Abstract
Paraguay’s economy recently experienced particularly large output swings. Economic policies will play a critical role in raising investment by making sure that macroeconomic stability is maintained. The spillovers from the agricultural sector to the rest of the economy are limited. The high level of bank excess reserves in Paraguay reflects a mixture of precautionary and involuntary factors. Large bank excess reserves weaken the monetary transmission channel, and cause inefficiency costs. Bank reforms should be undertaken to preserve financial and macroeconomic stability.
I. Paraguay—Potential Output Growth and Spillovers from Agriculture1
A. Introduction
1. Paraguay’s economy has recently experienced particularly large output swings, opening again the question of what its potential growth rate is. Notably, in only two years, the economy went from a half-a-century record-low GDP contraction of 3¾percent in 2009—after the severe and widespread drought of that year and the adverse international environment due to the global crisis—to a record-high GDP growth of 15 percent in the subsequent year, led by the unprecedented boom in the agricultural sector during 2010.
2. This paper assesses Paraguay’s potential growth and its associated output gap, and estimates the spillovers from agriculture to the rest of the economy. In assessing potential growth, the paper uses a number of econometric and statistical techniques, focusing on both the long-run trends of the economy and more recent developments. Owing to the large relevance of the agricultural sector in Paraguay, the paper also estimates the potential spillovers from the agricultural sector to the rest of the economy.
3. Three main results follow from the paper. First, it is shown that after the 2002 crisis the economy seems to have entered into a higher GDP growth path. Although Paraguay’s economy is still very volatile, estimations suggest that a GDP growth rate of 4½ to 5 percent per year might be achievable over the medium term, up from the historical 4 percent estimated in previous studies.2 To attain this, however, the economy will require larger investment ratios. Economic policies will be crucial to raising the investment ratio, including by maintaining a climate conducive to higher private investment and addressing well-known bottlenecks to higher long-run growth in Paraguay, such as the quality of infrastructure. Second, estimations show that Paraguay had a positive output gap in 2010. Additionally, a more disaggregated analysis indicates that the economic boom (and the output gap) of 2010 was driven by a broad-based economic expansion and not only by the historically large agricultural sector boom. Finally, estimations show that the spillovers from agriculture are limited, with construction being the sector that is more influenced by the swings of the Paraguayan agricultural sector.
4. The rest of the paper is organized as follows. Section B discusses Paraguay’s longrun output trends, including a growth accounting exercise to evaluate the determinants of growth. Section C focuses instead on the most recent developments, to evaluate the current potential growth of Paraguay and its associated output gap. Section D deals with a quantitative estimation of the spillovers from the agricultural sector to the rest of the economy, while Section E closes the paper presenting a number of concluding remarks.
B. Long-Run Growth and Total Factor Productivity (TFP) Estimates
5. Paraguay’s GDP growth has historically been very volatile. However, significant differences in the average growth rate arise depending on the specific period under consideration. During the belle époque of rapid growth, between 1963 and 1981, the economy grew on average 7.1 percent per year. This vigorous rate plummeted in the following 20 years to a modest 2 percent annual growth. From 2003 until 2010 the economy took off again, reaching an average growth rate of 4.9 percent. Interestingly, although Paraguay experienced important economic swings in the past 50 years, sharp and persistent output contractions have been relatively rare. In fact, the largest GDP fall in half a century occurred in 2009—3¾ percent—after the widespread drought that affected the agricultural sector and the adverse international environment due to the global crisis. Yet the downturn immediately reverted with the all-time record harvest of 2010, which substantially contributed to the 15 percent record-high GDP growth of that year.
Actual and Potential GDP Growth
(In percent)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.Actual and Potential GDP Growth
(In percent)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.Actual and Potential GDP Growth
(In percent)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.6. The upswings of the Paraguayan economy have also led to important changes in its potential growth assessment over half a century. According to a growth accounting exercise explained later in detail, Paraguay’s potential growth hovered around 6.7 percent per year between 1963 and 1981. This level later decreased to 2.7 percent between 1982 and 2002, in line with the large downturn in actual growth. As the economy’s actual GDP growth rate increased significantly from that point onwards, potential growth rose accordingly, reaching about 4.9 percent by end-2010.
7. The assessment of Paraguay’s long-run potential output growth builds on a standard growth accounting exercise. Specifically, the following Cobb-Douglas production function is estimated:
where Yt is total output, At is a technology parameter or TFP, Kt and Lt denote the capital stock and labor inputs, respectively, and α is the share of capital income in total output. In this setting, the capital stock evolves according to the following law of motion:
where It is total investment and δ is the depreciation rate of the capital stock. This methodology requires the estimation of Paraguay’s capital stock, as explained below.
8. Although Paraguay’s capital stock has generally trended upwards over the last 50 years, the capital-to-output ratio has decreased after a peak reached in the early 2000s. To estimate the capital stock an initial condition is required, which is obtained from the following steady state relation,
where g is the long-run growth rate of the capital stock. The capital stock of the year 1962, K1962, is set as the initial condition for the model. Specifically, this is obtained from Eq. (3) using investment figures for the year 1962, I1962, and setting δ=0.08—a capital depreciation rate of 8 percent per year—and g=0.04—the average GDP growth rate over the last 60 years, equal to 4.06 percent.3 Over the long run (in the steady state), the growth rate of the capital stock and GDP should coincide. The capital stock later evolves according to Eq. (2). Results of the estimation are summarized in Figure 2.4
Capital Stock Estimations
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.Capital Stock Estimations
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.Capital Stock Estimations
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.9. Estimations also suggest that Paraguay’s TFP has fluctuated significantly more than output, trending downward or remaining flat after the early 1980s’ peak. TFP estimations are obtained from Eq. (1), once capital and labor inputs are accounted for, and the parameter α is calibrated. The latter is set to 0.35 in this exercise.56
10. The model is also used to evaluate the dynamics of Paraguay’s per worker GDP growth, as it helps assess more thoroughly the long-run performance of the economy. Expressing Eq. (1), in per worker terms and applying logs gives,
where yt is output per worker, kt is capital per worker and at is TFP, all in logs. Differentiating Eq. (4), with respect to t provides a decomposition of per worker GDP growth between the contributions due to TFP growth (or productivity gains) and those due to capital per worker growth.
