Productivity and Growth in Peru1
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, but most notably, a decline in total factor productivity growth. In line with the macroeconomic trends, firm-level productivity has worsened, and the decline has been broad-based across the economy. Special corporate tax regimes and labor legislations and regulations have created barriers to productivity growth. To raise productivity, policies will need to focus on reforming regulations that impose excessive costs to formalizing or growing a business. Down the line, introducing greater labor market flexibility would ensure that workers can transition to productive sectors of the economy and reduce labor informality.
A. Macro-Level Measures of Potential Output
1. After a decade of high economic growth averaging over 6 percent per year, Peru’s economic growth has disappointed since 2014. There is a growing consensus that potential growth has declined since 2014.2 Recent estimates have ranged from 2.3 percent (Banco Central de Reserva del Perú, 2023) to 2.6 percent (Sánchez and Renato Vassallo, 2023). While many shocks have hit the Peruvian economy in recent years, there is a concern that the COVID-19 pandemic has exacerbated the slowdown in potential growth.
2. Three approaches—a univariate Hodrick-Prescott (HP) filter, a multivariate filter, and the production function approach—were used to estimate potential growth. For the HP filter, potential output was estimated by applying the HP filter to the annual real GDP series from 1990 to 2028 and using a smoothing parameter of 100. Using the smooth parameter suggested by Ravn and Uhlig (2002) yields similar results. For the multivariate filter, potential output was estimated following Alichi et al. (2017) and Blagrave et al. (2015) with the following specifications:
where Y is the output gap, π is the inflation rate, and u indicates the unemployment rate gap. For the production function approach, potential output was estimated using a Cobb-Douglas production function of the form
3. Using these three approaches, potential growth was estimated at 2.0 to 2.5 percent. All three approaches show over a decade of high potential growth averaging over 5.5 percent between 2002–2013 then a steady decline to an average of 2.0–2.5 percent between 2017–2022, excluding 2020 when potential output sharply declined due to the pandemic. Estimates from the production function approach are more volatile and show a lower level of potential growth of about 0.5 percentage points after the pandemic.
4. Potential growth has been declining in both the mining and non-mining economy. Potential growth in the mining and hydrocarbons sector is examined separately, as it may not exhibit the same growth patterns as the rest of the economy. Production relies on the stock of natural endowments, rather than accumulation of factors of production such as capital and labor, and is highly volatile due to commodity price cycles. Data on output, investment, and employment in the mining sector was collected to estimate potential output separately for the mining and non-mining economy. For the production function approach, a capital elasticity of 0.8 was used for the mining economy. While potential growth of the mining economy is estimated to be lower and more volatile than in the non-mining economy, it has been declining in both the mining and non-mining economy since 2014.
Peru: Actual and Potential GDP Growth
Citation: IMF Staff Country Reports 2024, 134; 10.5089/9798400276729.002.A001
Sources: INEI, MINEM, ENAHO, BCRP and staff calculations.5. More puzzling is the fact that TFP growth has been negative, even in the mining economy. TFP was the highest contributor to growth in the economy in 2002–2006, but its contribution rapidly declined and has been negative since 2012. Notably, while measured TFP growth in the mining and hydrocarbons sector is volatile and may reflect non-technological factors such as measurement error and capacity utilization, it has not been positive since 2007. The contribution of labor has remained stable in the economy, but its contribution is small and has been declining. Since 2007, capital has been the primary contributor to growth throughout the economy but as the rate of capital accumulation has been declining since 2012, so too has its contribution.
Peru: Growth Decomposition
Citation: IMF Staff Country Reports 2024, 134; 10.5089/9798400276729.002.A001
Sources: INEI, MINEM, ENAHO, BCRP and staff calculations.6. Augmenting the production function with mining resources affirms the negative growth in TFP. Estimates of TFP using a production function that does not account for mining resources as a factor of production may overestimate the contribution of capital and underestimate TFP growth (Bakker, 2023; Hamilton et al. 2019). Augmenting the production function with mining resources as a third factor or production, using data from World Bank (2021), increases estimated TFP growth, but it remains negative. While the mining and hydrocarbons sector is an important share of exports, mining resource rents account for less than 15 percent of GDP, thus not leading to a large bias in TFP estimation.
