## 1mexea2019002

## Source details

**Canonical URL:** [1mexea2019002](https://www.imf.org/-/media/files/publications/cr/2019/1mexea2019002.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2019/1mexea2019002.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2019/1mexea2019002.pdf.json)

---

### Expressway network — background and motivation
- Low productivity growth is a critical challenge for Mexico; upgrading basic public infrastructure, and in particular road infrastructure, raises productivity among firms including small and micro firms.
- Infrastructure quality has declined since around 2013 according to World Economic Forum indicators.
- Public gross fixed capital formation:
  - declined to 3 percent of GDP in 2018,
  - from a peak of 6 percent in 2009,
  - and compared to an average of 4.2 percent since 2000.
- Comparable investment ratios in 2018:
  - emerging market peers: 5.3 percent,
  - regional peers: 3.4 percent.
- Transport and roads specific indicators:
  - 26 percent of firms in Mexico consider transportation—both quality and access—a major constraint (World Bank enterprise survey 2010),
  - 23 percent in Latin America and the Caribbean,
  - 20 percent globally.
- Current spending trends particularly fall short of roads investment needs (Global Infrastructure Hub).

### Data and travel-time construction
- Firm-level panel:
  - Mexican Economic Census covering the universe of Mexican non-agricultural firms with fixed establishments, sample period 1993–2013 in five-year intervals.
  - After cleaning and focusing on manufacturing firms: 398,382 firm-year observations.
- Travel-time measure construction (Box 1):
  - Step 1 (end of sample cross-section, 2014 network):
    - Use 2014 digital GIS road network (INEGI) + average truck speeds from SCT on selected federal and state roads.
    - For federal and state roads without speed information, predict average truck speed by regression on road characteristics.
    - For non-federal/non-state roads with limited information, assume average truck speed = 90 percent of the statutory speed limit.
    - Compute time distance on each road section = distance / speed; apply Dijkstra’s optimal route algorithm to compute shortest time distances between all Mexican localities and the largest metropolitan areas.
  - Step 2 (historical reconstruction 1993–2013):
    - Reconstruct networks for each 5-year period by geocoding expressway construction/improvement (SCT Informe de Labores 2003–2013 + physical maps for 1989, 1995, 2000).
    - Start from 2014 network, eliminate roads built after the period under consideration, attribute a speed penalty to roads improved after the period under consideration.
    - Hold average truck speed constant and vary only length and quality of the federal road network.
    - Reapply optimal route algorithm to compute travel times for each period.
- Identification addresses endogeneity by excluding firms within radii around new/improved roads (radius ranging from 3 to 30 km) and using precise firm and road coordinates.

### Econometric specification and identification
- Baseline specification:
  - ln(Prod_ft) = α_f + α_l + α_stsec_y + β1 ln(TTimeM_ft) + γ X_ft + ε_ft
  - Extended specification includes interaction terms β2 ln(TTimeM_ft) X_ft.
- Definitions and controls:
  - ln(Prod_ft) = log(value added per worker).
  - ln(TTimeM_ft) = log of average travel time between a firm’s locality and the M largest metropolitan areas (baseline M = 20).
  - Controls X_ft include: logged firm age, multi-establishment dummy, share of social security contributions in wage bill (formality), small firm dummy (<10 employees).
  - Fixed effects: firm (α_f), locality (α_l), state-industry-year (α_stsec_y, 4-digit industry).
- Identification strategy:
  - Exclude firms located within radii (3 km, 5 km, 10 km, 20 km, 30 km) around newly built or improved roads to mitigate bias from roads being built where high-growth firms are located.

### Key empirical results (Table 1 — TTime20 coefficients and selected controls)
- TTime20, Logged coefficients (clustered standard errors in brackets):
  - All Firms: -1.267*** [0.449]
  - Drop Firms in 3km Radius: -1.213** [0.495]
  - Drop Firms in 5km Radius: -1.236*** [0.407]
  - Drop Firms in 10km Radius: -1.217** [0.478]
  - Drop Firms in 20km Radius: -1.208* [0.618]
  - Drop Firms in 30km Radius: -1.159*** [0.392]
- Selected control coefficients (All Firms column with standard errors in brackets):
  - Age, Logged: 0.241*** [0.023]
  - Multi-Establishment Firm Dummy: 0.009 [0.020]
  - Formality: 0.008*** [0.001]
  - Small and Micro Firm Dummy (<10 Workers): 0.037 [0.036]
- Sample sizes and R-squared:
  - Observations: 398,382 (All Firms); 364,792 (3km); 352,124 (5km); 286,799 (10km); 220,156 (20km); 180,519 (30km).
  - R-squared: 0.767 (All Firms); 0.770 (3km); 0.772 (5km); 0.779 (10km); 0.779 (20km); 0.786 (30km).
- Significance legend: *** p<0.01, ** p<0.05, * p<0.1.