11. Paraguay’s per worker GDP growth performance has been less than stellar from a long-run perspective, notably after the 1980s. This fact can be explained by the combination of moderate productivity gains in the economy and a relatively stagnant capital per worker growth rate. During the 1970s, the rapid per worker GDP growth was sustained by the significant increases in the capital per worker ratio and, to some extent, TFP gains. However, the stagnant or decreasing investment ratio that prevailed after the 1970s reduced the overall contribution of capital to explaining GDP growth. Indeed, during certain particular years the main engine for growth was related to increases in TFP. This occurred for instance during the late 1980s and early 2000s, yet in general productivity gains in Paraguay have been modest and intermittent, and thus they have not been the main source of economic growth in the last 30 years. Lately, in 2010, per worker GDP growth has surprised on the upside—about 13.1 percent—which the model particularly assigns to significant TFP gains. However, this probably reflects the effect of extremely favorable weather conditions on agricultural sector growth in that year and might thus be short-lived.
TFP Estimations
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.TFP Estimations
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.TFP Estimations
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.C. Potential Growth and the Output Gap: Recent Developments
12. A structural break affected the Paraguayan economy by end-2002, leading to significant differences in the assessment of its potential growth. According to standard statistical techniques, Paraguay’s potential growth rose on average from about 0.6 percent in the period 1994–2002 to an average of about 4½percent in the period 2003–2010 (Table 1 and Figure 4).
Potential Growth Estimations
(In percent)
(1): 1994Q1 to 2010Q4 period; (2): 1994Q1 to 2002Q4 period; and (3): 2003Q1 to 2010Q4 period.
To reduce the end-of-sample bias problem eight additional quarters are added to the 1994Q1-2010Q4 sample for these estimations using IMF staff projections.
Potential Growth Estimations
(In percent)
| Aggregate GDP1/2/ | Non-agricultural GDP1/2/ | |||||
|---|---|---|---|---|---|---|
| (1) | (2) | (3) | (1) | (2) | (3) | |
| Hodrick and Prescott | 2.5 | 0.8 | 4.3 | 2.1 | 0.2 | 4.0 |
| Baxter and King | 2.6 | 0.4 | 4.5 | 2.1 | -0.2 | 4.2 |
| Christiano and Fitzgerald | 2.6 | 0.6 | 4.3 | 2.1 | 0.0 | 4.0 |
| Average | 2.6 | 0.6 | 4.4 | 2.1 | 0.0 | 4.1 |
(1): 1994Q1 to 2010Q4 period; (2): 1994Q1 to 2002Q4 period; and (3): 2003Q1 to 2010Q4 period.
To reduce the end-of-sample bias problem eight additional quarters are added to the 1994Q1-2010Q4 sample for these estimations using IMF staff projections.
Potential Growth Estimations
(In percent)
| Aggregate GDP1/2/ | Non-agricultural GDP1/2/ | |||||
|---|---|---|---|---|---|---|
| (1) | (2) | (3) | (1) | (2) | (3) | |
| Hodrick and Prescott | 2.5 | 0.8 | 4.3 | 2.1 | 0.2 | 4.0 |
| Baxter and King | 2.6 | 0.4 | 4.5 | 2.1 | -0.2 | 4.2 |
| Christiano and Fitzgerald | 2.6 | 0.6 | 4.3 | 2.1 | 0.0 | 4.0 |
| Average | 2.6 | 0.6 | 4.4 | 2.1 | 0.0 | 4.1 |
(1): 1994Q1 to 2010Q4 period; (2): 1994Q1 to 2002Q4 period; and (3): 2003Q1 to 2010Q4 period.
To reduce the end-of-sample bias problem eight additional quarters are added to the 1994Q1-2010Q4 sample for these estimations using IMF staff projections.
Actual and Potential Output
(Billions of 1994 Guaraníes)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.Actual and Potential Output
(Billions of 1994 Guaraníes)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.Actual and Potential Output
(Billions of 1994 Guaraníes)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.13. Potential growth in the non-agricultural sector also increased after 2002, following a path broadly similar to that of total output. Specifically, potential growth in the non-agricultural sector rose from about 0 percent to about 4.1 percent on average between 1994–2002 and 2003–2010. These findings suggest that the increase in Paraguay’s potential growth during 2003–2010 has been driven by a widespread expansion of the different sectors of the economy and not only by the very dynamic agricultural sector.
14. Statistical techniques are also used to estimate Paraguay’s output gap in recent years. The statistical techniques used to evaluate Paraguay’s potential growth may also be applied to estimate its output gap. As summarized below, the methodologies permit to evaluate easily the output gap for the economy as a whole, and that specific to the non-agricultural sector.7 It is important to estimate the latter separately given the relatively large share (around one-fifth) of the agricultural sector in Paraguay and the inherent difficulties of estimating an output gap for agriculture. In this respect, and given the limited spillovers from agriculture to the rest of the economy (see section D), the output gap of the non-agricultural sector is arguably a more meaningful indicator of demand pressures—or their absence—in Paraguay.
Total GDP: The severe drought in combination with the global crisis led to a substantially negative output gap in 2009—estimations ranged from -6.7 percent according to the Hodrick-Prescott (HP) filter to about -6 percent according to the Christiano and Fitzgerald (CF) filter (Table 2). With the 2010 economic boom, the output gap moved into positive territory again, with estimates ranging from 1.6 percent (HP filter) to 2.4 percent (CF filter). Notwithstanding the strong economic boom in 2010, the output gap in that year remained below that estimated for 2008, a year in which the economy also experienced an important boom.
