. 18 Table 1. Descriptive statistics for transacted and non-transacted parcels Transacted parcels Not transacted parcels Difference Price per ha, USD (2024) 1097.99 (811.70) NMV per ha, USD (2024) 759.71 (354.02) 716.54 (916.75) -43.168*** Plot size, ha 3.12 (2.67) 2.75 (2.81) -0.374*** Dist. motorway, km 27.21 (23.15) 28.68 (24.09) 1.469*** Dist. primary road, km 12.09 (10.32) 12.52 (10.54) 0.430*** Dist. secondary road, km 6.85 (5.47) 7.01 (5.68) 0.157*** Dist. tertiary road, km 2.21 (1.88) 2.35 (2.21) 0.136*** Dist. port, km 338.48 (144.65) 344.34 (163.20) 5.867*** Dist. railway, km 10.05 (7.89) 10.35 (8.16) 0.301*** Dist. city, km 52.38 (33.06) 52.24 (33.16) -0.139 Dist. elevator, km 11.86 (7.48) 12.64 (7.67) 0.785*** Built-up % 0.20 (3.13) 0.27 (4.05) 0.073*** Cropped % 86.90 (29.02) 79.87 (35.67) -7.025*** Forested % 1.49 (8.62) 1.91 (10.65) 0.426*** Grassland % 11.02 (26.35) 17.21 (32.74) 6.185*** Soil density 0-60cm, cg/cm3 130.79 (4.74) 131.36 (4.84) 0.572*** Soil nitrogen 0-60cm, cg/kg 394.19 (82.00) 389.95 (87.79) -4.241*** Soil pH 0-60cm 68.66 (3.14) 68.32 (3.51) -0.341*** Soil organic content, t/ha 59.47 (8.30) 58.64 (8.70) -0.831*** War (0|1) 0.79 (0.41) 1.00 (0.00) 0.213*** Price below NMV (0|1) 0.29 (0.45) 0.00 (0.00) -0.292*** Missing NMV (0|1) 0.01 (0.08) 0.00 (0.00) -0.007*** N subjects 1.00 (0.07) 1.00 (0.00) -0.003*** No. of transactions (1-latest, 5-oldest) 1.02 (0.15) 1.00 (0.00) -0.022*** Transaction number 1.05 (0.22) 1.00 (0.00) -0.049*** Total area, 1000 ha 550.9 20,452.0 N non-missing NMV 175,386 1,130,661 No. of obs 176,609 7,448,637 Note: Statistics is reported as: Mean (SD), Column 3 reports the difference in means between parcels that were and were not transacted together with results of a two-sided Welch t-test assuming unequal variances for the significance of the difference. Significance levels are `*` < 0.05, '**' < 0.01, and '***' < 0.001 19 Table 2. Regression results (5) (2) Oblast (3) Rayon (4) Old rayon (6) Village (7) Cad. zone (8) Cad. (1) Pooled Community FE FE FE FE FE block FE FE Area HA (log) 13.239*** 15.144 15.465 15.123* 15.155** 11.828*** 13.473*** 11.360*** (0.716) (22.920) (11.489) (7.766) (6.135) (4.174) (3.992) (3.551) Area HA (log squared) 31.304*** 35.009*** 35.969*** 36.093*** 34.317*** 31.828*** 29.189*** 30.180*** (0.626) (7.673) (5.705) (5.067) (5.530) (5.227) (6.102) (6.874) War dummy (0|1) -0.067*** -0.167*** -0.172*** -0.163*** -0.165*** -0.157*** -0.149*** -0.152*** (0.008) (0.027) (0.023) (0.017) (0.016) (0.012) (0.011) (0.010) Year 2022 (0|1) -0.047*** -0.027* -0.023* -0.024** -0.023** -0.018** -0.017** -0.012 (0.007) (0.014) (0.013) (0.010) (0.011) (0.008) (0.008) (0.008) Year 2023 (0|1) -0.052*** -0.008 -0.013 -0.018 -0.014 -0.026** -0.035*** -0.038*** (0.009) (0.022) (0.022) (0.017) (0.016) (0.011) (0.011) (0.010) Year 2024 (0|1) -0.042*** -0.007 -0.015 -0.025 -0.020 -0.036*** -0.043*** -0.046*** (0.009) (0.023) (0.025) (0.019) (0.018) (0.013) (0.012) (0.011) Missing NMV (0|1) 0.297*** 0.222*** 0.173** 0.145* 0.140* 0.108* 0.087 0.067 (0.017) (0.077) (0.067) (0.081) (0.077) (0.056) (0.062) (0.046) Sales below NMV (0|1) -0.349*** -0.347*** -0.347*** -0.335*** -0.336*** -0.311*** -0.306*** -0.293*** (0.003) (0.046) (0.027) (0.014) (0.011) (0.008) (0.007) (0.006) N benef. 