### Interpretation, magnitude, and heterogeneity
- Magnitude:
  - Baseline M = 20 result interpreted as: if road construction reduces the average travel time from a firm’s locality to the 20 largest metropolitan areas in the country by 1 percent, the productivity of that firm goes up by 1.3 percent (coefficient ≈ 1.3, reported as -1.267 on ln(TTime) with ln(Prod) outcome).
- Robustness:
  - Negative and significant coefficients on ln(TTime) persist when excluding firms within up to a 30 km radius around new/improved roads, supporting identification.
- Firm-size heterogeneity (interaction terms with TTime20):
  - Micro firms (0–3 employees): predicted productivity increase ~1.2 percent for a 1 percent decline in travel times to M=20.
  - Large firms (>50 employees): productivity increases by some 1.8 percent (interaction for large firms noted as insignificant in text).
  - Aggregate interpretation: larger firms benefit more from road investment, but micro and small firms also gain significant productivity improvements.

### Identification strategy and robustness checks (additional points)
- Regressions 2–6 exclude firms within radii: Regression 2: 3 km; Regression 3: 5 km; Regression 4: 10 km; Regression 5: 20 km; Regression 6: 30 km.
- Observations decline from 398,382 to 180,519 (a decline of some 55 percent).
- Travel time coefficient stable across regressions, fluctuating between 1.16 and 1.27.
- Robustness to travel time definition:
  - TTimeM re-estimated for M = 20 (baseline), 15, 10, 5, 1 — TTimeM remains highly significant in each specification.
  - Coefficient magnitude decreases as M falls:
    - M=1 (Mexico City only): 1 percent decrease in travel time increases productivity by about 0.4 percent.
    - M=20 (baseline): 1 percent decrease increases productivity by 1.3 percent.
  - Explanation: reducing average travel time across many metro areas requires more road building than reducing time to a single metro area.

### Policy implication (expressway network)
- Greater government spending on road infrastructure will support efforts to raise productivity and growth over the medium term, benefiting both large firms and Mexico’s large number of small and micro firms.
- Contextual constraint: within the projected physical capital spending envelope, the share of basic non-energy infrastructure investment is unlikely to increase because of a shift toward Pemex investments starting with the 2019 budget and the envelope needing to account for large priority projects (Maya train, Trans-Isthmus Railway).

---

### Public spending trends and fiscal context (Mexico)
- Total Programmable Expenditure:
  - Average (2008-2017): 19.6 percent of GDP
  - 2018: 17.3 percent of GDP
  - 2019 (projected): 16.7 percent of GDP
  - 2020 (projected): 16.6 percent of GDP
- Selected programmable composition (Average; 2018; 2019; 2020 — percent of GDP):
  - Wages and Salaries: 5.7; 5.2; 5.0; 5.0
  - Subsidies & Transfers: 3.3; 2.7; 2.7; 2.7
  - Pensions: 2.9; 3.3; 3.6; 3.7
  - Other Current Expenditure: 2.8; 2.8; 2.7; 2.4
  - Physical Capital: 4.0; 2.6; 2.6; 2.5
  - Other: 0.9; 0.7; 0.2; 0.2
  - Interest: 2.1; 2.6; 2.7; 2.6
- Functional classification (selected items, percent of GDP: Average; 2018; 2019; 2020):
  - Economic Affairs: 6.4; 5.1; 4.6; 4.6
  - Housing and Community Services: 1.5; 1.0; 1.0; 0.9
  - Health: 2.5; 2.4; 2.4; 2.4
  - Education: 3.5; 3.0; 2.9; 2.9
  - Social Protection: 3.4; 3.8; 4.2; 4.3
  - Other (incl Stabilization Funds): 0.3; 0.2; 0.1; 0.0
- Recent outturns (January–August 2018 vs January–August 2019, billion pesos; Real % growth):
  - Total Programmable Expenditure: 2,609; 2,590; Real % growth: -4.6
  - Physical Capital: 476; 423; Real % growth: -14.6
  - Health: 343; 342; Real % growth: -4.2
  - Education: 441; 434; Real % growth: -5.3
  - Social Protection: 592; 679; Real % growth: 10.3
  - Interest: 389; 424; Real % growth: 4.9

