## _wp11182

## Source details

**Canonical URL:** [_wp11182](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11182.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11182.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11182.pdf.json)

---

### Introduction: scope and objectives
- Peru remained one of the world’s largest producers of coca leaves; UNODC data suggests that in 2009 a total of 59,996 hectares were cultivated in Peru to produce coca.
- Potential manufacture of up to 317 metric tons of cocaine in Peru in 2009 (UNODC-based estimate).
- Official Peruvian national statistics do not include an estimate for the impact of coca and cocaine production in GDP.
- Primary objective: empirically investigate the relationship between illicit coca GDP and non-coca (formal) GDP regionally within Peru using panel time series techniques and a newly developed database of illegal coca and coca derivatives (gross value added) at national and regional levels for 2001 to 2009 (annual frequency).
- Important limitations:
  - Largely unrecorded nature of activity complicates measurement.
  - Limitations in availability of regional coca and coca derivative prices and seizure data.

### Background, trends, and cross-country context
- Estimated production and shares:
  - Illegal coca and cocaine production represents the equivalent of 0.9 percent of total GDP in 2009 (computed as ratio of illegal coca GDP to national GDP measured in constant 1994 Nuevo Sol (N.S.) units) using UNODC/IDEI data.
  - Alternative source estimates range between approximately 0.6 and 1 (CNC and CADA-CORAH).
  - Bolivia: UDAPE (2010) data imply coca leaf represents between 1/8 percent and 1 and 1/4 percent of total GDP; UNODC suggests illicit coca production represented the equivalent of 21 percent and 14 percent of agricultural output in 2008 and 2009, respectively.
  - Colombia: DANE (2010) data imply the share of production of coca and cocaine ranged between ¾ and 3¾ percent of total GDP from 2000 to 2008.
- Historical Peru trends:
  - After peaking in 1992, cultivation declined until 1999 then recovered; from 1999 to the present officially recorded cultivation increased 58.84 percent.
  - Peak early 1992: about 219,375 metric tons of coca leaf; 2009: about 118,000 metric tons.
  - By 2009 Peruvian cultivation was as much as 88 percent the size of Colombia’s cultivation.
  - Alternate 2009 data point: 59,900 hectares → estimated supply 128,000 metric tons dry coca leaf; DEVIDA estimates 9,000 metric tons consumed for traditional purposes annually.
- Market evolution:
  - Transition from coca paste producer (1995) to cocaine producer (2009).
  - From 2000 to 2009, potential production of cocaine increased 124% (Garcia y Antezana, 2010).
  - Productivity/process changes and new markets (Europe, Asia, Latin America) and spillover from contraction in Colombian coca leaf production influenced growth.

### Data construction and methodological approach
- Database: constructed illegal coca and coca derivatives gross value added at national and regional levels for 2001–2009 (annual).
- GDP-equivalent construction:
  - Sum of gross value added estimates for four categories: illegal coca leaf, gross coca paste, washed coca paste, and cocaine.
  - For each category: gross value added = estimated tonnage × price per ton − intermediate input consumption.
  - Method follows INEI (2007) value added methodology.
- Regional allocation:
  - Used IDEI (2009) published estimates of cultivated hectares (illegal + legal) combined with published regional tonnage estimates to allocate illegal and legal coca GDP by region for 2001–2009.
  - Panel of 11 coca-producing regions: Ayacucho, Junín, Cuzco, Huánuco, San Martin, Ucayali, Puno, Loreto, Pasco, Amazonas, Cajamarca. La Libertad, Ancash and Madre de Dios dropped due to lack of data.
- Empirical/econometric strategy:
  - Panel time series techniques robust to heterogeneous region-specific dynamics.
  - Mix of panel VAR (estimated in log differences with one lag; impulse responses accumulated for log levels; orthogonalization via Cholesky) and panel cointegration methods.
  - Robustness checks: sectorally disaggregated value added (24 regions), banking-sector proxies, electricity and agriculture proxies, specification including government investment.

### Long-run (cointegration) findings
- Panel unit root tests confirm unit roots for all variables.
- Evidence of a long-run cointegrating relationship between regional coca production and formal sector production under both regional allocation approaches; evidence stronger using IDEI-derived estimates.
- Cointegration interpretation: historical sustainability of illicit coca activity levels during the sample period.
- Directional long-run causality:
  - Changes in formal sector activity cause changes in the coca sector with p-values between 2% and less than 1% depending on specification.
  - Reverse long-run causality (coca → formal sector) varied by specification with p-values ranging from 9% to 16%.
- Caveat: only nine annual observations per region; low power for long-run causality tests from illegal coca to GDP.

### Short-run dynamics — panel VAR key numeric findings
- Median regional response of formal sector GDP to an unexpected positive shock to illegal coca production is negative; illicit coca production crowds out formal sector production over a two year horizon across specifications.
- Crowd-out magnitude:
  - Coefficient magnitude at two years is negative but typically less than one in absolute value → crowding out is less than one-for-one, implying total economic activity (illicit + formal) increases after a positive coca shock.
  - Considerable regional heterogeneity: 25% quantile regional responses range to as large as -3.0%.
- Specification comparisons:
  - Figure 1a (UNODC ratios): median regional formal sector GDP response = -1.75% after two years.
  - Figure 1b (IDEI ratios): median response = -0.1% after two years.
- Example numeric scenario (from figure 1c):
  - Median one standard deviation shock to regional illegal coca production ≈ 42% initial first-year increase, damping to 31% net annual increase by the fifth year.
  - Ayacucho 2009 illegal coca production ≈ 123 million Soles:
    - 42% increase ≈ 51.5 million Soles → fall in Ayacucho formal sector GDP by 0.0025%, or roughly 4.7 million Soles.
  - Cusco 2009 illegal coca production ≈ 112 million Soles:
    - 42% increase ≈ 47 million Soles → fall in Cusco formal sector GDP by 0.0025%, or roughly 11.6 million Soles.
- Variance decomposition:
  - Vast majority of variation in formal sector activity is due to shocks originating in the formal sector; illicit coca sector explains very little variation.
- Reverse short-run causality (formal GDP → illegal coca):
  - Unanticipated regional GDP shock leads to increase in illegal coca production: initially ~20%, peaking a little over 30% by beginning of second year.
  - Example numeric impacts:
    - One standard deviation, or 3%, shock to GDP in Cusco ≈ 140 million Soles → initially ~22 million Soles increase in illegal coca production.
    - 3% shock to GDP in Ayacucho ≈ 55.9 million Soles → initially ~36.9 million Soles increase in illegal coca production.
  - When GDP shock originates at national level, median regional illegal coca response slightly negative but not statistically different from zero.

