Growth Volatility in Paraguay : Sources, effects, and options (Vol. 1 of 2)

Report No. 78318-PY Growth Volatility in Paraguay Sources, Effects, & Options June 5, 2014 Argentina, Paraguay and Uruguay Country Management Unit Poverty Reduction and Economic Management Latin America and the Caribbean Region Document of the World Bank  CURRENCY AND EXCHANGE RATE (As of June 5, 2014) CURRENCY UNIT = GUARANIES US$1.00 = 4421 Guaraníes FISCAL YEAR January 1 – December 31 Abbreviations and Acronyms ASERCA Supports and services for MATBA Futures and options exchange in agricultural marketing (Apoyos y Buenos Aires, Argentina servicios a la comercialización (Mercado a termino de Buenos agropecuaria.) Aires S.A.) ARM Agricultural risk management MERCOSUR Southern common market (Mercado Común del Sur) AYII Area-based Yield Index Insurance MEF Ministry of Finance (Ministerio de Hacienda) AxC Contract based agriculture MICs Middle Income Countries (Agricultura por contrato) BCP Central Bank of Paraguay (Banco OECD Organization for Economic Central de Paraguay) Cooperation and Development CADENA Component Attention to Natural PHEFA Hemispheric Plan of Disasters in the Agricultural and Eradication of Foot and Mouth Fisheries (Componente Atencion a Disease Desastres Naturales en el Sector Agropecuario y Pesquero) CBOT Chicago Board to Trade PEMEX Mexico's state-owned petrol company (Petróleos Mexicanos) CSF Chile’s Copper Stabilization Fund PIT Personal income tax CIT Corporate income tax PPP Public-private partnership EPH National Household Survey RENAMU National registry of (Encuesta Permanente de Hogares) municipalities (Registro Nacional de Municipalidades) ii FMD Foot and mouth disease ROFEX Futures and options exchange in Rosario, Argentina (Mercado a termino de Rosario S.A. GAP Good agriculture practices SENACSA National Service of Animal's Quality and Health (Servicio Nacional de Calidad y Salud Animal) GDP Gross Domestic Product SPF Norway’s Stabilization State Petroleum Fund IBLIP Index-Based Livestock Insurance SPS Sanitary and Phytosanitary Program IMAGRO Agricultural corporate income tax VAT Value Added Tax DGEEC National Institute of Statistics WTO World Trade Organization (Dirección Nacional de Encuestas Estadísticas y Censos) ISC Selective Consumption Tax y-o-y Year-on-year (Impuesto Selectivo al Consumo) LAC Latin America and the Caribbean MICs Middle Income Countries Vice President: Jorge Familiar Calderon Country Director: Jesko S. Hentschel Sector Director: J. Humberto Lopez Sector Manager: Auguste Tano Kouame Sector Leader: Zafer Mustafaoglu Task Team Leaders: Friederike (Fritzi) Koehler-Geib iii  Contents Abbreviations and Acronyms ......................................................................................................... ii Executive Summary ........................................................................................................................ 2 Chapter 1: The Sources of Volatility in Paraguay ..................................................................... 5 1.1. Stylized facts ................................................................................................................. 6 1.2. Sources of volatility .................................................................................................... 11 1.3. The role of factor markets ........................................................................................... 18 Chapter 2: The effects of growth volatility in Paraguay with a focus on volatility originating in the agricultural sector ................................................................................. 20 2.1 Propagation of shocks within agriculture ................................................................... 20 2.2 The impact of volatility originating in the agricultural sector on other sectors .......... 24 2.3 The impact of volatility originating in the agricultural sector on macroeconomic aggregates .................................................................................................................................. 28 Chapter 3: Policy options for the management of growth volatility in Paraguay ................ 32 3.1 The macroeconomic toolbox to address growth volatility.......................................... 33 3.2 The agricultural risk management toolbox ................................................................. 38 3.3 Combining the macroeconomic and the agricultural risk management toolbox ........ 55 List of Figures Chapter 1 Figure 1.1: Real GDP growth ....................................................................................................................... 6 Figure 1.2: Volatility over time, Paraguay in regional comparison .............................................................. 7 Figure 1.3: GDP—breakpoints of volatility of quarterly y-o-y GDP growth using Inclan, Tiao (1994)...... 8 Figure 1.4: Agricultural GDP—breakpoints of volatility of quarterly y-o-y GDP growth using Inclan, Tiao (1994).................................................................................................................................................... 8 Figure 1.5: Share of agriculture in GDP ....................................................................................................... 9 Figure 1.6: Growth volatility by economic sector ........................................................................................ 9 Figure 1. 7: Volatility of rainfall ................................................................................................................. 10 Figure 1. 8: Rainfall and agriculture GDP .................................................................................................. 