Non-agricultural GDP: The output gap for the non-agricultural sector followed a path similar to that of the economy as a whole, yet the size of the gap has been lower. According to the results of the estimation, the output gap excluding agriculture was on average about 0.8 percent in 2008, decreasing to around -2.5 percent in 2009. After the widespread economic boom of 2010, the output gap is estimated to have turned positive in the second quarter of the year and was slightly positive for the year as a whole.
Output Gap Estimations
(In percent)
HP: Hodrick and Prescott methodology; BK: Baxter and King methodology; and CF: Christiano and Fitzgerald methodology.
To reduce the end-of-period bias problem eight additional quarters are added to the 1994Q1- 2010Q4 sample for these estimations using the IMF staff projections
Output Gap Estimations
(In percent)
| Aggregate GDP1/2/ | Non-agricultural GDP1/2/ | |||||
|---|---|---|---|---|---|---|
| HP | BK | CF | HP | BK | CF | |
| 2008Q1 | 3.14 | 2.66 | 4.75 | 0.71 | 1.63 | 2.53 |
| 2008Q2 | 4.28 | 3.54 | 5.52 | 2.61 | 1.70 | 2.97 |
| 2008Q3 | 0.42 | 2.44 | 3.00 | -0.05 | 0.63 | 0.27 |
| 2008Q4 | 1.64 | -0.21 | -0.70 | -1.74 | -0.69 | -1.26 |
| 2008 | 2.37 | 2.11 | 3.14 | 0.38 | 0.82 | 1.13 |
| 2009Q1 | -7.56 | -4.13 | -6.24 | -3.24 | -2.20 | -2.81 |
| 2009Q2 | -8.33 | -5.67 | -7.79 | -5.03 | -2.67 | -3.93 |
| 2009Q3 | -7.14 | -4.09 | -5.31 | -3.57 | -1.67 | -2.57 |
| 2009Q4 | -3.73 | -2.25 | -4.77 | -0.39 | -0.77 | -1.60 |
| 2009 | -6.69 | -4.03 | -6.03 | -3.06 | -1.83 | -2.73 |
| 2010Q1 | 1.88 | 0.91 | 1.18 | -1.03 | -0.26 | -0.80 |
| 2010Q2 | 0.67 | 1.72 | 2.51 | -0.38 | 0.22 | 0.70 |
| 2010Q3 | -0.64 | 1.40 | 2.41 | 0.15 | 0.40 | 1.00 |
| 2010Q4 | 4.64 | 1.65 | 3.55 | 1.95 | 1.45 | 2.45 |
| 2010 | 1.64 | 1.42 | 2.41 | 0.17 | 0.45 | 0.84 |
HP: Hodrick and Prescott methodology; BK: Baxter and King methodology; and CF: Christiano and Fitzgerald methodology.
To reduce the end-of-period bias problem eight additional quarters are added to the 1994Q1- 2010Q4 sample for these estimations using the IMF staff projections
Output Gap Estimations
(In percent)
| Aggregate GDP1/2/ | Non-agricultural GDP1/2/ | |||||
|---|---|---|---|---|---|---|
| HP | BK | CF | HP | BK | CF | |
| 2008Q1 | 3.14 | 2.66 | 4.75 | 0.71 | 1.63 | 2.53 |
| 2008Q2 | 4.28 | 3.54 | 5.52 | 2.61 | 1.70 | 2.97 |
| 2008Q3 | 0.42 | 2.44 | 3.00 | -0.05 | 0.63 | 0.27 |
| 2008Q4 | 1.64 | -0.21 | -0.70 | -1.74 | -0.69 | -1.26 |
| 2008 | 2.37 | 2.11 | 3.14 | 0.38 | 0.82 | 1.13 |
| 2009Q1 | -7.56 | -4.13 | -6.24 | -3.24 | -2.20 | -2.81 |
| 2009Q2 | -8.33 | -5.67 | -7.79 | -5.03 | -2.67 | -3.93 |
| 2009Q3 | -7.14 | -4.09 | -5.31 | -3.57 | -1.67 | -2.57 |
| 2009Q4 | -3.73 | -2.25 | -4.77 | -0.39 | -0.77 | -1.60 |
| 2009 | -6.69 | -4.03 | -6.03 | -3.06 | -1.83 | -2.73 |
| 2010Q1 | 1.88 | 0.91 | 1.18 | -1.03 | -0.26 | -0.80 |
| 2010Q2 | 0.67 | 1.72 | 2.51 | -0.38 | 0.22 | 0.70 |
| 2010Q3 | -0.64 | 1.40 | 2.41 | 0.15 | 0.40 | 1.00 |
| 2010Q4 | 4.64 | 1.65 | 3.55 | 1.95 | 1.45 | 2.45 |
| 2010 | 1.64 | 1.42 | 2.41 | 0.17 | 0.45 | 0.84 |
HP: Hodrick and Prescott methodology; BK: Baxter and King methodology; and CF: Christiano and Fitzgerald methodology.
To reduce the end-of-period bias problem eight additional quarters are added to the 1994Q1- 2010Q4 sample for these estimations using the IMF staff projections
Output Gap Estimations
(In percent)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.Output Gap Estimations
(In percent)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.Output Gap Estimations
(In percent)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.D. Are the Spillovers from the Agricultural Sector Relevant in Paraguay?