0.325*** 0.319*** 0.315*** 0.302*** 0.291*** 0.265*** 0.254*** 0.246*** (0.021) (0.065) (0.056) (0.055) (0.045) (0.044) (0.044) (0.046) Repeated (1-last/5-old) 0.033*** 0.015 0.015 0.009 0.009 0.002 0.001 -0.008 (0.013) (0.043) (0.030) (0.021) (0.019) (0.017) (0.013) (0.011) N sales 0.016* 0.056* 0.041 0.041 0.031 0.042** 0.041** 0.044*** (0.009) (0.028) (0.038) (0.029) (0.026) (0.019) (0.018) (0.014) City, km (log) -0.038*** -0.037** -0.047*** -0.044*** -0.039*** -0.039** -0.008 0.000 (0.001) (0.014) (0.012) (0.014) (0.014) (0.018) (0.007) (0.007) Motorway, km (log) -0.020*** 0.001 0.004 -0.008 -0.011 -0.030*** -0.022*** -0.014** (0.001) (0.011) (0.012) (0.008) (0.007) (0.009) (0.009) (0.006) Primary road, km (log) -0.020*** -0.020* -0.024*** -0.018*** -0.019*** -0.026*** -0.026*** -0.021*** (0.001) (0.010) (0.008) (0.006) (0.006) (0.007) (0.009) (0.006) Secondary road, km (log) -0.004*** -0.006 -0.005 0.001 0.001 -0.014*** -0.018*** -0.017*** (0.001) (0.007) (0.007) (0.005) (0.005) (0.004) (0.005) (0.005) Tertiary road, km (log) -0.035*** -0.013** -0.011*** -0.015*** -0.014*** -0.015*** -0.013*** -0.015*** (0.001) (0.005) (0.004) (0.003) (0.003) (0.003) (0.003) (0.003) Railway, km (log) 0.004*** 0.000 -0.005 -0.003 -0.007 -0.014*** -0.019*** -0.017*** (0.002) (0.007) (0.006) (0.006) (0.005) (0.005) (0.005) (0.005) Grain storage, km (log) -0.055*** -0.052** -0.031** -0.030** -0.011 0.000 0.014 0.010 (0.003) (0.025) (0.015) (0.013) (0.009) (0.009) (0.009) (0.010) Port terminal, km (log) 0.114*** 0.062 0.091* 0.094** 0.092* -0.054 -0.160** 0.011 (0.004) (0.051) (0.049) (0.040) (0.054) (0.078) (0.081) (0.092) Builtup area share (L1) 1.580*** 1.364*** 1.331*** 1.157*** 1.049*** 0.889*** 0.669*** 0.444*** (0.055) (0.159) (0.144) (0.126) (0.123) (0.115) (0.097) (0.085) Crop area share (L1) 1.116*** 0.814*** 0.782*** 0.731*** 0.679*** 0.615*** 0.569*** 0.408*** (0.032) (0.095) (0.077) (0.071) (0.073) (0.066) (0.062) (0.061) Forest area share (L1) 0.928*** 0.710*** 0.628*** 0.572*** 0.497*** 0.437*** 0.376*** 0.221*** (0.036) (0.165) (0.120) (0.093) (0.089) (0.073) (0.068) (0.067) Grassland area share (L1) 0.728*** 0.512*** 0.495*** 0.454*** 0.411*** 0.372*** 0.336*** 0.208*** (0.032) (0.119) (0.094) (0.077) (0.074) (0.069) (0.064) (0.063) Soil density, cg/cm3 (log) 2.564*** -0.761 -1.607* -1.880*** -1.641*** -1.746*** -1.595*** -1.216*** (0.050) (0.979) (0.963) (0.543) (0.342) (0.250) (0.279) (0.221) Soil nitrogen, cg/kg (log) -36.031*** -26.127*** -10.537* -10.694*** -8.829*** -7.704*** -7.501*** -11.117*** (0.830) (8.288) (6.021) (3.680) (2.552) (2.018) (1.971) (1.922) Soil nitrogen, cg/kg (log sq.) -17.673*** -16.586*** -15.578*** -10.036*** -7.427*** -4.476*** -4.413*** -3.527*** (0.662) (5.243) (4.139) (3.235) (1.996) (1.417) (1.447) (1.352) Soil pH (log) 16.007*** 40.774*** 45.003*** 49.831*** 42.129*** 38.342*** 33.620*** 24.998*** (1.030) (9.599) (9.748) (9.604) (5.121) (3.727) (3.881) (2.641) Soil pH (log sq.) -10.413*** -5.928 0.065 5.756 3.729 5.352** 4.929** 3.001* (0.768) (8.746) (5.101) (4.900) (2.878) (2.152) (2.026) (1.807) Soil carbon, t/ha (log) 29.688*** 36.760*** 23.774*** 18.026*** 18.790*** 17.962*** 18.832*** 