10 Figure 1. 9 Exports by product.................................................................................................................... 11 Figure 1.10: Correlation between trade balance and world interest rate ..................................................... 13 Figure 1.11: Impulse response of trade balance to a shock to the world interest rate ................................. 13 Figure 1.12: Correlation between TOT and world interest rate .................................................................. 14 iv Figure 1.14: Business cycle fluctuations in Paraguay—Government investment versus GDP .................. 15 Figure 1.15:Contribution of public and private demand components to real GDP growth ........................ 16 List of Figures Chapter 2 Figure 2.1: Wavelet analysis of rainfall and agricultural GDP ................................................................... 21 Figure 2.2: Impulse response functions linking Paraguay’s agricultural GDP to world............................. 23 Figure 2.3: Impulse response functions linking Paraguay’s agricultural GDP to the construction sector .. 25 Figure 2.4: Impulse response functions linking Paraguay’s agricultural GDP to the services sector......... 26 Figure 2.5: Wavelet analysis of agriculture and non-agricultural GDP ...................................................... 28 Figure 2.6: Wavelet analysis of agriculture and private consumption ........................................................ 29 Figure 2.7: Wavelet analysis of non-agriculture and private consumption................................................. 29 List of Figures Chapter 3 Figure 3.1: The World Bank Agricultural Risk Management Framework ................................................. 39 List of Tables Chapter 1 Table 1.1: Export by destination ................................................................................................................. 11 Table 1.2 Variance decomposition of GDP volatility ................................................................................. 12 Table 1.3 Correlations across variables ...................................................................................................... 14 Table 1.4: Variance Decomposition of Agricultural and Non-Agricultural GDP volatility ....................... 17 List of Tables Chapter 3 Table 3.1: Instruments for Managing Production Risk ............................................................................... 39 Table 3.2 Colombia’s study case. Benefits, Challenges and Considerations for Paraguay ........................ 41 Table 3.3: Mexico’s study case. Benefits, challenges and considerations for Paraguay ............................ 44 Table 3.4: Malawi’s study case. Benefits, challenges and considerations for Paraguay ........................... 45 Table 3.5: Mongolia’s study case. Benefits, challenges and considerations for Paraguay ......................... 47 Table 3.6: Instruments for Managing Market Risk ..................................................................................... 48 Table 3.7: Peru’s study case. Benefits, Challenges and considerations for Paraguay ................................ 49 Table 3.8: Subsidy Components, AxC Program, ASERCA ....................................................................... 52 Table 3.9: Mexico’s study case. Benefits, Challenges and considerations for Paraguay ............................ 52 Table 3.10 : Comparison CBOT, MATBA, & ROFEX.............................................................................. 53 Table 3.11: Argentina’s study case. Benefits, challenges and considerations for Paraguay ....................... 54 List of Annexes v . Annex 1.1: Volatility over time, international comparison......................................................................... 63 Annex 1.2: Volatility breaks of macroeconomic variables in Paraguay ..................................................... 64 Annex 1.3: Graphs on volatility breakpoints Inclan Tiao (1994) by variable............................................. 66 Annex 1.4: Sectoral GDP correlations ........................................................................................................ 70 Annex 3.1: Traditional measures for agricultural risk management ........................................................... 72 Annex 3.2: Insurance products.................................................................................................................... 72 vi This report was prepared by a team led by Friederike (Fritzi) Koehler-Geib (LCSPE) under the overall supervision and guidance of Zafer Mustafaoglu (Lead Economist and Sector Leader, LCSPR), Auguste T. Kouame (Sector Manager, LCSPE), J. Humberto Lopez (Sector Director, LCSPR), Rodrigo A. Chaves (former Sector Director, LCSPR) and C. Penelope Brook (Country Director, LCC7C). The peer reviewers were Aristomene Varoudakis (Advisor, DECOS), Cesar Calderon (Senior Economist, DECWD), Julie Dana (Lead Financial Officer, FABLO), and Norbert Fiess (Principal Economist/Credit Risk Head, CFRCR). The core team included Elida Caballero Cabrera, Diana Lachy, Rei Odawara, Guillermo Cabral, Jorge Araujo, Miriam Beatriz Villarroel, Marcelo Echague, Patricia Chacon Holt, Peter Siegenthaler, Silvia Gulino (all LCSPE), Dante Mossi, (Country Manager, LCCPY), Gloria Dure, Rosa Arestivo de Cuentas Zavala, (all LCCPY), and Rossana Polastri (former Country Manager, LCCPY). Inputs and background papers were also received from Sophie Storm Theis (LCSSD), Diego Arias Carballo (LCSAR), Hakan Berument, Julio Ramirez, Viktoria Hnatkovska (all consultants), Andres Lajer Baron, Carolina Saizar, Hannah Nielsen, Nathalie Picarelli, Pia Maria Zanetti, and Sona Varma (all LCSPE), Oscar Calvo-Gonzalez (LCSPR), Julian Lampietti (LCSSD), David Gould (SARCE). Comments and inputs were also received from many colleagues working in the Paraguay country team, including Andrew Follmer, Carla Cutolo, Elena Feeney, Mariela Alvarez, Sabine Hader (all LCC7C). The team is thankful for the excellent collaboration with the Ministry of Finance, in particular with the vice ministry of economics including the departments of economic studies, Macro Fiscal Policies, and Debt Policy. 