### Robustness checks and alternative specifications
- Legal coca (nine-region subset):
  - Shocks to national GDP stimulate legal coca production.
  - Shocks to legal coca production have relatively small and statistically insignificant effects on formal sector GDP.
  - Legal coca production does not appear to significantly crowd out formal sector GDP regionally.
- Specification 4 (three-variable panel VAR: formal GDP, government investment, illicit coca):
  - After controlling for government investment, shocks to illicit coca production initially yield similar negative median effects on formal sector GDP; after three years the effect becomes positive and then eventually goes to zero.
- Proxy-based cross-checks:
  - Banking-sector proxy (deposits):
    - Hypothesis: illicit coca exports disproportionately increase foreign-currency bank deposits relative to domestic-currency deposits.
    - Results: shocks to foreign-currency deposits lead to decrease in domestic deposits regionally — consistent with illicit coca crowding out formal GDP if deposit proxy conjecture holds.
    - Caveats: deposit dollarization influenced by macro factors; mining and tourism can also affect foreign-currency inflows.
  - Agriculture and electricity proxies:
    - Hypothesis: coca cultivation crowds out other agriculture; electricity demand reflects total economic activity but not coca itself.
    - Results: shocks to agriculture have very small and statistically insignificant effects on electricity regionally. Net impact of illegal coca production crowds out formal sector GDP in some regions and induces net positive effects in others; median regional response close to zero.
    - Caveat: electricity demand in major producing regions may reflect non-local demand.

### Key data series and numeric highlights (selected figures preserved)
- Cultivated hectares and tonnage:
  - Table 4 total annual hectares (coca valleys): 2001: 46,232; 2002: 46,721; 2003: 44,230; 2004: 50,300; 2005: 48,242; 2006: 51,416; 2007: 53,681; 2008: 56,060; 2009: 59,926.
  - Table 6 estimated illicit coca cultivation hectares (TOTAL): 2001: 39,625; 2002: 40,114; 2003: 37,815; 2004: 43,873; 2005: 41,815; 2006: 44,877; 2007: 47,014; 2008: 49,096; 2009: 53,388.
  - Table 8 estimated tonnage of illicit coca leaf (TOTAL): 2001: 43,600; 2002: 43,700; 2003: 63,800; 2004: 101,000; 2005: 96,999; 2006: 105,905; 2007: 107,798; 2008: 111,873; 2009: 117,585.
- Illegal Coca GDP (Thousands of 1994 N.S) — Table 3 totals:
  - Illegal Coca GDP by year: 2001: 630,224; 2002: 659,060; 2003: 833,915; 2004: 1,451,468; 2005: 1,309,987; 2006: 1,159,408; 2007: 1,158,253; 2008: 1,281,812; 2009: 1,539,966.
- Legal Coca GDP (Thousands of 1994 N.S) — Table 3 totals:
  - Legal Coca GDP by year: 2001: 8,323; 2002: 11,583; 2003: 9,582; 2004: 9,326; 2005: 9,381; 2006: 9,224; 2007: 10,260; 2008: 10,460; 2009: 9,654.
- Potential cocaine production and conversion scenarios — Table 1b (examples):
  - 2007 UNODC/DEVIDA hectares 53,682.00 → Cocaine (44%) 308.82; Cocaine (72%) 505.34.
  - 2009 UNODC/DEVIDA hectares 59,926.00 → Cocaine (44%) 352.13; Cocaine (72%) 576.22.
  - Note: Table 1b aligns 317 metric tons figure with a cocaine extraction efficiency assumption of approximately 44 percent; if extraction efficiency assumed 72% GDP production varies accordingly.
- Prices (selected nominal series) — Table 12:
  - Illegal Coca Leaf Annual Average Farm Dry Coca Leaf (US$/kg): 2001: 2.30; 2002: 2.50; 2003: 2.10; 2004: 2.80; 2005: 2.90; 2006: 2.50; 2007: 2.50; 2008: 3.40; 2009: 3.20.
  - Farm Dry Coca Leaf Price (NS/Ton): 2001: 8,068; 2002: 8,794; 2003: 7,306; 2004: 9,558; 2009: 9,638.
  - Gross Coca Paste Average Price (US$/Kg): 2001: 560; 2002: 590; 2003: 530; 2004: 632; 2009: 778.
  - Cocaine Hydrochloride Price (NS/Kg): 2001: 2,793; 2009: 3,075.
  - Exchange rate examples: 2001: 3.51; 2002: 3.52; 2003: 3.48; 2004: 3.41; 2009: 3.01.
- Other notable numeric items:
  - From 2000 to 2009, potential production of cocaine increased 124%.
  - DEVIDA reports 9,000 metric tons of coca leaf consumed for traditional purposes annually (2009).
  - Table 11.1 Intermediate Consumption of Legal Coca Leaf per 1 hectare (2007) — Total: 526.50 (S/.), 100.00%.
  - Table 11.2 Intermediate Consumption of Illegal Coca Leaf per 1 hectare (2007) — Total: 642.6, 100%.

### Main empirical conclusions
- On balance, illicit coca production tends to crowd out formal sector production at the regional level; crowding out is typically less than one-for-one, so total production and income nevertheless increase following a positive illicit coca shock.
- When controlling for government investment:
  - Negative effect on formal production over a three year period; after three years illicit coca production induces higher formal sector production, but the effect dissipates over time.
- Variance decomposition and robustness checks:
  - Formal sector shocks explain the vast majority of formal sector variation; illicit coca explains little variation.
  - Results consistent across multiple specifications, proxies, and robustness checks, though regional heterogeneity exists.

### Policy relevance, implications, and recommended data improvements
- Policy implications:
  - Policy design must account for possible regional heterogeneity in crowding out versus spillover effects.
  - Market dynamics (international demand shifts, spillovers from neighboring countries’ eradication efforts) materially affect incentives and production patterns.
  - Structural features (micro-producer payments, transportation tolls, regional processing hubs, access to precursor chemicals) are important for effective policy targeting.
- Recommended data improvements:
  - Invest in collection of more reliable data on coca and coca derivative production quantities, costs, prices, and seizure data at the political-region level.
  - Coordinate best practices among institutions estimating local coca cultivation to reduce methodological differences in regional allocation of cultivated hectares and tons.
  - Consider sensitivity analyses for intermediate consumption and transformation coefficients to reflect modern versus traditional production methods.
- Monitoring and evaluation caveats:
  - Estimation of illicit coca production GDP is difficult and imprecise; short time series limit inference.
  - Improved regional price and seizure data would enhance ability to infer causal relationships and fine-tune policy.

*Source: IMF working paper content compiled from _wp11182 (2001–2009 regional analysis, methodology and results).*

### REFERENCES .............................................................................................................