19.657*** (0.783) (10.884) (8.065) (4.556) (3.250) (2.112) (2.098) (2.162) Soil carbon, t/ha (log squared) 23.458*** 16.566*** 11.481** 6.709** 7.593*** 5.322*** 6.034*** 6.275*** (0.621) (3.531) (5.288) (2.914) (2.417) (1.520) (1.553) (1.592) Num.Obs. 176609 176609 176609 176609 176609 176609 176609 176609 R2 Adj. 0.215 0.293 0.327 0.383 0.422 0.534 0.572 0.610 R2 Within Adj. 0.188 0.163 0.151 0.143 0.125 0.119 0.108 RMSE 0.59 0.56 0.54 0.52 0.50 0.44 0.41 0.37 Significance levels are: `*` < 0.1, '**' < 0.05, and '***' < 0.01 Robust standard errors clustered at the level of fixed effects are reported in brackets. 20 Table 3. Frequency of fixed effects used for in-sample predictions Region Total Cad. block Cad. zone Village Community Old rayon Rayon Oblast Center 82,133 40,955 25,298 10,162 5,640 72 6 0 East 10,261 4,960 557 3,134 1,486 99 23 2 North 38,339 19,383 10,984 5,947 1,979 27 19 0 South 22,710 14,284 1,937 4,015 2,349 103 22 0 West 17,793 7,054 2,658 3,500 4,162 363 35 21 Country 171,236 86,636 41,434 26,758 15,616 664 105 23 Percent 50.6% 24.2% 15.6% 9.1% 0.4% 0.1% 0.0% 21 Table 4. Predictive power of different models by cross-validation split Model Fold1 Fold2 Fold3 Fold4 Fold5 Oblast 0.284 0.292 0.293 0.290 0.306 Rayon 0.318 0.328 0.328 0.322 0.337 Old rayon 0.376 0.381 0.382 0.372 0.392 Community 0.410 0.416 0.417 0.407 0.427 Cad. zone 0.506 0.509 0.511 0.493 0.513 Village 0.527 0.531 0.531 0.514 0.534 Cad. block 0.544 0.551 0.547 0.525 0.550 Ensamble 5 0.523 0.528 0.532 0.511 0.537 Ensamble 10 0.508 0.513 0.515 0.497 0.520 Ensamble 40 0.438 0.441 0.442 0.428 0.452 Reg. averages 0.446 0.447 0.453 0.434 0.458 22 Table 5: Prediction table by region Total Sample Prediction Center East North South West No. of parcels 7,619,873 171,236 7,448,637 2,746,940 740,713 1,394,287 1,154,939 1,582,994 w. NMV 1,301,405 170,062 1,131,343 491,003 73,427 246,137 178,450 312,388 Area, 1000 ha 20,989.21 537.24 20,451.97 8,065.02 3,139.05 2,849.45 5,227.99 1,707.69 Plot size, ha 2.75 3.14 2.75 2.94 4.24 2.04 4.53 1.08 Actual price 1,066 1,066 1,122 1,025 1,065 876 1,379 Pred. price 994 1,041 993 1,093 909 907 873 1,201 Reg avg price 944 981 943 1,007 928 873 832 1,135 NMV 802 801 803 819 923 727 829 667 Δ Reg. avg/NMV 17.6 22.4 17.5 22.9 0.6 20.2 0.4 70.2 Δ Pred/NMV 23.9 29.9 23.7 33.4 -1.5 24.8 5.3 80.2 Δ Actual/NMV 32.8 33.0 37.0 11.0 46.6 5.7 106.9 Note: Δ is the percentage increase of the nominator (regional average price, predicted price by regression, and actual sales prices) compared to the normative monetary value (NMV) as explained in the text. 23 Figure 1. Actual vs. predicted prices by model 24 Appendix figures and tables Figure A1. Parcel price per ha versus size relationship 25 Figure A2. Plot area, NMV and soil characteristics distribution in regression and out-of-sample data Figure A3. Distances to infrastructure distribution in regression and out-of-sample data 26 Figure A4. Transactions frequency by cadastral block Figure A5. Transactions frequency by cadastral zone 27 Figure A6. Transactions frequency by village Figure A7. Transactions frequency by community 28 Figure A8. Transactions frequency by ‘old’ rayon Figure A9. Transactions frequency by ‘new’ rayon 29 Figure A10. Frequency of NMV observations