1 Paraguay’s real GDP growth has been one of the most volatile in the region in recent years. Between 2000 and 2011, real GDP growth in Paraguay fluctuated by 5.5 percentage points, exceeding the volatility of most Latin American peer countries. This was not always the case. During the period 1960-2011, growth volatility in Paraguay was lower than in other countries in the region. It is too early to tell whether high volatility in Paraguay is temporary or permanent, even though some structural changes as for example the increase in the weight of the agricultural sector in GDP are in line with the idea that volatility is there to stay. The high level of volatility is concerning because of the significant costs associated with it in terms of welfare, economic growth, and equality. For developing countries, macroeconomic volatility, summarized by output volatility, is reflected disproportionately in consumption volatility, and welfare gains from reducing consumption volatility can be substantial (Loayza, Ranciere, Serven, and Ventura (2007)). No less important is the negative impact of volatility on economic growth. The impact arises through a decrease in productivity and various forms of uncertainty such as economic, political, policy-related, as well as a tightening of binding investment constraints.1 The negative link between macroeconomic volatility and equality has also been established in the literature.2 Designing policies that help mitigate the impact of shocks to the economy and that help increase the country’s resilience is particularly relevant in this light, also because Paraguay still has a low per capita income compared to its neighbors and continues to suffer from a high degree of inequity and poverty. The high volatility of GDP growth has coincided with a volatile macroeconomic environment. A large number of relevant economic variables and variables with economic significance have shown high levels of volatility in recent years, including the world interest rate, Paraguay’s nominal exchange rate, its current account balance, public consumption and investment, credit to the private sector, agricultural GDP, rainfall and, soy prices. While the agricultural sector was particularly affected, preceded by an increase in the volatility of soy prices and rainfall, non-agricultural GDP actually registered a decrease in volatility. External shocks explain over 50 percent of Paraguay’s GDP growth volatility. A key factor behind volatility in Paraguay is the strong dependence on agriculture and its concentration on few products and few export destinations, both of which have increased over time. Of the external shocks, foreign demand for Paraguayan output accounts for about 30 percent of GDP volatility, the world interest rate for 20 percent, and terms of trade for 3 percent. The impact of the world interest rate runs through its impact on portfolio reallocation, commodity prices, and economic conditions of main trading partners. Fluctuations in commodity prices and world real interest rates have hit all countries in the region and the world. However, their output response reflects the interaction of these shocks with country-specific conditions that range from the strong dependence on a few goods and services or a narrow tax base and economic policies. 1 A large body of literature has addressed this topic from various perspectives Acemoglu et al (2003), Aizenman and Pinto (2005), Berument, Dincer, and Mustafaoglu (2011), Ramey and Ramey (1995) and Wolf (2005). 2 See for example Breen and Garcia-Penalosa (2004), Garcia-Penalosa and Turnovsky (2004) or Huang, Fang, and Miller (2012). 2 Domestic variables explain the remaining share. 25 percent of stemming from shocks to real GDP, 15 percent from shocks to investment, and 3 percent from pro-cyclical fiscal and monetary policy. Pro-cyclicality is defined as a positive response of government spending to an exogenous expansionary business cycle shock. In developing countries it is usually linked to limited access to credit in downturns, to lax fiscal stances in good times, to and burdensome bureaucratic processes. While Paraguay shares this pattern there is an indication that the fiscal stance was counter-cyclical during the contractions in 2009 and 2012. In contrast to overall GDP volatility, fluctuations of agricultural GDP originate to a large extent from domestic shocks, with weather related shocks to agricultural output itself accounting for more than a third. Despite a significant decrease, persistent rigidities in factor markets and limited mobility across sectors reduce the economy’s capacity to buffer shocks and exacerbate business cycle fluctuations and hence volatility. While labor market distortions have declined, firms’ access to credit have improved, and agricultural efficiency has increased, important challenges remain which suggest that shocks hitting Paraguay may rather be exacerbated than buffered. In particular, labor and capital returns between agriculture and non-agriculture remain large, suggesting limited factor mobility across sectors, financing constraints facing