15. To analyze empirically the spillovers from the agricultural sector to the rest of the economy, the following vector autoregressive (VAR) model is estimated:
where A(L) and B(L) are a n x n and a n x k polynomial matrices in the lag operator L, respectively, Yt is a n x 1 vector of endogenous variables, Xt is a k x 1 vector of exogenous variables, and Ut is a n x 1 vector of estimated residuals. The VAR model is further specified as follows:
and
16. In this specification the vector of endogenous variables Yt includes the output of Paraguay disaggregated by sectors. Specifically, Yt includes agricultural output (Agriculturet) as the most exogenous variable, followed by industrial output (Industryt), construction output (Constructiont) and services output (Servicest).8 According to this ordering, a shock to the agricultural sector at period t affects contemporaneously all the variables of the model, whereas an innovation to the services sector at time t affects the remaining variables of the model only with a lag. The Dummyt variable in vector Xt is included to control for the drought that affected the economy during 2009.
17. The VAR model includes the main sectors of Paraguay, which represent around 90 percent of 2010 GDP. In terms of the relevance of each particular sector, agriculture roughly represents 22 percent of 2010 GDP, industry about 13 percent, construction about 4 percent, and services about 51 percent (Figure 6).9
Main Sectors of the Paraguayan Economy
(In percent of total output)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.Main Sectors of the Paraguayan Economy
(In percent of total output)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.Main Sectors of the Paraguayan Economy
(In percent of total output)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.18. The VAR model is estimated considering quarterly data for the period 2003Q1–2010Q4. Therefore, this data set covers only the period that follows the structural break of the Paraguayan economy that took place by end-2002. All variables in Yt are measured at constant prices, are seasonally adjusted and expressed in logs. Using standard techniques, 2 lags are chosen for the estimation of the VAR model.
19. VAR results suggest little spillovers from the agricultural sector to the rest of the economy, with the construction sector responding more significantly to changes in agricultural output. Specifically, Figure 7. depicts the impulse-response functions to a 10 percent increase in the output of the agricultural sector. In the case of construction, its output rises on impact about 2 percent, yet the effect vanishes about 4 quarters after the shock. Although there is a somewhat mild increase in the output of the services sector, the effect is not very significant. Interestingly, industrial output does not seem to be affected by the increase in the output of the agricultural sector. In general, these results might reflect the fact that agriculture in Paraguay relies heavily on imports of intermediate and capital goods, it is a land- and capital-intensive sector, and pays relatively little taxes.10
Impulse-Response Functions to a 10 Percent Increase in Agricultural Output
(In percent)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Impulse-Response Functions to a 10 Percent Increase in Agricultural Output
(In percent)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Impulse-Response Functions to a 10 Percent Increase in Agricultural Output
(In percent)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
20. A forecast error variance decomposition analysis confirms the limited spillover effect from the agricultural sector to the rest of the economy. At most, an agricultural sector specific shock contributes to explaining about 22.5 percent of the output volatility in the construction sector—the sector with the largest spillover from agriculture—around 4 quarters after the shock. For the other two sectors the contribution of agricultural shocks in explaining their sectoral volatility is rather scant, with a slightly larger relevance in the services sector (Table 3).
Forecast Error Variace Decomposition Due to an Agricultural Sector Shock
(In percent)
Forecast Error Variace Decomposition Due to an Agricultural Sector Shock
(In percent)
| Quarters | Industry | Construction | Services |
|---|---|---|---|
| 1 | 0.1 | 9.8 | 4.2 |
| 4 | 2.3 | 22.5 | 5.6 |
| 8 | 2.6 | 20.0 | 3.6 |
| 12 | 2.7 | 17.5 | 3.0 |
| 24 | 2.5 | 12.5 | 2.4 |
Forecast Error Variace Decomposition Due to an Agricultural Sector Shock
(In percent)
| Quarters | Industry | Construction | Services |
|---|---|---|---|
| 1 | 0.1 | 9.8 | 4.2 |
| 4 | 2.3 | 22.5 | 5.6 |
| 8 | 2.6 | 20.0 | 3.6 |
| 12 | 2.7 | 17.5 | 3.0 |
| 24 | 2.5 | 12.5 | 2.4 |
21. To contrast results with those provided by the VAR model, a set of simple OLS regressions are also considered in the analysis. The estimated univariate OLS model is represented by the following equation:
where Yj,t is output in either industry, construction or services as indicated by the sub-index j, At is agricultural sector output and Dummyt, as in the VAR specification, is a dummy variable that corrects for the effects of the 2009 drought. Also as in the VAR model, the estimations consider quarterly data for the period 2003Q1 to 2010Q4, with all sectoral output data expressed in logs and seasonally adjusted. In this specification, the short-run impact of agriculture on the other sectors is given by c(3), whereas the long-run impact is accordingly given by
22. OLS regressions for each sector support the findings of the VAR model: the largest short-run spillover effect of agriculture is on the construction sector, and the effects on industry and services are limited. Estimations also show that over the long run a rise in agricultural output produces a significant and positive effect on construction and services, yet the effect on industry remains particularly small (Table 4).