B. Measuring Productivity Using Firm-Level Data
7. Firm level data was used to estimate productivity at the microeconomic level. The analysis uses the Encuesta Ecónomica Annual (EEA), an annual survey of formal firms in Peru conducted by the INEI. The sample used is that of Schiffbauer et al. (2022), which restricts the sample to firms observed across multiple years and which are single establishments (which excludes many firms in the agricultural and mining sectors). This results in underrepresentation in microenterprises and overrepresentation in the commerce and manufacturing industries. Notwithstanding these issues, the sample is a panel of firms, allowing to follow productivity dynamics over time within a firm, and captures a sizeable portion of the economy (total value-add represents 14 to 19 percent of GDP). Total factor productivity was estimated following the integrated control function approach of Ackerberg, Caves, and Frazer (2015) and De Loecker and Warzynski (2012) and using a translog production function and including controls for firm age and municipality. The production function was estimated separately by industry.
8. Since 2014, in line with the macroeconomic trends, firm-level productivity has worsened. Patterns of productivity growth at the firm level match the aggregate trends. Between 2007 and 2013, as the Peruvian economy expanded, firm-level productivity growth accelerated from about 4 percent to about 6 percent per year. Vostroknutova et al. (2015) found that during this period, many firms in Peru significantly narrowed their productivity gap with the United States. Between 2014 and 2017, productivity growth reversed, in line with the macroeconomic trends, averaging about -5 percent per year.
9. The slowdown in productivity growth has been broad-based across the economy. Whether examining by industry, by firm size, or by region, firm-level productivity growth has worsened between 2014 and 2017 compared to earlier periods.4 In 2014, total factor productivity of the median firm in the most productive sectors – firms in construction and other services, large firms, and firms located in the coastal regions – was over two times that in the rest of the economy. Nevertheless, firms in these sectors also experienced negative productivity growth between 2014 and 2017. Following Olley and Pakes (1996), a decomposition of productivity growth shows that most of the decline is accounted for by less productive firms growing faster in market share and the composition of firms entering and exiting the market, rather than a shift in the overall productivity distribution. This suggests that barriers to firm entry (which have prevented productive firms from entering and unproductive firms from exiting) and to firm growth (which have prevented productive firms from growing as fast as they could) may have played an important role in productivity dynamics.
Peru: Firm-Level Total Factor Productivity Growth
Percent annualized growth, weighted by value added.
Citation: IMF Staff Country Reports 2024, 134; 10.5089/9798400276729.002.A001
Sources: EEA and staff calculations.10. Labor and tax legislations and regulations can create barriers to firm growth and productivity. Several labor and tax legislations and regulations in Peru are implemented using eligibility thresholds, which can create distortions in two ways. First, threshold-based regulations can create incentives for firms to maintain smaller sizes to take advantage of different regulations (or lack thereof). Second, threshold-based regulations impose a non-linear application of a regulation, which then translates into non-linear costs on firms. For example, Agosin et al. (2010) notes that moving between the various special corporate tax regimes in Peru can reduce profits by 30 to 50 percent. The result is that firms are not incentivized to formalize and grow, either to evade regulations or to maintain a certain level of sales and profits. These barriers to formalize and grow in Peru have likely contributed to the lack of medium-sized firms and high inequality in productivity between large and small firms (Agosin et al., 2010; Vostroknutova et al., 2015). The lack of incentives to formalize are particularly important, as informality is tightly linked with productivity.
Industry Productivity and Labor Informality
2007–2017
Citation: IMF Staff Country Reports 2024, 134; 10.5089/9798400276729.002.A001
Source: EEA, ENAHO, and staff calculations11. To investigate these potential barriers, two labor and tax regulations are analyzed. The first is a profit-sharing legislation. Decreto Legislativo No. 892 was introduced in 1996 and requires firms with more than 20 employees to share a certain percentage of profits with their workers. Firms in fishing, manufacturing, and telecommunications industries are required to share 10 percent of profits with workers. Firms in commerce, hospital, and mining industries are required to share 8 percent of profits with workers. Firms in all other sectors are required to share 5 percent of profits with workers. The second legislation is one of the special corporate tax regimes, the Régimen Especial del Impuesto a la Renta (RER). The RER was introduced in 2007 and applies to firms in extractive, manufacturing, trade, service, and agricultural industries. Firms in the RER pay a 1.5 percent tax on gross income, relative to a 10–29.5 percent tax on profits for firms in the Régimen MYPE Tributario (RMT) and Régimen General (RG). Firms in the RER are also only required to maintain purchases and sales registers, while firms in the RG are required to also maintain accounting journals, ledgers, inventories, and balance sheets. While not currently analyzed, there are two other tax regimes in Peru that are implemented using thresholds. One is the RMT for firms with annual net income below 1,700 UIT. The second is the Nuevo Régimen Unico Simplificado (NRUS) for firms with total sales below 96,000 soles. Firms in the NRUS play a single flat fee and have no accounting obligations.
