## Cadasters, Asset Tangibility, and Growth — (wpiea2026034-source-pdf)

## Source details

**Canonical URL:** [Cadasters, Asset Tangibility, and Growth — (wpiea2026034-source-pdf)](https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026034-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2026/english/wpiea2026034-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2026/english/wpiea2026034-source-pdf.pdf.json)

---

### Overview and hypothesis
- Cadasters formalize property rights, improve transparency of ownership, and help the state raise tax revenues.
- Central hypothesis: The more a firm’s assets are tangible (e.g., buildings and machinery), the more relevant cadasters become; therefore, development of cadasters should stimulate growth more in activities that depend more on tangible assets.

### Mechanisms and channels
- Asset tangibility channel (primary focus):
  - Security of tenure and enforceable ownership records reduce risks of confiscation or illegal seizure, particularly for firms with high tangible assets.
  - Improved access to credit: tangible assets can be used as collateral when property rights are formalized.
  - Improved functioning of land and real estate markets by lowering transaction costs via visible ownership information.
- Additional channels (not primary focus):
  - Strengthening the state’s capacity to tax through information on taxable assets.
  - Enhancing administrative capacity and rule-based governance, mitigating corruption risks.

### Data and identification strategy
- Cadastral data:
  - Main dataset: D’Arcy et al. (2024); time-varying cadaster index ranges from 0 to 1; index available until 2015.
  - Index components: (i) existence of a state-administered cadaster, (ii) narrative vs cartographic, (iii) land coverage.
- Industry data:
  - Two-digit manufacturing industries (ISIC Rev. 3, 15-36) from UNIDO, covering the last six decades; dependent variables constructed as 10-year log differences (winsorized at 1st and 99th percentiles).
- Asset tangibility benchmark:
  - Industry-level dependence on tangible assets from large, listed US firms (Compustat), defined as share of net property, plant and equipment in total book-value of assets; firm-level ratios averaged over 1980-1999; median within two-digit industry used as time-invariant proxy.
- Identification:
  - Rajan and Zingales (1998) style within-country, cross-industry strategy exploiting exogenous cross-industry variation in asset tangibility.
  - Fixed effects: country-industry, country-year, industry-year.
  - Standard errors clustered at country-year level.

### Key empirical findings — growth
- Baseline interaction estimates indicate cadasters increase growth more in high-tangibility industries:
  - Table 2 (baseline, column 1) — 퐶푎푑푎푠푡푒푟푠×퐼 = 50.090*** (18.754); 푉퐴 share = -6.731*** (0.143); Observations: 61,473; R-squared: 0.745.
- Quantified reform scenarios (using cadaster index 25th→75th percentiles: 0.459→1):
  - Rubber products (ISIC 25, tangibility 0.363) vs Furniture (ISIC 36, tangibility 0.204): additional value-added growth = 4.3 percentage points cumulatively over 10 years.
  - Top three highest-tangibility industries (ISIC 26, 21, 23) vs bottom three (ISIC 18, 19, 30): additional growth boost = 10.6 percentage points on average over 10 years.
- Full reform scenario (0→1):
  - Rubber products gain 8.0 percentage points growth over a decade versus furniture; top three highest-tangibility industries gain 19.5 percentage points on average versus lowest three.
- Dynamics:
  - Local projections show interaction becomes statistically significant by year 3 and remains significant through the decade.
- Context:
  - Sample average (median) of the 10-year changes in industry value = 71.6 (61.3) percent.

### Key empirical findings — investment
- Investment results support asset tangibility mechanism:
  - Table 6 (dependent variable: 10-year change in gross fixed capital formation):
    - Column 1 (Tangibility baseline): 퐶푎푑푎푠푡푒푟푠×퐼 = 162.548*** (77.768); 푉퐴 share = -4.703*** (0.257); Observations: 33,438; R-squared: 0.606.
  - Reform scenario in investment sample (25th→75th percentiles: 0.693→1):
    - Rubber products vs furniture: additional 7.9 percentage points growth in investment over 10 years.
  - Interpretation: Cadastral reforms disproportionately boost investment in physical capital in industries with high asset tangibility.

### Robustness and alternative explanations
- Alternative industry proxies:
  - Physical capital intensity measures produce consistent interaction estimates (examples: 15.398*** (4.816); 14.890*** (4.866); 8.421*** (3.217)).
  - Correlations: physical capital intensity measures correlate with asset tangibility at 0.84 and 0.76.
- Country-level confounders (specification adding Xc,t−10×Ii):
  - Democracy interaction example: 퐶푎푑푎푠푡푒푟푠×퐼 = 42.491*** (19.336); X×I (democracy×I) = 75.584*** (19.716); Observations: 59,018; R-squared: 0.744.
  - Multiple country-level variables tested: population density, urban population, agriculture share, infant mortality, GDP per capita, trade—main 퐶푎푑푎푠푡푒푟푠×퐼 remains positive and often significant.
- Financial channel:
  - Correlation between dependence on tangible assets and external finance = -0.17.
  - Horserace regressions show financial development benefits industries dependent on external finance but do not negate cadasters’ effect.
- Placebo and subsample tests:
  - Lead cadaster index (10-year forward) yields insignificant interaction.
  - Human capital intensity interaction is insignificant.
  - Results hold when dropping advanced economies, using only countries with at least one cadaster reform, varying winsorization, sample periods, weighting schemes, clustering choices, inclusion/exclusion of US, excluding countries with full reversal of cadasters in 1991, and dropping outliers (Kuwait and Qatar).

