## PRODUCTIVITY AND GROWTH IN PERU

## Source details

**Canonical URL:** [PRODUCTIVITY AND GROWTH IN PERU](https://www.imf.org/-/media/files/publications/cr/2024/english/1perea2024002.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2024/english/1perea2024002.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2024/english/1perea2024002.pdf.json)

---

### Executive summary
- After a decade of high economic growth averaging over 6 percent per year, potential growth has been falling since 2014.
- The decline has been driven by a much slower pace of investment and human capital accumulation, and most notably, a decline in total factor productivity (TFP) growth.
- Firm-level productivity has worsened and the decline has been broad-based across industries, firm sizes, and regions.
- Special corporate tax regimes and labor legislations and regulations implemented with eligibility thresholds have created barriers to productivity growth by discouraging formalization and firm expansion.
- Policy focus: reform threshold-based regulations and labor-market rigidities to reduce distortions that keep firms small and impede productive reallocation.

### A. Macro-level measures of potential output
- Estimation approaches applied:
  - Univariate Hodrick-Prescott (HP) filter: annual real GDP 1990–2028, smoothing parameter of 100; Ravn and Uhlig (2002) smooth parameter yields similar results.
  - Multivariate filter: follows Alichi et al. (2017) and Blagrave et al. (2015) using an output gap, a Phillips Curve, and Okun’s Law specifications.
  - Production function approach: Cobb-Douglas Yt = At ⋅ Kt^α ⋅ Lt^(1−α) with elasticity of capital = 0.6, elasticity of labor = 0.4, depreciation rate = 5 percent, steady-state growth rate = 3.8 percent (average GDP growth 1951–2022) used to calculate initial capital stock in 1950.
- Estimated potential growth: 2.0 to 2.5 percent using the three approaches.
- Historical pattern:
  - Over a decade of high potential growth averaging over 5.5 percent between 2002–2013.
  - Steady decline to an average of 2.0–2.5 percent between 2017–2022 (excluding 2020).
  - Production-function estimates are more volatile and show about 0.5 percentage points lower potential growth after the pandemic.
- Mining vs. non-mining:
  - Potential growth declined in both mining and non-mining economies since 2014.
  - For mining, a capital elasticity of 0.8 was used in the production function.
  - Mining resource rents account for less than 15 percent of GDP.

### B. TFP and growth decomposition
- TFP contribution:
  - TFP was the highest contributor to growth in 2002–2006, then rapidly declined and has been negative since 2012.
  - Measured TFP growth in mining and hydrocarbons is volatile and has not been positive since 2007.
- Factor contributions since 2007:
  - Labor contribution has remained stable but small and declining.
  - Capital was the primary contributor to growth, but its contribution has fallen as capital accumulation slowed since 2012.
- Augmenting the production function to include mining resources increases estimated TFP growth but TFP remains negative.

### C. Measuring productivity using firm-level data
- Data and method:
  - Uses the Encuesta Ecónomica Annual (EEA) panel of formal firms (sample of Schiffbauer et al., 2022): restricts to multi‑year observations and single establishments; underrepresents microenterprises and agricultural and mining firms; sample covers total value‑add representing 14 to 19 percent of GDP.
  - TFP estimated with integrated control function approach of Ackerberg, Caves, and Frazer (2015) and De Loecker and Warzynski (2012), using a translog production function and controls for firm age and municipality; estimated separately by industry.
- Firm-level TFP trends:
  - Between 2007 and 2013, firm-level productivity growth accelerated from about 4 percent to about 6 percent per year.
  - Between 2014 and 2017, productivity growth reversed, averaging about −5 percent per year.
- Broad-based slowdown:
  - Worsening productivity observed across industries, firm sizes, and regions.
  - In 2014 the median firm in the most productive sectors (construction and other services), large firms, and coastal firms had TFP over two times that of the rest of the economy, yet these also experienced negative productivity growth 2014–2017.
- Within-industry dynamics:
  - Olley and Pakes (1996) decomposition indicates most decline arises from less productive firms gaining market share and from firm entry/exit composition, rather than a uniform downward shift in the productivity distribution.
  - Suggests barriers preventing productive firms from entering and unproductive firms from exiting, and obstacles to firm growth.

### D. Regulatory distortions and micro evidence
- Threshold-based labor and tax regulations create non-linear costs and incentives to remain small.
- Profit-sharing legislation (Decreto Legislativo No. 892, 1996):
  - Applies to firms with more than 20 employees.
  - Profit-sharing rates: 10 percent for fishing, manufacturing, telecommunications; 8 percent for commerce, hospital, mining; 5 percent for all other sectors.
  - Evidence: histogram of firm density by number of employees shows a clear discontinuity at 20 employees.
  - Local linear regression discontinuity (sample limited to firms with 10–30 employees) estimate: firms just below 20 employees are about 40 percent less productive than firms just above 20 employees.
- Régimen Especial del Impuesto a la Renta (RER, introduced 2007):
  - RER firms pay 1.5 percent tax on gross income versus 10–29.5 percent tax on profits for RMT and RG firms.
  - RER firms only required to maintain purchases and sales registers; RG firms must maintain accounting journals, ledgers, inventories, and balance sheets.
  - Eligibility: firms with annual sales below 525,000 soles can be in the RER.
  - Histogram by annual sales does not show clear massing just below the RER threshold, but regression evidence shows distortion:
    - Local linear regression discontinuity (sample limited to firms with annual sales between 325,000 and 725,000 soles) estimate: firms just outside the RER are about 40 percent less productive than firms just within the RER.
- Other thresholds:
  - RMT applies to firms with annual net income below 1,700 UIT.
  - Nuevo Régimen Unico Simplificado (NRUS) applies to firms with total sales below 96,000 soles; NRUS firms pay a single flat fee and have no accounting obligations.
- Administrative and compliance costs:
  - Multiple tax regimes increase compliance and administrative costs and opportunities for tax arbitrage.
  - Agosin et al. (2010) note moving between special corporate tax regimes can reduce profits by 30 to 50 percent.