OLS Estimation of Spillover Effects From the Agricultural Sector
OLS Estimation of Spillover Effects From the Agricultural Sector
| c(1) | c(2) | c(3) | c(4) | c(5) | Short-run effect | Long-run effect | R-squared | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Industry | 1.72 | 0.70 | 0.00 | 0.02 | 0.03 | 0.00 | 0.09 | 0.75 | |||
| t-Statistic | 2.39 | 5.09 | 0.02 | 0.46 | 2.33 | ||||||
| Construction | 1.64 | 0.18 | 0.23 | 0.13 | 0.13 | 0.23 | 0.44 | 0.85 | |||
| t-Statistic | 3.51 | 1.83 | 2.38 | 1.37 | 5.01 | ||||||
| Services | 0.25 | 0.93 | 0.05 | 0.00 | 0.02 | 0.05 | 0.61 | 0.97 | |||
| t-Statistic | 0.68 | 12.22 | 1.10 | -0.07 | 1.18 | ||||||
OLS Estimation of Spillover Effects From the Agricultural Sector
| c(1) | c(2) | c(3) | c(4) | c(5) | Short-run effect | Long-run effect | R-squared | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Industry | 1.72 | 0.70 | 0.00 | 0.02 | 0.03 | 0.00 | 0.09 | 0.75 | |||
| t-Statistic | 2.39 | 5.09 | 0.02 | 0.46 | 2.33 | ||||||
| Construction | 1.64 | 0.18 | 0.23 | 0.13 | 0.13 | 0.23 | 0.44 | 0.85 | |||
| t-Statistic | 3.51 | 1.83 | 2.38 | 1.37 | 5.01 | ||||||
| Services | 0.25 | 0.93 | 0.05 | 0.00 | 0.02 | 0.05 | 0.61 | 0.97 | |||
| t-Statistic | 0.68 | 12.22 | 1.10 | -0.07 | 1.18 | ||||||
23. Estimations also show little change in the spillovers from the agricultural sector over time, but suggest an increase in the spillover to the construction sector around 2002. Using the VAR stated previously, a set of rolling VARs is estimated using different subsamples. Each estimated VAR comprises a 32-window period, starting in 1994Q1. The impulse response functions to a one standard deviation agricultural sector shock are then computed, and the associated one quarter response of each sector is stored. Each subsequent subsample adds one quarter to the previous one, with the latest subsample starting in 2003Q1 (going through 2010Q4) being equal to that used in the previous VAR estimations. Figure 8 plots the one-quarter responses of the different sectors considering the rolling VAR estimations. It follows from the figure that only the construction sector output had a significant and positive increase in the spillover from agriculture starting around 2002, roughly coinciding with the boom in the production of soybeans in Paraguay.
Rolling VARs
(In percent)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.Rolling VARs
(In percent)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.Rolling VARs
(In percent)
Citation: IMF Staff Country Reports 2011, 239; 10.5089/9781462320387.002.A001
Source: IMF staff calculations.E. Concluding Remarks
24. The analysis of the paper suggests that Paraguay’s potential growth increased substantially in the years that followed the 2002 crisis, reaching about 5 percent in 2010. To sustain this growth level over the medium term, however, the economy will require larger investment levels. Economic policies will play a critical role in raising investment, notably by making sure that macroeconomic stability and a favorable climate for private investment are maintained, while well-known bottlenecks to long-run growth in the country (e.g., infrastructure quality) are effectively addressed.
25. Estimations also suggest that the output gap turned positive in mid-2010. Containing, and eventually eliminating, the overheating pressures associated with a positive output gap is also an important challenge for the authorities to sustain a stable macroeconomic environment. Finally, although the agricultural sector represents about one fifth of Paraguay’s GDP, the paper suggests that the spillovers from this sector to the rest of the economy are limited. While strengthening the linkages of agriculture with the other sectors of the economy is a big challenge, economic policy measures that raise the tax contribution of the agricultural sector in an efficient manner could help in this regard. This would also contribute to the ultimate goal of providing the basis for Paraguay to grow at its potential level over the medium run by providing the state with additional resources to improve infrastructure and raise human capital.
References
Brauman, B., 2009, “Long-Run Potential Growth” in Paraguay: Addressing the Stagnation and Instability Trap, ed. By Alejandro Santos (Washington: International Monetary Fund)
Fernandez Valdovinos, C. and A. Monge Naranjo, 2004, “Economic Growth in Paraguay,” Economic and Social Study Series, (Washington: Inter-American Development Bank)
Magud, N. and L. Medina, 2011, “The Chilean Output Gap”, IMF Working Paper No 11/2 (Washington: International Monetary Fund).
Rodriguez, P., 2009, “The Challenge of Sustaining Growth” in Paraguay: Addressing the Stagnation and Instability Trap, ed. By Alejandro Santos (Washington: International Monetary Fund).
Prepared by Santiago Acosta-Ormaechea.
See, for example, Brauman (2009).
The depreciation rate for the capital stock is taken from Fernández Valdovinos and Monge Naranjo (2004)
The capital-to-output ratio in Paraguay still remains well below the level of industrialized countries. For instance, in the case of the U.S., the average of this ratio for the period 1962 to 2008 reached 2.9 whereas in Paraguay it reached 1.8.
The chosen value for α is in line with previous growth accounting exercises. For instance, Brauman (2009) estimates from Paraguay’s national accounts an average capital income share on output of about 35 percent considering data from the late 1990s and early 2000s. He then sets α=0.35 in his analysis, as it is also done here.
Results are robust to different parameterizations of the depreciation rate 5 and the capital income share a. In addition to the calibration used here of δ=0.08 and α=0.35, other estimations have been considered taking instead δ=0.1 (from Brauman, 2009) and α=0.44 (from a number of unpublished estimations of the Central Bank of Paraguay), without showing major differences in results.
The use of different statistical filters to calculate potential output and its associated output gap is widely considered in the literature. Magud and Medina (2011), for instance, calculate these unobservable variables following an approach similar to that considered here.
Strictly speaking, due to data availability Industry also includes Mining, yet the latter is a very minor activity in Paraguay—about 0.1 percent of 2010 GDP.
The other relevant sectors not included in the model are livestock (about 8 percent of 2010 GDP) and electricity (about 2 percent of 2010 GDP).
The stronger link with the construction could be related to the tendency of agricultural booms to generate property booms in countries where agriculture is important as a source of income.