households have remained pronounced and time-varying and the efficiency in the non-agricultural sector has shown no signs of improvement, to the contrary has been deteriorating. These remaining frictions reduce efficiency of the Paraguayan economy and prevent its capacity to buffer shocks that hit the economy. Growth volatility impacts Paraguay’s economy in various ways, it impacts : i) the agricultural sector; ii) other sectors; and iii) macroeconomic aggregates such as investment, tax revenues, or poverty and equity. The main sources of volatility in the agricultural sector can be categorized into shocks to production and shocks to markets. Shocks to production include variations in rainfall, investment levels, and disease outbreaks (e.g., foot and mouth disease). Shocks to markets include commodity price variations, the closing of markets in the case of disease outbreaks, and fluctuations of prices of imported inputs like fertilizers and pesticides. Overall, market participants in Paraguay report that a lack of information and knowledge on the patterns and impacts of volatility on the economy is the biggest challenge to operating within this environment. Within the agricultural sector sources of volatility manifest themselves through shocks that impact the level of infrastructure and R&D investment; cause payment delays; and trigger the use of diversification strategies. In terms of other sectors, the volatility of agricultural GDP mainly affects those economic activities that provide inputs such as machinery or storage and transport services, but it also affects the financial services and construction sectors. In terms of macroeconomic aggregates, the exchange rate and levels of employment fluctuate as a consequence of shifts in agricultural exports. There is some indication that private consumption plays a role in propagating the impact of agricultural GDP through the economy, impacting non-agricultural GDP. Investment levels are lower, fiscal revenues are indirectly affected, and the reduction of poverty and inequity are generally slower than in countries with lower levels of volatility. In terms of managing volatility, it is important to develop a comprehensive macroeconomic risk management framework that takes all different sources of volatility and risks into account. Sources of volatility are interrelated and taking a broader perspective allows finding optimal ways to manage observed volatility and risks. Any policy option needs to be assessed in 3 terms of its fiscal implications; be it in terms of its effects on sustainability, on redistribution, or on potential contingent liabilities. Policy options are presented as: i) a macroeconomic tool set; and ii) an agricultural risk management tool set, which need to be aligned with one another in an overall framework. Macroeconomic policy options include the development of a strategy on the role of agriculture in the economy and its structure; policies that render factor markets flexible; and fiscal policies such as the introduction of fiscal rules and stabilization funds. The agricultural risk management tool box is designed to address the shocks to production and to markets specific to the agricultural sector. First, four case studies are presented on new tools and approaches to mitigate, cope with, and transfer agricultural production risks: i) building animal health capacity to prevent foot and mouth disease in Colombia; ii) introducing weather derivatives based on a rainfall index for severe drought in Malawi; iii) establishing a weather contingency fund for the agricultural sector (CADENA) in Mexico; and iv) implementing an index-based livestock insurance project in Mongolia. Second, three case studies provide examples of measures to mitigate and transfer agricultural market risks: i) developing the asparagus market in Peru; ii) introducing subsidies for commodity price hedging contracts in Mexico; and iii) introducing agricultural commodity exchanges in Argentina. While all case studies have been selected based on their relevance for Paraguay, a careful assessment of their applicability to Paraguay would be required as part of an overall assessment of agricultural risks. Government and the World Bank have been engaging in a dialogue on this topic through the preparation of this study and with a joint agricultural risk management assessment. 4  Chapter 1: The Sources of Volatility in Paraguay Paraguay’s real GDP growth has been one of the most volatile in the region in recent years. This was not always the case. During the period 1960-2011, growth volatility in Paraguay was lower than in other countries in the region and many of these countries managed to reduce volatility (see Table Annex 1). The high level of volatility in Paraguay is concerning because of the significant costs associated with it in terms of welfare, economic growth, and equality. 