### E. Policy implications and recommendations
- Core recommendations:
  - Remove or reform threshold-based regulations that create incentives to remain small or to split firms, including the profit-sharing threshold at 20 employees and tax-regime eligibility rules that impose non-linear costs.
  - Reduce legal and administrative barriers to formalization and firm growth to encourage entry of productive firms and exit of unproductive firms, promote growth of medium-sized firms, and reduce productivity dispersion between large and small firms.
  - Revive structural reforms and improve the business environment by reversing counter-reforms since 2006 that have deteriorated permitting, licensing, tax codes, and labor regulations; restore reforms that previously supported investment and productivity (macroeconomic framework, trade and financial liberalization, infrastructure, healthcare, education, civil service, pensions, taxes, and sectoral development).
  - Introduce greater labor market flexibility to enable worker transitions to productive sectors and reduce labor informality.
- Expected outcomes:
  - Reforms that lower costs of formalizing and growing would reduce regulatory uncertainty, promote investment, and support a rebalancing away from a dual economy where a low‑productivity informal sector coexists with a highly productive formal sector.

### Climate risks, adaptation, and macro-fiscal implications
- Exposure and disaster history:
  - Peru experienced 61,708 emergencies linked to natural phenomena between 2003-2019.
  - Economic losses and damages from disasters in 1982-83, 1997-98, and 2017 amounted to 11.6, 6.2, and 1.6 percent of GDP, respectively.
  - Approximately 40 percent of total damages in 2017 were inflicted on the road network.
- El Niño Costero impacts:
  - Occurs every 4-5 years; strong events typically reduce fish production by 70 percent and reduce agricultural output by 11 percent.
  - Non-core CPI increases by 2.2 p.p. and non-core food prices by 8.1 p.p.; core CPI increases by 1 p.p. and core food prices by 2.3 p.p. in the second year following a strong event.
  - Historical sectoral effects:
    - Fisheries: output dropped by 70 percent within the first year on average; in March-November 2023, fishing output was 27.3 percent below trend.
    - Agriculture: output dropped by about 11 percent within the same time span; in March-November 2023, agricultural output was 4.9 percent below trend.
    - Construction: in March-November 2023, construction output was 9.2 percent lower relative to trend.
    - Manufacturing: in March-November 2023, manufacturing output was 6.6 percent lower relative to trend.
  - Typical duration: strong events typically last slightly longer than a year.
  - Real GDP impact: El Niño Costero events typically reduced real GDP by 5 percentage points within the first year, relative to a baseline real GDP trajectory.
  - Central government primary balance would typically fall by about 2 percent of GDP for the same period.
  - Consistent staff estimates: the annual central government primary balance fell by around 1.2 percent of GDP from 2022 to 2023, reflecting about half a percentage point increase in nominal primary expenditures and 6 percent lower nominal revenues during a strong El Niño Costero year.
- Modeling and scenarios:
  - IMF climate FGG DSGE model used with Markov-switching disaster regimes; calibrations include strong El Niño shocks impacting physical capital by 2.6 percent of the capital stock per quarter; weak and moderate shocks inflict ¼ of damages of a strong event; other disasters calibrated as average El Niño impact divided by 3.13.
  - Baseline scenarios consider that investments in adaptation will offset 1/3 of the climate’s impact on natural capital and productivity.
- Long-run losses under warming scenarios:
  - Current recurring natural hazards reduce potential output by 4.1 percent in 2023 relative to the no climate counterfactual.
  - Scenarios (SSP1-2.6, SSP2-4.5, SSP3-7.0):
    - Hothouse world assumes a 59 percent rise in the number of strong El Niño events by the second half of the century.
    - Potential output losses relative to no climate change counterfactual:
      - By 2050: up to 16.1-22.8 percent lower.
      - By 2100: 28.2-102.3 percent lower.
    - Equivalent effect on average annual potential growth rate: 0.1-0.3 percentage point reduction in the average annual potential growth rate due to climate and climate change.
- Benefits of adaptation and resilience:
  - Policy package simulated:
    - (a) 0.8 percent of GDP in public adaptation investments per year until 2030;
    - (b) climate proofing 80 percent of Peru’s transport infrastructure;
    - (c) increasing public investment efficiency by 5 percentage points.
  - Simulated potential output gains from the policy package:
    - By 2050: as high as 9.3-12.3 percent.
    - By 2100: 12.4-31 percent.
  - Residual output losses after the policy package:
    - By 2050: 6.8-10.5 percent.
    - By 2100: 15.8-71.3 percent.
  - Improvements in public investment efficiency alone account for a 0.8 percent positive shift in the potential GDP trajectory.
- Fiscal returns and cost-effectiveness:
  - Between 2024 and 2100, combined benefits produce, on average, an annual 2.3-4.6 percent of GDP in fiscal savings.
  - Investments in adaptation and resilience are cost-effective under all three global warming scenarios, even when accounting for discounting; baseline discount rate used in calculations is 6 percent.
  - Table 1 summary (selected entries preserved exactly as reported; scenario groupings correspond to SSP1-2.6 | SSP2-4.5 | SSP3-7.0):
    - Total Return (a): 120.25 | 262.36 | 122.34 | 286.80 | 143.05 | 336.35
    - Stock saving — Decrease in reconstruction cost after NDs: 4.53 | 6.84 | 4.71 | 6.89 | 4.78 | 6.96
    - Flow saving — Moderation of tax base decline after NDs: 16.28 | 31.00 | 16.58 | 33.92 | 19.52 | 39.97
    - Potential growth — Increase in the tax base trajectory: 99.43 | 224.52 | 101.04 | 245.99 | 118.75 | 289.42
    - Total Cost (b) 3/: 10.53 | 11.65
    - Adaptation — Investments in Water, Agriculture, Forestry, and Fisheries: 4.84 | 4.92
    - Resilience — Investment to climate proof and maintain transport infrastructure: 5.69 | 6.73
    - Net annual saving (a) - (b): 109.72 | 250.71 | 111.81 | 275.15 | 132.52 | 324.70
  - Figure 8 highlights cumulative discounted fiscal savings by 2100 under alternative discount rates (5% | 6% (Baseline) | 7% | 10%):
    - Hothouse world (SSP3-7.0): 1.60 (2024-2050) | 4.60 (2024-2100)
    - Intermediate GHG emissions (SSP2-4.5): 1.29 (2024-2050) | 3.21 (2024-2100)
    - Delayed Net Zero (SSP1-2.6): 1.21 (2024-2050) | 2.29 (2024-2100)

### Financial protection and insurance gaps
- Insuring public assets is mandatory but "subject to the availability of funds."
- Absence of a centralized inventory of public assets complicates monitoring of public sector insurance coverage.
- The existing system requires each public entity to individually insure its managed assets; a centralized inventory could enable collective negotiation and consolidation of indemnity insurances into a single policy.
- Catastrophe insurance penetration in the private sector remains low: "only 8 percent of properties insured against natural disasters as of 2017."
- Among insured properties, policies often exclude flood risk.
- A significant share of uninsured public and private assets represents an implicit contingent liability for the government that could overwhelm the risk retention system if a tail risk materializes.
- Progress with adopting nontraditional insurance instruments has stalled since the 2018 Pacific Alliance issuance of cat-bonds linked to earthquake risk.