by community 30 Appendix table A1: Sizes and prices outliers in regression data and by cross validation split Fold1 Fold2 Fold3 Fold4 Fold5 Full sample Actual price (winsorized) 2024 USD/ha Min 109.89 109.89 109.89 109.89 109.89 109.89 Q1 111.78 109.89 109.89 117.69 110.83 111.73 Q5 245.26 244.14 244.86 252.63 242.84 246.14 Median 912.27 914.14 915.34 919.02 914.65 914.94 Q99 5100.16 5308.69 5228.31 5297.77 5293.28 5281.99 Max 5332.58 5332.58 5332.58 5332.58 5332.58 5332.58 Actual price 2024 USD/ha Min 0.42 0.75 0.01 0.00 7.85 0.00 Q1 111.78 109.44 108.47 117.69 110.83 111.73 Q5 245.26 244.14 244.86 252.63 242.84 246.14 Median 912.27 914.14 915.34 919.02 914.65 914.94 Q99 5100.16 5308.69 5228.31 5297.77 5293.28 5281.99 Max 706084.72 814811.52 425830.51 438432.94 878804.17 878804.17 Area, ha Min 0.00 0.00 0.00 0.00 0.00 0.00 Q1 0.11 0.11 0.11 0.11 0.11 0.11 Q5 0.33 0.33 0.33 0.34 0.34 0.33 Median 2.72 2.71 2.72 2.72 2.71 2.72 Q99 10.05 10.18 10.23 10.16 10.09 10.15 Max 129.05 374.60 223.16 104.86 147.98 374.60 31 Appendix table A2: Covariates contribution to variance explained in percent N Oblas Old Commu Cad. Cad. Pooled Rayon Village variables t rayon nity zone block All variables 30 21.5% 18.8% 16.3% 15.1% 14.3% 12.5% 11.9% 10.8% Area, HA 2 1.2% 1.6% 1.7% 1.7% 1.5% 1.2% 1.0% 1.1% War dummy 1 0.0% 0.2% 0.2% 0.2% 0.2% 0.3% 0.3% 0.3% Year dummy 3 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Missing/below NMV dummy 2 5.6% 6.1% 6.3% 6.0% 6.3% 5.5% 5.6% 5.4% Transaction details 3 0.1% 0.2% 0.2% 0.2% 0.1% 0.1% 0.1% 0.1% Settlements proximity 1 0.3% 0.3% 0.4% 0.2% 0.1% 0.1% 0.0% 0.0% Roads and storage 7 1.2% 0.6% 0.4% 0.3% 0.2% 0.3% 0.2% 0.1% Land use, lag 1 year 4 2.5% 1.7% 1.7% 1.6% 1.5% 1.4% 1.2% 0.9% Soil features 7 4.6% 3.2% 2.0% 1.6% 1.2% 1.1% 0.9% 0.6% 32 Appendix table A3: Frequencies of the FE used by region in out-of-sample prediction Region Total Cad. block Cad. zone Community Village Old rayon Rayon Oblast Center 2,664,807 785,649 705,679 618,335 484,531 56,100 14,513 0 East 730,452 96,780 23,220 293,970 158,136 91,267 45,822 21,257 North 1,355,948 317,615 335,137 343,133 331,968 10,457 17,638 0 South 1,132,229 295,911 106,980 473,511 167,701 71,770 16,356 0 West 1,565,201 151,624 113,964 918,466 221,800 133,967 18,063 7,317 Country 7,448,637 1,647,579 1,284,980 2,647,415 1,364,136 363,561 112,392 28,574 Percentage 22.1% 17.3% 35.5% 18.3% 4.9% 1.5% 0.4% 33 References: Ahlfeldt, G. M., S. Heblich and T. Seidel. 2023. "Micro-Geographic Property Price and Rent Indices." Regional Science and Urban Economics 98: 103836. Ali, D. and K. Deininger. 2021. Can Mobile Reminders and Electronic Payment Options Increase Africa’s Property Tax Collection? Evidence from a Randomized Experiment in Zambia. The World Bank, Washington, DC. Ali, D. A., K. Deininger and M. Wild. 2020. "Using Satellite Imagery to Create Tax Maps and Enhance Local Revenue Collection." Applied Economics 52 (4): 415-429. Almy, R. 2015. "Property Valuation and Taxation in Lithuania." Land Tenure Journal 15 (2): 29-48. Arnott, R. J. and J. E. Stiglitz. 1981. "Aggregate Land Rents and Aggregate Transport Costs." Economic Journal 91 (362): 331- 347. Awasthi, R., M. Nagarajan and K. W. Deininger. 