## _wp09264 ## Source details **Canonical URL:** [_wp09264](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09264.pdf) ## Other formats - [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09264.pdf.md) - [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09264.pdf.json) --- ### Introduction and observed productivity trends - Total factor productivity (TFP) in Chile at end-2007 was about 2 percent lower than at end-1997 (Ministerio de Hacienda, 2008). - During 1986–97 productivity grew by a cumulative 30 percent. - Since 1998 Chile experienced a clear change in trend: average GDP growth decelerated and measured productivity growth weakened. - Post-1997 slowdown occurred despite a marked pickup in investment from 2004 onward: - Investment in machinery and equipment has almost doubled (as percentage of GDP) from 2004, and by end-2008 amounted to about half of total investment. - More than 80 percent of Chile’s machinery and equipment (M&E) purchases are imported. ### Investment in machinery and equipment (M&E) and investment-specific productivity (ISP) - New M&E typically embodies investment-specific productivity (ISP) improvements that raise the productivity of the new stock relative to the existing stock of M&E. - Technological advances caused a secular decrease in the relative price of M&E in terms of consumption goods. - Increases in the rate of ISP improvements can produce decreases in measured productivity performance because of learning and quality mismeasurements associated with new technologies (Hornstein and Krussell, 1996). - Empirical patterns in several high-income and emerging economies show similar quantity rises and relative price declines in M&E. ### Economic model and analysis framework - Framework: adapted Greenwood, Hercowitz and Krusell (GHK, 1997) with: - Exogenous increases in labor-specific productivity. - An index of utilization of the capital stock (u), assumed known and exogenous and identical for equipment and structures. - Production technology: Cobb-Douglas with two capital types: - Machinery and equipment (k_{e,t}) and non-housing structures (k_{s,t}). - Neutral (total-factor) productivity is z; ISP is q (an index of investment-specific productivity). - Dynamics: - Structures: k_{s,t+1} = (1 - δ_s) k_{s,t} + i_{s,t}, with growth components including H and N (human capital and population). - Equipment: k_{e,t+1} = (1 - δ_e) k_{e,t} + q_t i_{e,t}; q evolves at gross growth rate γ_q. - Government taxes labor and capital and returns revenues as a lump-sum transfer to the representative household. ### Balanced growth path (BGP) conditions and calibration approach - Variables normalized per effective available labor. Along a BGP output, consumption, and investment grow at the same gross rate γ_y (output per available effective hour). - Equipment stock grows at gross rate γ_e > γ_y due to growing q. - Key BGP restrictions used (equations referenced in source): - Relations between γ_y, γ_q, γ_e, and factor shares (α_e, α_s). - Stationary transformed problem yields conditions linking calibrated parameters and BGP observables. - Calibration procedure: - Calibrates 12 parameters so that unknowns in the BGP coincide with their average observed values over the sample period. - ISP (γ_q) proxied by the ratio of the implicit price deflator for personal consumption expenditures to the implicit price deflator for equipment and machinery. - Dataset and program codes used in the calibration are available upon request. ### Calibration results and quantitative contributions to growth - Chile (1986-2008 sample) calibrated averages: - ISP (γ_q): 3.8 percent per year. - Neutral factor productivity (TNP) growth: 0.7 percent per year. - ISP accounted for about 45 percent of the long-term growth in output per effective hour in Chile. - ISP contribution to long-term growth of output per hour: 28 percent. - TNP contribution to output per hour: 37 percent. - Growth in human capital (HK) contributed the remaining 35 percent to output per hour. - Comparative averages: - ISP growth averaged about 3 percent per year across the other countries analyzed; Chile’s ISP growth (3.8 percent) was similar to Norway (3.6 percent). - TNP growth was lower than ISP growth in all countries in the sample. - Contribution of ISP to output growth increased significantly since the mid-1990s, especially in Chile and other net commodity exporters. - For net commodity exporters (Australia, Canada, Chile, Norway) TNP growth was positive until the late 1990s or early 2000s and turned negative afterward; TNP decreased in Australia, Canada, Chile and Norway since about the mid-2000s. ### Key quantitative summary (selected