### Non-linearity and regional heterogeneity
- Non-linearity (Table 4):
  - 퐶푎푑푎푠푡푒푟푠×퐼 = 313,327*** (73,898); 퐶푎푑푎푠푡푒푟푠^2×퐼 = -262,936*** (70,752); 푉퐴 share = -6.749*** (0.143); Observations: 61,473; R-squared: 0.745.
  - Interpretation: Marginal impact larger where cadastral institutions are weaker; joint effect remains positive and significant even when index is close to 1.
- Regional heterogeneity (Table 5) — 퐶푎푑푎푠푡푒푟푠×퐼 by region:
  - Sub-Saharan Africa: 479.124*** (166.096); Observations: 6,638; R-squared: 0.745.
  - Latin America & Caribbean: 228.219** (98.441); Observations: 9,258; R-squared: 0.783.
  - South Asia & combined: 12.400 (58.398); Observations: 10,274; R-squared: 0.753.
  - East Asia & Pacific: 25.468 (33.591); Observations: 6,104; R-squared: 0.719.
  - Middle East & North Africa: 90.743** (42.567); Observations: 5,762; R-squared: 0.714.
  - Europe & Central Asia: 80.553 (84.050); Observations: 23,196; R-squared: 0.820.
- SSA economic implication example:
  - Median SSA country (2015) cadaster index increases from 0.03 to 0.197 (moves to 25th percentile of Latin America & Caribbean) → additional growth in rubber products vs furniture over next decade = 14.8 percentage points.
  - Context: mean 10-year industry growth rates in SSA = 79.2 percent.

### Legal origin heterogeneity
- Triple-interaction results (equation (4), Table 7 and conclusion summary):
  - Cadastral reforms are more effective in common-law countries (English-origin).
  - Conclusion section example estimates:
    - 퐶푎푑푎푠푡푒푟푠×퐿
      푈퐾
      ×퐼: 300.400*** (75.464); 219.173** (92.807).
  - Among 122 countries in this analysis, 32 have English common-law tradition (La Porta et al. 2008).
  - Effect remains when controlling for democracy index.

### Stylized facts and summary statistics
- Cadaster index distribution (main sample: 126 countries, 22 industries, period 1963-2023):
  - Mean = 0.729; median = 1; standard deviation = 0.351; 25th percentile = 0.459; 75th percentile = 1.
  - Note: US excluded from main sample.
- Evolution of cadasters (1960–2015, D’Arcy et al. 2024):
  - Share with cadaster index below 0.5: 62 percent in 1960 → 29 percent in 2015.
  - Share with cadaster index above 0.5: 38 percent in 1960 → 61 percent in 2015.
  - Countries with no cadasters: 27 percent in 1960 → 9 percent in 2015.
  - Share with a full cadaster: 32 percent → 45 percent between 1960 and 2015.
- Regional cadaster index means (2015):
  - Sub-Saharan Africa: 0.16.
  - Latin America & Caribbean: 0.51.
  - South Asia & East Asia & Pacific: close to 0.75.
  - Middle East & North Africa: close to 0.75.
  - Europe & Central Asia: close to 0.75.
  - High-income countries: 0.98.
- Associations (non-causal) with macro outcomes:
  - Per capita GDP growth:
    - Average = 2.5 percent when cadaster index is above median.
    - Average = 1.1 percent when cadaster index is below median.
  - Private investment (gross fixed capital formation as percent of GDP):
    - 18.2 percent where cadaster index is above median.
    - 14.6 percent where cadaster index is below median.

### Conclusions and policy recommendations
- Main conclusions:
  - Development of cadastral institutions fosters long-term economic growth, especially in industries with greater asset tangibility.
  - Increased investment in physical capital in those industries is a principal channel for the observed growth differential.
- Policy implications:
  - Strengthening cadastral institutions can act as a catalyst for economic development by unlocking investment in asset-tangible industries.
  - Effective cadastral reforms typically require administrative capacity and financing and operate within broader legal and institutional environments.
  - Legal and institutional settings that protect property rights and enforce contracts (e.g., features of some common-law systems) amplify cadastral reform impacts.
- Implementation considerations:
  - Reforms can take time to yield full effects.
  - Complementary improvements in legal frameworks and administrative capacity enhance the effectiveness of cadastral reforms.

*Source: IMF Working Paper "Cadasters, Asset Tangibility, and Growth" (wpiea2026034-source-pdf).*

### 1. Introduction

### 1. Introduction

### Overview and hypothesis
- Cadasters—records of land and property ownership—are an important pillar of a country's institutional framework: they formalize property rights, improve transparency of ownership, and help the state raise tax revenues.
- Central hypothesis: The more a firm’s assets are tangible (e.g., buildings and machinery), the more relevant cadasters become; therefore, development of cadasters should stimulate growth more in activities that depend more on tangible assets.