12. Firms have maintained smaller sizes to avoid the application of a profit-sharing legislation, which has resulted in lower productivity. A histogram of firm density by number of employees exhibits a clear discontinuity at 20 employees, the cutoff at which the profit-sharing regulation is applied, indicating that firms have been maintaining smaller sizes. Firms can maintain smaller sizes by limiting expansion or splitting growing firms into multiple smaller firms to avoid the regulation. This has translated into lower productivity. To estimate the impact on productivity, a local linear regression discontinuity was estimated with the following specification:
yit = β1 1(employees ≤ 20) + β2 employees + β3 1(employees ≤ 20) x employees + δis + θt + εit
Where δis are industry fixed effects and θt are year fixed effects. Because regression discontinuity estimates are only identified within a narrow window around the threshold, the sample was limited to firms between 10 and 30 employees, but the estimates are robust to using other bandwidths. The estimates imply that firms just below 20 employees are about 40 percent less productive than firms just above 20 employees.
Peru: Profit-Sharing Regulation
Citation: IMF Staff Country Reports 2024, 134; 10.5089/9798400276729.002.A001
Sources: EEA and staff calculations.13. Special corporate tax regimes have imposed non-linear costs on firms, impacting productivity. The existence of numerous tax regimes for businesses also increases compliance and administrative costs and opportunities for tax arbitrage. Firms with annual sales below 525,000 soles can be in the RER, which allows them to pay a lower tax rate on gross income and have much lower accounting obligations compared to firms in the RMT and RG. A histogram of firm density by annual sales does not indicate that firms have been maintaining a smaller size to remain in the RER. However, the RER has distorted productivity. To estimate the impact on productivity, a local linear regression discontinuity was estimated with the following specification:
yit = β1 1(sales ≤ 525000) + β2 sales + β3 1(sales ≤ 525000) x sales + Sis + 8t + sit
where δis are industry fixed effects and θis are year fixed effects. Because regression discontinuity estimates are only identified within a narrow window around the threshold, the sample was limited to firms between 325,000 and 725,000 in annual sales, but the estimates are robust to using other bandwidths. Firms just outside the RER are about 40 percent less productive than firms just within the RER.
Peru: Régimen Especial del Impuesto de la Renta
Citation: IMF Staff Country Reports 2024, 134; 10.5089/9798400276729.002.A001
Sources: EEA and staff calculations.14. Reforming policies that impose excessive costs to formalizing or growing a business will be key to boosting productivity. While the high productivity growth between 2002–2013 coincided with the commodity price boom, there was a surge in private investment that was stimulated by significant reforms. These included a revamp of the macroeconomic policy framework (the introduction of inflation targeting and fiscal rules); trade and financial liberalization reforms; improvements in infrastructure; reforms in healthcare, education, civil service, pensions, and taxes; and development of the mining, agriculture, and tourism sectors (IMF, 2022; Ortiz and Winkelried, 2022; Vostroknutova et al., 2015). The years since have failed to yield comparable reforms. Notably, several business environment indicators – permitting and licensing requirements, tax codes, and labor regulations – have deteriorated since 2006. These counter-reforms should be reversed, and efforts will need to focus on reforming legislations and regulations that impose excessive costs to formalizing or growing a business, which would also reduce regulatory uncertainty and promote investment. As firm-level productivity increases, reforms will also need to ensure that workers can transition to productive sectors of the economy or risk exacerbating the dual economy: an unproductive informal economy coexisting with a highly productive formal economy.
Business Environment Indicators
Citation: IMF Staff Country Reports 2024, 134; 10.5089/9798400276729.002.A001
Source: World Bank Enterprise Surveys and staff calculations.References
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.
Prepared by Moya Chin (WHD) and Daria Kolpakova (WHD), with contributions from Matteo Ghilardi (WHD) on estimating potential output and aggregate productivity.
See, for example, estimates from Castillo and Hoyle (2019) and IMF (2022).
where I1 is investment, g is the average GDP growth rate between 1950 and 2022, and d is the depreciation rate.
Small firms are defined as microenterprises and small enterprises, which SUNAT defines as firms with below 1,700 UIT in annual sales. Large firms are defined as medium and large enterprises, which SUNAT defines as firms with above 1,700 UIT in annual sales.
