Journal Issue

El Salvador: Selected Issues

International Monetary Fund. Western Hemisphere Dept.
Published Date:
January 2015
  • ShareShare
Show Summary Details

Assessing Potential Output1

Based on various filters and the production function approach, El Salvador’s potential growth is estimated at about 2 percent for the period of 1999–2015, and the output gap is now virtually closed. Potential growth after the global financial crisis has fallen as a result of lower capital accumulation and total factor productivity (TFP). Going forward, it is critical to undertake structural reforms to strengthen capital and TFP to raise potential growth.

1. The level and growth of potential output are non-observable and are commonly defined in the literature as: (i) the long-run rate of growth of real GDP after removing cyclical factors (statistical definition), which may be estimated through various de-trending methods; or (ii) the full-employment level of output and its corresponding maximum growth that is sustainable without rising or slowing inflation (Okun, 1970). This definition requires estimating the gap between actual and potential output, based on equilibrium employment and capacity utilization.

2. El Salvador’s potential growth is the lowest in the Central American region and has been declining over time. On average, El Salvador’s potential growth is 2 percent for the period 1999–2015, compared with an average of about 4 percent in the region, excluding Panama. Factor accumulation has been the main contributor to potential growth in El Salvador, while TFP growth has been weakening it during 1999–2013.

Potential Growth, 2001-2014


Sources: WEO and Fund staff estimates.

Central America: Contribution to Real GDP Growth

(Period average, percent)

Source: REO, and Fund staff estimates.

3. Potential growth slowed during 1991–2015, possibly due to several structural changes.2 More recently, it declined markedly from 2.6 percent before the global financial crisis (GFC) of 2008–09 to 1.4 percent for 2011–14. Reduced contribution from capital formation (1.6 percent and 0.8 percent before and after the GFC, respectively) and negative TFP after the crisis (-1 percent) were the main drivers of the decline. Labor contribution to potential growth was higher after the crisis (1.3 percent vs. 1 percent before the crisis). For 2014, potential growth was estimated at about 1.7 percent.

Contributions to Potential GDP Growth


Source: WEO, ILO, UN, and Fund staff estimates.

Structural Breaks


Sources: Fund staff estimates.

Table 1.Potential Output Growth and Output Gap Estimates
Potential GDP Growth RateOutput gap
Production Function Approach1991–20151999–20152014201520142015
Potential GDP Growth Rate
Cycle Extraction Filters1991–20151999–20152014201520142015
Univariate and MultivariatePotential GDP Growth RateOutput gap
Kalman Filters (UVF and MVF)1991–20151999–20152014201520142015
MVF: Phillips Curve and Okun’s Law1.951.531.990.000.40
Average of All Models2.901.931.722.000.030.32
Source: Fund staff estimates.
Source: Fund staff estimates.

4. TFP growth depends on technological progress, as well as the institutional, regulatory, and legal environment in which businesses operate. TFP captures the efficiency with which labor and capital are combined to generate output, which, in turn, depends on businesses’ ability to innovate, as well as an environment that fosters competition, removes unnecessary administrative burden, provides modern and efficient infrastructure, and allows easy access to finance. Productivity shortfalls in El Salvador may reflect, inter alia, lags in investment in R&D and adoption and development of new technologies (chart 7 in Figure 1). In addition, productivity gains are also hindered by a lack of competition and high market concentration as determined by the Herfindahl-Hirschman Index (also charts 6–8 in Figure 1). Weak business environment, including political and economic uncertainty, poor security, high red tape and corruption, lack of legal/judicial stability, poor infrastructure, and lack of access to financing (charts 4–5 in Figure 1) are additional factors. Fostering human capital and advanced education (which averages only 1.7 years) and the return of high-skilled El Salvadorans from abroad can also contribute to TFP growth.

Figure 1.El Salvador: Investment, Competitiveness, and Human Capital

Source: World Economic Forum, Fusades, FM Global Resilience Index, and Fund staff estimates.

1/ Asia, excluding China.

Figure 1.El Salvador: Investment, Competitiveness, and Human Capital (continued)

Source: Global Innovation Index, OECD Product Market Regulation Database, OECD-WBG Product Market Regulation Indicators for LAC, REO, World Bank Indicators, World Economic Forum, and Fund staff estimates.

New Firm and Market Concentration Density Across Countries

Source: World Bank.

5. From a cyclical perspective, the economy is assessed to be operating at potential and labor market conditions also appear to be broadly neutral. The non-accelerating inflation rate of unemployment (NAIRU) is estimated at 6.3 percent during 1999–2015 and the unemployment gap appears to have closed in 2014 (charts 5–6 in Figure 2). Supplementary indicators from the World Economic Forum-based surveys suggest certain labor market rigidities, including inefficiencies in wage determination, alignment of pay with productivity, capacity to retain talent (Figure 1), mismatches between skills and jobs, and high informality. Such indicators have informed estimates for the NAIRU, and the estimate for a closed unemployment gap in 2014. A positive output gap of one percent of potential output is associated with about a quarter percentage point reduction in the unemployment gap, which could create pressures on inflation and the external balance. A positive output gap of one percent is associated with about a 0.1 percentage point increase in inflation.