### Mechanisms linking cadasters to growth (focus on asset tangibility channel)
- Security of tenure and ownership rights through legally recognized and enforceable records reduces risks of confiscation or illegal seizure, particularly relevant for firms with high tangible assets.
- Improved access to credit: tangible assets can be used as collateral when property rights are formalized.
- Improved functioning of land and real estate markets by lowering transaction costs via visible ownership information.
- Additional channels (not the primary focus here) include:
  - Strengthening the state’s capacity to tax by providing information on taxable assets and liable parties, enabling investment in growth-enhancing public projects such as physical infrastructure.
  - Enhancing administrative capacity and rule-based governance, mitigating corruption risks, and improving transparency and implementation of public policies.

### Data sources and identification strategy
- Main cadastral dataset: D’Arcy et al. (2024), which codes (i) whether there exists a state-administered cadaster, (ii) whether the cadaster is narrative or cartographic, and (iii) how much of the country’s land is covered by the cadaster.
- Industry-level data: two-digit manufacturing industries from the UNIDO database covering the last six decades.
- Benchmarking asset tangibility: industry-level dependence on tangible assets calculated using data from large, listed firms from the US (a country with a strong institutional framework and highly developed financial markets) to ensure the measure reflects production needs rather than local institutional frictions.
- Empirical approach: a Rajan and Zingales (1998) style within-country, cross-industry identification that exploits exogenous cross-industry variation in asset tangibility.
  - Controls and fixed effects:
    - Country-industry fixed effects to capture persistent industry characteristics within countries.
    - Country-year fixed effects to absorb common impacts of policy changes, reforms, and supply/demand shocks.
    - Industry-year fixed effects to absorb industry-specific global trends or opportunities.
- Limits of the approach: with granular fixed effects the method identifies differential industry-level effects (mechanism) rather than the aggregate effect of cadastral reforms on national growth.

### Key empirical findings
- More developed cadasters boost long-term growth, particularly in industries with high asset tangibility.
- Differential growth estimates:
  - Moving from the 25th to 75th percentile of the cadaster index, a typical rubber products industry (high asset tangibility) sees an additional value-added growth of 4.3 percentage points cumulatively over the 10-year period compared to the furniture industry (low asset tangibility).
  - Following the same reform, the top three industries with the highest asset tangibility experience an additional growth boost of 10.6 percentage points on average, compared to the three industries with the lowest asset tangibility.
- Cadasters have a disproportionate effect on investment in physical capital in industries with high asset tangibility, consistent with the proposed mechanism that cadastral reforms boost investment in tangible assets.

### Robustness checks and alternative explanations considered
- The analysis rules out several alternative explanations:
  - Cadasters as a proxy for broader institutional quality or state capacity improvements.
  - Confounding from population density, urbanization, or greater share of non-traditional production raising both cadastral need and growth in tangible-asset industries.
  - Confounding from economic development and trade raising the need for cadastral reforms and changing asset composition.
  - Exclusive testing of a financial channel (credit/collateral) does not overturn the main result; the result remains similar.
- Reverse causality concern mitigated by using industry-level growth (growth in a two-digit manufacturing industry is unlikely to trigger a nationwide cadastral reform).

### Heterogeneity: region and legal origin
- Differential effects by region and initial cadastral strength:
  - Cadastral development is on average lowest in Sub-Saharan Africa (SSA) in the last year of the sample; Latin America & Caribbean follows.
  - Example scenario: if the median country in SSA goes through a cadastral reform reaching the country at the 25th percentile in Latin America & Caribbean, the additional growth in rubber products compared to furniture industry over a decade is estimated to be 14.8 percentage points.
- Legal-origin heterogeneity:
  - Cadastral reforms are more effective in common-law countries (originating in English law), consistent with common-law systems having lower formalism of judicial procedures and greater judicial independence, features associated with better contract enforcement and protection of property rights.
  - The effectiveness in common-law countries suggests that legal frameworks conducive to investment (property right protection and contract enforcement) amplify cadastral reform impacts.

### Contribution to literature and scope
- Advances the literature on property rights and growth by focusing specifically on cadasters and identifying an asset tangibility channel in a cross-country setting using granular industry-level data.
- Methodological contribution: applies an identification strategy from Rajan and Zingales (1998) to help address omitted variables and reverse causality concerns, making a step toward causal inference linking cadasters to growth via asset tangibility.
- Novelty claim: to the author’s knowledge, this is the first study examining causality between cadasters and growth using granular industry-level data in a cross-country setting.

