## Potential Growth in LAC5 ## Source details **Canonical URL:** [Potential Growth in LAC5](https://www.imf.org/-/media/files/publications/wp/2017/wp1735.pdf) ## Other formats - [Markdown version](/-/media/files/publications/wp/2017/wp1735.pdf.md) - [Structured JSON version](/-/media/files/publications/wp/2017/wp1735.pdf.json) --- ### III. Major findings — introduction and motivation - Potential growth in LAC5 is on a declining trend; with the exception of Mexico, potential growth in LAC5 fell in the last decade and under current policies is projected to decline further in most countries the next five years. - IMF (2016) identifies four common reasons for expected weak potential growth in the next five years: - (i) shortcomings in the quality of education; - (ii) low export diversity and complexity; - (iii) lower commodity prices for commodity exporters; and - (iv) inadequate infrastructure. - This paper focuses on the last reason — inadequate infrastructure — and asks whether improvements in infrastructure could lift growth in the region. - LAC5 scores relatively weakly on the quality of infrastructure dimension of the World Economic Forum Global Competitiveness Report (GCR) in 2015, relative to major emerging markets, with the notable exception of Chile. - In contrast, the region compares favorably on measures of the stock of economic infrastructure such as power generation capacity (as discussed in IMF (2016)). ### III. Empirical strategy and data - Identification approach: - Uses a Rajan and Zingales (1998) difference-in-difference framework exploiting variation across sectors in dependence on infrastructure and across countries in infrastructure quality/quantity. - Central assumption: sectors that depend relatively more on infrastructure will grow relatively faster when infrastructure improves. - Value added and dependence measures: - Sectoral value added growth for 34 sectors in 61 countries for 1995‒2011 obtained from OECD input-output tables; Latin American countries in the sample: Argentina, Brazil, Chile, Colombia, Costa Rica, and Mexico. - Dependence on infrastructure proxied by transportation inputs relative to gross output (transportation inputs scaled by gross output). - Quality of infrastructure data: - World Economic Forum Global Competitiveness Report indicators (infrastructure pillar); focus on overall quality of infrastructure and quality of roads. - Indicators are survey-based and range from 1 to 7 (higher values denote better infrastructure). - In regressions, ܨܧܹ ௖ is defined as the country average of the infrastructure or road indicator over 2006‒11. - Quantity of infrastructure data: - Length of the road network from the International Road Federation and Eurostat covering 2001‒10 (with gaps). - Panel specification uses lagged log change in length of road network (Δlog(KmRoads)). - Key econometric specifications: - Cross-sectional sector-country specification (equation (1)) uses average annual growth rate of real value added sector s in country c over 2006‒11 and interaction of sectoral dependence on transportation with country-level infrastructure quality. - Panel sector-country-time specification uses log change in real value added and interaction of sectoral dependence with lagged log change in length of road network; includes country-sector and country-year fixed effects. - Caveats and identification limits: - Framework identifies effects operating through dependence on transportation but does not capture all channels (e.g., aggregate demand multiplier from construction activity). - Assumes dependence on transportation is exogenous to infrastructure quality; in some cases sectors could adjust transportation use endogenously. ### III. Key statistics and summary dataset descriptors - Quality of infrastructure regression (Table 1): - V A growth: 0.024 0.067 2,051 - Share in total V A: 0.029 0.038 2,051 - TranspDep * WEF infrastructure: 0.104 0.166 2,051 - TranspDep * WEF roads: 0.099 0.170 2,051 - Quantity of infrastructure regression: - V A growth: 0.049 0.180 13,930 - Share in total V A: 0.029 0.035 13,930 - TransDep * Δlog(KmRoads): 0.000 0.008 13,930 ### III. Results — sectoral growth and aggregate implications - Main empirical finding: - Improvements in the quality of infrastructure raise sectoral growth; the coefficient on the interaction between sectoral dependence on transportation and country infrastructure quality (from GCR indicators) is positive and statistically significant for the full sample and for the subsample of emerging markets. - Illustrative quantitative magnitudes: - If the quality of roads in Colombia improved to the sample median, a sector with median dependence on transportation (hotels and