numbers reported exactly) - Chile: TFP at end-2007 was about 2 percent lower than at end-1997. - Investment in M&E as share of GDP: almost doubled since 2004; by end-2008 M&E investment amounted to about half of total investment. - ISP growth (Chile, average percent per year): 3.8 percent. - TNP growth (Chile, average percent per year): 0.7 percent. - Contribution to long-run output per hour (Chile, percent): - ISP: 28 percent - TNP: 37 percent - Human capital (HK): 35 percent - ISP accounted for about 45 percent of long-term growth in output per effective hour in Chile. ### What could explain the productivity slowdown? - Overestimation of the effective capital stock: - Growth accounting often assumes current investment is incorporated immediately into the capital stock; multi-year projects should not be fully added until completion. - Chile: contemporaneous investment added to the effective capital stock decreased from an average of about 75 percent during 2001-2005, to less than 60 percent in 2008. - Increase in number and relative importance of projects with long maturity, particularly in the Energy and Mining Sectors (Ministerio de Hacienda, 2007). - Similar patterns noted for other net commodity exporters undertaking sizable commodity-related projects with maturity periods exceeding one year. - Output measurement issues (unmeasurable vs measurable sectors): - Certain activities (construction, trade, finance, insurance, real estate, other services, government) are harder to measure in output and quality improvements than agriculture, mining, manufacturing, transportation, communications. - Productivity slowdowns can reflect mismeasurement if the “unmeasurable sector” grows relative to the economy. - Chile: the proportion of the unmeasurable sector in total GDP (at factor cost) has remained relatively stable. - Costly learning and adoption of new technologies: - Adoption involves learning costs; during transition output and TFP growth could decline. - Wage dispersion and skill premia would be expected to increase during the adjustment. - Chile: premium paid to the most skilled workers (upper firm/public sector management and liberal professions) with respect to unskilled workers increased by about 10 percent during 1997-2008. - Relative wage of technical workers (machine operators and artisans) to unskilled workers has remained fairly constant. - Interpretation: some increase in wage dispersion, but stability of technical-to-unskilled relative wages suggests costly learning has not been the primary cause of the productivity slowdown. - Business regulations and investment climate: - Excessive or inadequate regulation can weaken productivity by impairing the investment climate. - World Bank’s Doing Business (2009): Chile outranks Latin America but ranks below the OECD average on the general “Ease of doing business” indicator. - Chile has not made significant gains in several Doing Business categories during 2004-2009. - Increase in the average rate of unemployment after 1998 suggests regulatory rigidities may still limit the economy’s capacity to absorb shocks (see Restrepo, 2008). - Declining stage of traditional (commodity) sectors: - For economies with large non-renewable resource shares, exploitation can lead to rising marginal costs and eventual declines in production and productivity. - Chile: mining GDP has fluctuated around a constant level since 2004 (in line with copper production), while “core” GDP (excluding mining, electricity gas and water and fishing) increased at an annual average rate of 5.2 percent. - Timing: productivity slowdown in Chile started around 1998, when mining GDP and physical copper production were expanding strongly. - During 1998-2004, mining GDP increased at an average annual rate of 4.9 percent. - Physical copper production expanded at an annual average rate of 6.9 percent. - For commodity exporters in the sample (Australia, Canada, Chile and Norway), large increases in commodity prices of 2005-2008 coincide with decreases in productivity; part of the decrease could reflect expansion of production into areas of lower marginal productivity. - Market concentration and competition: - Limits on product market competition can impede convergence to frontier technology. - Chile: Global Competitiveness Report (2008) ranks Chile 28 among 137 countries, better than any country in Latin America but below the OECD average. - Corporate activity in Chile dominated by a limited number of conglomerates; Chile ranks 57th in the “market dominance” indicator. - Scarce competition in financial sector (banks and pension fund managers) may constrain access to funds for middle and small