Appendix I: Legal Reserve Requirements in Selected Latin American Countries
| Country | Local Currency | Foreign Currency | Marginal and Additional Rates | Remuneration | Eligible Liabilities | Compliance Assets | Averaging | Penalty | Purpose/ Last change |
|---|---|---|---|---|---|---|---|---|---|
| Paraguay | 18 percent sight | 25 percent sight deposits; and 16.5 percent time and saving deposits | No | Local currency: Interbank interest rate Foreign currency: one- month Libor | Deposits in local and foreign currency | Cash deposits at the central bank in the some currency than the deposits | No allowed | To be determined by the board of the central bank. | Monetary policy; reduced in 2009 to counter the crisis and increased back in 2011 to rein in credit growth. |
| Peru | Minimum 9 percent of deposits in local currency | Minimum 9 percent of deposits in foreign currency | Yes; 25 percent for deposit of residents in local currency, 120 percent for deposits of nonresidents in foreign currency; 55 percent for deposits in foreign currency. | Only the additional requirements are remunerated at the overnight rate less 1 percentage point for reserve in local currency and 60 percent of the one- month Libor for reserves in foreign currency. | Deposits, bonds, mortgages securities, subordinated debt and external borrowing with maturity less than two years. | Cash in vault, deposits at the central bank in the some currency than the deposits | Yes, over a one-month maintenance period | 1.5 times the lending rate in local and foreign currency over the average deficiency. | Monetary policy, macro-prudential tool. |
| Colombia | 11 percent sight deposits; 4.5 percent time deposits | Not applicable | Discontinued | No remunerated | Deposits | Cash in vault and deposits at the central bank | Yes; two weeks maintenance period. | 3.5 percent over the average deficiency | Monetary policy |
| Brazil | 43 percent sight deposits 20 percent time and saving deposits (except if allocated to fund rural loans) | Not applicable | Yes; 12 percent on sight and time deposits and 10 percent on savings deposits | Only reserves corresponding to saving deposits and additional requirements, at 6.17 percent. | Deposits and interbank loans, excluding government deposits in public banks | Cash in vault, deposits in the central bank, federal government bonds and notes. | Yes, only for demand deposits for a two-week maintenance period. | Policy rate+ 14 percent over the average deficiency | Monetary policy, incentivize credit, macro-prudential tool |
| Dominican Republic | 17 percent commercial | 20 percent Foreign | No | Reserves in local currency are | Deposits | Deposits at the central | Yes | 0.1 percent over the average | Monetary policy |
| banks; 12.5 percent credit and saving associations; 10 percent other financial associations. | Currency Deposits | unremunerated. Foreign currency reserves are remunerated at the Fed Fund rat minus 2 percentage points. | bank in the some currency than the deposits | deficiency | |||||
| Chile | 9 percent sight deposits, 3.6 percent time deposits, 0 percent saving deposits. | 9 percent sight deposits, 3.6 percent time deposits, 0 percent saving deposits. | No | No remunerated | Deposits and external borrowing. | Cash in vault and deposits at the central bank | Yes for over a one-month maintenance period | 2 times the 3- month interest rate over the average deficiency. | Monetary Policy |
| Bolivia | 2 percent for cash, 10 percent for deposits below one-year maturity and 12 percent for two- year maturity deposits | 2 percent for cash, 10 percent for deposits below one-year maturity and 12 percent for two-year maturity deposits | Additional 30 percent requirement for deposits in foreign currency | Reserves maintained as deposits at the central bank are no remunerated. If deposited in the “Liquid Asset Requirement Fund”, the reserves are invested by the central bank of behalf of institutions and remunerated at the market interest rates. | Deposits and securities in local and foreign currency, except government deposits, short-term external credit lines, prepaid warranty deposits of credit letters. | Cash in vault, deposits at the central bank in the some currency than the deposits | Yes, for a 14 days maintenance period | 2 percent over average deficiency or twice the lending rate. The penalty increase in 2 percentage points for any additional subsequent period of deficiency | Monetary Policy and incentivize credit (the total requirement could be reduce to 6 percent if the institution increase loans in local currency) |
| Costa Rica | 15 percent | 15 percent | No | No remunerated | Deposits and trusts | Deposits at the central bank (only those at the reserve account) in the some currency than the deposits | Yes, for a 15-day maintenance period | Interest rate of discount window over the reserve deficiency | Monetary policy |
| The Bahamas | 5 percent | No applicable | No | No remunerated | Deposits, including non-residents’ accounts. | Cash in vault, deposits at the central bank | Yes, over a one-month maintenance period | 2 times the annual discount rate for every day of reserve deficiency | |
| Argentina | 19 percent sight deposits, 0-14 percent time deposits depending on maturity | 20 percent sight deposits, 0-20 percent time deposits depending on maturity | No | No remunerated | Deposits | Cash in vault and deposits at the central bank in the some currency than the deposits | Yes, over a one-month maintenance period | Two times the market deposits rate for both local and foreign currency deposits, and two times the 30-day LIBOR for US$ deposits if it is above market rate.. | |
| Guatemala | 31 percent of deposits | 31 percent of deposits | No | 0.6 percent | Deposits | Cash in vault and deposits at the central bank in the some currency than the deposits | No | n.a | Monetary policy |
| Honduras | 6 percent of deposits | 12 percent of deposits | No | Only compulsory investments are remunerated al V of the policy rate. | Deposits, expired term deposits, reduced capital contracts and savings stamps and others | Cash in vault, deposits at the central bank, and government bonds in the case of compulsory investment in local currency in the some currency than the deposits | Yes, over a two-week period | Penalties depend on the currency of denomination and the type of institution. | |
| Uruguay | 12 percent on private deposits; and 100 percent on government deposits | 9 percent on private deposits; and 100 percent on government deposits | No | 2 percent for local currency, 0.025% for US$ currency, 0.20 percent for Euros, and zero percent for government deposits | Deposits and external loans | Cash in vault, but only for deposits in local currency, deposits in the some currency than the deposits | Yes, but only for reserves on local currency. No more than 4 days of reserve deficit are allowed per month. | On local currency 0.007 percent over the reserve deficiency with a minimum penalty of US$40.000. For foreign currency, 0.003 percent over the daily deficiency. |