restaurants) would grow 0.15 percentage points faster. - If the quality of infrastructure in Argentina improved to the sample median, GDP growth would increase 0.13 percentage points. - Quantity of infrastructure: - Panel results using changes in the length of the road network indicate that increases in the quantity of infrastructure benefit relatively more sectors that depend more on infrastructure. - Full-sample example: a 10 percent growth in Colombia's road network implies a sector with median dependence on transportation would grow 0.05 percentage points faster. - LAC-subsample example: a 10 percent growth in Colombia's road network implies a sector with median dependence on transportation would grow 1.2 percentage points faster. - Aggregate GDP impact (LAC sample, using LAC coefficient): increasing road networks by 1 percent could raise growth rates by 0.1‒0.2 percentage points. ### IV. Firm-level investment results - Data and specification: - Orbis sample of about 600,000 firms in 18 countries for the period 2008‒14. - Latin American coverage: Colombia (>10,000 firms), Brazil and Chile (under 100 firms each). - Dependent variable: average net investment rate of firm i in sector s in country c over a 3-year window. - Main regressor: interaction of sectoral dependence on transportation and country-level infrastructure quality (WEF indices or LPI). - Controls: firm-specific controls lagged one period (ROA, equity, cash), sector and country fixed effects; standard errors clustered at the country level. - Summary statistics (2011‒13 cross-section N Obs where reported): - Net investment rate mean: 0.092; StDev: 0.425; N Obs: 795,396 - TransDep * WEF infrastructure mean: 0.189; StDev: 0.260; N Obs: 795,396 - TransDep * WEF roads mean: 0.175; StDev: 0.253; N Obs: 795,396 - ROA mean: 0.036; StDev: 0.088; N Obs: 732,975 - Equity mean: 0.380; StDev: 0.298; N Obs: 728,644 - Cash mean: 0.147; StDev: 0.159; N Obs: 727,741 - Main firm-level results: - Positive and significant interaction effects: firms in sectors more dependent on transportation invest relatively more when infrastructure quality improves. - Typical coefficient magnitudes (TransDep * WEF infrastructure, selected cross-sections): - 2008-10: 0.053*** - 2009-11: 0.035** - 2010-12: 0.056*** - 2011-13: 0.056*** - 2012-14: 0.029 - Using TransDep * WEF roads yields similar but slightly smaller coefficients (examples): - 2008-10: 0.050*** - 2009-11: 0.034** - 2010-12: 0.054** - 2011-13: 0.053** - 2012-14: 0.027 - ROA and cash generally have positive and significant coefficients; equity sometimes displays a negative and occasionally significant coefficient. - Firm-level aggregate implications for Colombia: - If the quality of infrastructure in Colombia improved to the sample median (1.7 point improvement in the GCR score), the investment rate of a sector with median dependence on transportation (hotels and restaurants) would increase 0.18 percentage points (using the 2011‒13 cross-section coefficient). - Aggregating across firms in the Orbis sample for Colombia (about 11,000 firms) implies the corporate investment rate would increase 0.43 percentage points if infrastructure quality improved to the sample median. ### Robustness and auxiliary findings - Alternative infrastructure quality measure: World Bank Logistics Performance Index (LPI). - LPI correlation with GCR quality of infrastructure index: 81 percent. - Using averaged LPI (2007 and 2010) and value added growth over 2006‒10: TransDep * LPI infrastructure (full sample): 0.032*; r2 = 0.35; N = 1,917. - Using the World Bank LPI (averaged 2010 and 2012) and average investment rate over 2011‒13: TransDep * LPI Infrastructure: 0.088***; ROA: 0.253***; Equity: -0.015; Cash: 0.200***; r2 = 0.03; N = 623,404. - Robustness notes: - Sectoral results are robust for the full sample though significance is lower for some alternative measures; results are not significant for some emerging-market subsamples due to limited cross-country variation. - Firm-level results hold for emerging markets and are on average 37 percent larger than baseline; two-year windows yield smaller coefficients while four-year windows are comparable to three-year baseline. ### V. Policy-relevant implications - Targeted infrastructure improvements (quality and quantity) can yield measurable increases in sectoral growth and aggregate GDP, with larger effects concentrated in sectors that are relatively more transport-dependent. - Improving road quality to median sample levels can produce non-trivial sectoral and aggregate gains: - Example: 0.15 percentage points faster sector growth for a median-dependent sector in Colombia if roads reach the sample median. - Example: 0.13 percentage points higher GDP growth for Argentina if infrastructure quality reaches the sample median. - Infrastructure policy evaluation should account for heterogeneity across sectors in transportation dependence when estimating growth and investment benefits. - Private-sector complementarity: better infrastructure quality can stimulate higher corporate investment, particularly in transport-intensive sectors (illustrated by a 0.43 percentage point increase in the aggregate corporate investment rate for Colombia when road quality reaches the sample median). *Source: IMF working paper “Potential Growth in LAC5” (wp1735) — extracted chapter/section content.