### Technology diffusion, digitalization, AI, and sectoral gains
- Productivity divergence and diffusion barriers:
  - Key barriers to technology diffusion include weak competition, constrained access to finance, proliferation of small firms, low R&D investment, and inadequate human capital.
- Digital access and GovTech:
  - Peru has lower internet access than other LA5 economies; National Policy of Digital Transformation for 2023-2030 aims to increase proportion of adults using digital technologies from about 52 percent to 85 percent by 2030.
  - Multilateral support: IDB and World Bank initiatives to connect 3.5 million students and train over 265,000 teachers in digital skills.
  - Government classified in the "very high" group of the United Nations' E-Government Development Index in 2022.
- AI exposure and growth implications:
  - Peru’s labor force AI exposure: Approximately 30 percent of the labor force in Peru is highly exposed to AI; over 60 percent in the U.S. and the UK are highly exposed.
  - In the formal economy, AI exposure and complementarity are on par with advanced economies.
  - Growth simulations:
    - If Peru halves its technological and AI exposure gap with the U.S. beginning in 2028 (optimistic scenario), the annual average difference in GDP growth between optimistic and pessimistic scenarios from 2040 to 2050 is projected to be 0.21 percentage points.
    - Atlas growth simulations indicate that reducing the technological gap would increase Peru’s long-term growth by 1.1 percent.
- Sectoral gains and digital prerequisites:
  - Sectors likely to gain most from AI: finance, government, trade, IT, education, health, and real estate; over 50 percent of the workforce in these sectors is significantly exposed to AI.
  - Digitalization gaps:
    - Only a third of hospitals are connected to the internet.
    - E-invoicing reform (2013) led to over a 5 percent increase in reported sales, purchases, and value-added for firms in the first year after adoption.
    - Tax administration: 85 percent of tax payments by value were made electronically (ISORA 2021).
  - FinTech developments:
    - In 2021, 57.5 percent of Peruvians aged 15 and older had financial accounts; 22 percent borrowed from formal financial institutions.
    - In 2022, approximately 45 percent of Peruvians used digital wallets, a 12-percentage point increase from the year before.
    - Digital wallets helped boost financial inclusion to 65 percent in late 2023.
    - About 300 FinTech startups; Peru has one of the highest numbers of FinTech startups per capita in the region.
    - Regulatory steps: BCRP permitted the first e-money provider to deposit funds in its accounts; 2022 regulation ensured interoperability among digital wallets; authorities working on an Open Banking law; virtual asset providers required to comply with AML regulations; regulatory sandbox refinement ongoing.
- Policy priorities:
  - Facilitate formalization and firm growth to attract workers from informal to formal sectors.
  - Develop human capital and digital skills, expand internet access, and improve competition and access to finance to narrow the technology and AI exposure gaps.
  - Support transitions for the workforce: about 17 percent of the workforce is high exposure to AI and low complementarity, requiring policies to mitigate income loss during job transitions.

*Source: Excerpt from "PRODUCTIVITY AND GROWTH IN PERU" (IMF country chapter).*

### References ___________________________________________________________________________ 10

### PRODUCTIVITY AND GROWTH IN PERU

### Executive summary
- After a decade of high economic growth averaging over 6 percent per year, potential growth has been falling since 2014.
- The decline has been driven by a much slower pace of investment and human capital accumulation, and most notably, a decline in total factor productivity (TFP) growth.
- Firm-level productivity has worsened and the decline has been broad-based across industries, firm sizes, and regions.
- Special corporate tax regimes and labor legislations and regulations implemented with eligibility thresholds have created barriers to productivity growth by discouraging formalization and firm expansion.
- Policy focus: reform threshold-based regulations and labor-market rigidities to reduce distortions that keep firms small and impede productive reallocation.

### A. Macro-Level measures of potential output
- Multiple estimation approaches were applied: a univariate Hodrick-Prescott (HP) filter, a multivariate filter, and a production function approach (Cobb-Douglas).
- HP filter specification: applied to annual real GDP 1990–2028 with smoothing parameter of 100; Ravn and Uhlig (2002) smooth parameter yields similar results.
- Multivariate filter specification: follows Alichi et al. (2017) and Blagrave et al. (2015) using:
  - Output gap 푌t = 휑푌t−1 + 휀Y
  - Phillips Curve 휋t = 휆휋t+1 + (1−휆)휋t−1 + 훽푌t + 휀π
  - Okun’s Law ugap,t = 휏1ugap,t−1 + 휏2푌t + 휀ugap
- Production function approach: Cobb-Douglas 푌t = 𝐴t ⋅ 𝐾t^α ⋅ 𝐿t^(1−α) with:
  - elasticity of capital = 0.6
  - elasticity of labor = 0.4
  - depreciation rate = 5 percent
  - steady-state growth rate = 3.8 percent (average GDP growth 1951–2022) used to calculate initial capital stock in 1950
- Estimated potential growth: 2.0 to 2.5 percent using the three approaches.
- Historical pattern: over a decade of high potential growth averaging over 5.5 percent between 2002–2013; steady decline to an average of 2.0–2.5 percent between 2017–2022 (excluding 2020).
- Production-function estimates are more volatile and show about 0.5 percentage points lower potential growth after the pandemic.
- Mining vs. non-mining:
  - Potential growth has declined in both mining and non-mining economies since 2014.
  - For mining, a capital elasticity of 0.8 was used in the production function.
  - Mining resource rents account for less than 15 percent of GDP.