| Country | Local Currency | Foreign Currency | Marginal and Additional Rates | Remuneration | Eligible Liabilities | Compliance Assets | Averaging | Penalty | Purpose/ Last change |
|---|---|---|---|---|---|---|---|---|---|
| Paraguay | 18 percent sight | 25 percent sight deposits; and 16.5 percent time and saving deposits | No | Local currency: Interbank interest rate Foreign currency: one- month Libor | Deposits in local and foreign currency | Cash deposits at the central bank in the some currency than the deposits | No allowed | To be determined by the board of the central bank. | Monetary policy; reduced in 2009 to counter the crisis and increased back in 2011 to rein in credit growth. |
| Peru | Minimum 9 percent of deposits in local currency | Minimum 9 percent of deposits in foreign currency | Yes; 25 percent for deposit of residents in local currency, 120 percent for deposits of nonresidents in foreign currency; 55 percent for deposits in foreign currency. | Only the additional requirements are remunerated at the overnight rate less 1 percentage point for reserve in local currency and 60 percent of the one- month Libor for reserves in foreign currency. | Deposits, bonds, mortgages securities, subordinated debt and external borrowing with maturity less than two years. | Cash in vault, deposits at the central bank in the some currency than the deposits | Yes, over a one-month maintenance period | 1.5 times the lending rate in local and foreign currency over the average deficiency. | Monetary policy, macro-prudential tool. |
| Colombia | 11 percent sight deposits; 4.5 percent time deposits | Not applicable | Discontinued | No remunerated | Deposits | Cash in vault and deposits at the central bank | Yes; two weeks maintenance period. | 3.5 percent over the average deficiency | Monetary policy |
| Brazil | 43 percent sight deposits 20 percent time and saving deposits (except if allocated to fund rural loans) | Not applicable | Yes; 12 percent on sight and time deposits and 10 percent on savings deposits | Only reserves corresponding to saving deposits and additional requirements, at 6.17 percent. | Deposits and interbank loans, excluding government deposits in public banks | Cash in vault, deposits in the central bank, federal government bonds and notes. | Yes, only for demand deposits for a two-week maintenance period. | Policy rate+ 14 percent over the average deficiency | Monetary policy, incentivize credit, macro-prudential tool |
| Dominican Republic | 17 percent commercial | 20 percent Foreign | No | Reserves in local currency are | Deposits | Deposits at the central | Yes | 0.1 percent over the average | Monetary policy |
| banks; 12.5 percent credit and saving associations; 10 percent other financial associations. | Currency Deposits | unremunerated. Foreign currency reserves are remunerated at the Fed Fund rat minus 2 percentage points. | bank in the some currency than the deposits | deficiency | |||||
| Chile | 9 percent sight deposits, 3.6 percent time deposits, 0 percent saving deposits. | 9 percent sight deposits, 3.6 percent time deposits, 0 percent saving deposits. | No | No remunerated | Deposits and external borrowing. | Cash in vault and deposits at the central bank | Yes for over a one-month maintenance period | 2 times the 3- month interest rate over the average deficiency. | Monetary Policy |
| Bolivia | 2 percent for cash, 10 percent for deposits below one-year maturity and 12 percent for two- year maturity deposits | 2 percent for cash, 10 percent for deposits below one-year maturity and 12 percent for two-year maturity deposits | Additional 30 percent requirement for deposits in foreign currency | Reserves maintained as deposits at the central bank are no remunerated. If deposited in the “Liquid Asset Requirement Fund”, the reserves are invested by the central bank of behalf of institutions and remunerated at the market interest rates. | Deposits and securities in local and foreign currency, except government deposits, short-term external credit lines, prepaid warranty deposits of credit letters. | Cash in vault, deposits at the central bank in the some currency than the deposits | Yes, for a 14 days maintenance period | 2 percent over average deficiency or twice the lending rate. The penalty increase in 2 percentage points for any additional subsequent period of deficiency | Monetary Policy and incentivize credit (the total requirement could be reduce to 6 percent if the institution increase loans in local currency) |
| Costa Rica | 15 percent | 15 percent | No | No remunerated | Deposits and trusts | Deposits at the central bank (only those at the reserve account) in the some currency than the deposits | Yes, for a 15-day maintenance period | Interest rate of discount window over the reserve deficiency | Monetary policy |
| The Bahamas | 5 percent | No applicable | No | No remunerated | Deposits, including non-residents’ accounts. | Cash in vault, deposits at the central bank | Yes, over a one-month maintenance period | 2 times the annual discount rate for every day of reserve deficiency | |
| Argentina | 19 percent sight deposits, 0-14 percent time deposits depending on maturity | 20 percent sight deposits, 0-20 percent time deposits depending on maturity | No | No remunerated | Deposits | Cash in vault and deposits at the central bank in the some currency than the deposits | Yes, over a one-month maintenance period | Two times the market deposits rate for both local and foreign currency deposits, and two times the 30-day LIBOR for US$ deposits if it is above market rate.. | |
| Guatemala | 31 percent of deposits | 31 percent of deposits | No | 0.6 percent | Deposits | Cash in vault and deposits at the central bank in the some currency than the deposits | No | n.a | Monetary policy |
| Honduras | 6 percent of deposits | 12 percent of deposits | No | Only compulsory investments are remunerated al V of the policy rate. | Deposits, expired term deposits, reduced capital contracts and savings stamps and others | Cash in vault, deposits at the central bank, and government bonds in the case of compulsory investment in local currency in the some currency than the deposits | Yes, over a two-week period | Penalties depend on the currency of denomination and the type of institution. | |
| Uruguay | 12 percent on private deposits; and 100 percent on government deposits | 9 percent on private deposits; and 100 percent on government deposits | No | 2 percent for local currency, 0.025% for US$ currency, 0.20 percent for Euros, and zero percent for government deposits | Deposits and external loans | Cash in vault, but only for deposits in local currency, deposits in the some currency than the deposits | Yes, but only for reserves on local currency. No more than 4 days of reserve deficit are allowed per month. | On local currency 0.007 percent over the reserve deficiency with a minimum penalty of US$40.000. For foreign currency, 0.003 percent over the daily deficiency. |
Appendix II: The CBP’s Operational and Legal Framework
The CBP’s operational framework includes tools in line with conventional practices.