2020. "Property Taxation in India: Issues Impacting Revenue Performance and Suggestions for Reform." Land Use Policy (110): 104539. Bahl, R. and S. Wallace. 2008. "Reforming the Property Tax in Developing Countries: A New Approach." International Studies Program, Andrew Young School of Policy Studies, Georgia State University, International Studies Program Working Paper Series, at AYSPS, GSU. Balan, P., A. Bergeron, G. Tourek and J. Weigel. 2020. "Land Formalization in Weak States: Experimental Evidence from Urban Property Titling in the D.R. Congo." Harvard University, Cambridge, MA. Basri, M. C., M. Felix, R. Hanna and B. A. Olken. 2019. "Tax Administration Vs. Tax Rates: Evidence from Corporate Taxation in Indonesia." National Bureau of Economic Research, Inc, NBER Working Papers: 26150. Ben-Shahar, D., R. Golan and E. Sulganik. 2020. "Tax Evasion in the Housing Market: Identification and Exploration." Journal of Real Estate Research 42 (3): 315-340. Besley, T. and T. Persson. 2014. "Why Do Developing Countries Tax So Little?" Journal of Economic Perspectives 28 (4): 99- 120. Best, M. C. and H. J. Kleven. 2018. "Housing Market Responses to Transaction Taxes: Evidence from Notches and Stimulus in the U.K." Review of Economic Studies 85 (1): 157-193. Blochliger, H., C. Hilber, O. Schoni and M. von Ehrlich. 2017. "Local Taxation, Land Use Regulation, and Land Use: A Survey of the Evidence." OECD Publishing, OECD Economics Department Working Papers: 1375. Bonnet, O., G. Chapelle, A. Trannoy and E. Wasmer. 2021. "Land Is Back, It Should Be Taxed, It Can Be Taxed." European Economic Review 134: 103696. Breuillé, M.-L., P. Duran-Vigneron and A.-L. Samson. 2018. "Inter-Municipal Cooperation and Local Taxation." Journal of Urban Economics 107: 47-64. Brockmeyer, A., A. Estefan, K. Ramirez Arras and J. C. Suarez Serrato. 2020. "Taxing Property in Developing Countries: Theory and Evidence from Mexico." World Bank, Washington DC. Cabral, M. and C. Hoxby. 2012. "The Hated Property Tax: Salience, Tax Rates, and Tax Revolts." National Bureau of Economic Research, Inc, NBER Working Papers: 18514. Carolini, G. Y., F. Gelaye and K. Khan. 2020. "Modelling Improvements to Property Tax Collection: The Case of Addis Ababa." in A. P. T. Initiative (ed.), ICTD Working Paper 103, International Centre for Tax and Development, Brighton. Carrillo, P., E. O. Castro Mendez and C. Scartascini. 2017. "Do Rewards Work? Evidence from the Randomization of Public Works." IDB WORKING PAPER SERIES No IDB-WP-794. Casaburi, L. and U. Troiano. 2016. "Ghost-House Busters: The Electoral Response to a Large Anti-Tax Evasion Program." Quarterly Journal of Economics 131 (1): 273-314. Castro, L. and C. Scartascini. 2015. "Tax Compliance and Enforcement in the Pampas Evidence from a Field Experiment." Journal of Economic Behavior and Organization 116: 65-82. Chirico, M., et al. 2017. "Procrastination and Property Tax Compliance: Evidence from a Field Experiment." National Bureau of Economic Research, Inc, NBER Working Papers: 23243. Christensen, D. and F. Garfias. 2020. "The Politics of Property Taxation: Fiscal Infrastructure and Electoral Incentives in Brazil." The Journal of Politics. Coury, M., T. Kitagawa, A. Shertzer and M. Turner. 