*Source: wpiea2026034-source-pdf - 1. Introduction.*

### 2. Data

### 2. Data

### Industry-level variables
- Data on industry value-added and gross fixed capital formation (in US dollars) for two-digit manufacturing industries (ISIC Rev. 3, 15-36) are from the UNIDO database, available starting from the early 1960s.
- The 10-year change in value added is calculated as the log difference between years t and t-10, and winsorized at the 1st and 99th percentiles. The 10-year change in gross fixed capital formation for each industry is similarly calculated.
- Industry value-added share in each country’s manufacturing sector (in percent) is used as a control variable.
- To measure industries’ use of tangible assets (asset tangibility), the paper adopts a proxy from publicly listed US firms (Compustat), defined as the share of net property, plant and equipment in total book-value of assets (following Baker and Wurgler 2002, Braun 2005, Braun and Larrain 2005; adopted from Igan et al. 2022).
  - The firm-level annual ratios are averaged over the period 1980-1999 to smooth temporal fluctuations. The median within each two-digit industry is used as the time-invariant industry-level proxy.
  - Rationale: detailed historical data on tangible assets at two-digit industry level are not available for most countries and using country-specific measures would risk endogeneity (e.g., cadastral reforms affecting asset use). The US benchmark is used following Rajan and Zingales (1998).
  - The ordering of industries in terms of asset tangibility is assumed to be similar across countries and over time for identification; a robustness test splits industries into subsamples with high and low asset tangibility (dummy).
- Alternative and robustness industry-level proxies:
  - Physical capital dependence: ratio of total real capital stock to total value-added (following Nunn 2007 and Ciccone and Papaioannou 2009), calculated and adopted from Igan et al. (2022).
  - Physical capital intensity alternative from Erman and Kaat (2019), based on NBER manufacturing database in 1980.
  - Human capital intensity (HCI): ratio of non-production worker wages to total wages, averaged over 1980–1999 (from Igan et al. 2022).
  - Dependence on external finance: share of firm’s capital expenditures not financed with cash flow from operations (Rajan and Zingales 1998).
- Summary statistics and key measures from Table 1:
  - Asset tangibility: mean 0.28; standard deviation 0.13.
  - Top industries (highest asset tangibility): Coke, refined petroleum, nuclear fuel (ISIC 23): 0.617; Paper and paper products (ISIC 21): 0.510; Non-metallic mineral products (ISIC 26): 0.421; Basic metals (ISIC 27): 0.397; Food and beverages (ISIC 15): 0.363; Rubber and plastics products (ISIC 25): 0.363.
  - Lowest industries (lowest asset tangibility): Wearing apparel, fur (ISIC 18): 0.116; Leather, leather products, footwear (ISIC 19): 0.128; Office, accounting, computing machinery (ISIC 30): 0.133; Medical, precision, optical instruments (ISIC 33): 0.155.
  - Selected entries from Table 1 (Industry — ISIC — Asset tangibility — Physical capital intensity — Physical capital intensity (alternative)):
    - Food and beverages — 15 — 0.363 — 1.466 — 1.81
    - Tobacco products — 16 — 0.219 — 1.047 — 0.86
    - Textiles — 17 — 0.329 — 1.287 — 1.95
    - Wearing apparel, fur — 18 — 0.116 — 0.440 — 0.49
    - Leather, leather products, footwear — 19 — 0.128 — 0.630 — 0.61
    - Wood products (excl. furniture) — 20 — 0.293 — 1.126 — 2.08
    - Paper and paper products — 21 — 0.510 — 1.674 — 2.23
    - Printing and publishing — 22 — 0.267 — 0.911 — 1.00
    - Coke, refined petroleum, nuclear fuel — 23 — 0.617 — 2.496 — 2.42
    - Chemicals and chemical products — 24 — 0.206 — 1.768 — 2.31
    - Rubber and plastics products — 25 — 0.363 — 1.253 — 2.14
    - Non-metallic mineral products — 26 — 0.421 — 1.600 — 2.63
    - Basic metals — 27 — 0.397 — 2.272 — 3.35
    - Fabricated metal products — 28 — 0.278 — 1.137 — 1.36
    - Machinery and equipment n.e.c. — 29 — 0.216 — 0.995 — 1.39
    - Office, accounting, computing machinery — 30 — 0.133 — 0.752 — 0.62
    - Electrical machinery and apparatus — 31 — 0.245 — 0.938 — 1.04
    - Radio, TV, communication equipment — 32 — 0.179 — 1.108 — 1.01
    - Medical, precision, optical instruments — 33 — 0.155 — 0.549 — 0.69
    - Motor vehicles, trailers, semi-trailers — 34 — 0.270 — 1.365 — 2.28
    - Other transport equipment — 35 — 0.250 — 0.914 — 0.81
    - Furniture; manufacturing n.e.c. — 36 — 0.204 — 0.764 — 0.97
  - Correlation of physical capital intensity measures with asset tangibility: 0.84 and 0.76.

### Cadasters
- Definition: A cadaster is a public record which includes information on a land or real estate asset and the party that holds interests (rights, restrictions, responsibilities) over the asset; it compiles information on what, who, and which (interest), and includes fiscal cadasters and legal cadasters.
- Cadastral index (from D’Arcy et al. 2024):
  - Constructed via weighting answers to: (i) whether there exists a state-administered cadaster, (ii) whether the cadaster is narrative or cartographic, and (iii) how much of the country’s land is covered by the cadaster.
  - Generates a time-varying index ranging from 0 to 1: 0 = no state-administered cadaster; 1 = a mapped cadaster that covers at least 90 percent of the territory (full cadaster).
  - The index is available until 2015.

### Other country-level variables
- Institutional environment proxies:
  - Index on liberal democracy from the V-Dem dataset by Coppedge et al. (2019), ranges from 0 to 1.
  - Alternative proxies from Polity V dataset (Center for Systemic Peace): policy score (ranging from -10 to 10), index on democracy (ranging from 0 to 10), index on the degree of institutionalized constraints on political executives (ranging from 1 to 7). Higher values represent stronger institutions.
- Macroeconomic and demographic controls (World Bank WDI unless noted): GDP (constant 2015 US dollars, in log), GDP per capita (constant 2015 US dollars, in log), population (in log), population density (people per square kilometer of land area, in log), trade (in percent of GDP), value-added in agriculture (in percent of GDP), infant mortality rate (per 1,000 live births, in log), urban population (in percent of total population), gross fixed capital formation by the private sector (in percent of GDP).
- Financial development proxy: IMF’s composite index (Svirydzenka 2016) ranging from 0 to 1, constructed via principal component analysis; higher values indicate more developed financial systems.
- Legal origin: data from La Porta et al. (2008); countries separated by a dummy equal to 1 for English-origin common-law countries and 0 for civil-law countries. Scandinavian legal-origin countries are dropped in the relevant part of the analysis.