* ### 1. Potential Growth in LAC5 .......................................................................................... ### 1. Potential Growth in LAC5 ### III. Major findings — introduction and motivation - Potential growth in LAC5 is on a declining trend; with the exception of Mexico, potential growth in LAC5 fell in the last decade and under current policies is projected to decline further in most countries the next five years. - IMF (2016) identifies four common reasons for expected weak potential growth in the next five years: - (i) shortcomings in the quality of education; - (ii) low export diversity and complexity; - (iii) lower commodity prices for commodity exporters; and - (iv) inadequate infrastructure. - This paper focuses on the last reason — inadequate infrastructure — and asks whether improvements in infrastructure could lift growth in the region. - LAC5 scores relatively weakly on the quality of infrastructure dimension of the World Economic Forum Global Competitiveness Report (GCR) in 2015, relative to major emerging markets, with the notable exception of Chile. - In contrast, the region compares favorably on measures of the stock of economic infrastructure such as power generation capacity (as discussed in IMF (2016)). ### III. Empirical strategy and data - Identification approach: - Uses a Rajan and Zingales (1998) difference-in-difference framework exploiting variation across sectors in dependence on infrastructure and across countries in infrastructure quality/quantity. - Central assumption: sectors that depend relatively more on infrastructure will grow relatively faster when infrastructure improves. - Value added and dependence measures: - Sectoral value added growth for 34 sectors in 61 countries for 1995‒2011 obtained from OECD input-output tables; Latin American countries in the sample: Argentina, Brazil, Chile, Colombia, Costa Rica, and Mexico. - Dependence on infrastructure proxied by transportation inputs relative to gross output (transportation inputs scaled by gross output). - Quality of infrastructure data: - World Economic Forum Global Competitiveness Report indicators (infrastructure pillar); focus on overall quality of infrastructure and quality of roads. - Indicators are survey-based and range from 1 to 7 (higher values denote better infrastructure). - In regressions, ܨܧܹ ௖ is defined as the country average of the infrastructure or road indicator over 2006‒11. - Quantity of infrastructure data: - Length of the road network from the International Road Federation and Eurostat covering 2001‒10 (with gaps). - Panel specification uses lagged log change in length of road network (Δlog(KmRoads)). - Key econometric specifications: - Cross-sectional sector-country specification (equation (1)) uses average annual growth rate of real value added sector s in country c over 2006‒11 and interaction of sectoral dependence on transportation with country-level infrastructure quality. - Panel sector-country-time specification uses log change in real value added and interaction of sectoral dependence with lagged log change in length of road network; includes country-sector and country-year fixed effects. - Caveats and identification limits: - Framework identifies effects operating through dependence on transportation but does not capture all channels (e.g., aggregate demand multiplier from construction activity). - Assumes dependence on transportation is exogenous to infrastructure quality; in some cases sectors could adjust transportation use endogenously. ### III. Key statistics and summary dataset descriptors - Table 1 summary statistics (as presented): - Quality of infrastructure regression - V A growth0.0240.0672,051 - Share i n total V A0.0290.0382,051 - TranspDe p * WEF i nf rastructure0.1040.1662,051 - TranspDe p * WEF roads0.0990.1702,051 - Quantity of infrastructure regression - V A growth0.0490.18013,930 - Share i n total V A0.0290.03513,930 - TransDep * Δl og( KmRoads)0.0000.00813,930 ### III. Results — sectoral growth and aggregate implications - Main empirical