Figure 2.El Salvador: Potential Output and Output Gap 1999–2014

Source: Fund staff estimates.

6. The estimated potential growth and NAIRU results should be interpreted with caution. There are serious data limitations with respect to the labor market and capacity utilization. Also, the statistical filters have several shortcomings—identifying the appropriate value of the detrending parameter is difficult and estimates have an endpoint bias. As for the TFP measure, it is by definition a residual—the difference between potential growth and the quantity (and quality) of inputs. Thus, any measurement errors in the labor and capital series are automatically imputed to TFP. For instance, employment shifts from the formal to the informal sector, migration of skilled labor, changes in the quality of the capital and labor stocks which are not correctly accounted for, and changes in the level of capital utilization and the use of land would be reflected in TFP.

7. Strengthening capital and TFP growth going forward is critical to achieve the authorities’ goal of raising potential growth to 3 percent over the medium term. Structural reforms should prioritize mobilizing domestic savings to invest and build a higher capital stock, enhancing R&D/technological diffusion and competition in product and labor markets, strengthening institutions to secure property rights and reduce red tape, improving infrastructure, facilitating access to financing, and fostering human capital to boost TFP growth.

Box 1.Methodology Underlying Potential Growth Estimates

Following Benes and others (2010), potential output is estimated using a Bayesian methodology, namely the regularized maximum likelihood. The multivariate Kalman filter incorporates relevant empirical relationships between actual and potential GDP, including unemployment, and headline inflation. The model is applied to annual data from 1999-2019. The model is a standard macroeconomic model built on an output gap and an unemployment gap. These gaps are pinned down by a number of identifying equations, including an inflation equation that relates inflation to the output gap through a Phillips curve relationship, and an unemployment equation that estimates an Okun’s law relationship.

1) Stochastic process for output:

2) Phillips curve equation

3) Okun’s equation

4) Inflation and growth expectations

where Y¯t is the log of potential GDP at time t, Gt is an unobserved slope component given by a fixed growth rate in the steady-state, GSS, and one of its lags. yt is the output gap (YtY¯t),pt is the headline inflation rate, ut is the unemployment gap given by the NAIRU (U¯t) and the actual unemployment rate (Ut),GtU¯ is an unobserved slope component and USS is a fixed steady-state unemployment rate. Finally, pt+jC and GROWTHt+jC are the inflation expectations and output growth expectations at time t for the j-periods ahead. Shock terms include: to the level of potential output òtY¯, to the growth rate of potential output òtG, to the output gap òty, to inflation òtp, to the unemployment gap òtu, to the level of NAIRU òtU¯, to the growth rate of the NAIRU òtGU¯.

The methodology requires taking a stance on prior beliefs regarding a number of variables. A key assumption fed into the model’s estimation is that supply shocks are the primary source of real GDP fluctuations in El Salvador. The prior belief that supply is more volatile than demand leads the model to assign much of the observed volatility of real GDP to potential GDP fluctuations. In addition to the prior distributions of parameters, values for the steady-state (long-run) unemployment rate and potential GDP growth rates are provided, which were set at 5.3 percent (based on the trend decline in the unemployment rate since the GFC) and 2 percent, respectively.

After obtaining estimates of potential output and NAIRU from the multivariate Kalman filter, potential TFP is calculated as a residual in the Cobb-Douglas function:

where Yt is potential output, Kt and Lt are capital and labor inputs, while At is the contribution of technology or TFP. Output elasticities (α is the capital share in the production function and is set at 0.35) sum up to one. Data on the working age population is obtained from the UN and the labor force participation rate is obtained from the ILO up to 2013 and assumed to grow at the 1999-2013 average annual rate thereafter. A labor force participation trend is calculated.

The capital stock series is constructed using a perpetual inventory method: Kt =(1 – δ) Kt—1 + It where the depreciation rate δ is set as 0.05, while the initial capital stock is computed as K0 = I*/(g + δ). I* is the benchmark investment and g is the average economic growth over 1999–2013.


Prepared by Iulia Teodoru.

Three different structural breakpoints were identified using an algorithm based on Bai (1997) and Bai & Perron (1998) to test for existence of multiple unknown structural breaks. The breakpoints were in 1997, 2001, and 2009. Although no causal inferences can be drawn from this exercise, these years correspond to the end of the economic rebound after civil war, the earthquake, and the global financial crisis, respectively. Natural disasters, such as the earthquake in 2001, can lower potential growth (direct and indirect costs over the period 1999–2013 have been estimated at over 20 percent of GDP, much higher compared to other countries in the region).

Other Resources Citing This Publication