### _wp11182 - REFERENCES .............................................................................................................

### Introduction: scope and objectives
- Peru remains one of the world’s largest producers of coca leaves; UNODC data suggests that in 2009 a total of 59,996 hectares were cultivated in Peru to produce coca.
- This implies the potential to manufacture up to 317 metric tons of cocaine in Peru in 2009.
- Official Peruvian national statistics do not include an estimate for the impact of coca and cocaine production in GDP.
- Illicit cultivation and trafficking of coca leaf products are thought to account for a sizeable portion of the informal (unrecorded) sector of the Peruvian economy.
- Primary objective: empirically investigate the relationship between illicit coca GDP and non-coca GDP regionally within Peru, using panel time series techniques and a newly developed database of illegal coca and coca derivatives (gross value added) at national and regional levels for 2001 to 2009 (annual frequency).
- Important limitations identified:
  - The largely unrecorded nature of activity makes it difficult to access relationships between the illicit coca sector and other economic activity.
  - Limitations in the availability of coca and coca derivative prices and seizures at the regional level.

### Background and related literature: trends, cross-country context, and evolution
- Estimated production and shares:
  - Using UNODC data, illegal coca and cocaine production represents the equivalent of 0.9 percent of total GDP in 2009 (computed as the ratio of illegal coca GDP to national GDP measured in constant 1994 Nuevo Sol (N.S.) units).
  - Corresponding estimates based on CNC and CADA-CORAH produce estimates that range between approximately 0.6 and 1 respectively.
  - Bolivia: UDAPE (2010) data imply coca leaf represents between 1/8 percent and 1 and 1/4 percent of total GDP; UNODC suggests illicit coca production represented the equivalent of 21 percent and 14 percent of agricultural output in 2008 and 2009, respectively, with values ranging between 3.6% and 8.25% from 1990 to 2008.
  - Colombia: DANE (2010) data imply the share of production of coca and cocaine ranged between ¾ and 3¾ percent of total GDP from 2000 to 2008, declining toward the end of the period due to eradication efforts.
- Historical cultivation and production trends in Peru:
  - After peaking in 1992, Peruvian cultivation estimates sharply declined until bottoming out in 1999, then recovered; from 1999 to the present officially recorded cultivation increased 58.84 percent.
  - Estimated hectares in 2009 (reported elsewhere in the text) also appear as 59,900 hectares, resulting in an estimated supply of 128,000 metric tons of dry coca leaf; DEVIDA estimates only 9,000 metric tons are consumed for traditional purposes annually.
  - Estimated metric tons of coca leaf production declined from a peak in early 1992 of about 219,375 metric tons (more than 70% percent of world production) to barely 118,000 metric tons (more than 37.72 percent of world coca bush cultivation) in 2009.
  - By 2009 Peruvian cultivation of coca was as much as 88 percent the size of Colombia’s cultivation.
- Market and sector evolution:
  - From 1995 (coca paste producer) to 2009 (cocaine producer), Peru evolved with an estimated 317 metric tons of cocaine production in 2009.
  - From 2000 to 2009, potential production of cocaine increased 124% (Garcia y Antezana, 2010).
  - During 2000–2009, the increase in Peruvian coca cultivation and cocaine production reflected increased demand from new markets (Europe, Asia, Latin America) and, since 2009, spillover from contraction in Colombian coca leaf production (Diaz and Antezana, 2010) which widened the gap between global demand and supply and increased Peru’s incentive to produce.
  - Productivity and process changes: productivity changes in coca leaf cultivation and coca paste production; micro-production processes for coca paste have become predominant and extended to Andean and Amazonian producers.
- Production and trafficking structure:
  - Most coca farmers are micro-entrepreneurs; regional cocaine producers and traffickers often pay farmers in advance of the coca harvest.
  - Established transportation routes protected and controlled by drug trafficking organizations levy tolls on producers and transporters of coca, cocaine paste and cocaine HCL.
  - Firms generally refine cocaine paste into HCL in regional trafficking hubs; increasingly, traffickers transport coca paste directly to the metropolitan Lima area for further refinement because access to precursor chemicals is easier in Lima.
- Extraction efficiency notes:
  - The 317 metric tons figure aligns with Table 1b when cocaine extraction efficiency is assumed to be approximately 44 percent; if extraction efficiency is assumed to be 72%, GDP production at the national level will vary accordingly.

### Hypotheses and mechanisms: crowding out versus spillovers
- Two competing mechanisms described:
  - Crowding out: resources devoted to illicit coca cultivation and trafficking may be drained from other productive opportunities, decreasing economic activity in other sectors.
  - Spillover effects: illicit coca sector activities may generate economic spillovers that increase economic activity in other sectors, potentially enhancing legal sectors.
- These mechanisms can operate in opposite directions, may differ regionally within Peru, and may vary over different time horizons.
- The illegal unobserved economy may interact with the legal unobserved economy; illegal unobserved activities include tax evasion, noncompliance with employment standards, noncompliance with administrative procedures, and illegal production (goods/services whose distribution or possession is illegal, or otherwise legal activities when undertaken by unauthorized producers), per SNA 2008 definitions.

### Data construction, empirical approach, and contribution
- Database: constructed estimates of illegal coca and coca derivatives gross value added at national and regional levels for 2001–2009 (annual).
- Methodology: panel time series techniques to assess extent of crowding out versus spillover effects regionally and to examine long-run sustainability of illicit coca production.
- Contribution: first attempt (to authors’ knowledge) to simultaneously consider national and regional data for Peru in this context; overcomes previous studies’ reliance on national aggregates by using panel/regional approach.
- Empirical challenges reiterated:
  - Unobserved and illicit nature of activity complicates measurement.
  - Regional price and seizure data limitations.

### Key quantitative findings and statistics reported
- 59,996 hectares cultivated in Peru in 2009 (UNODC).
- Potential to manufacture up to 317 metric tons of cocaine in 2009.
- Illegal coca and cocaine represents the equivalent of 0.9 percent of total GDP in 2009 (constant 1994 N.S. units, UNODC/IDEI data).
- Alternative source estimates range between approximately 0.6 and 1 (CNC and CADA-CORAH).
- From 1999 to the present, officially recorded cultivation increased 58.84 percent.
- Early 1992 peak: about 219,375 metric tons of coca leaf; 2009: about 118,000 metric tons.
- By 2009, Peruvian cultivation was as much as 88 percent the size of Colombia’s cultivation.
- Estimated 59,900 hectares under cultivation in 2009 in another passage, yielding an estimated supply of 128,000 metric tons of dry coca leaf; DEVIDA reports 9,000 metric tons consumed for traditional purposes annually.
- Potential production of cocaine HCL estimated by UNODC at 302 metric tons (text break indicates continuation elsewhere).
- Cocaine extraction efficiency scenarios mentioned: approximately 44 percent and 72%.
- From 2000 to 2009, potential production of cocaine increased 124%.