3 Designing policies that help mitigate the impact of shocks to the economy and that increase the country’s resilience is particularly relevant in this light, and also because Paraguay still has a low per capita income compared to its neighbors and suffers from a persistently high degree of inequality and poverty. The purpose of the current study is to contribute to a deeper understanding of growth volatility in Paraguay and to provide an input into the discussion on how to better manage it. In particular, the study will ask three questions: i) what are the sources of volatility in Paraguay? ii) How does growth volatility, in particular that arising from the strong dependence on the agricultural sector, impact the rest of the economy? iii) What are optimal policies for managing the types of volatility observed in Paraguay? This study’s quantitative analyses mainly rely on quarterly data available since the first quarter of 1994 (earliest available) and allow insights primarily into business cycle volatility. Wherever possible the study also shows a longer-term perspective based on yearly data. However, data restrictions do not allow for a rich analysis of these long-term volatility trends for Paraguay, and from a policy perspective business cycle volatility appears more relevant. The current study seeks to provide answers to the questions identified in three chapters: i) the first chapter covers the sources of volatility. It provides a description of stylized facts and an analysis of the sources based on structural vector autoregression (SVAR) analysis and a business cycle accounting exercise. ii) The second chapter addresses the effects of volatility with a particular focus on volatility arising from a strong dependence on the agricultural sector. This chapter is based on a qualitative analysis relying on 25 structured interviews with key players in the Paraguayan economy, as well as on a quantitative approach based on VAR analysis and a wavelet approach. iii) The third chapter presents policy options for managing volatility. In particular, it provides an overview and a discussion of cases of other Governments that have successfully managed volatility, similar to that observed in Paraguay. 3 See for example Loayza, Ranciere, Serven, and Ventura (2007), Athanasoulis and van Wincoop (2000), World Bank (2000) on the impact of volatility on welfare, Hnatkovska and Loayza (2005) and Calderon and Schmitt- Hebbel (2003) and Berument, Dincer, and Mustafaoglu (2011) on the growth impact, and Breen and Garcia Penalosa (2004), Garcia-Penalosa and Turnovsky (2004), or Huang, Fang and Miller (2012) for the impact on equality. 5  1.1. Stylized facts Paraguay’s GDP growth has been one of the most volatile in the region. The high volatility of GDP growth has coincided with a volatile macroeconomic environment. A large number of relevant economic variables have shown high levels of volatility in recent years, including the world interest rate, 4 Paraguay’s nominal exchange rate, its current account balance, public consumption and investment, credit to the private sector, agricultural GDP, rainfall and soy prices. While the agricultural sector was particularly affected, preceded by the increase in the volatility of soy prices and rainfall, non-agricultural GDP actually registered a decrease in volatility. Figure 1.1: Real GDP growth 14 12 10 8 6 Percent 4 2 0 -2 -4 -6 1951 1953 1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013* Source: Central Bank of Paraguay. Note:*Projection. Real growth in Paraguay has been more volatile than most other countries in Latin America and other regions in the past decade. One commonly used definition of volatility of economic growth is the standard deviation of real GDP growth rates or of the cyclical components of GDP. 5 According to these measures, Paraguay’s growth volatility was much lower than that of Latin American peer countries in the period from 1960 to 2000, whereas growth has been more volatile than most other countries in Latin America in the last decade. Real GDP growth varied by 4 percentage points in the period from 1960 to 2000, in the past decade it had a standard deviation of 5.5 (Table Annex 1.1). In contrast, many other countries in the region managed to reduce volatility, explaining the drop in the regional mean and median from 4.7 and 4.5 to 3.1 and 2.8 respectively. Paraguay’s MERCOSUR neighbor Brazil managed to reduce the variation of its real GDP from a standard deviation of 4.5 from the 1960 to 2000 period to 2.3 in the past decade, a pattern that the country shares with Bolivia, Chile, Colombia, Ecuador, Mexico and Peru. Also, the East Asia and Pacific region, Middle East and North 4 Measured by the U.S. 3-month treasury bill rate. 5 See for example Loayza, Ranciere, Serven, and Ventura (2007), Perry and Fiess (2006), Alouini and Hubert (2010), or Furceri and Karras (2007). 6 Africa, Sub-Saharan Africa, South Asia, and Paraguay’s peer group of lower middle income countries have all seen a reduction in growth volatility in the past decade (Figure 1.2). While countries in Europe and Central Asia, OECD members, and higher middle income countries share the trend of higher volatility in the period from 2000 to 2011, Paraguay’s growth volatil ity exceeds the ones in these regions. In fact, Venezuela, Argentina, and Uruguay are the only three countries in the period from 2000 to 2011 that have observed higher levels of growth volatility than that experienced by Paraguay. Figure 1.2: Volatility over time, Paraguay in regional comparison Panel a: Standard deviation (GDP growth)—1960-1999 Panel b: Standard deviation (GDP gap)— 1960-1999 Paraguay Paraguay LAC mean (excl. Paraguay) LAC mean (excl. Paraguay) LAC median (excl. Paraguay) LAC median (excl. Paraguay) East Asia & Pacific* East Asia & Pacific (all income… Europe & Central Asia* Europe & Central Asia* Middle East & North Africa* Middle East & North Africa* South Asia South Asia Sub-Saharan Africa* Sub-Saharan Africa* Lower middle income Lower middle income Upper middle income Upper middle income OECD members OECD members 0 1 2 3 4 5 6 0 1 2 3 4 5 6 Panel c: Standard deviation (GDP growth)—2000-2011 Panel d: Standard deviation (GDP gap)— 2000-2011 Paraguay Paraguay LAC mean (excl. Paraguay) LAC mean (excl. Paraguay) LAC median (excl. Paraguay) LAC median (excl. Paraguay) East Asia & Pacific* East Asia & Pacific (all income… Europe & Central Asia* Europe & Central Asia* Middle East & North Africa* Middle East & North Africa* South Asia South