### Sample
- Main analysis sample:
  - 22 manufacturing industries (two-digit ISIC, see Table 1).
  - 126 countries, of which 36 are advanced economies.
  - Period 1963-2023. The sample is restricted by availability of industry-level variables and the cadaster index.
  - The US is excluded from the sample because industry-level proxies are calculated using US data.
  - While cadaster data ends in 2015, the regression analysis extends to 2023 because it is based on 10-year windows.
- Cadaster index distribution in the main sample:
  - Mean value 0.729; median 1.
  - Standard deviation 0.351.
  - 25th percentile 0.459; 75th percentile 1.

### Stylized facts — Evolution of cadasters (1960–2015)
- Global changes documented (D’Arcy et al. 2024):
  - Share of countries with a low or medium-low cadaster index (below 0.5): 62 percent in 1960 → 29 percent in 2015.
  - Share with a high or medium-high cadaster index (above 0.5): 38 percent in 1960 → 61 percent in 2015.
  - Progress appears steadier starting from the early 1990s.
  - Countries with no cadasters at all: 27 percent of the global sample in 1960 → 9 percent in 2015.
  - Share of countries with a full cadaster rose from 32 percent to 45 percent between 1960 and 2015.
  - Note: Several ex-Soviet countries saw a full reversal of cadasters in 1991, based on the dataset.

### Stylized facts — Cadasters across regions (2015)
- Large heterogeneities across and within regions; Sub-Saharan Africa lags markedly.
- Regional mean cadaster index values (2015):
  - Sub-Saharan Africa: 0.16.
  - Latin America & Caribbean: 0.51.
  - South Asia & East Asia & Pacific: close to 0.75.
  - Middle East & North Africa: close to 0.75.
  - Europe & Central Asia: close to 0.75.
  - High-income countries: 0.98.
- Sub-Saharan Africa specifics:
  - Of 39 countries in the region, 31 score below 0.25 on the index.
  - 7 of these 31 lack a cadaster altogether.
  - Only 4 countries have an index above 0.75, with 2 of them having a full cadaster.
- Dispersion: minimal among high-income countries; much wider within other regions (gap between 25th and 75th percentiles large, except for Sub-Saharan Africa where most countries lack strong cadastral systems).

### Stylized facts — Cadasters and aggregate outcomes
- Associations (non-causal) between cadaster index and macro variables:
  - Per capita GDP growth:
    - Average per capita GDP growth rate = 2.5 percent when the cadaster index is high (above the median).
    - Average per capita GDP growth rate = 1.1 percent when the cadaster index is below the median.
  - Private investment (gross fixed capital formation) as percent of GDP:
    - 18.2 percent where the cadaster index is above the median.
    - 14.6 percent where the cadaster index is below the median.
- Robustness note: These patterns remain similar when using residual growth rates and investment ratios after regressing each on lagged per capita GDP level and the index on liberal democracy (to account for level of economic development and institutional environment).

*Source: Section 2 (Data) of the IMF Working Paper "Cadasters, Asset Tangibility, and Growth" (data and tables as described in the source PDF).*

### 4. Methodology

### 4. Methodology

### 4.1. Cadasters, Asset Tangibility, and Growth
- Objective:
  - Examine the differential impact of cadasters on industry growth based on variation in industries’ asset tangibility.
- Main specification (equation (1)):
  - Δlog(푉퐴)푐,푖,푡−10→푡 = 훽1 퐶푎푑푎푠푡푒푟푐,푡−10 × 퐼푖 + 훼 푉퐴 푠ℎ푎푟푒푐,푖,푡−10 + 휃푐,푡 + 휃푐,푖 + 휃푖,푡 + 휖푐,푖,푡
  - Dependent variable: 10-year change in industry value-added from year 푡−10 to 푡 (in percent).
  - Alternative dependent variables: 10-year change in investment in physical capital variables in separate regressions.
  - Key interaction: 퐶푎푑푎푠푡푒푟푐,푡−10 × 퐼푖; coefficient 훽1 captures the asset tangibility channel of the growth impact of cadasters.
- Variables and measures:
  - 퐼푖: proxy for asset tangibility in each two-digit industry (alternative: industries’ physical capital intensity in separate estimations).
  - 퐶푎푑푎푠푡푒푟푐,푡−10: index from D’Arcy et al. (2024), representing development of cadastral institutions in each country, adopted from year 푡−10.
  - 푉퐴 푠ℎ푎푟푒푐,푖,푡−10: industry value-added share in total manufacturing value-added in its country (10-year lagged) to capture potential convergence patterns.
- Identification strategy and fixed effects:
  - Country-year fixed effects (휃푐,푡): isolate variation common across two-digit industries in a country in each year; control for time-variant economic, political, and institutional developments and country-specific shocks.
  - Country-industry fixed effects (휃푐,푖): account for unobserved reasons causing a distinct average growth pattern in an industry in a country.
  - Industry-year fixed effects (휃푖,푡): absorb variation from industry-specific global trends, shocks, or growth opportunities at an annual frequency.
  - Note on inflation: year-fixed effects (country-year and industry-year) absorb role of inflation dynamics; industry-year fixed effects eliminate common global (US dollar) inflation for each industry, country-year fixed effects absorb country-specific inflation components.
- Standard errors and reverse causality:
  - Standard errors are clustered at the country-year level where the policy variable changes.
  - Using two-digit industry-level data mitigates reverse causality concerns (growth in a single industry is unlikely to trigger nationwide institutional reform).
- Trade-off / limitation:
  - Inclusion of country-year fixed effects captures differences within each country-year cell but prevents gauging the effect of cadasters on aggregate growth (overall effect captured by fixed effects).