firms. - Progress: creation of an “Anti-monopoly Court” in 2004; Chile ranks 19th in the “intensity of local competition” indicator. - Conclusion: market concentration may have constrained dynamism, but does not preclude intense competition in some areas. ### Concluding observations and policy implications - Positive: marked increase in level and quality of new capital goods in Chile is favorable for future productivity growth. - Negative: slow growth in neutral technological change suggests the need for additional reforms. - Multifactor explanation for slowdown: - Regulatory rigidities that limited the economy’s capacity to absorb the external shocks of 1998 and 2001 likely contributed to the initial phase of the slowdown. - From 2004, factors that probably compounded the problem include the slowdown in the mining sector and some overestimation of the effective capital stock. - Other factors that may have played a role and warrant further investigation: costly learning following adoption of new technologies, and market concentration. - Policy focus points: - Improve measurement and accounting for multi-year investments to avoid overestimating contemporaneous capital stocks. - Continue reforms to improve the regulatory environment and labor market flexibility to enhance shock absorption and reduce rigidities associated with higher unemployment. - Address market concentration and barriers to finance for middle and small firms to boost competition and dynamism. - Investigate sectoral productivity trends (especially mining/commodity production at the margin) and the allocation of investment across projects with differing maturities and marginal productivities. *Source: Fund Staff calculations, _wp09264 - References.* ### References.............................................................................................................. ### _wp09264 - References ### Introduction and observed productivity trends - Total factor productivity (TFP) in Chile at end-2007 was about 2 percent lower than at end-1997 (Ministerio de Hacienda, 2008). - During 1986–97 productivity grew by a cumulative 30 percent. - Since 1998 Chile experienced a clear change in trend: average GDP growth decelerated and measured productivity growth weakened. - The post-1997 slowdown occurred despite a marked pickup in investment from 2004 onward: - Investment in machinery and equipment has almost doubled (as percentage of GDP) from 2004, and by end-2008 amounted to about half of total investment. - More than 80 percent of Chile’s machinery and equipment (M&E) purchases are imported. ### Investment in machinery and equipment (M&E) and investment-specific productivity (ISP) - New M&E typically embodies investment-specific productivity (ISP) improvements that raise the productivity of the new stock relative to the existing stock of M&E. - Technological advances caused a secular decrease in the relative price of M&E in terms of consumption goods. - As argued by Hornstein and Krussell (1996), increases in the rate of ISP improvements can produce decreases in measured productivity performance because of learning and quality mismeasurements associated with new technologies. - Empirical patterns in several high-income and emerging economies show similar quantity rises and relative price declines in M&E (Figure 1). ### Economic model and analysis framework - The paper adapts the Greenwood, Hercowitz and Krusell (GHK, 1997) framework, with modifications: - Allows for exogenous increases in labor-specific productivity. - Incorporates an index of utilization of the capital stock (u), assumed known and exogenous and identical for equipment and structures. - Production is modeled with a Cobb-Douglas technology with two capital types: - Machinery and equipment (ket) and non-housing structures (kst). - Neutral (total-factor) productivity is z; ISP is q (an index of investment-specific productivity). - Dynamics: - Structures: k_{s,t+1} = (1 - δ_s) k_{s,t} + i_{s,t}, with growth components including H and N (human capital and population). - Equipment: k_{e,t+1} = (1 - δ_e) k_{e,t} + q_t i_{e,t}; q measures the quality of new equipment and evolves at gross growth rate γ_q. - Government taxes labor and capital and returns revenues as a lump-sum transfer to the representative household. ### Balanced growth path (BGP) conditions and calibration approach - Variables are normalized per effective available labor. Along a BGP output, consumption, and investment grow at the same gross rate γ_y (output per available effective hour). - Equipment stock grows at gross rate γ_e > γ_y due to growing q. - Key BGP restrictions used (as presented): - Equations (8) and (9) impose relationships between γ_y, γ_q, γ_e, and factor shares (α_e, α_s). - The stationary transformed problem yields conditions (11)–(16) that link calibrated parameters and BGP observables. - Calibration procedure: - Calibrates 12 parameters so that unknowns in the BGP coincide with their average observed values over the sample period. - ISP (γ_q) proxied by the ratio of the implicit price deflator for personal consumption expenditures to the implicit price deflator for equipment and machinery. - Dataset and program codes used in the calibration are available upon request. ### Calibration results and quantitative contributions to growth - Calibration outputs (high-level results summarized in the text and tables): - For Chile (1986-2008 sample): - ISP (γ_q) averaged about 3.8 percent per year. - Neutral factor productivity (TNP) growth averaged 0.7 percent per year. - ISP accounted for about 45 percent of the long-term growth in output per effective hour in Chile. - ISP contribution to long-term growth of output per hour was 28 percent in Chile. - TNP contribution to output per hour in Chile was 37 percent. - Growth in human capital (HK) contributed the remaining 35 percent to output per hour in Chile. - Comparative averages (from calibrated sample): - ISP growth averaged about 3 percent per year across the other countries analyzed; Chile’s ISP growth (3.8 percent) was similar to Norway (3.6 percent). - TNP growth was lower than ISP growth in all countries in the sample. - The contribution of ISP to output growth increased significantly since the mid-1990s, especially in Chile and other net commodity exporters. - For net commodity exporters (Australia, Canada, Chile, Norway) TNP growth was positive until the late 1990s or early 2000s and turned negative afterward; TNP decreased in Australia, Canada, Chile and Norway since about the mid-2000s (Figure 2). ### Key quantitative summary (selected numbers reported exactly) - Chile: TFP at end-2007 was about 2 percent lower than at end-1997. - Investment in M&E as share of GDP: almost doubled since 2004; by end-2008 M&E investment amounted to about half of total investment. - ISP growth (Chile, average percent per year): 3.8 percent. - TNP growth (Chile, average percent per year): 0.7 percent. - Contribution to long-run output per hour (Chile, percent): - ISP: 28 percent - TNP: 37 percent - Human capital (HK): 35 percent - ISP accounted for about 45 percent of long-term growth in output per effective hour in Chile. *Source: Fund Staff calculations, _wp09264 - References.* ### 2004. In all cases, the observed ### WHAT COULD EXPLAIN THE PRODUCTIVITY SLOWDOWN? ### Overestimation of the effective capital stock - Growth accounting typically assumes current investment is incorporated immediately into the capital stock; multi-year projects should not be fully added until completion, otherwise capital’s contribution to growth is overestimated and TFP/TNP is underestimated. - Chile: official estimates indicate the amount of contemporaneous investment that should be added to the effective capital stock decreased from an average of about 75 percent during 2001-2005, to less than 60 percent in 2008. - Increase in number and relative importance of projects with long maturity, particularly in the Energy and Mining Sectors (Ministerio de Hacienda, 2007). - Similar patterns noted for other net commodity exporters undertaking sizable commodity-related projects with maturity periods exceeding one year. ### Output measurement issues (unmeasurable vs measurable sectors) - Certain activities (construction, trade, finance, insurance, real estate, other services, government) are harder to measure in output and quality improvements than agriculture, mining, manufacturing, transportation, communications (Hornstein and Krusell, 1996; Griliches, 1994). - Productivity slowdowns can reflect mismeasurement if the “unmeasurable sector” grows relative to the economy. - Chile: the proportion of the unmeasurable sector in total GDP (at factor cost) has remained relatively stable. ### Costly learning and adoption of new technologies - Adoption of new, more productive technologies involves learning costs; during transition output and TFP growth could decline (Hornstein and Krusell, 1996; Greenwood and Yorukoglu, 1997). - Wage dispersion and skill premia would be expected to increase during the adjustment. - Chile: premium paid to the most skilled workers (upper firm/public sector management and liberal professions) with respect to unskilled workers increased by about 10 percent during 1997-2008. - Relative wage of technical workers (machine operators and artisans) to unskilled workers