2021. "The Value of Piped Water and Sewers: Evidence from 19th Century Chicago." National Bureau of Economic Research Working Paper 29718, Cambridge, MA. D'Arcy, M. and M. Nistotskaya. 2018. "The Early Modern Origins of Contemporary European Tax Outcomes." European Journal of Political Research 57 (1): 47-67. D’Arcy, M., M. Nistotskaya and O. Olsson. 2024. "Cadasters and Economic Growth: A Long-Run Cross-Country Panel." Journal of Political Economy 132 (11): 3785-3826. Dachis, B., G. Duranton and M. A. Turner. 2012. "The Effects of Land Transfer Taxes on Real Estate Markets: Evidence from a Natural Experiment in Toronto." Journal of Economic Geography 12 (2): 327-354. Deininger, K. W., D. A. Ali and R. Neyter. 2023. "Impacts of a Mandatory Shift to Decentralized Online Auctions on Revenue from Public Land Leases in Ukraine." Journal of Economic Behavior & Organization 213: 432-450. delCarpio, L. 2015. "Are the Neighbors Cheating? Evidence from a Social Norm Experiment on Property Taxes in Peru." Princeton, NY. 34 Diewert, W. E., C. Shimizu, T. Watanabe and K. G. Nishimura. 2020. Property Price Index: Theory and Practice. Springer, Tokyo. Diewert, W. E. and C. Shimizu. 2022. "Residential Property Price Indexes: Spatial Coordinates Versus Neighborhood Dummy Variables." Review of Income and Wealth 68 (3): 770-796. Dincecco, M. and G. Katz. 2016. "State Capacity and Long-Run Economic Performance." Economic Journal 126 (590): 189-218. Eerola, E., O. Harjunen, T. Lyytikainen and T. Saarimaa. 2021. "Revisiting the Effects of Housing Transfer Taxes." Journal of Urban Economics 124: 103367. Elinder, M. and L. Persson. 2017. "House Price Responses to a National Property Tax Reform." Journal of Economic Behavior and Organization 144: 18-39. Englund, P. 2003. "Taxing Residential Housing Capital." Urban Studies 40 (5-6): 937-952. Ermini, B. and R. Santolini. 2017. "Urban Sprawl and Property Tax of a City's Core and Suburbs: Evidence from Italy." Regional Studies 51 (9): 1374-1386. European_Comission. 2013. Handbook on Residential Property Price Indices. Publications Office of the European Union, Luxembourg. Feir, D. L., M. E. C. Jones and D. Scoones. 2023. "When Do Nations Tax? The Adoption of Property Tax Codes by First Nations in Canada." Public Choice: 1-34. Franzsen, R. and W. McCluskey. 2017. Property Tax in Africa. Lincoln Institute of Land Policy, Cambridge, MA. Garza, N. and I. Gonzalez. 2021. "An Urban System Assessment of Land Value Capture: The Colombian Case." Land Use Policy 109: 105598. Gevaert, C. M., C. Persello, R. Sliuzas and G. Vosselman. 2020. "Monitoring Household Upgrading in Unplanned Settlements with Unmanned Aerial Vehicles." International Journal of Applied Earth Observation and Geoinformation 90: 102117. Gonzalez-Navarro, M. and C. Quintana-Domeque. 2016. "Paving Streets for the Poor: Experimental Analysis of Infrastructure Effects." Review of Economics and Statistics 98 (2): 254-267. Gouriéroux, C. and A. Laferrère. 2009. "Managing Hedonic Housing Price Indexes: The French Experience." Journal of Housing Economics 18 (3): 206-213. Grover, R., M.-P. Torhonen, P. Munro-Faure and A. Anand. 2017. "Achieving Successful Implementation of Value-Based Property Tax Reforms in Emerging European Economies." Journal of European Real Estate Research 10 (1): 91-106. Gupta, A., S. Van Nieuwerburgh and C. Kontokosta. 