### Local projections and robustness to other country-level variables
- Local projections (Jorda, 2005) specification (equation (2)):
  - Δlog(푉퐴)푐,푖,푡−1→푡+푝 = 훽1푝 퐶푎푑푎푠푡푒푟푐,푡−1 × 퐼푖 + 훼 푉퐴 푠ℎ푎푟푒푐,푖,푡−1 + 휃푐,푡 + 휃푐,푖 + 휃푖,푡 + 휖푐,푖,푡
  - Implementation: run for 10 years (푝 = 0,...,9) and report 훽1푝 to capture how impact evolves and accumulates over time.
- Accounting for other country-level variables (equation (3)):
  - Δlog(푉퐴)푐,푖,푡−10→푡 = 훽1 퐶푎푑푎푠푡푒푟푐,푡−10 × 퐼푖 + 훽2 푋푐,푡−10 × 퐼푖 + 훼 푉퐴 푠ℎ푎푟푒푐,푖,푡−10 + 휃푐,푡 + 휃푐,푖 + 휃푖,푡 + 휖푐,푖,푡
  - 푋푐,푡−10: set of country-level variables, including macroeconomic, institutional and financial development.
  - Interpretation: 훽2 captures the impact of those country-level variables varying with industries’ asset tangibility. If cadasters are not a proxy for these factors, 훽1 should remain similar to the baseline.

### 4.2. Legal Origin
- Objective:
  - Investigate whether a country’s legal origin influences the differential impact of cadasters on industries with greater asset tangibility.
- Triple-interaction specification (equation (4)):
  - Δlog(푉퐴)푐,푖,푡−10→푡 = 훽1 퐶푎푑푎푠푡푒푟푐,푡−10 × 퐼푖 + 훽2 퐶푎푑푎푠푡푒푟푐,푡−10 × 퐿푈퐾 × 퐼푖 + 훼 푉퐴 푠ℎ푎푟푒푐,푖,푡−10 + 휃푐,푡 + 휃푐,푖 + 휃푖,푡 + 휖푐,푖,푡
  - 퐿푈퐾: dummy variable representing common-law countries (1 for common-law countries, 0 otherwise).
  - Interpretation: positive 훽2 implies the differential impact of cadasters on growth in industries with greater asset tangibility is more pronounced in common-law countries.
- Estimation note:
  - Interactions between (i) legal origin and asset tangibility and (ii) legal origin and cadasters cannot be included due to the presence of fixed effects (휃푐,푖 and 휃푐,푡).

*IMF Working Paper — 4. Methodology*

### 5. Results

### 5. Results

### 5.1. Main Results
- Table 2 (baseline, column 1) — interaction term 퐶푎푑푎푠푡푒푟푠×퐼: 50.090*** (18.754). 푉퐴 share: -6.731*** (0.143). Observations: 61,473. R-squared: 0.745.
- Interpretation: Positive and statistically significant interaction implies cadasters lead to higher growth, particularly in industries with greater asset tangibility (baseline proxy from Igan et al. 2022).
- Quantified reform scenarios (using 25th and 75th percentiles of cadaster index across sample: 0.459 to 1):
  - Rubber products (ISIC 25, 75th percentile tangibility) versus Furniture products (ISIC 36, 25th percentile tangibility): additional growth of 4.3 percentage points over the 10-year period. Calculation uses coefficient 50.090 and tangibility gap (0.363 versus 0.204) and cadaster jump 0.459→1.
  - Top three industries with highest tangibility (ISIC 26, 21, 23) versus bottom three (ISIC 18, 19, 30): additional growth boost of 10.6 percentage points on average over 10 years for the reform 0.459→1.
  - Transition from no cadaster to full cadaster (0→1): rubber products gain 8.0 percentage points growth over a decade compared to furniture; top three highest-tangibility industries gain 19.5 percentage points on average versus lowest three.
- Context: Sample average (median) of the 10-year changes in industry value is 71.6 (61.3) percent.
- Robustness to alternative industry proxies (columns 3–4): Physical capital intensity proxies (Igan et al. 2022; Erman and Kaat 2019) produce consistent results. For baseline table entries: columns 2–4 interaction estimates: 15.398*** (4.816); 14.890*** (4.866); 8.421*** (3.217). 푉퐴 share remains ~-6.73.
- Dynamics: Local projections (Figure 4) based on equation (2) — interaction effect becomes statistically significant by the third year and remains significant throughout the decade. Dependent variable: change in industry value-added; dashed lines depict 90 percent confidence intervals.
- Convergence: Negative coefficient estimate on lagged value-added share indicates smaller industries grow faster.