### Policy relevance and implications (derived from paper objectives and findings)
- Improved empirical understanding of regional interactions between illicit coca GDP and non-coca GDP is valuable for policymakers designing strategies to redirect economic activity toward legal sectors.
- Policy design must account for:
  - Possible regional heterogeneity in crowding out versus spillover effects.
  - The role of market dynamics (international demand shifts, spillovers from neighboring countries’ eradication efforts).
  - Structural aspects of the coca economy (micro-producer payments, transportation tolls, processing hubs, access to precursor chemicals).
- Data gaps (regional prices, seizures, and unobserved activity measures) constrain precise policy targeting and monitoring.

*Italic source attribution line.*

### 2008. The UNODC has not estimated the potential production for 2009 because it is in the

### 2008. The UNODC has not estimated the potential production for 2009 because it is in the

### Estimates, uncertainties, and background
- The UNODC has not estimated the potential production for 2009 because it is in the process of revising the conversion factors used in estimating potential production.
- Estimates of any illegal activity are highly speculative; the economic effects of the drug sector have been widely debated and sharp differences of opinion exist.
- Scholarly difficulties:
  - Isolating the specific effect of cocaine on the economy and measuring indirect impacts on activities such as construction and public services is difficult.
  - Effects depend on social groups profiting, economic structure, money laundering ease, laundering methods, government policies, and government implementation capacity.
- Published historical revenue estimates for Peru from coca/cocaine exports:
  - Shams (1992): around $800 million a year in foreign revenue from exports of coca base and cocaine.
  - Alvarez (1995): in 1993-94 exports represented between 23 and 40% of legal exports, amounting to between $0.9 and $1.6 billion.
  - Nadelman (1989): $800 million.
  - Thobani (1994): approximately $550 million annual net revenue.

### Data challenges and scope
- Very nature of activity being largely unrecorded makes it difficult to access the relationship between the illicit coca sector and other economic activity in Peru.
- Due to limitations in availability of coca and coca derivative prices and seizures at the local level, estimates are produced only for national coca production from 2001 to 2009.
- With very few data points, inferring causal relationships between the coca sector and other economic sectors based on national data alone is difficult.
- The mechanisms of interaction between the coca sector and legal informal/formal sectors are largely unrecorded.
- Regional heterogeneity is expected because economic composition differs by region.
- Data sources and limitations:
  - Panel of regional coca production constructed for 11 coca producing regions (Ayacucho, Junín, Cuzco, Huánuco, San Martin, Ucayali, Puno, Loreto, Pasco, Amazonas, Cajamarca); La Libertad, Ancash and Madre de Dios dropped due to lack of data.
  - Coca leaf production yields are based on the Breakthrough Operation of 2003-2004 and may differ from current yields.
  - UNODC is revising conversion factors used in estimating potential production.
  - Regional prices for coca leaf and derivatives are generally unavailable; national average prices used for each year (table 12 referenced).
  - Intermediate consumption coefficients and transformation coefficients are taken from DIRANDRO PNP and INEI (2009); assumed similar across regions due to data limitations.
  - Seizures of coca leaf, cocaine, gross and washed coca paste are not available at the regional level; seizures per region assumed proportional to national estimates.

### Methodology — constructing GDP equivalents and panel strategy
- GDP equivalent construction:
  - Sum gross value added estimates for four categories: illegal coca leaf, gross coca paste, washed coca paste, and cocaine.
  - For coca leaf, coca paste and cocaine categories, gross value added = estimated tonnage × price per ton − costs of intermediate inputs (consumption).
  - Method follows National Statistics Institute of Peru (INEI) methodology for 2007 value added.
- Regional allocation:
  - Use IDEI (2009) published estimates of number of cultivated hectares of illegal and legal coca combined with published estimates of tons produced by region to compute regional GDP equivalents from 2001 to 2009.
- Econometric approach:
  - Use panel time series techniques robust to heterogeneous relationships across regions (heterogeneous region-specific dynamics).
  - Employ mix of panel VAR and panel cointegration methods to allow for endogenous relationships and differing short- and long-run dynamics.
  - Panel VAR estimated in log differences with one lag due to short panel; impulse responses accumulated for log levels.
  - Orthogonalization of shocks via reduced form Cholesky triangularization (implies contemporaneous effect from formal sector to illicit sector but not vice versa).
  - Robustness checks include sectorally disaggregated value added data (24 regions) and banking sector proxies.

### Long-run cointegration findings
- Panel unit root tests confirm presence of unit roots for all variables.
- Evidence of a long run cointegrating relationship between regional coca production and formal sector production under both regional allocation approaches; evidence stronger using IDEI-derived estimates.
- Cointegration implies historical sustainability of illicit coca activity levels during the sample period.
- Direction of long-run causality (panel cointegration tests):
  - Changes in formal sector activity cause changes in the coca sector at statistical p-values between 2% and less than 1%, depending on data approach.
  - Reverse long-run causality (coca → formal sector) varied by specification, with p-values ranging from 9% to 16%.
- Caveat: only nine annual observations per region; low power for long-run causality tests from illegal coca to GDP.

### Short-run dynamic panel VAR results — key numeric findings
- Panel VAR results summarized (figures 1a–1d referenced):
  - Median regional response of formal sector GDP to an unexpected positive shock to illegal coca production is negative; illicit coca production crowds out formal sector production over a two year horizon in all specifications.
  - Coefficient magnitude at two years is negative but typically less than one in absolute value, implying crowding out is less than one-for-one and total economic activity (illicit + formal) increases following a positive coca shock.
  - Considerable heterogeneity across regions: 25% quantile regional responses of formal sector GDP to regional coca shocks range to as large as -3.0%.
- Comparison across specifications:
  - Figure 1a (UNODC ratios): median regional formal sector GDP response to regional illicit coca shock = -1.75% after two years.
  - Figure 1b (IDEI ratios): smallest median response = -0.1% after two years.
- Specific numerical example from figure 1c:
  - Median one standard deviation shock to regional illegal coca production ≈ 42% initial first-year increase, damping to 31% net annual increase by the fifth year.
  - Ayacucho 2009 estimate illegal coca production ≈ 123 million soles:
    - 42% increase ≈ 51.5 million Soles → fall in Ayacucho formal sector GDP by 0.0025%, or roughly 4.7 million Soles.
  - Cusco 2009 estimate illegal coca production ≈ 112 million Soles:
    - 42% increase ≈ 47 million Soles → fall in Cusco formal sector GDP by 0.0025%, or roughly 11.6 million Soles.
  - Interpretation: illicit coca production crowds out formal sector production by less than one-for-one; total production increases.
- Variance decomposition:
  - Vast majority of variation in formal sector economic activity is due to shocks originating in the formal sector; illicit coca sector explains very little variation (consistent with illicit sector being small relative to formal sector).
- Reverse causality (formal GDP → illegal coca):
  - Unanticipated regional GDP shock leads to increase in illegal coca production: initially ~20%, peaking a little over 30% by beginning of second year.
  - Example: a one standard deviation, or 3%, shock to GDP in Cusco ≈ 140 million Soles → initially ~22 million Soles increase in illegal coca production.
  - A 3% shock to GDP in Ayacucho ≈ 55.9 million Soles → initially ~36.9 million Soles increase in illegal coca production.
  - When GDP shock originates at national level, regional illegal coca responses are mixed; median regional response slightly negative but not statistically different from zero.