has remained fairly constant. - Interpretation: some increase in wage dispersion, but stability of technical-to-unskilled relative wages suggests costly learning has not been the primary cause of the productivity slowdown. ### Business regulations and investment climate - Excessive or inadequate regulation can weaken productivity by impairing the investment climate (World Bank, 2004). - World Bank’s Doing Business (2009): Chile outranks Latin America but ranks below the OECD average on the general “Ease of doing business” indicator. - Chile has not made significant gains in several Doing Business categories during 2004-2009. - Evidence implies that a better regulatory environment might not have prevented the productivity slowdown completely, but there is scope to improve regulations so markets function better, especially during economic stress. - Increase in the average rate of unemployment after 1998 suggests regulatory rigidities may still limit the economy’s capacity to absorb shocks (see Restrepo, 2008). ### Declining stage of traditional (commodity) sectors - For economies with large non-renewable resource shares, exploitation can lead to rising marginal costs and eventual declines in production and productivity. - Australia evidence: declining productivity in the mining sector (Ewing et al., 2007). - Chile: mining GDP has fluctuated around a constant level since 2004 (in line with copper production), while “core” GDP (excluding mining, electricity gas and water and fishing) increased at an annual average rate of 5.2 percent. - Timing: productivity slowdown in Chile started around 1998, when mining GDP and physical copper production were expanding strongly. - During 1998-2004, mining GDP increased at an average annual rate of 4.9 percent. - Physical copper production expanded at an annual average rate of 6.9 percent. - Commodities 2005-2008: for commodity exporters in the sample (Australia, Canada, Chile and Norway), large increases in commodity prices of 2005-2008 coincide with decreases in productivity; part of the decrease could reflect expansion of production into areas of lower marginal productivity. ### Market concentration and competition - Limits on product market competition can impede middle-income countries’ convergence to frontier technology (Acemoglu, Aghion and Zilibotti, 2002). - Chile: Global Competitiveness Report (2008) ranks Chile 28 among 137 countries, better than any country in Latin America but below the OECD average. - Corporate activity in Chile dominated by a limited number of conglomerates; Chile ranks 57th in the “market dominance” indicator (Engel and Navia, 2006). - Scarce competition in financial sector (banks and pension fund managers) may constrain access to funds for middle and small firms, perpetuating market concentration. - Progress: creation of an “Anti-monopoly Court” in 2004; Chile ranks 19th in the “intensity of local competition” indicator. - Conclusion: market concentration may have constrained dynamism, but does not preclude intense competition in some areas. ### Concluding observations and policy implications - The marked increase in level and quality of new capital goods in Chile is favorable for future productivity growth. - Slow growth in neutral technological change suggests the need for additional reforms. - The productivity slowdown in Chile is also observed in other net commodity exporters and may reflect multiple factors: - Regulatory rigidities that limited the economy’s capacity to absorb the external shocks of 1998 and 2001 likely contributed to the initial phase of the slowdown. - From 2004, factors that probably compounded the problem include the slowdown in the mining sector and some overestimation of the effective capital stock. - Other factors that may have played a role and warrant further investigation: costly learning following adoption of new technologies, and market concentration. - Policy focus points implied by the analysis: - Improve measurement and accounting for multi-year investments to avoid overestimating contemporaneous capital stocks. - Continue reforms to improve the regulatory environment and labor market flexibility to enhance shock absorption and reduce rigidities associated with higher unemployment. - Address market concentration and barriers to finance for middle and small firms to boost competition and dynamism. - Investigate sectoral productivity trends (especially mining/commodity production at the margin) and the allocation of investment across projects with differing maturities and marginal productivities. *Source: _wp09264 - 2004.* --- _Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2009/_wp09264.pdf_