### 5.2. Robustness
- Specification (3) tests country-level confounders. Table 3 reports interaction coefficients and additional country-level interactions:
  - Column 1 (Democracy interaction): 퐶푎푑푎푠푡푒푟푠×퐼 = 42.491*** (19.336). X×I (democracy×I) = 75.584*** (19.716). 푉퐴 share = -6.897*** (0.137). Observations: 59,018. R-squared: 0.744.
  - Column 2 (Population density): 퐶푎푑푎푠푡푒푟푠×퐼 = 50.543*** (19.018). X×I = -30.529* (16.661). Observations: 59,560. R-squared: 0.744.
  - Column 3 (Urban population): 퐶푎푑푎푠푡푒푟푠×퐼 = 45.108** (18.780). X×I = 1.064* (0.578). Observations: 60,658. R-squared: 0.744.
  - Column 4 (Agriculture share): 퐶푎푑푎푠푡푒푟푠×퐼 = 88.514*** (29.243). X×I = 0.552 (0.897). Observations: 47,575. R-squared: 0.749.
  - Column 5 (Infant mortality): 퐶푎푑푎푠푡푒푟푠×퐼 = 50.717*** (19.525). X×I = -14.362*** (12.062). Observations: 59,170. R-squared: 0.744.
  - Column 6 (GDP per capita): 퐶푎푑푎푠푡푒푟푠×퐼 = 60.387*** (19.628). X×I = 42.562*** (10.257). Observations: 59,146. R-squared: 0.744.
  - Column 7 (Trade): 퐶푎푑푎푠푡푒푟푠×퐼 = 50.834** (21.215). X×I = 0.333* (0.179). Observations: 54,517. R-squared: 0.750.
- Institutional quality: Including liberal democracy (V-Dem) interaction does not remove main effect; growth in high-tangibility industries also appears higher under more liberal democracies. Appendix tests with Polity V proxies show similar robustness.
- Financial development channel:
  - Correlation between proxies for dependence on tangible assets and external finance is -0.17 (negative and small).
  - Horserace regressions (Appendix Table A.3) show no evidence that financial development undermines cadasters' effect via the asset tangibility channel; financial development disproportionately benefits industries dependent on external finance but does not negate cadasters’ role.
- Placebo tests (Appendix Table A.4):
  - Replacing 10-year lagged cadaster index with 10-year forward value (lead) yields statistically insignificant interaction — future index values do not predict growth in high-tangibility industries.
  - Replacing asset tangibility with human capital intensity yields insignificant interaction.
- Subsample tests (Appendix Table A.5):
  - Dropping advanced economies; using only countries with at least one cadastral reform — results hold.
- Additional unreported robustness checks confirm results hold when varying winsorization, sample periods (starting in 1980s or 1990s, or ending in 2019), weighted regressions (weights: 10-year lagged log real GDP, GDP per capita, industry value-added or inverse), clustering choices, inclusion/exclusion of US, excluding countries with full reversal of cadasters in 1991, and dropping outliers Kuwait and Qatar.

### 5.3. Non-linearity and Heterogeneity across Regions
- Non-linearity (Table 4): specification adds interaction with squared cadaster index.
  - 퐶푎푑푎푠푡푒푟푠×퐼 = 313,327*** (73,898).
  - 퐶푎푑푎푠푡푒푟푠^2×퐼 = -262,936*** (70,752).
  - 푉퐴 share = -6.749*** (0.143). Observations: 61,473. R-squared: 0.745.
  - Interpretation: Impact of similar improvements is more pronounced where cadastral institutions are weaker; joint evaluation indicates impact remains positive and statistically significant at the 1 percent level even when index is close to 1.
- Regional heterogeneity (Table 5) — interaction 퐶푎푑푎푠푡푒푟푠×퐼 by region:
  - Sub-Saharan Africa: 479.124*** (166.096). 푉퐴 share = -5.079*** (0.234). Observations: 6,638. R-squared: 0.745.
  - Latin America & Caribbean: 228.219** (98.441). 푉퐴 share = -6.517*** (0.347). Observations: 9,258. R-squared: 0.783.
  - South Asia & (combined column): 12.400 (58.398). Observations: 10,274. R-squared: 0.753.
  - East Asia & Pacific: 25.468 (33.591). Observations: 6,104. R-squared: 0.719.
  - Middle East & North Africa: 90.743** (42.567). Observations: 5,762. R-squared: 0.714.
  - Europe & Central Asia: 80.553 (84.050). Observations: 23,196. R-squared: 0.820.
- Economic implication for SSA: Reform scenario — median SSA country as of 2015 increases cadaster index from 0.03 to 0.197 (moves to 25th percentile of Latin America & Caribbean). Result: additional growth in rubber products compared to furniture over next decade = 14.8 percentage points. Context: mean 10-year industry growth rates in SSA = 79.2 percent.