### B. TFP and growth decomposition
- TFP contribution:
  - TFP was the highest contributor to growth in 2002–2006, then rapidly declined and has been negative since 2012.
  - Measured TFP growth in mining and hydrocarbons is volatile and has not been positive since 2007.
- Factor contributions since 2007:
  - Labor contribution has remained stable but small and declining.
  - Capital was the primary contributor to growth, but its contribution has fallen as capital accumulation slowed since 2012.
- Augmenting the production function to include mining resources increases estimated TFP growth but TFP remains negative.

### C. Measuring productivity using firm-level data
- Data and method:
  - Uses the Encuesta Ecónomica Annual (EEA) panel of formal firms (sample of Schiffbauer et al., 2022): restricts to multi‑year observations and single establishments, underrepresenting microenterprises and agricultural and mining firms; sample covers total value‑add representing 14 to 19 percent of GDP.
  - TFP estimated with integrated control function approach of Ackerberg, Caves, and Frazer (2015) and De Loecker and Warzynski (2012), using a translog production function and controls for firm age and municipality; estimated separately by industry.
- Firm-level TFP trends:
  - Between 2007 and 2013, firm-level productivity growth accelerated from about 4 percent to about 6 percent per year.
  - Between 2014 and 2017, productivity growth reversed, averaging about −5 percent per year.
- Broad-based slowdown:
  - Worsening productivity observed across industries, firm sizes, and regions.
  - In 2014 the median firm in the most productive sectors (construction and other services), large firms, and coastal firms had TFP over two times that of the rest of the economy, yet these also experienced negative productivity growth 2014–2017.
- Within-industry dynamics:
  - Olley and Pakes (1996) decomposition indicates most decline arises from less productive firms gaining market share and from firm entry/exit composition, rather than a uniform downward shift in the productivity distribution.
  - Suggests barriers preventing productive firms from entering and unproductive firms from exiting, and obstacles to firm growth.

### D. Regulatory distortions and micro evidence
- Threshold-based labor and tax regulations create non-linear costs and incentives to remain small.
- Profit-sharing legislation (Decreto Legislativo No. 892, 1996):
  - Applies to firms with more than 20 employees.
  - Profit-sharing rates: 10 percent for fishing, manufacturing, telecommunications; 8 percent for commerce, hospital, mining; 5 percent for all other sectors.
  - Evidence: histogram of firm density by number of employees shows a clear discontinuity at 20 employees, indicating firms maintain smaller sizes to avoid the threshold.
  - Local linear regression discontinuity (sample limited to firms with 10–30 employees) estimate: firms just below 20 employees are about 40 percent less productive than firms just above 20 employees.
- Régimen Especial del Impuesto a la Renta (RER, introduced 2007):
  - Applies to firms in extractive, manufacturing, trade, service, and agricultural industries.
  - Firms in RER pay 1.5 percent tax on gross income versus 10–29.5 percent tax on profits for firms in the Régimen MYPE Tributario (RMT) and Régimen General (RG).
  - Firms in RER only required to maintain purchases and sales registers; RG firms must maintain accounting journals, ledgers, inventories, and balance sheets.
  - Eligibility: firms with annual sales below 525,000 soles can be in the RER.
  - Histogram by annual sales does not show clear massing just below the RER threshold, but regression evidence shows distortion:
    - Local linear regression discontinuity (sample limited to firms with annual sales between 325,000 and 725,000 soles) estimate: firms just outside the RER are about 40 percent less productive than firms just within the RER.
- Other thresholds (not analyzed in detail here):
  - RMT applies to firms with annual net income below 1,700 UIT.
  - Nuevo Régimen Unico Simplificado (NRUS) applies to firms with total sales below 96,000 soles; NRUS firms pay a single flat fee and have no accounting obligations.
- Administrative and compliance costs:
  - Multiple tax regimes increase compliance and administrative costs and opportunities for tax arbitrage.
  - Agosin et al. (2010) note moving between special corporate tax regimes can reduce profits by 30 to 50 percent.

### E. Policy implications and recommendations
- Remove or reform threshold-based regulations that create incentives to remain small or to split firms:
  - Address profit-sharing threshold at 20 employees and tax-regime eligibility rules that impose non-linear costs.
- Reduce legal and administrative barriers to formalization and firm growth to:
  - Encourage entry of productive firms and exit of unproductive firms.
  - Promote growth of medium-sized firms and reduce productivity dispersion between large and small firms.
- Revive structural reforms and improve the business environment:
  - Reverse counter-reforms since 2006 that have deteriorated permitting, licensing, tax codes, and labor regulations.
  - Restore reforms that previously supported investment and productivity (macroeconomic framework, trade and financial liberalization, infrastructure, healthcare, education, civil service, pensions, taxes, and sectoral development).
- Labor market:
  - Introduce greater labor market flexibility to enable worker transitions to productive sectors and reduce labor informality.
- Expected outcomes:
  - Reforms that lower costs of formalizing and growing would reduce regulatory uncertainty, promote investment, and support a rebalancing away from a dual economy where a low‑productivity informal sector coexists with a highly productive formal sector.