Asia Sub-Saharan Africa* Sub-Saharan Africa* Lower middle income Lower middle income Upper middle income Upper middle income OECD members OECD members 0 1 2 3 4 5 6 0 1 2 3 4 5 6 Source: World Development Indicators, and Central Bank of Paraguay, staff calculations. Note:*All income levels. While Paraguay has experienced clusters of high volatility in the past, fluctuations never quite reached the levels observed in recent years. In 2009, the economy contracted by 4 percent, the worst outcome in Paraguay’s recorded history, and rebounded to 13.1 percent in 2010, the best outcome ever observed. Shortly thereafter, Paraguay has been experiencing a similar pattern of extreme fluctuation, with a -1.2 contraction in 2012 and a projected recovery of 10.5 percent in 2013. This translates into extreme year-on-year (y-o-y) changes in terms of percentage points of growth: drops of 10 and 9 percentage points in 2009 and 2011 respectively, and increases of 17 and 14 percentage points in 2012 and 2013 respectively. The period in the past that comes closest to these extreme fluctuations is the end of the 1970s and the beginning of 7 the 1980s, which coincides with the construction and then completion of the Itaipú dam and hydro power plant (from 1975 to 1982). Growth dropped from 9.2 percent in 1981 to -1.4 in 1982. This period represented a transition between a period of high growth due to the impulse of the Brazilian Paraguayan construction project, approximately 4 times the size of Paraguay’s GDP at the time, and the period of low growth thereafter. Yet, GDP did not oscillate between sharp contractions versus fast expansions from year to year. Figure 1.3: GDP—breakpoints of volatility Figure 1.4: Agricultural GDP—breakpoints of quarterly y-o-y GDP growth using Inclan, of volatility of quarterly y-o-y GDP growth Tiao (1994) using Inclan, Tiao (1994) 8 25 Stddev. of agriculture GDP growth 12 7 36 20 St ddev. of GDP growth 6 26 7 Percent growth rate 5 15 16 Percent GDP growth 4 2 6 3 10 -4 2 -3 5 1 -14 -8 0 -24 0 Jan/ 95 Jan/ 96 Jan/ 97 Jan/ 98 Jan/ 99 Jan/ 00 Jan/ 01 Jan/ 02 Jan/ 03 Jan/ 04 Jan/ 05 Jan/ 06 Jan/ 07 Jan/ 08 Jan/ 09 Jan/ 10 Jan/ 11 Jan/ 12 Jan/ 95 Jan/ 96 Jan/ 97 Jan/ 98 Jan/ 99 Jan/ 00 Jan/ 01 Jan/ 02 Jan/ 03 Jan/ 04 Jan/ 05 Jan/ 06 Jan/ 07 Jan/ 08 Jan/ 09 Jan/ 10 Jan/ 11 Jan/ 12 GDP breaks stddev. Agriculture GDP breaks stddev. Source: World Bank calculations based on Central Source: World Bank calculations based on Central Bank, Paraguay. Bank, Paraguay. Three other past periods displayed significant y-o-y fluctuations, however not at the levels observed in the most recent past. First, between 1955 and 1961 GDP growth rates dropped twice by over 5 percentage points and also expanded twice by over 5 percentage points at the beginning of the Stroessner dictatorship (1954–1989) and after a period of high inflation (Cubas, Escobar, Franco, Olmedo, and Smith (2011)). Second, between 1966 and 1968, political uncertainties and some elements of democratization contributed to strong variations in economic growth. Between 1995 and 2002 Paraguay underwent a period of recurrent financial crises that went hand in hand with substantial fluctuations in growth. The biggest y-o-y change occurred between 1995 and 1996 when growth dropped from 6.8 percent in the first year to only 1.5 in the latter. A closer look at quarterly y-o-y real growth rates reveals that high volatility is a very recent phenomenon, with a significant increase in the fourth quarter of 2008 (Figure 1.3). The study relies on Inclan and Tiao (1994) to identify structural breaks in the volatility of the analyzed time series, a method that performs well with the type of data used.6 Before the break, the standard deviation of quarterly y-o-y growth since the first quarter of 1994 amounts to 4 percentage points, after the break it shifts up to 7 percentage points. 6 The Inclan-Tiao test is characterized by its simplicity and independence from estimated long-run variance, which make the test robust to time period selection, and it also performs well with shorter time series compared to other tests such as Kokoszka-Leipus (2000) or Quandt (1960) and Andrews (1993). 8 The dynamic of GDP growth mirrors an increase in the volatility of agricultural GDP in the fourth quarter of 2008, a sector whose weight has increased over time. The Inclan Tiao test applied to agricultural GDP growth reveals a break point in volatility also in the fourth quarter of 2008 (Figure 1.4). Before the fourth quarter of 2008, the standard deviation of real agricultural GDP growth was 6 percentage points, afterwards it shot up to 22. While the share of agriculture in total GDP amounted to about 12 percent in the second half of the 1990s it increased significantly to over 18 percent in 2010 and 2011 (Figure 1.5). Figure 1.5: Share of agriculture in GDP Figure 1.6: Growth volatility by economic sector 25 25 20-qurater standard deviation of quarterly y-o-y growth 20 20 Percent of GDP 15 15 10 10 5 5 0 0 2002:I 2005:I 2008:I 2011:I 2000:III 2003:III 2006:III 2009:III 2012:III 1999:IV 2001:II 2002:IV 2004:II 2005:IV 2007:II 2008:IV 2010:II 2011:IV 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Paraguay LAC Primary Secondary Tertiary Source: WDI. Source: BCP. An increase in rainfall volatility preceded the breakpoint of agricultural GDP volatility, this is relevant given that only 2 percent of agricultural surfaces are cultivated using irrigation. The instability of rainfall increased in 2007 and remained high until 2013 (Figure 1.7).7 Given the agricultural production cycle in Paraguay, rainfall from December the previous year and January of the current year are particularly relevant for the harvest. Figure 1.6 displays annual data aggregating rain data from those months. The correlation between the cyclical component of rainfall and agriculture GDP is 0.69, showing the high dependence of agriculture production on weather conditions (see Figure 1.8.). Despite the high and increasing correlation between the climate cycle and agricultural production, only 2 percent of the surface used for agricultural is irrigated.8 7 No formal volatility break test was applied because the length of the yearly time series does not allow for it. 8 United Nations Development Program (2006). 