### Analysis of spending composition and sustainability
- Between 2016 and 2018 budgetary expenditures were cut by almost 2.8 percentage points of GDP; decline in programmable budgetary spending amounted to 3.4 percentage points of GDP.
- Decline largely explained by reduction in capital expenditure and, to a lesser extent, falls in wages and salaries and in subsidies and transfers (mostly in education, health, and housing and community services).
- Spending increases concentrated in pensions and interest payments.
- Projected fiscal adjustment:
  - Ongoing adjustment projected to further shrink programmable expenditure by another ¾ percent of GDP from 2018 to 2020.
  - Pension spending expected to continue increasing significantly; non-PEMEX related physical capital spending set to decline further.
  - Cuts of 0.7 percent of GDP envisaged in other current spending (implying a decline by a quarter).
- Sustainability concerns:
  - Legally-mandated expenditures represent 2/3 of the budget; another 20 percent comprises technically discretionary but inflexible expenditures.
  - Continued cuts focus on a small share of effectively discretionary expenditures, including vital capital investments.
- Policy recommendation (fiscal composition):
  - Reallocate expenditure toward capital spending and education while improving efficiency across areas to achieve a more growth-friendly and inclusive spending mix consistent with medium-term objectives.

### Social spending, benchmarking, and SDG costing (health and education)
- Social spending definition: education, health, social protection, housing and community services, and environmental protection.
- Historical and current levels:
  - Social spending increased by 2.5 percent of GDP from 2007 to 2015, reaching a maximum of 12.1 percent of GDP, then shrank to 10.3 percent in 2018.
  - Social protection and education together absorb more than 60 percent of total social spending; health amounts to close to a fourth.
  - Mexico currently has one of the lowest levels of social spending among peers.
- Health sector efficiency opportunities:
  - Administrative costs account for almost 10 percent of total health spending; aligning administrative and insurance costs with the OECD average of 3 percent would generate savings of at least 0.15 percent of GDP.
  - Eliminating beneficiary overlaps and inconsistencies across insurance schemes could yield further fiscal savings (more than 0.1 percent of GDP).
- SDG costing (health, selected values and scenario results for 2030):
  - GDP per capita (2016 or latest): 8,909.5; 8,115.8; 12,623.2; 8,814.9; 10,234.4
  - Doctors per 1,000 population (scenarios): 2.0; 1.9; 2.7; 2.2; 2.7
  - Other medical personnel per 1,000 population: 6.4; 6.3; 7.7; 9.6; 7.7
  - Doctor wages (ratio to GDP per capita): 4.1; 4.2; 4.0; 3.5; 4.0
  - Private share (% total spending) (scenarios): 39.5; 47.8; 31.6; 47.8; 31.6
  - Health spending (percent of GDP) — All/Public/Private across scenarios:
    - All: 6.4; 5.9; 7.1; 5.9; 6.4
    - Public: 3.9; 3.1; 4.9; 3.1; 4.4
    - Private: 2.5; 2.8; 2.3; 2.8; 2.0
  - Per capita spending (USD 2018) in scenarios: 570.4; 482.3; 899.5; 523.8; 653.6
  - Conclusion: Mexico could face an increase of more than 0.5 percent of GDP in total health expenditure by 2030 to meet the SDGs; fiscal pressures would be higher if public sector share in health provision increases.
- Education SDG costing: additional spending needs estimated at 0.5 percent of GDP by 2030.
- Policy directions (education):
  - Shift composition toward capital spending (equipment, facilities, IT, infrastructure).
  - Improve transparency and accountability in the education payroll.
  - Improve early-childhood education quality.
  - Increase access in low-coverage regions and for disadvantaged children.
  - Recalibrate salaries/personnel to emulate high-performing countries (lower teacher wages in percent of GDP but smaller classes).
  - Increase enrollment rates.