### Robustness checks and additional specifications
- Legal coca production (nine-region subset):
  - Shocks to GDP at national level stimulate legal coca production.
  - Shocks to legal coca production have relatively small and statistically insignificant effects on formal sector GDP.
  - Legal coca production does not appear to significantly crowd out formal sector GDP regionally.
- Specification 4: three-variable panel VAR (formal sector GDP, government investment, estimated illicit coca production):
  - After controlling for government investment, shocks to illicit coca production initially have similar negative median effect on formal sector GDP, but after three years the effect becomes positive and then eventually goes to zero.
- Proxy-based cross-checks:
  - Banking-sector proxy:
    - Hypothesis: illicit coca exports disproportionately increase foreign-currency bank deposits relative to domestic-currency deposits.
    - Results (figure 3): shocks to international denominated deposits (regional and national) lead to decrease in domestic deposits regionally — consistent with illicit coca crowding out formal GDP if deposit proxy conjecture holds.
    - Caveats: deposit dollarization is influenced by macro factors; mining and tourism could also affect foreign-currency inflows.
  - Sectoral GDP and electricity proxy:
    - Hypothesis: coca cultivation crowds out other agriculture; electricity demand reflects total economic activity but not coca itself.
    - Results (figure 4): shocks to agriculture have very small and statistically insignificant effects on electricity regionally. Net impact of illegal coca production crowds out formal sector GDP in some regions and induces net positive effects in others; median regional response close to zero.
    - Caveat: electricity demand in major producing regions may reflect non-local demand.

### Main conclusions and policy-relevant implications
- Primary empirical conclusion:
  - On balance, illicit coca production tends to crowd out formal sector production at the regional level, but the crowding out is typically less than one-for-one, so total production and income nevertheless increase.
- Dynamics when controlling for government investment:
  - Negative effect on formal production over a three year period; after three years illicit coca production induces higher formal sector production, but this effect dissipates over time.
- Data and methodological caveats:
  - Estimation of illicit coca production GDP is difficult and imprecise; short time series limit inference.
  - Results are supported by numerous robustness checks, alternative specifications, and proxy-based cross-checks, but further data would improve confidence.
- Recommendations for data improvement (Peru-specific and general):
  - Invest in collection of more reliable data on coca and coca derivative production quantities, cost and price estimates, and seizure data at the level of Peruvian political regions to better match with regional economic data.
  - Coordinate best practices among institutions estimating local coca cultivation to reduce methodological differences in regional allocation of cultivated hectares and tons.
  - Consider sensitivity analyses for intermediate consumption and transformation coefficients to reflect modern versus traditional production methods.

*Source: IMF working paper content (2001–2009 regional analysis, methodology and results).*