*Source: Excerpt from "PRODUCTIVITY AND GROWTH IN PERU" (IMF country chapter).*

### References

### References

### Cited Works
- Ackerberg, Daniel A., Kevin Caves, and Garth Frazer (2015). “Identification properties of recent production function estimators,” Econometrica 83, pp. 2411-2451.
- Agosin, Manuel, Juan Pablo Atal, Juan Blyde, Matías Busso, Eduardo Cavallo, Alberto Chong, Christian Daude, Eduardo Fernández-Arias, Arturo Galindo, Pablo Ibarrarán, Alejandro Izq and Hugo Ñopo (2010). “La era de la productividad: Cómo transformar las economías desde sus cimientos,” Inter-American Development Bank Publications Number 56538, edited by Carmen Pagés.
- Alichi, Ali, Olivier Bizimana, Douglas Laxton, Kadir Tanyeri, Hou Wang, Jiaxiong Yao, and Fan Zhang (2017). “Multivariate Filter Estimation of Potential Output for the United States,” IMF Working Paper 17/106.
- Bakker, Bas B. (2023). “Unveiling the Hidden Impact of Urban Land Rents on Total Factor Productivity,” IMF Working Paper 23/170.
- Banco Central de Reserva del Perú (2023). “Reporte de Inflación: Panorama actual y proyecciones macroeconómicas 2023-2024,” September 2023.
- Blagrave, Patrick, Roberto Garcia-Saltos, Douglas Laxton, and Fan Zhang (2015). “A simple multivariate filter for estimating potential output,” IMF Working Paper 15/79.
- Castillo, Luis E. and David Florián Hoyle (2019). “Measuring the output gap, potential output growth, and natural interest rate from a semi-structural dynamic model for Peru,” Banco Central de Reserva del Perú Working Paper 2019-012.
- Castillo, Paul and Youel Rojas (2014). “Términos de intercambio y productividad total de factores: Evidencia empírica de los mercados emergentes de América Latina,” Revista Estudios Económicos 28, pp. 27-46.
- Céspedes, Nikita, María Apuije, Alan Sánchez, and Rafael Vera-Tudela (2014). “Productividad sectorial en el Perú: un análisis a nivel de firmas,” Banco Central de Reserva del Perú Working Paper 2014-013.
- De Loecker, Jan and Frederic Warzynski (2012). “Markups and firm-level export status,” American Economic Review 102, pp. 2437-2471.
- Hamilton, Kirk, Esther Naikal and Glenn-Marie Lange (2019). “Natural Resources and Total Factor Productivity Growth in Developing Countries: Testing a New Methodology,” World Bank Policy Research Working Paper 8704.
- International Monetary Fund (2022). “Peru: Staff Report for the 2022 Article IV Consultation,” IMF Country Report 22/137.
- Olley, G. Steven and Pakes, Ariel (1996). “The Dynamics of Productivity in the Telecommunications Industry.” Econometrica, 64, pp. 1263-1298.
- Ortiz, Marco and Diego Winkelried (2022), editors, Hitos de la Reforma Macroeconomica and El Peru, Universidad del Pacifico.
- Ravn, Morten O. and Harald Uhlig (2002). “On Adjusting the Hodrick-Prescott Filter for the Frequency of Observations,” The Review of Economics and Statistics, 84(2), pp. 371-380.
- Sánchez, William and Renato Vassallo (2023). “Midiendo de PBI potencial peruano tas la pandemia del COVID-19: un filtro multivariado extendido,” Consejo Fiscal Documento de Investicación No 03-2023.
- Schiffbauer, Marc, James Sampi, and Javier Coronado (2022). “Competition and Productivity: Evidence from Peruvian Municipalities,” The Review of Economics and Statistics 1-45.
- World Bank (2021). “The Changing Wealth of Nations 2021: Managing Assets for the Future,” Washington, DC, World Bank.
- Vostroknutova, Ekaterina, Alberto Rodriguez, Pablo Saavedra, and John Panzer (2015). “Peru – Building on Success: Boosting Productivity for Faster Growth,” World Bank Group Report No. 99400-PE.

### Economic benefits of building resilience to natural hazards and climate change — Key findings and projections
- Exposure and recent disaster history
  - Peru experienced 61,708 emergencies linked to natural phenomena between 2003-2019 (World Bank, 2022a).
  - The climate-related physical risk profile is dominated by floods, landslides, droughts, and storms.
  - Economic losses and damages from disasters in 1982-83, 1997-98, and 2017 amounted to 11.6, 6.2, and 1.6 percent of GDP, respectively (World Bank, 2016).
  - Approximately 40 percent of total damages in 2017 were inflicted on the road network (Macroconsult, 2017).

- El Niño Costero: frequency, channels, and impacts
  - El Niño Costero occurs every 4-5 years and is a recurring warming of sea surface temperature along Peru’s coast (Niño 1+2 region).
  - Strong to extreme El Niño Costero events typically:
    - Reduce fish production by 70 percent.
    - Reduce agricultural output by 11 percent.
    - Increase non-core CPI by 2.2 p.p. and non-core food prices by 8.1 p.p.; these effects generally take over a year to recover to pre-El Niño levels.
    - Increase core CPI by 1 p.p. and core food prices by 2.3 p.p. in the second year following a strong event, indicating inflation passthrough.
    - Worsen the fiscal balance via lower tax collections and higher disaster recovery and reconstruction spending.

- Long-term climate change impacts
  - Peru is the most vulnerable country to climate change among the LA5 group and third most vulnerable in Latin America (IMF-adapted ND-GAIN, 2021).
  - Agriculture and fish production jointly constitute 7.6 percent of GDP and employ 28 percent of the workforce (OECD, 2023).
  - These sectors account for over 18 percent of total exports.
  - Climate change could result in large potential output losses in the long run, in the range of -26.2 to -102.3 percent of GDP depending on the global warming scenario (see Section C for details).

- Drivers of vulnerability and resilience gaps
  - Low adaptive capacity driven by insufficient water management capacity (e.g., dams, water treatment plants) and relatively low quality of infrastructure (e.g., ports, railroads, roads, information technology).
  - Weak public investment management, lack of coordination between various levels of government, and capacity gaps in the civil service undermine adaptive capacity and constrain public investment efficiency.
  - Chronic under-execution of capital budgets is a major obstacle to effective investments in structural resilience.

- Institutional and policy frameworks — progress and gaps
  - Legal and strategic frameworks:
    - 2018 Framework Law on Climate Change integrates adaptation principles into national planning and public investment management processes.
    - 2021 National Adaptation Plan (NAP) outlines sectoral adaptation priorities across 5 thematic areas (Health, Water, Agriculture, Aquaculture, and Forests) and details 84 associated adaptation measures scheduled for completion by 2030.
    - As of July 2023, implementation had commenced for 33 out of 84 measures, 18 of which are concentrated in the water thematic area.
  - Financing and implementation gaps:
    - Current annual adaptation expenditure is $0.6 billion (as of 2023) and is insufficient to meet NAP targets.
    - If the current annual adaptation expenditure level of $0.6 billion is maintained, it would take approximately 20 years more (until 2052) to implement costed adaptation plans originally set for 2030.
    - Business as Usual (BAU) projections assume annual adaptation expenditure at the 2023 level: $0.6 billion in nominal terms and 0.2 percent as a share of GDP.