9 Figure 1. 7: Volatility of rainfall Figure 1. 8: Rainfall and agriculture GDP 1800 2500 15 1600 2000 10 1400 1500 1000 5 Percent of GDP 1200 Standard deviation Milimeters 1000 500 0 0 800 -500 -5 600 -1000 -10 400 -1500 200 -15 -2000 0 -2500 -20 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 Rain (cycle) Agriculture cycle (cycle as a share of GDP, RHS) Source: DINAC and BCP. Rain is considered for the months of December from the previous year and January of the same year because of their importance in the main agricultural soy product. All weather stations of the Eastern part of Paraguay are considered. Paraguay’s agricultural sector concentrates on a few products and export destinations. Soy and beef alone made up an average of 34 percent of total exports over the past 5 years; and exports to Brazil and Argentina alone reached almost 50 percent of total exports in the period since 2008 (Figure 1.9 and Table 1.1). Overall, agricultural GDP in Paraguay is much more volatile than the aggregate or than non-agricultural GDP. Quarterly y-o-y agricultural GDP growth has fluctuated by 12 percentage points since the first quarter of 1994, contrasting with overall GDP that showed a standard deviation of 4.7 percentage points, and with non-agricultural GDP which varied by 4 percentage points (Figure 1.5). In particular, the wedge between agricultural and non-agricultural GDP volatility has increased in recent years because non-agricultural GDP is one of the few macroeconomic aggregates that have not become more volatile but have remained relatively stable. The Inclan Tiao test did not identify any structural break in the volatility of this variable. Despite these facts the growth of agricultural GDP has exceeded the growth of aggregate GDP and of non-agricultural GDP. Quarterly y-o-y growth of agricultural GDP amounted to 4 percent compared to 2.7 percent of aggregate GDP and 2.5 percent of non-agricultural GDP (Table Annex 1.2). Paraguay’s macro-economic environment has become more volatile with shifts in the volatility of soy prices and world interest rates preceding those of other variables. Most macroeconomic variables underwent an increase in volatility during 2007 and 2008, some even increased before that (Table Annex 1.2). In light of the potential interaction between different economic aggregates, an interesting sequence is that soy prices became more volatile in the third quarter of 2003, and world real interest rates became more volatile in the fourth quarter of 2007, and that these increases preceded the increases of volatility in the nominal exchange rate in the first quarter of 2008; of public investment in the second quarter of 2008; of overall GDP and agricultural GDP in the fourth quarter of 2008; and of public consumption in the first quarter of 2009. Also the increase in the volatility of the current account balance in the first quarter of 2007 was preceded by an increase in volatility in soy prices. It is also important to note that there were additional increases, like that of inflation in the second quarter of 1995; public consumption in 10 the second quarter of 2000; or credit to the private sector in the fourth quarter of 2002. This illustrates the need to carefully assess causalities and to take into account relevant additional factors in an analytical assessment when searching for the sources of growth volatility in Paraguay. Such analysis is provided in the second section of this chapter. Figure 1. 9 Exports by product Table 1.1: Export by destination 100 50 Percent of total exports 80 40 60 30 Percent of GDP Total Continental Rest of the Argentina Brazil Uruguay MERCOSUR China World 40 20 average 13 47 6 66 1 33 since 2000 20 10 average 11 37 1 49 1 50 since 2008 0 0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Other Grains Beef Soy beans Electricity Total in percent of GDP(RHS) Source: Central Bank, Paraguay. There are only a few exceptions to this increase in volatility; the main exception is tax revenues. Variables whose volatility has not changed over the period since the first quarter of 1994 include total and private investment, private consumption, oil prices, non-agricultural GDP, Paraguay’s real interest rate, the real exchange rate, total Government revenues, beef prices, and terms of trade. The only variable that displayed a decrease in volatility that was not followed by a later increase is tax revenues. The single breakpoint is the third quarter of 2004 coinciding with the 2004 tax policy and administration reform in Paraguay. While it is too early to tell whether high volatility in Paraguay is temporary or permanent, and there is no indication of a causal link explaining the increase in volatility, some structural changes are in line with the idea that volatility is there to stay. In particular, the increased weight of the agricultural sector and its concentration on a few products exposes the economy more to fluctuations in rainfall which seem to be increasing in the context of climatic change. It also renders the economy vulnerable to fluctuations in commodity prices. Yet, it is unclear whether commodity prices will remain as volatile as they have displayed clusters of heightened volatility in the past.9 1.2. Sources of volatility External shocks are the main source of growth volatility in Paraguay, with foreign demand explaining around 30 percent of GDP volatility, 20 percent is explained by the world interest rate, and 3 percent comes from terms of trade. Domestic variables also contribute to the volatility with 25 percent stemming from shocks to real GDP, 15 percent from shocks to 9 See Calvo-Gonzalez, Shankar, and Trezzi (2010). 