finding: - Improvements in the quality of infrastructure raise sectoral growth; the coefficient on the interaction between sectoral dependence on transportation and country infrastructure quality (from GCR indicators) is positive and statistically significant for the full sample and for the subsample of emerging markets. - Illustrative quantitative magnitudes reported: - If the quality of roads in Colombia improved to the sample median, a sector with median dependence on transportation (hotels and restaurants) would grow 0.15 percentage points faster. - If the quality of infrastructure in Argentina improved to the sample median, GDP growth would increase 0.13 percentage points. - Quantity of infrastructure: - Panel results using changes in the length of the road network indicate that increases in the quantity of infrastructure benefit relatively more sectors that depend more on infrastructure (positive expected sign on interaction term in panel specification). ### IV. Firm-level investment results - Difference-in-difference applied to firm-level investment using Orbis panel (Bureau van Dijk): - Firms that depend more on infrastructure invest relatively more when the quality of infrastructure improves. - Example magnitude: the investment rate of a firm in the lodging and restaurant sector in Colombia would increase 0.43 percentage points if the quality of infrastructure in the country improved to the sample median. - Relationship to literature: - Adds to literature on public investment and corporate investment determinants, including IMF (2014), Bom and Ligthart (2014), Romp and de Haan (2007), Arslanalp et al. (2010), Bom and Ligthart (2010), Gupta et al. (2011), Fazzari et al. (1988), Love and Zicchino (2006), and Magud and Sosa (2015). ### V. Policy-relevant implications - Targeted infrastructure improvements (quality and quantity) can yield measurable increases in sectoral growth and aggregate GDP, with larger effects concentrated in sectors that are relatively more transport-dependent. - Improving road quality to median sample levels can produce non-trivial sectoral and aggregate gains (examples: 0.15 percentage points for a median-dependent sector in Colombia; 0.13 percentage points for Argentina’s GDP). - Infrastructure policy evaluation should account for heterogeneity across sectors in transportation dependence when estimating growth and investment benefits. - Firm-level investment responses suggest private-sector complementarity: better infrastructure quality can stimulate higher corporate investment, particularly in transport-intensive sectors (illustrated by a 0.43 percentage point increase for lodging and restaurant firms in Colombia when road quality reaches the sample median). *Source: IMF working paper “Potential Growth in LAC5” (chapter/section content provided).* ### 1.7 point improvement in the GCR score, which would put Colombia on a par with the Czech ### wp1735 - 1.7 point improvement in the GCR score, which would put Colombia on a par with the Czech Republic ### Sectoral effects of infrastructure quality - Regression evidence at the sectoral level shows positive interactions between sectoral dependence on transportation and infrastructure indicators: - TransDep * WEF infrastructure: 0.024**, 0.026* - TransDep * WEF roads: 0.026**, 0.028* - Aggregate interpretation (using sector shares in total value added): - Improving the quality of roads to the sample median and upper quartile produces non-negligible GDP effects, especially for Colombia, Argentina, and Costa Rica. - If the quality of infrastructure in Colombia improved to the sample median (an improvement of 1.7 points in the GCR score), GDP growth would increase by about 0.1 percentage points. - Country/subsample notes: - Results for the LAC sample are not significant in some specifications, likely because the variation in the GCR indicators across just four countries is very limited. ### Sectoral effects of infrastructure quantity - Effects of quantity (road network size) on sectoral growth are statistically weaker overall but larger for the LAC subsample: - Full-sample example: a 10 percent growth in Colombia's road network implies a sector with median dependence on transportation would grow 0.05 percentage points faster. - LAC-subsample example: a 10 percent growth in Colombia's road network implies a sector with median dependence on transportation would grow 1.2 percentage points faster. - Aggregate GDP impact (LAC sample, using LAC coefficient): - Increasing road networks by 1 percent could raise growth rates by 0.1‒0.2 percentage points. ### Robustness of sectoral findings - Alternative infrastructure quality measure: World Bank Logistics Performance Index (LPI). - LPI correlation with GCR quality of infrastructure index: 81 percent. - Using averaged LPI (2007 and 2010) and value added growth over 2006‒10: - TransDep * LPI infrastructure (full sample): 0.032*; r2 = 0.35; N = 1,917 - Results are robust for the full sample though significance is lower; results are not significant for the subsample of emerging markets. ### Firm-level investment: specification and data - Empirical setup: - Dependent variable: average net investment rate of firm i in sector s in country c over a 3-year window. - Main regressor: interaction of sectoral dependence on transportation and country-level infrastructure quality (WEF indices or LPI). - Controls: firm-specific controls lagged one period (ROA, equity, cash), sector and country fixed effects; standard errors clustered at the country level. - Data: - Orbis sample of about 600,000 firms in 18 countries for the period 2008‒14. - Latin American coverage: Colombia (>10,000 firms), Brazil and Chile (under 100 firms each). - Net investment rate mean: 0.092; StDev: 0.425; N Obs: 795,396 (summary for 2011‒13 cross-section). - TransDep * WEF infrastructure mean: 0.189; StDev: 0.260; N Obs: 795,396. - TransDep * WEF roads mean: 0.175; StDev: 0.253; N Obs: 795,396. - ROA mean: 0.036; StDev: 0.088; N Obs: 732,975. - Equity mean: 0.380; StDev: 0.298; N Obs: 728,644. - Cash mean: 0.147; StDev: 0.159; N Obs: 727,741. ### Firm-level investment: main results - Positive and significant interaction effects: firms in sectors more dependent on transportation invest relatively more when infrastructure quality improves. - Typical coefficient magnitudes (selected cross-sections, TransDep * WEF infrastructure): - 2008-10: 0.053*** - 2009-11: 0.035** - 2010-12: 0.056*** - 2011-13: 0.056*** - 2012-14: 0.029 - r2 values around 0.02–0.03 in baseline cross-sections. - Using TransDep * WEF roads gives similar but slightly smaller coefficients (examples): - 2008-10: 0.050*** - 2009-11: 0.034** - 2010-12: 0.054** - 2011-13: 0.053** - 2012-14: 0.027 - Firm controls: - ROA and cash generally have positive and significant coefficients. - Equity sometimes displays a negative and occasionally significant coefficient, suggesting possible omitted-variable influences. - Aggregated firm-level implications for Colombia: - If the quality of infrastructure in Colombia improved to the sample median (1.7 points in the GCR score), the investment rate of a sector with median dependence on transportation (hotels and restaurants) would increase 0.18 percentage points (using the 2011‒13 cross-section coefficient). - Aggregating across firms in the Orbis sample for Colombia (about 11,000 firms) implies the corporate investment rate would increase 0.43 percentage points if infrastructure quality improved to the sample median. ### Firm-level results for emerging markets and timing - Emerging markets subsample: - Results hold and the interaction term often displays stronger statistical significance. - Estimated effects on average are 37 percent larger than in the baseline. - Time-window robustness: - Two-year windows generally yield smaller coefficients, suggesting effects materialize over relatively longer periods. - Four-year windows show coefficients comparable to the baseline (3-year windows), supporting persistence of effects over several years. ### Robustness of firm-level findings - Using the World Bank LPI (averaged 2010 and 2012) and average investment rate over 2011‒13: - TransDep * LPI Infrastructure: 0.088***; ROA: 0.253***; Equity: -0.015; Cash: 0.200***; r2 = 0.03; N = 623,404. - Finding that improvements in infrastructure have a positive impact on corporate investment is robust to the LPI. ### Conclusions and policy-relevant magnitudes - Sectoral-level conclusion: - Improvements in the quality of infrastructure are positively associated with sectoral value added growth; aggregate gains can be economically meaningful. - Example magnitude: improving Colombia's infrastructure quality to the study median → GDP growth up by about 0.1 percentage points. - Firm-level conclusion: - Improvements in infrastructure quality raise corporate investment. - Example magnitude: improving Colombia's infrastructure quality to the study median → aggregate corporate investment rate up by 0.43 percentage points. - Broader implication: - Infrastructure improvements—both in quality and quantity—can lift growth and investment, with larger measured effects in LAC-specific regressions for quantity and notable firm-level responses where firm coverage is adequate. *Source: IMF working paper (wp1735) – extracted content from the provided PDF chapter/section.* --- _Source: https://www.imf.org/-/media/files/publications/wp/2017/wp1735.pdf_