- Structural protection and disaster risk management
  - National frameworks:
    - SINAGERD (National Framework for Disaster Risk Management) established in 2011.
    - National Disaster Risk Management Policy by 2050 and an updated National Plan for Disaster Risk Management (PLANAGERD 2022-2030) released in 2021; PLANAGERD includes 21 strategic actions.
  - Sectoral and infrastructure efforts:
    - Ministry of Transport and Communications (MTC) builds bridges with provisions for riverine floods, maintains road infrastructure ahead of El Niño, and reconstructs ex-post.
    - National Infrastructure Agency (ANIN), established in 2023, will assume a fraction of subnational projects aiming to strengthen structural resilience of public assets by adopting international engineering standards and focusing on DRM in construction planning.
  - Despite these efforts, natural disasters continue to inflict sizeable damages to public infrastructure as resources are often diverted to other development objectives.

- Fiscal risk retention and insurance mechanisms
  - The 2016 Comprehensive Strategy for Financial Protection against Natural Disasters envisages a multi-layered approach with risk retention (fiscal stabilization fund, contingency credit lines, and reserves) and risk transfer (indemnity insurance and cat-bonds).
  - By 2024, all elements of the strategy’s risk retention component were operational and had been utilized in managing recent natural disasters.
  - The adequacy of financial safeguards had not been assessed against a wider set of potential natural hazard risks; currently only earthquake risk is included.

*Source: 1perea2024002 - References (IMF).*

### 9.      Significant financial protection gaps remain in the country’s risk transfer system which

### 1perea2024002 - 9.      Significant financial protection gaps remain in the country’s risk transfer system which

### Financial protection gaps in the risk transfer system
- Insuring public assets is mandatory but "subject to the availability of funds."
- Absence of a centralized inventory of public assets complicates monitoring of public sector insurance coverage.
- The existing system requires that each public entity individually insures its managed assets; creating a centralized inventory could enable collective negotiation and consolidation of indemnity insurances into a single policy.
- Catastrophe insurance penetration in the private sector remains low: "only 8 percent of properties insured against natural disasters as of 2017."
- Among insured properties, insurance policies often exclude flood risk.
- A significant share of uninsured public and private assets represents an implicit contingent liability for the government that could overwhelm the risk retention system if a tail risk materializes.
- Progress with adopting nontraditional insurance instruments has stalled since the 2018 Pacific Alliance issuance of cat-bonds linked to earthquake risk.

### Estimating the impact of El Niño on economic activity — methodology
- Empirical approach: local projection method proposed by Jordà (2005) to estimate impulse response functions for strong El Niño events.
- Dependent frequency: quarterly data for sectoral output and monthly data for price indices.
- Controls include oil price indices for Peru’s main oil import partners (U.S. and Nigeria), a global fertilizer index, and a local production index for monthly price regressions.
- For quarterly dependent variables the production index is omitted.

### Empirical findings: inflation and food prices
- Strong El Niño Costero events historically associated with inflationary pressures: both core and headline inflation rates rose following events.
- Non-core inflation path:
  - Reached a "4.4 percentage point increase in the year over year inflation rate by the end of the first year," then gradually declined.
- Non-core food inflation:
  - Registered an "8.1 percentage point (p.p.) increase by the end of the first year."
- Core food inflation:
  - Registered a "2.3 p.p. increase in the inflation rate 20 months after the onset of El Niño Costero."
- 2023 El Niño Costero episode specific observations:
  - Fish and Seafood component: "11 p.p. increase between February and June."
  - Fruits component: "22.6 p.p. increase between February and September."

### Empirical findings: sectoral output and persistence
- Fisheries:
  - Historically, output dropped "by 70 percent within the first year since the onset of El Niño, on average," followed by recovery as sea surface temperatures normalized.
  - In March-November 2023, fishing output was "27.3 percent below trend."
- Agriculture:
  - Historically, output dropped "by about 11 percent within the same time span," with production still "about 3.4 percent below pre-El Niño levels 1.5 years after the shock."
  - In March-November 2023, agricultural output was "4.9 percent below trend."
- Construction and manufacturing:
  - In March-November 2023, construction output was "9.2 percent lower" and manufacturing output was "6.6 percent lower" relative to trend.

### Macroeconomic and fiscal effects of strong El Niño events
- Typical duration: "strong El Niño events typically last slightly longer than a year."
- Real GDP impact: El Niño Costero events typically reduced real GDP "by 5 percentage points within the first year, relative to a baseline real GDP trajectory."
- Central government primary balance:
  - Would typically fall "by about 2 percent of GDP for the same period" due to increased expenditures and lower collections.
  - Consistent staff estimates: "the annual central government primary balance fell by around 1.2 percent of GDP from 2022 to 2023," reflecting "about half a percentage point increase in nominal primary expenditures and 6 percent lower nominal revenues during a strong El Niño Costero year."

### Modeling long-term impacts and adaptation: FGG DSGE model extensions
- Model used: IMF’s climate FGG DSGE model (Fernandez-Corugedo, Gonzalez, and Guerson, 2023), a Markov-switching dynamic small open economy model.
- Key model features:
  - Economy oscillates between "disaster" and "non-disaster" regimes.
  - Two types of public capital: standard capital (vulnerable to disasters) and resilient capital (immune to disasters).
  - Model extended to include natural capital in the production function and to account for El Niño impact on natural capital.
  - A fraction of public investment is assumed to transform into productive public capital, capturing public investment inefficiencies (Buffie et al. 2012).
- Calibrations and assumptions:
  - Impact of strong/very strong El Niño shocks on physical capital set to "2.6 percent of the capital stock per quarter."
  - Weak and moderate El Niño shocks assumed to inflict "¼ of damages and losses resulting from a strong/very strong event."
  - Impact of other natural disasters calibrated by dividing the average El Niño impact by "3.13."
  - Transition probabilities calibrated using history of El Niño occurrences since 1980.
  - Baseline scenarios consider that investments in adaptation will offset "1/3 of the climate’s impact on natural capital and productivity."

### Long-run losses from existing climate hazards and projections under warming scenarios
- Current impact: recurring natural hazards reduce potential output "by 4.1 percent in 2023, relative to the no climate counterfactual."
- Global warming scenarios evaluated: Delayed Net Zero (SSP1-2.6), Intermediate GHG emissions (SSP2-4.5), Hothouse world (SSP3-7.0).
- Hothouse world assumption: "assumes a 59 percent rise in the number of strong El Niño events by the second half of the century."
- Potential output losses relative to no climate change counterfactual:
  - By 2050: "up to 16.1-22.8 percent lower."
  - By 2100: "28.2-102.3 percent lower."
- Equivalent effect on average annual potential growth rate: "0.1-0.3 percentage point reduction in the average annual potential growth rate due to climate and climate change."
- In all scenarios except Hothouse world, losses in potential growth are front-loaded due to anticipated deceleration of global warming under SSP1-2.6 and SSP2-4.5.