2022. "Take the Q Train: Value Capture of Public Infrastructure Projects." Journal of Urban Economics 129: 103422. Han, L., R. Ngai and K. Sheedy. 2022. "To Own or to Rent? The Effects of Transaction Taxes on Housing Markets." London School of Economics, London. Hawley, Z., J. J. Miranda and W. C. Sawyer. 2018. "Land Values, Property Rights, and Home Ownership: Implications for Property Taxation in Peru." Regional Science and Urban Economics 69: 38-47. Hilber, C. A. L. and T. Lyytikainen. 2017. "Transfer Taxes and Household Mobility: Distortion on the Housing or Labor Market?" Journal of Urban Economics 101: 57-73. Hill, R. J. 2013. "Hedonic Price Indexes for Residential Housing: A Survey, Evaluation and Taxonomy." Journal of Economic Surveys 27 (5): 879-914. Hoff, K. 1991. "Land Taxes, Output Taxes, and Sharecropping: Was Henry George Right?" World Bank Economic Review 5 (1): 93-111. Hoj, A. K., M. Rahbek Jorgensen and P. Schou. 2018. "Land Tax Changes and Full Capitalisation." Fiscal Studies 39 (2): 365-380. Hollenbach, F. M. and T. N. Silva. 2019. "Fiscal Capacity and Inequality: Evidence from Brazilian Municipalities." The Journal of Politics 81 (4): 1434-1445. Huang, H., G. Y. Miller, B. J. Sherrick and M. I. Gomez. 2006. "Factors Influencing Illinois Farmland Values." American Journal of Agricultural Economics 88 (2): 458-470. Kalkuhl, M. and O. Edenhofer. 2017. "Ramsey Meets Thunen: The Impact of Land Taxes on Economic Development and Land Conservation." International Tax and Public Finance 24 (2): 350-380. Knoll, K., M. Schularick and T. Steger. 2017. "No Price Like Home: Global House Prices, 1870-2012." American Economic Review 107 (2): 331-353. Kresch, E. P., et al. 2023. "Sanitation and Property Tax Compliance: Analyzing the Social Contract in Brazil." Journal of Development Economics 160: 102954. Kuiiper, M. and R. Kaathman. 2015. "Property Valuation and Taxation in the Netherlands." Land Tenure Journal 15 (2): 47-62. Larreguy, H., J. Marshall and L. Trucco. 2018. "Breaking Clientelism or Rewarding Incumbents? Evidence from an Urban Titling Program in Mexico." Harvard University, Cambridge, MA. Le Roux, J. and M. Roma. 2019. "Accuracy and Determinants of Self-Assessed Euro Area House Prices." European Central Bank, Working Paper Series: 2328, Frankfurt. Liu, S. and X. Yang. 2020. "Property Tax Limits and Female Labor Supply: Evidence from the Housing Boom and Bust." Journal of Housing Economics 50: 101714. Livy, M. R. 2018. "Intra-School District Capitalization of Property Tax Rates." Journal of Housing Economics 41: 227-236. Lopez-Luzuriaga, A. and C. Scartascini. 2019. "Compliance Spillovers across Taxes: The Role of Penalties and Detection." Journal of Economic Behavior & Organization 164: 518-534. Lutz, B. 2015. "Quasi-Experimental Evidence on the Connection between Property Taxes and Residential Capital Investment." American Economic Journal: Economic Policy 7 (1): 300-330. 35 Martínez, L. R. 2023. "Natural Resource Rents, Local Taxes, and Government Performance: Evidence from Colombia." The Review of Economics and Statistics: 1-28. McCluskey, W., R. Franzsen, M. Kabinga and C. Kasese. 