### REFERENCES

### _wp11182 - REFERENCES

### Key bibliographic sources
- Alvarez, Elena H., 1995, “Economic Development, Restructuring and the Illicit Drug Sector in Bolivia and Peru: Current Policies” Journal of InterAmerican Studies and World Affairs, Vol. 37, No.3 (Autumn), pp. 125-149 (Miami: University of Miami).
- Alvarez, Elena H., 1992, “Reflexiones en tomo a la economía ilegal de la coca: los programas de sustitución y el desarrollo alternativo: El caso de Perú y Bolivia,” Paper presented at Peru: El Problema Agrario en Debate – Sepia IV, Lima, Peru.
- Ball, Maggie, Escobar, Ramon, Grin, Steven, and MacKeen, Leslie, 2006, “A Socioeconomic Evaluation of Alternative Development in the Tropics of Cochabamba, Bolivia: Findings, Observations and Policy Recommendations,” (MA. Dissertation; New York: Columbia University).
- Banco Central de la Reserva del Peru, 2010, Series históricas de la inflación, tipos de cambio, índices de precios de productos químicos y petróleo.
- Canning, D. and P. Pedroni, 2008, “Infrastructure, Long Run Economic Growth and Causality Tests for Cointegrated Panels” The Manchester School, 76, 504-527.
- Comision Nacional para el Desarrollo y Vida sin drogas (DEVIDA), 2010. Available via Internet: http://www.devida.gob.pe/.
- De Franco, Mario, and Godoy, Ricardo, 1992, “The Economic Consequences of Cocaine Production in Bolivia: Historical, Local, and Macroeconomic Perspective” Journal of Latin American Studies, Vol. 24, No. 2 (May), pp. 375-406 (Cambridge: Cambridge University).
- Doerr, Jessica E., 2003, “Realizing the Potential of Crop Substitution: An Analysis of the Andean Coca Trade,” thesis submitted to the Department of Economics at Duke University.
- Fondo Nacional de Financiamiento de la Actividad Empresarial del Estado - Fonafe, Perú: Oferta de Hoja de Coca, June, 2005.
- Garcia-Escribano, Mercedes, 2010, “Peru: Drivers of De-dollarization,” IMF Working Paper 10/169 (Washington: International Monetary Fund).
- Garcia, Jaime and Antezana, Jaime, 2009, Diagnóstico de la situación de desvío de insumos químicos al narcotráfico, Comisión Nacional para el Desarrollo y la Vida sin Drogas.
- Garcia, Jaime and Antezana, Jaime, 2010, Estudio Comparativo de la lucha antidroga en Perú y Colombia: La situación de la coca y la cocaína. Instituto de Estudios Internacionales de la Pontificia Universidad Católica del Perú.
- Instituto Nacional de Estadística e Informática, 2009, Cambio Ano Base de las cuentas nacionales, Estudio especial de la hoja de coca y sus derivados. Metodología de estimaciones de la cuenta de producción de la hoja de coca y sus derivados, 2007.
- Instituto Nacional de Estadística e Informática, 2008, Resumen Metodológico del Cálculo del Producto Interno por Departamento 2001-2008.
- Instituto Nacional de Estadística e Informática, 2010, Series Históricas del PIB regional y por ramas de actividad 2001-2009.
- Instituto de Estudios Internacionales de la Universidad Pontificia Católica del Perú, 2009, El mapa del narcotráfico en el Perú.
- Im, K., Pesaran, H., and Shin, Y., 2003, “Testing for Unit Roots in Heterogeneous Panels” Journal of Econometrics, 115, 53-74.
- Kennedy, Michael, Reuter, Peter, and Riley, Kevin Jack, 1994, A Simple Economic Model of Cocaine Production (Santa Monica: RAND Corporation).
- Kutan, Ali M, and Cespedes, Jose L., 2006, Socio-Political Determinants of Bank Spreads and Deposit Growth: The case of Bolivia. Retrieved from http://www.siue.edu/business/econfin/pdf/Microsoft_Word_-_Kutan_JMCB_07-24-06.pdf.
- Macroconsult, 2008, Narcotráfico: amenaza al crecimiento sostenible del Perú.
- Mejia, Daniel, and Esteban Pasada, Carlos, 2008, “Cocaine Production and Trafficking: What Do We Know?” World Bank Policy Research Working Paper No. 4618 (Washington: World Bank).
- Melvin, Michael and Ladman, Jerry, 1991, “Coca Dollars and the Dollarization of South America,” Journal of Money, Credit and Banking, Vol. 23, No. 4 (Nov), pp. 752-763.
- Melvin, Michael and Peiers, Bettina, 1996, “Dollarization in developing countries: Rational remedy or domestic dilemma? Contemporary Economic Policy, Vol.14, No. 3, pp 30-40.
- Moreno-Sanchez, Rocio, Kraybill, David, and Thompson, Stanley, 2003, “An Econometric Analysis of Coca Eradication Policy in Colombia,” World Development, Vol. 31, No. 2, pp. 375-383.
- Nadelman, E.A., 1989, Latin America and the Drug Trade, pp.144-145, Britannica Book of the Year.
- Novak Fabian, Garcia Jaime and Namihas Sandra, Ensayo El impacto del Narcotráfico a nivel Regional, en el Libro El Mapa del narcotráfico en el Perú – Amenaza a la Seguridad, Instituto de Estudios Internacionales de la Pontificia Universidad Católica del Perú, July, 2009.
- Office of National Drug Control Policy, 2008, Cocaine Smuggling in 2007.
- Oficina de las Naciones Unidas contra la Droga y el Delito (ONUDD), Perú: Monitoreo de cultivos de hoja de coca, June, 2002-2010.
- Oficina de las Naciones Unidas contra la Droga y el Delito (ONUDD), 2010, The Globalization of Crime: A Transnational Organized Crime Threat Assessment.
- Pedroni, P., 1999, “Critical Values for Cointegration Tests in Heterogeneous Panels with Multiple Regressors” Oxford Bulletin of Economics and Statistics, 61, pp. 653-670.
- Pedroni, P., 2000, “Fully Modified OLS for Heterogeneous Cointegrated Panels,” Advances in Econometrics; Nonstationary Panels, Panel Cointegration and Dynamic Panels, 15, pp. 93-130.
- Pedroni, P., 2001, “Purchasing Power Parity Tests in Cointegrated Panels,” Review of Economics and Statistics, 83, pp. 1371-75.
- Pedroni, P., 2004, “Panel Cointegration; Asymptotic and Finite Sample Properties of Pooled Time Series Tests, With an Application to the PPP Hypothesis” Econometric Theory, 20, pp. 597-625.
- Pedroni, P., 2008, “A note on the Econometrics of a Panel SVAR,” working paper, Williams College.
- Pesaran, H. and R. Smith, 1995, "Estimating long-run relationships from dynamic heterogeneous panels," Journal of Econometrics, 68, pp. 79-113.
- Rocha, R., 1997, “Aspectos económicos de las drogas ilegales en Colombia”, In Drogas ilícitas en Colombia, ed UNDP and Dirección Nacional de Estupefacientes. Ariel Ciencia Política, Bogotá.
- Rodriguez, Argos A., 1965, “Possibilities of Crop substitution for the Coca Bush in Bolivia,” Bulletin on Narcotics, Issue 3.
- Rojas Farfan, Fernando, 2002, “La Economia de la Coca,” Instituto de Investigaciones Socio-Economicas, Universidad Catolica Boliviana, La Paz, Bolivia.
- Shams, R., 1992, “The Drugs Economy and Anti-Drug Policy in Developing Countries” Intereconomics, May/June, pp. 139-144.
- Superintendencia de Banca, Seguros, y AFP, 2010, Depósitos en Moneda Nacional y en Moneda Extranjera.
- Steiner, Roberto, 1998, “Colombia’s Income from the Drug Trade”, World Development, Vol. 26, No. 6, pp. 1013-1031.
- Tammen, Melanie S., 1991, “The Drug War vs. Land Reform in Peru,” Cato Institute Policy Analysis No. 156 (Washington: Cato Institute).
- Toledo, 2003, The Illicit Coca-Cocaine Economy and the Potential Impact of Free Trade on the Licit Economy of Alternative Crops in Bolivia, European Trade Study Group.
- Thobani, M., 1994, Peru: Agricultural Policies for Economic Efficiency. Paper prepared for the Workshop on the Illicit Drug Trade in Latin America and the Caribbean, the World Bank, Washington DC, May 18 and 19.
- Thoumi, Francisco E., 2003, “Illegal Drugs in Colombia: From Illegal Economic Boom to Social Crisis,” Annals of the American Academy of Political and Social Science, 582 (July).
- Thoumi, Francisco E., 2005, “The Numbers Game: Let’s All Guess the Size of the Illegal Drug Industry!” The Journal of Drug Issues, Vol. 0022-0426 (January), pp. 185-200.
- Torres, Ana, 2011, “La politica commercial de la UE en la lucha contra las drogas en America Latina: 20 anos de preferencias comerciales,” (Madrid: Real Instituto Elcano).
- United Nations, Office on Drugs and Crime, 2008, The Threat of Narco-Trafficking in the Americas (Vienna).
- United States, Department of Justice, 2007, National Drug Threat Assessment 2008 (Washington).