### Benefits of adaptation, resilience, and improved public investment efficiency
- Policy package simulated:
  - (a) "0.8 percent of GDP in public adaptation investments per year until 2030;"
  - (b) climate proofing "80 percent of Peru’s transport infrastructure;"
  - (c) increasing public investment efficiency by "5 percentage points."
- Simulated potential output gains from the policy package:
  - By 2050: "as high as 9.3-12.3 percent."
  - By 2100: "12.4-31 percent."
- Residual output losses after the policy package:
  - By 2050: "6.8-10.5 percent."
  - By 2100: "15.8-71.3 percent."
- Cost-effectiveness: the policy package is described as cost-effective under all three global warming scenarios because it "simultaneously lifts the potential output trajectory and produces fiscal savings."
- Note: improvements in public investment efficiency alone account for a "0.8 percent positive shift in the potential GDP trajectory" (i.e., not the sole driver of gains).

*Source: IMF staff calculations and analysis from the provided content.*

### 22.      Public investments in adaptation and resilience are also associated with large fiscal

### 1perea2024002 - 22.      Public investments in adaptation and resilience are also associated with large fiscal

### Fiscal returns from public investments in adaptation and resilience
- Making public infrastructure resilient to El Niño and other natural disasters leads to fiscal savings by preventing repeated reconstruction of damaged roads and bridges and by mitigating transitory reductions in government revenues associated with these events.
- Investments in adaptation partially offset the impact of global warming on the potential growth rate, producing a higher trajectory for the tax base and tax receipts.
- Between 2024 and 2100, the combined benefits produce, on average, an annual 2.3-4.6 percent of GDP in fiscal savings.
- Fiscal benefits materialize over time, while fiscal costs are immediate; therefore, discounted fiscal savings must be calculated to compare benefits and costs.
- Investments in adaptation and structural resilience are cost-effective under all three global warming scenarios even when accounting for the time value of money and under a range of discount rates.

### Quantified returns and costs (Table 1 and Figure 8 summary)
- Table 1: Peru: Cumulative Discounted Fiscal Savings from Investment in Resilience and Adaptation (Present value, in percent of 2023 GDP) — selected entries preserved exactly as reported:
  - Total Return (a): 120.25 | 262.36 | 122.34 | 286.80 | 143.05 | 336.35
  - Stock saving — Decrease in reconstruction cost after NDs: 4.53 | 6.84 | 4.71 | 6.89 | 4.78 | 6.96
  - Flow saving — Moderation of tax base decline after NDs: 16.28 | 31.00 | 16.58 | 33.92 | 19.52 | 39.97
  - Potential growth — Increase in the tax base trajectory: 99.43 | 224.52 | 101.04 | 245.99 | 118.75 | 289.42
  - Total Cost (b) 3/: 10.53 | 11.65
  - Adaptation — Investments in Water, Agriculture, Forestry, and Fisheries: 4.84 | 4.92
  - Resilience — Investment to climate proof and maintain transport infrastructure: 5.69 | 6.73
  - Net annual saving (a) - (b): 109.72 | 250.71 | 111.81 | 275.15 | 132.52 | 324.70
  - Scenario labels: SSP1-2.6 | SSP2-4.5 | SSP3-7.0 (values correspond to scenario groupings in table)
- Figure 8: Cumulative Discounted Fiscal Savings by 2100 under Alternative Discount Rates (Present value, in percent of 2023 GDP) — highlighted values:
  - Discount rates analyzed: 5% | 6% (Baseline) | 7% | 10%
  - Cumulative discounted fiscal savings by 2100 under scenarios (examples extracted from chart labels):
    - Hothouse world (SSP3-7.0): 1.60 (2024-2050) | 4.60 (2024-2100)
    - Intermediate GHG emissions (SSP2-4.5): 1.29 (2024-2050) | 3.21 (2024-2100)
    - Delayed Net Zero (SSP1-2.6): 1.21 (2024-2050) | 2.29 (2024-2100)
- Conclusion from quantified analysis: Even after discounting (baseline discount rate set to 6 percent in calculations), building resilience and adapting to climate change yields positive, cost-effective fiscal outcomes across SSP1-2.6, SSP2-4.5, and SSP3-7.0 scenarios.

### Policy recommendations to enhance effectiveness and financial resilience
- Improve coordination within government:
  - Enhance coordination among various levels and branches of the government to increase effectiveness of public adaptation and disaster risk management efforts.
  - The MEF could estimate the costs of measures outlined in the national adaptation and DRM plans as a first step toward effective budgeting of climate policies, or adopt green budget tagging.
- Strengthen financial protection and fiscal risk management:
  - Generate a comprehensive framework for quantification of fiscal risks that includes all major natural hazards.
  - Ensure comprehensive insurance coverage for public assets.
  - Foster development of a disaster insurance market.
  - Enhance insurance uptake by the private sector to diminish macroeconomic volatility and mitigate contingent risks for the public sector.
- Enforce DRM-linked territorial planning and zoning:
  - Develop a robust and holistic strategy for territorial planning that includes DRM considerations.
  - Imperative for subnational governments to diligently enforce zoning regulations to prevent continued exposure of the most-at-risk communities (e.g., rural, low-income) to future natural hazards.