The framework includes:
Reserve requirements, that, for deposits in guaranies, stand at 18 percent for deposits up to 360 days and 0 percent for longer terms, and foreign currency-denominated deposits at 25 percent for deposits up to 360 days, 16.5 percent for time deposits, and 0 for deposits up to 540 days or above.
Instrumentos de Regulation Monetaria (IRM), which are sold to commercial banks in competitive and non-competitive auctions, at different maturities (from 14, 35, 61, 91, and 180 days) and can be redeemed at the CBP at anytime. Since 2008 banks are not allowed to transfer their IRM to third private non-banking parties.12 The “policy rate” is defined as the weighted average of the IRM issued for lower-maturity IRM.
A short-term collateralized liquidity facility, called Facilidad de Liquidez de Corto Plazo con Reporto de Instrumentos de Regulaciȑn Monetaria (FLIR). The FLIR, which was established during the crisis in 2009, works as a discount window where banks can obtain liquidity for up to 60 days on repo operations collateralized with public sector securities, with a 5 percent haircut applied on market prices. The FLIR has two stages. The first one, during which banks can borrow liquidity for up to 30 consecutive days at the penalty interest rate determined by the CBP. After 30 days the FLIR can be renewed for up to an additional 30 consecutive days with a 1 percentage point penalty over the interest rate of the first stage. The current interest rate for the FLIR is 9 percent, which implies a 1 percentage point penalty from the CBP’s 14-day rate of 8 percent.
Exchange rate interventions which are conducted in a very ad hoc basis in the spot market.
References
Agenor, P.R., Aizenman, J., and Hoffmaister A.W, “The Credit Crunch in East Asia: What Can Bank Excess Liquid Assets Tell Us? Journal of International Money and Finance, Vol. 23, pp. 27–49.
Agenor, P.R. and Aynaoui (2010), “Excess liquidity, bank pricing rules, and monetary policy”, Journal of Banking and Finance, Vol 34, pp. 923–933
Chailloux, A. and Hakura, D., 2009, “Systemic Liquidity Management in the U.A.E.: Issues and Options”, Working Paper 09/261, International Monetary Fund, Washington D.C.
Freedman, C., and I. Otker-Robe (2009), “Country Experiences with the Introduction and Implementation of Inflation Targeting,” Working Paper 09/161, International Monetary Fund, Washington DC.
Gray, S. (2011), “Central Bank Balance and Reserve Requirements”, Working Paper WP/11/36, International Monetary Fund, Washington D.C.
Mlachila, M. (2009), “Recurrent Financial Crisis: causes, Costs, and Consequences”, in Alejandro Santos (ed.), Paraguay: Addressing the Stagnation and Instability Trap, Internacional Monetary Fund, Washington, DC.
O’Brien, Yueh-Yun C., 2007 “Reserve Requirement Systems in OECD Countries”, Finance and Economics Discussion Series, Divisions of Research & Statistics and Monetary Affairs, Federal Reserve Board, Washington DC.
Saxegaard, M., “Excess Liquidity and Effectiveness of Monetary Policy: Evidence from Sub-Saharan Africa”, Working Paper WP/06/115, International Monetary Fund, Washington DC.
Teresa Dabán Sánchez. This paper draws on the Technical Note “Monetary Policy and Systemic Liquidity Management Issues” prepared by the author for the recently conducted Financial Sector Assessment Program mission.
Bank excess reserves are a feature shared by several developing countries, including Paraguay (Agenor and Aynaoui, 2010 and Gray, 2011).
Banks are allowed to use their legal reserves only after experiencing a 20 percent reduction in their deposits
In October 2008, the reduction focused on the ratios for sight deposits which fell by 2 percentage points in the case of deposits in guaranies and 5.5 percentage points for foreign currency deposits. In February 2009, the reduction focused on the requirements ratios for time deposits, which fell by 7 percentage points. The increase adopted in early 2011 rolled back the reductions in the ratios adopted in 2008.
This does not implies that banks in those countries do not hold liquidity buffers above and beyond the reserve requirements. They do, but in form of non-cash and remunerated liquid assets, such as government securities.
Bank failures were caused by several factors, including poor supervision and regulation, excessive risk taking, lending to related enterprises and, in some cases, outright fraud (see IMF Staff Country Reports No. 90/10 and No 01/88). Altogether, 13 banks and 35 financial companies were closed or liquidated during the period. See Mlachila (2009), Recurrent Financial Crisis: Causes, Costs, and Consequences.
Deposit insurance in Paraguay benefits deposits for amounts equivalent to 75 times the minimum wage or lower. See Appendix III for the detail breakdown of banking deposits in Paraguay by size, currency and holders.
Results as regards involuntary variables are very robust to changes in the sample period, while results on precautionary variables are very sensitive to the time span used in the estimations. In particular, we tested the impact of considering a breaking point in 2003, to reflect the end of the years of crises. To do so we followed two approaches, by estimating the equations in Table 8 for the two periods 1995-2003 and 2004-10, and by introducing dummies variables.
The issuance of Treasury bonds for sterilization purposes is successfully used in many countries where the fiscal situation does not require the existence of a public sector securities market (e.g. Australia and Singapore), or where sterilization costs are covered by the Treasury (e.g. Uganda and Tanzania). See Chailloux and Hakura (2009) for further details.
The design of a well-designed payments system is underway, which the authorities expect to implement in October 2011. At present, a Draft Law on Payments Systems is at Congress under discussion, and the authorities expect it be approval shortly.
In 2005 there were two types of IRM, theCartas de Compromiso (kept at the BCP and only transferable among banks) and theLetras de Regulacion Monetaria (transferable with third non-banking parties). After 2008, the BCP only issues Cartas de Compromiso, which in a generic way are called Instrumentos de Regulacion Monetaria (IRM).