### Figures and empirical outputs referenced
- Figure 1a–1d: Peru Impulse Responses: Illicit Coca GDP vs. Formal GDP (Specifications 1–4). Sources cited: UNODC; Authors’ estimates; IDEI 2009 and 2010; MEF. Notes: 1/ Formal GDP refers to non-coca GDP. 2/ Illicit coca GDP series are based on UNODC regional coca cultivation estimates for 2008 and 2009 or IDEI estimates; robustness check includes a 10% adjustment for 2001 and 2002 regional baseline tonnage.
- Figure 2: Peru Impulse Responses: Legal Coca GDP vs. Formal GDP. Sources: IDEI 2009 and 2010; authors’ estimates.
- Figure 3: Peru Impulse Responses - Deposits in Domestic Currency vs. Deposits in Foreign Currency. Sources: SBS; Authors’ estimates.
- Figure 4: Peru Impulse Responses - Agriculture GDP vs. Electricity GDP. Sources: INEI; Authors’ estimates.
- Figure 5–6: Peru: Coca Leaf Cultivation in Hectares and Metric Tons. Sources: UNODC (2010); IDEI (2009). Additional sources: Garcia and Antezana (2010); Banco Central de Reserva del Peru (2010).
- Figure 7: Illicit Coca and Derivatives Ratio (As a percent of Non-Coca GDP). Source: Authors’ estimates. Note: estimates computed as ratio of illegal coca GDP to national GDP measured in constant 1994 nuevo sol (N.S.) units.

### Data construction, methodology, and variable definitions
- Illegal Coca and Derivatives Gross Value Added (“Illegal Coca”)
  - Measures the gross value added of illegal coca leaf and derivatives production (coca paste, coca base, and cocaine). Variable equals sum of illegal coca leaf, paste, base, and cocaine gross values added.
  - Gross value production: Table 8 and Table 12. Intermediate Consumption: Table 11.2 and Table 6. Price: Table 12.
- Illegal Coca Leaf Gross Value Added
  - Computed as estimated tonnage of illegal coca leaf cultivation multiplied by price per ton at point of cultivation minus intermediate input consumption (intermediate inputs per hectare × estimated hectares).
  - Gross value production: Table 8 and Table 12. Intermediate Consumption: Table 11.2. Price: Table 12.
- Gross Coca Paste Gross Value Added
  - Computed as estimated tonnage of gross coca paste × price per ton minus intermediate input costs.
  - Intermediate Consumption: Table 11.3.
- Washed Coca Paste Gross Value Added
  - Computed as estimated tonnage of coca base × coca paste price per ton minus intermediate inputs for coca base production.
  - Intermediate Consumption: Table 11.3.
- Cocaine Value Added
  - Difference between gross value added of cocaine production and intermediate consumption.
  - Intermediate Consumption: Table 11.3.
- Legal Coca Leaf Gross Value Added
  - Computed as estimated tons of legal coca leaf × price per ton minus intermediate consumption per hectare × estimated hectares.
  - Intermediate Consumption: Table 11.1.
- Regional variables used in panel analyses:
  - Agriculture (by region): gross value added of agriculture sector by region (Instituto Nacional de Estadística e Informática del Perú (2010)).
  - Electricity (by region): gross value added of electricity sector by region (INEI (2010)).
  - Deposits in Domestic Currency (by region): volume of deposits in N.S. by region (Superintendencia de Bancos y Seguros del Perú (2010)).
  - Deposits in Foreign Currency (by region): volume of deposits in foreign currency by region (Superintendencia de Bancos y Seguros del Perú (2010)).
  - Government Investment (by region): national, regional, and local government investment (accrued) by region (Ministerio de Economia y Finanzas (2010)).
- Note on two-step approach:
  - First, estimate national GDP equivalent of coca and derivatives production annually from 2001-2009 using INEI methodology for 2007.
  - Second, use IDEI (2009) regional hectares and published regional tonnage estimates to allocate illegal and legal coca GDP by region for 2001–2009.