2018. "The Role of Information Communication Technology to Enhance Property Tax Revenue in Africa: A Tale of Four Cities in Three Countries." in A. P. T. Initiative (ed.), ICTD Working Paper 88, International Centre for Tax and Development, Brighton. Norregaard, J. 2013. "Taxing Immovable Property Revenue Potential and Implementation Challenges." International Monetary Fund, IMF Working Papers: 13/129. Oates, W. E. and R. M. Schwab. 1997. "The Impact of Urban Land Taxation: The Pittsburgh Experience." National Tax Journal 50 (1): 1-21. Palmquist, R. B. 1989. "Land as a Differentiated Factor of Production: A Hedonic Model and Its Implications for Welfare Measurement." Land Economics 65 (1): 23-28. Palmquist, R. B. 2005. "Property Value Models." in K. G. Maler and J. R. Vincent (eds.), Handbook of Environmental Economics. Volume 2. Valuing Environmental Changes, North-Holland, Elsevier, Amsterdam and San Diego. Perez-Truglia, R. and U. Troiano. 2018. "Shaming Tax Delinquents." Journal of Public Economics 167: 120-137. Poggio, L., et al. 2021. "Soilgrids 2.0: Producing Soil Information for the Globe with Quantified Spatial Uncertainty." SOIL 7 (1): 217-240. Presbitero, A. F., A. Sacchi and A. Zazzaro. 2014. "Property Tax and Fiscal Discipline in Oecd Countries." Economics Letters 124 (3): 428-433. Regan, T. and P. Manwaring. 2023. "Public Disclosure and Tax Compliance." Georgetown University, Washington DC. Rosen, S. 1974. "Hedonic Prices and Implicit Markets: Product Differentiation in Pure Competition." Journal of Political Economy 82 (1): 34-55. Sanchez-Talanquer, M. 2020. "One-Eyed State: The Politics of Legibility and Property Taxation." Latin American Politics and Society 62 (3): 65-93. Schwerhoff, G., O. Edenhofer and M. Fleurbaey. 2020. "Taxation of Economic Rents." Journal of Economic Surveys 34 (2): 398- 423. Segú, M. 2020. "The Impact of Taxing Vacancy on Housing Markets: Evidence from France." Journal of Public Economics 185: 104079. Slack, E. and R. Bird. 2014. "The Political Economy of Property Tax Reform." in OECD (ed.), OECD Working Papers on Fiscal Federalism No. 18, Paris. Slemrod, J., O. Ur Rehman and M. Waseem. 2022. "How Do Taxpayers Respond to Public Disclosure and Social Recognition Programs? Evidence from Pakistan." Review of Economics and Statistics 104 (1): 116-132. Song, Y. and Y. Zenou. 2006. "Property Tax and Urban Sprawl: Theory and Implications for Us Cities." Journal of Urban Economics 60 (3): 519-534. Stöcker, C., et al. 2017. "Review of the Current State of Uav Regulations." Remote Sensing 9 (5). Surico, P. and R. Trezzi. 2019. "Consumer Spending and Property Taxes." Journal of the European Economic Association 17 (2): 606-649. Tourek, G., P. Balan, A. Bergeron and J. Weigel. 2020. "Local Elites as State Capacity: How City Chiefs Use Local Information to Increase Tax Compliance in the Dr Congo." Harvard University, Cambridge, MA. Weigel, J. L. 2020. "The Participation Dividend of Taxation: How Citizens in Congo Engage More with the State When It Tries to Tax Them." The Quarterly Journal of Economics 135 (4): 1849-1903. Wong, F. 2020. "Mad as Hell: Property Taxes and Financial Distress." National Bureau for Economic Research, Cambridge MA. 36