### 5.4. Investment
- Hypothesis: Cadastral reforms increase industries' investment in tangible assets, driving the growth differential.
- Table 6 (dependent variable: 10-year change in gross fixed capital formation; sample shrinks due to investment data availability):
  - Column 1 (Tangibility baseline): 퐶푎푑푎푠푡푒푟푠×퐼 = 162.548*** (77.768). 푉퐴 share = -4.703*** (0.257). Observations: 33,438. R-squared: 0.606.
  - Column 2 (Tangibility dummy): 퐶푎푑푎푠푡푒푟푠×퐼 = 32.699** (16.096). 푉퐴 share = -4.681*** (0.260).
  - Column 3 (Physical capital intensity): 퐶푎푑푎푠푡푒푟푠×퐼 = 36.862* (20.040). 푉퐴 share = -4.691*** (0.258).
  - Column 4 (Physical capital intensity alternative): 퐶푎푑푎푠푡푒푟푠×퐼 = 7.729 (12.167). 푉퐴 share = -4.676*** (0.260).
- Quantified investment scenario (25th to 75th percentiles of cadaster index in investment sample: 0.693 to 1):
  - Rubber products versus furniture: additional 7.9 percentage points growth in investment over 10 years for the cadaster reform 0.693→1.
  - Context: 10-year change in gross fixed capital formation in the sample = 63.6 percent.
- Interpretation: Investment dynamics in industries with greater asset tangibility likely drive the growth findings; cadastral reforms particularly important in less developed countries with limited capital deepening.

### 5.5. Legal Origin
- Estimations based on equation (4); Table 7 examines role of legal origin (퐿푈퐾 dummy = 1 for common-law countries).
- Main finding: Positive coefficient on triple interaction (cadasters × asset tangibility × common-law dummy) in column 1 implies cadastral reforms are more effective for growth via the asset tangibility channel in common-law countries.
- Controlling for democracy index (column 2) does not alter this pattern.
- Note: Among 122 countries in this analysis, 32 have English common-law tradition (data from La Porta et al. 2008).

*Source: IMF Working Paper — chapter "5. Results" (wpiea2026034-source-pdf).*

### 6. Conclusion

### 6. Conclusion

### Main findings
- The development of cadastral institutions fosters long-term economic growth.
- The growth effect is especially pronounced in industries with greater asset tangibility.
- Increased investment in physical capital in those industries as a result of more developed cadasters contributes to the observed growth pattern.

### Empirical evidence (selected regression results from conclusion tables)
- Variable estimates (standard errors in parentheses) reported in the conclusion section:
  - 퐶푎푑푎푠푡푒푟푠×퐼: 22.009 (18.877); 26.375 (19.682)
  - 퐶푎푑푎푠푡푒푟푠×퐿
    푈퐾
    ×퐼: 300.400*** (75.464); 219.173** (92.807)
  - 퐷푒𝑚𝑜𝑐𝑟𝑎𝑐𝑦×퐼: 64.975*** (19.996)
  - 푉퐴 𝑠ℎ푎𝑟𝑒: -6.712*** (0.144); -6.876*** (0.139)
- Regression sample statistics (from the same tables):
  - Observations: 57,879; 55,424
  - R-squared: 0.743; 0.742
  - Fixed effects included: Country-year F.E., Country-industry F.E., Industry-year F.E.
- Notes provided in the tables:
  - Results are based on equation (4), with the second column adding interaction with the index on democracy.
  - Dependent variable: change in industry value-added.
  - 퐿
    푈퐾 is the dummy variable representing common law countries.
  - Democracy is the index adopted from the V-Dem dataset.
  - Standard errors are in parentheses. *** p<0.01, ** p<0.05, * p<0.1.

### Heterogeneity and robustness
- The growth impact of cadasters is more pronounced in:
  - Regions which lack strong cadasters, such as Sub-Saharan Africa.
  - Common-law countries.
- Robustness checks and additional tests summarized elsewhere in the document (tables reported in the appendix) show:
  - Alternative country-level variables (Polity score, Democracy (Polity V), Constraints on executives (Polity V), GDP, Population) included in extended specifications produce significant 퐶푎푑푎푠푡푒푟푠×퐼 coefficients in several cases (e.g., 40.687** (19.201); 56.164*** (19.500); see table notes for exact columns and specifications).
  - Tests accounting for financial development as an alternative explanation report a consistent positive 퐶푎푑푎푠푡푒𝑟𝑠×퐼 coefficient (e.g., 56.441*** (20.627); observations 42,823; R-squared 0.727), and interactions involving proxies for financial development and industries’ dependence on external finance are reported (e.g., 퐹퐷×퐸퐹퐷: 5.910** (2.816)).
  - Placebo tests and subsample analyses are reported: placebo cadaster index (10-year ahead) and human capital intensity interactions produce non-significant or differing coefficients (e.g., 퐶푎푑푎푠푡𝑒𝑟𝑠×퐼: -56.946 (35.900); 29.111 (39.151)), while subsamples such as EMDEs and countries with cadaster reform show strong positive coefficients (e.g., 55.953*** (20.941); 86.292*** (21.634)).

### Policy implications
- Strengthening cadastral institutions can act as a catalyst for economic development by:
  - Unlocking investment, particularly investment in physical capital in asset-tangible industries.
  - Bolstering long-term growth through enhanced property-rights-related institutions.
- Implementation considerations:
  - Effective cadastral reforms can take time.
  - Reforms typically require administrative capacity and financing.
  - Cadastral systems operate within broader institutional and legal environments; progress toward an environment conducive to investment helps cadastral reforms achieve their objectives.

*IMF Working Paper — Cadasters, Asset Tangibility, and Growth (Working Paper No. WP/2026/034)*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2026/english/wpiea2026034-source-pdf.pdf_