11 investment, and 3 percent from pro-cyclical fiscal and monetary policy. Pro-cyclicality is defined as a positive response of government spending to an exogenous expansionary business cycle shock. In developing countries it is usually linked to limited access to credit in downturns, lax fiscal stances in good times, and burdensome bureaucratic processes. While Paraguay shares this pattern there is indication that the fiscal stance was counter-cyclical during the contractions in 2009 and 2012. In contrast to overall GDP volatility, fluctuations of agricultural GDP originate to a large extent from domestic shocks, with weather related shocks to agricultural output itself accounting for more than 50 percent. About half of Paraguay’s GDP growth volatility stems from external shocks while the other half originates in domestic shocks. These findings emerge from a variance decomposition based on a structural VAR analysis guided by economic theory for model specification. Hnatkovska and Koehler-Geib (2013) presents a detailed description of the model specification and a rational for variable selection. The variables included in the VAR are chosen to capture those factors identified in the literature as important determinants of business cycles in developing countries. External variables are: world interest rate (as measured by the US 3 month treasury bill rate), and foreign demand (as measured by the trade weighted GDP of Argentina, Brazil, Chile and Uruguay, the remaining export share is allocated to the US and its GDP).10 Domestic variables are: real GDP, total investment, Government consumption as measure of fiscal policy, domestic short-term interest rate as a measure of monetary policy, and the real trade balance to GDP ratio. The weights of the different variables reflect the variance decomposition at 12 quarters which is where the percentages for most cases stabilize and is the midpoint of the of the range between 6 and 32 quarters that researchers typically define as frequencies for business cycle movements (Baxter and King (1999)) (Table 1.2).11 Quarters TOT r_us Log(yf) Log(gc) Log(inv) Tby r Log(y) External Domestic 1 0.043 0.058 0.071 0.004 0.248 0.047 0.022 0.506 0.173 0.827 4 0.026 0.143 0.308 0.011 0.163 0.062 0.012 0.274 0.477 0.523 8 0.024 0.191 0.300 0.019 0.148 0.058 0.011 0.248 0.515 0.485 12 0.025 0.195 0.299 0.020 0.147 0.058 0.011 0.247 0.518 0.482 16 0.025 0.195 0.299 0.020 0.147 0.058 0.011 0.247 0.518 0.482 20 0.025 0.195 0.299 0.020 0.147 0.058 0.011 0.247 0.518 0.482 Source: Hnatkovska and Koehler-Geib (2013) 10 Kose, Otrok, and Prasad (2012) and Kose, Otrok, and Whiteman (2008) find a stronger role of domestic factors. This difference appears to mainly stem from a sample period that only goes up until 2005 only. Raddatz (2007) also finds a predominant role of external variables in a paper that examines whether the differences in output volatility between Latin America and other regions result from volatility of external shocks or from a more pronounced response to these. Podpiera and Tulin (2012), focusing on the role of financial external variables find a relevant role of external factors. 11 Table 1.2 presents the results of the variance decomposition based on the estimated SVAR system at different horizons. The variance decomposition allows quantifying the contribution of each shock to the variance of forecasting error for output. 12  The most important external factor, foreign demand, explains around 30 percent of Paraguayan growth volatility, and terms of trade another 3 percent, a fact that may be related to Paraguay’s agricultural sector. The significance of these shocks may arise from the sector’s weight in the economy and its concentration on a few export products (see Section 1.1). As argued in Hnatkovska and Koehler-Geib (2013) concentration in terms of products and export destinations renders countries more vulnerable to terms of trade and foreign demand shocks. Figure 1.10: Correlation between trade balance Figure 1.11: Impulse response of trade balance and world interest rate to a shock to the world interest rate .15 .06 r_us shock .004 .04 .1 Reat int. rate, world .002 .02 TB/GDP .05 0 0 0 -.02 -.002 -.04 -.05 0 5 10 15 20 quarters 1995q1 2000q1 2005q1 2010q1 95% CI IRF TB/GDP Reat int. rate, world Graphs by irfname, impulse variable, and response variable Source: Hnatkovska and Koehler-Geib (2013) The second most important external factor, the world interest rate, explains around 20 percent of Paraguayan growth volatility through its impact on portfolio reallocation, on commodity prices, the economic condition of trade partners, and remittances. At a first glance it may appear surprising that GDP growth of a country which is not heavily represented in international capital markets varies this much with the international interest rate. Yet, there are mainly four channels that explain the link. First, when international interest rates go up, foreign investors may shift out of Paraguayan assets inducing a contraction or even a reversal of capital inflows. Calvo, Leidermann, and Reinhart (1993) and Gavin, Hausmann, and Leidermann (1995) show these “pull