### Technology diffusion, digitalization, and AI (productivity context)
- Peru’s productivity divergence and role of technology diffusion:
  - Peru has been diverging from advanced economies due to sluggish productivity growth; constrained technology and knowledge diffusion is one contributor.
  - Key barriers to diffusion include weak competition, constrained access to finance, proliferation of small firms, low R&D investment, and inadequate human capital.
- Digital access and government digitalization efforts:
  - Peru has lower internet access than other LA5 economies and some regions and parts of society are largely excluded from the digital world.
  - The government has modernized digital services: classification in the "very high" group of the United Nations' E-Government Development Index in 2022; National Policy of Digital Transformation for 2023-2030 aims to increase proportion of adults using digital technologies from about 52 percent to 85 percent by 2030.
  - Multilateral donor collaboration: IDB and World Bank initiatives to connect 3.5 million students and train over 265,000 teachers in digital skills; IDB aiding GovTech reforms.
- AI exposure, complementarity, and growth implications:
  - Peru’s labor force AI exposure: Approximately 30 percent of the labor force in Peru is highly exposed to AI; over 60 percent in the U.S. and the UK are highly exposed.
  - In the formal economy, AI exposure and complementarity are on par with advanced economies, so larger formal firms have more opportunities to use AI to boost productivity.
  - Policies that facilitate formalization and upgrade human capital can help attract workers from informal to formal sectors and raise aggregate productivity.
- Growth simulation scenarios and numerical projection:
  - If Peru halves its technological and AI exposure gap with the U.S. beginning in 2028 (optimistic scenario), it would move toward economic convergence with the U.S.
  - The annual average difference in GDP growth between optimistic and pessimistic scenarios from 2040 to 2050 is projected to be 0.21 percentage points.
  - The Atlas growth simulations indicate that reducing the technological gap would increase Peru’s long-term growth by 1.1 percent.
- Recommended structural reforms to narrow the technology gap:
  - Improve the business environment and foster competition.
  - Expand access to finance.
  - Develop human capital and enhance access to higher-quality education.
  - Address labor and tax regulations that incur costs for formalization.

*IMF staff calculations and analysis as presented in the source document.*

### 21.      The sectors that stand to gain the most

### The sectors that stand to gain the most

### Sectoral exposure and complementarities
- Sectors identified as likely to gain most in productivity from AI: finance, government, trade, IT, education, health, and real estate.
- Over 50 percent of the workforce in these sectors is significantly exposed to AI (Figure 4).
- AI is complementary for a substantial share of the workforce in these sectors, implying many employees could integrate AI to augment existing skills and increase productivity.
- Potential sectoral outcomes:
  - Healthcare and education: AI automation could enhance quality-to-price ratio, lead to price declines, and enable new and improved services, expanding these sectors.
  - Many employees may adopt AI to augment tasks rather than be displaced.

### Digitalization as a prerequisite and current gaps
- Digitalization is a crucial prerequisite for implementing AI technologies across sectors.
- Examples of current digitalization enabling AI:
  - Government: fully digitalized national tax system and GovTech platforms reveal AI opportunities.
  - Finance: evolving digital payment and lending ecosystem demonstrates capacity for AI integration.
  - Trade: marketplace platforms linking small shops to large consumer goods producers indicate progress.
- Notable connectivity gap in health:
  - Only a third of hospitals are connected to the internet, constraining AI adoption in healthcare and representing an opportunity for public–private partnerships.

### Price elasticity, demand, and employment implications
- Price elasticity of demand in the mentioned sectors tends to be relatively high.
- Implication: if automation reduces prices, demand should increase, prompting sector growth.
- Because AI shows high complementarity in most of these areas, increased production due to AI could result in job creation.

### Workforce risk, transition, and policy implications
- About 17 percent of the workforce in Peru is classified as high exposure to AI and low complementarity — tasks susceptible to automation that reduce the need for human involvement.
- Policy priorities to support sustainable growth and transitions:
  - Encourage current and future workers to develop skills for integrating AI into their work.
  - Implement measures to mitigate risk of income loss during job transitions.

### Advancements in FinTech
- Structural context:
  - Low financial inclusion and limited competition among incumbents create opportunities for FinTech disruption.
  - High 7-percent net interest spread indicates restricted access to finance.
- Financial inclusion statistics:
  - In 2021, 57.5 percent of Peruvians aged 15 and older had financial accounts.
  - In 2021, 22 percent borrowed from formal financial institutions.
- Digital payments and wallets:
  - About half of the adult population is engaged in digital payments.
  - In 2022, approximately 45 percent of Peruvians used digital wallets, a 12-percentage point increase from the year before.
  - Digital wallets helped boost financial inclusion to 65 percent in late 2023.
- FinTech ecosystem breadth:
  - Ecosystem includes cryptoasset exchanges, insurtech, enterprise technology provisioning, regtech, and wealthtech.
  - About 300 startups; Peru has one of the highest numbers of FinTech startups per capita in the region.
- Enabling policies implemented by authorities:
  - BCRP permitted the first e-money provider to deposit funds in its accounts, granting access to the fast payment clearing house.
  - A 2022 regulation ensured interoperability among different digital wallets and payment systems.
  - Authorities are working on an Open Banking law to foster competition by enabling data sharing.
  - Virtual asset providers are required to comply with Anti-Money Laundering (AML) regulations.
  - Authorities are engaging with market participants to refine the regulatory sandbox for FinTech testing.

### GovTech progress and limitations
- Peru ranks highly in GovTech maturity and scores very high across GovTech Maturity Index categories relative to Latin America, the OECD, and emerging market economies.
- Central government adoption:
  - Leveraging cloud computing, big data analytics, blockchain, artificial intelligence, and APIs to improve interoperability of government digital applications.
  - Extensive digitization of revenue administration at the central level; Peru is a high achiever in the region.
- Sub-national gaps:
  - Progress concentrated at central level; sub-national implementation lags.
  - Tax administration: 85 percent of tax payments by value were made electronically (ISORA 2021).
  - E-invoicing (2013 reform) led to over a 5 percent increase in reported sales, purchases, and value-added for firms in the first year after adoption; effects stronger for smaller firms and sectors with higher non-compliance.
  - VAT collection is hindered by shortcomings in the VAT refund mechanism.
  - Regional tax systems are predominantly non-digital, especially land taxes which lack a comprehensive registry—an obstacle to land tax digitalization.
- Public financial management (PFM) digitalization:
  - Authorities are undertaking comprehensive digitalization of budget preparation, accounting, procurement, public investment, debt management, and payroll to enhance transparency and control.
  - Prior systems were fragmented, hindering interoperability; most progress at central level.
- Digital technology in social sectors:
  - GovTech is used to collect education data at the community level, track distribution of educational materials, and monitor attendance.
  - Health sector: efforts to evaluate and enhance healthcare services and health information management systems; telemedicine expanded during the COVID-19 pandemic (teleconsultations for cancer and diabetic patients).
  - Expanding internet access beyond the current one-third of hospitals, and to households, could significantly improve access to and productivity of healthcare services.

*Source: 1perea2024002 - 21.      The sectors that stand to gain the most*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2024/english/1perea2024002.pdf_