### Key tables and selected numeric highlights (preserve original figures)
- Table 1a: Coca Leaf Cultivated Hectares according to CNC, UNODC, and CADA-CORAH (2001-2009)
  - CNC totals (selected years): 2001: 32,100; 2002: 34,700; 2009: 40,000.
  - UNODC-DEVIDA totals (selected years): 2001: 46,200; 2009: 59,900.
  - CADA-CORAH (2005–2009 available): 2005: 49,481; 2007: 64,716; 2009: 61,629.
- Table 1b: Potential Cocaine Production (2007-2009) — conversion assumptions 44% and 72%
  - 2007 UNODC/DEVIDA hectares 53,682.00 → Cocaine (44%) 308.82; Cocaine (72%) 505.34.
  - 2007 CADA/CORAH hectares 64,717.11 → Cocaine (44%) 380.55; Cocaine (72%) 622.71.
  - 2009 UNODC/DEVIDA hectares 59,926.00 → Cocaine (44%) 352.13; Cocaine (72%) 576.22.
  - Reported MIN and MAX ranges per year (e.g., 2007 MIN hectares 37,340.00 → 211.36 and 345.86; MAX 64,717.11 → 380.55 and 622.71).
- Table 2: Variable definitions — refers to Tables 3, 6, 8, 11.x, 12 for inputs and prices.
- Table 3: Coca and Coca Derivatives Sector Gross Value Added Estimates (2001-2009) (Thousands of 1994 N.S)
  - Total Illegal Coca Leaf and Derivatives (Illegal Coca GDP) — Value Added by year:
    - 2001: 630,224
    - 2002: 659,060
    - 2003: 833,915
    - 2004: 1,451,468
    - 2005: 1,309,987
    - 2006: 1,159,408
    - 2007: 1,158,253
    - 2008: 1,281,812
    - 2009: 1,539,966
  - Total Legal Coca (Legal Coca GDP) — Value Added by year:
    - 2001: 8,323
    - 2002: 11,583
    - 2003: 9,582
    - 2004: 9,326
    - 2005: 9,381
    - 2006: 9,224
    - 2007: 10,260
    - 2008: 10,460
    - 2009: 9,654
  - Note: Negative cocaine GDP in 2008 captures estimated increase in cleaned coca paste price (~14 percent) and important increase in chemical products prices; estimated cocaine price increased by 3 percent.
- Table 4: Estimated Coca Cultivation in Hectares by Coca Valleys (2001-2009) — TOTAL annual hectares:
  - 2001: 46,232
  - 2002: 46,721
  - 2003: 44,230
  - 2004: 50,300
  - 2005: 48,242
  - 2006: 51,416
  - 2007: 53,681
  - 2008: 56,060
  - 2009: 59,926
- Table 5: Estimated Coca Cultivation in Hectares by Region (2001-2009) — TOTAL:
  - 2001: 46,232
  - 2002: 46,721
  - 2003: 44,230
  - 2004: 50,300
  - 2005: 48,242
  - 2006: 51,416
  - 2007: 53,690
  - 2008: 56,070
  - 2009: 59,938
- Table 6: Estimated Illicit Coca Cultivation in Hectares by Region (2001-2009) — TOTAL:
  - 2001: 39,625
  - 2002: 40,114
  - 2003: 37,815
  - 2004: 43,873
  - 2005: 41,815
  - 2006: 44,877
  - 2007: 47,014
  - 2008: 49,096
  - 2009: 53,388
- Table 7: Estimated Yield of Illicit Coca Cultivation by Region (2001-2009) — Average Yield:
  - 2001: 1.02
  - 2002: 1.01
  - 2003: 1.39
  - 2004: 1.90
  - 2005: 1.90
  - 2006: 1.90
  - 2007: 1.90
  - 2008: 1.93
  - 2009: 1.93
- Table 8: Estimated Tonnage of Illicit Coca Leaf by Region (2001-2009) — TOTAL:
  - 2001: 43,600
  - 2002: 43,700
  - 2003: 63,800
  - 2004: 101,000
  - 2005: 96,999
  - 2006: 105,905
  - 2007: 107,798
  - 2008: 111,873
  - 2009: 117,585
- Table 9: Estimated Legal Coca Cultivation in Hectares by Region (2001-2009) — TOTAL:
  - 2001: 6,607
  - 2002: 6,607
  - 2003: 6,415
  - 2004: 6,427
  - 2005: 6,427
  - 2006: 6,539
  - 2007: 6,676
  - 2008: 6,974
  - 2009: 6,539
- Table 10: Eradication by Region in Hectares (2001-2009) — TOTAL annual eradication:
  - 2001: 6,436
  - 2002: 7,135
  - 2003: 7,022
  - 2004: 7,606
  - 2005: 8,966
  - 2006: 10,136
  - 2007: 11,056
  - 2008: 10,144
  - 2009: NA
- Table 11.1: Intermediate Consumption of Legal Coca Leaf (cost structure per 1 hectare in 2007) — Total: 526.50 (S/.), 100.00%
- Table 11.2: Intermediate Consumption of Illegal Coca Leaf (cost structure per 1 hectare in 2007) — Total: 642.6 (units consistent with table), 100%
- Table 11.3: Intermediate Consumption composition per kilogram produced in 2007
  - Gross Coca Paste: Coca Leaf kg 100 → contribution 7.53; Sulfuric Acid kg 1.4 → 3.66; Sodium Carbonate kg 1 → 1; Kerosene kg 12.8 → 4.3; TOTAL 1.00 (table formatting preserved).
  - Washed Coca Paste: Gross Coca Paste kg 1; Sulfuric Acid kg 3.66; Sodium Carbonate kg 1 → 0.90; Potassium Permanganate kg 0.14 → 214.90 (as listed); TOTAL 100.
  - Cocaine Hydrochloride: Washed Coca Paste kg 1.25 → 1,825; Acetone kg 12.00 → 13.85; Hydrochloric Acid kg 0.46 → 10.60; Ethanol/Thinner kg 0.24 → 0.60; TOTAL 1.00.
  - Source: INEI (2009). Note: authors adopt INEI methodology for intermediate consumption.
- Table 12: Summary of Prices (selected series, nominal)
  - Illegal Coca Leaf Annual Average Farm Dry Coca Leaf (US$/kg): 2001: 2.30; 2002: 2.50; 2003: 2.10; 2004: 2.80; 2005: 2.90; 2006: 2.50; 2007: 2.50; 2008: 3.40; 2009: 3.20.
  - Exchange rate (selected years): 2001: 3.51; 2002: 3.52; 2003: 3.48; 2004: 3.41; 2009: 3.01.
  - Farm Dry Coca Leaf Price (NS/Ton): 2001: 8,068; 2002: 8,794; 2003: 7,306; 2004: 9,558; 2009: 9,638.
  - Gross Coca Paste Average Price (US$/Kg): 2001: 560; 2002: 590; 2003: 530; 2004: 632; 2009: 778.
  - Gross Coca Paste Average Price (NS/Kg): 2001: 1,964; 2002: 2,075; 2003: 1,843; 2004: 2,157; 2009: 2,343.
  - Washed Coca Paste (Coca Base NS/Kg) 2001: 2,547; 2009: 3,038.
  - Cocaine Hydrochloride Price (NS/Kg): 2001: 2,793; 2009: 3,075.
  - Legal Coca Leaf Selling Price (NS/Kg): 2001: 3.23; 2002: 4.28; 2003: 3.84; 2004: 3.97; 2009: 4.64.
  - Price Indexes (End of Period) — Chemical Products PI examples: 2001: 1.37; 2008: 30.18; 2009: -14.46.
  - Petroleum PI examples: 2001: -43.50; 2005: 46.95; 2009: -57.42.
  - CPI (end of period) examples: 2001: -0.13; 2007: 3.93; 2008: 6.65; 2009: 0.25.
  - Additional price notes:
    - Peruvian experts cite coca leaf price of $2.90/kg; UNODC 2009 estimates average price $3.20/kg in 2009.
    - State-owned ENACO paid $1.80/kg in 2009.
    - Average price of cocaine paste in cultivating regions ~ $778/kg in 2009.
    - UNODC (2009) estimated cocaine HCL price in Alto Huallaga at $1,020/kg and $1,500/kg in Lima; DEVIDA and other experts indicated export-ready prices up to $4,050/kg in 2009.
    - Cost of bribery estimated at approximately 75% of proceeds generated by total cocaine production.
- Table 13: Robustness Check of Coca and Coca Derivatives Sector Gross Value Added Estimates (2001-2009) (Thousands of 1994 N.S)
  - Robustness Value Added totals for Illegal Coca GDP by year (selected):
    - 2001: 584,741
    - 2002: 657,099
    - 2003: 832,015
    - 2004: 1,449,613
    - 2009: 1,538,103
  - Price index average-period series (2000–2009) listed for Chemical Products CPI, Petroleum CPI, Other Materials, CPI — e.g., CPI 2000: 3.8; CPI 2009: 2.93.

### Methodological and data caveats highlighted in the source
- Taxes and subsidies are not relevant to calculations for illegal coca and derivatives; Gross Value Added (GVA) considered same as GDP for these estimates.
- Negative cocaine GDP in 2008 reflects interaction of increased cleaned coca paste price (~14 percent) and large increases in chemical products prices; estimated cocaine price increase of 3 percent.
- The approach does not account for price of imported chemicals; smuggling/imported chemical prices could affect cocaine production costs.
- Price series are nominal; authors construct a deflator using chemical products price index, petroleum price index, other materials price index, and CPI to adjust years other than 2007 (2007 prices used for 2007).
- Substantial uncertainty exists regarding price data; multiple sources report differing coca leaf and cocaine prices.
- Robustness checks presented (Table 13) explore use of period average CPI values vs. end-of-period CPI values.

*Italic: Source compiled from _wp11182 - REFERENCES (PDF content provided).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp11182.pdf_