### Social protection and pensions
- Fragmentation and targeting:
  - More than 8,000 programs at federal, state and municipal levels, leading to duplication and fragmentation.
  - A social census covered about 20 million households (out of 30 million) — recommendation to use it to clean SISI and create a single registry to reduce inclusion/exclusion errors.
- Size and composition:
  - Administration spent about 1.8 percent of GDP on social assistance programs in 2018.
  - Transfers and PAM represented about 40 percent of the total social assistance received by an average household.
  - 2019–2020 budgets increased social assistance by about 0.5 percent of GDP with composition changes (increase in social/disability pensions; modest decrease in transfers).
  - Prospera canceled; resources redistributed.
  - PAM benefits about 8 million individuals; 60 percent of those aged 65+, and 100 percent of those aged 68+.
- Pensions adequacy and reform options:
  - Poverty rate among people over 65 is more than 30 percent.
  - Average contribution rate of 6.5 percent in IMSS DC scheme — may at best lead to a replacement rate of 26 percent for a full career average earner (second lowest replacement rate among OECD countries).
  - Suggested adjustments:
    - Increase contribution rate (possibly offset by reducing Infonavit wage-based contribution).
    - Increase effective retirement age linked to life expectancy gains.
    - Tighten early-retirement schemes.
    - Increase contribution period required for a full pension in old public-sector DB scheme.
    - Increase the age limit to get a full pension in the public sector faster.

### Fiscal strategy conclusions
- Low levels of discretionary spending imply raising revenues will be indispensable to put public debt on a downward path.
- Substantial room exists for tax policy and revenue administration reforms.
- Strengthen spending efficiency to make room for priority expenditures.
- Enhance value for money in social spending given SDG-related additional needs in education and health.

---

### PEMEX taxation regime — objectives and recent reforms
- Objective: using IMF FARI project-level cashflow modeling to evaluate key characteristics of PEMEX’s fiscal regime and compare to announced reforms and production sharing regime (PSCs).
- Recent reform actions (2019):
  - SHCP announced plans to ease tax burden by loosening cost deduction caps under PEMEX’s ‘entitlement’ fiscal regime, aligning them with cost recovery limits of PSCs concluded in 2015-18 and PEMEX’s Ek Balam area.
  - PEMEX business plan (July 2019) announced plans to reduce the profit-sharing rate from 65 percent to 58 and 54 percent in 2020 and 2021, respectively.
- Short-term implications:
  - Increase in the cost cap and a reduction in the profit-sharing rate will reduce the overall tax burden for PEMEX and the regressivity of the regime, and by increasing return to PEMEX, may release funds for further investment.
  - Even with these changes, the regime lacks sufficient progressive instruments to allow the government to share in upside from new developments.
- Longer-term recommendation: migration of entitlement assets to newer more balanced contractual regimes would be beneficial.

### Entitlement fiscal regime — principal components and parameters
- Principal components (prior to changes contemplated by SHCP):
  - Profit Sharing Fee: 65 percent of production value less cost deductions (subject to annual cost cap).
  - Hydrocarbons Extraction Fee: price-linked ad-valorem royalty.
  - Corporate Income Tax: Rate 30 percent of taxable income.
  - Annual fixed surface area fees: Hydrocarbon Exploration Fee and Tax on Hydrocarbon Exploration and Extraction Activity.
  - Profit-sharing fee and hydrocarbons fees are deductible for corporate income tax.
- Cost caps (post-2014 expressed as percent of revenue):
  - Onshore Oil: 12.50 percent
  - Offshore Oil Shallow: 12.50 percent
  - Offshore Oil Deep: 60 percent
  - Natural Gas: 80 percent
  - Chicontepec: 60 percent
- Hydrocarbon Extraction Fee (percent of value of hydrocarbons):
  - When oil price is less than $48 per barrel, 7.5 percent.
  - When oil price greater than or equal to $48 per barrel, (12.5 percent*Petroleum price) + 1.5 percent
- Corporate Income Tax:
  - Rate: 30 percent
  - Depreciation: 100 percent immediate expensing of exploration costs, 4 year straight-line depreciation of development costs
- Effective royalty rates (as computed in Table 2):
  - Onshore: 60.1 percent
  - Offshore Shallow Water: 60.1 percent
  - Offshore Deepwater: 31.6 percent
  - Natural Gas: 19.5 percent
  - Chicontepec Paleochannel: 31.6 percent
  - Proposed Reform (Chicontepec/Paleochannel): 27.5 percent
- Note: at current price levels of approximately $60/barrel, the royalty rate would be higher than the minimum rate, at around 9 percent ((0.125*60)+1.5). On July 3, 2019, the Mexican basket price was $59.33 per barrel.
- International comparison: median effective rate in a sample of 56 countries surveyed is approximately 20 percent. Mexico’s effective royalty rate for oil under the entitlement regime is far above international norms.
- Ringfencing: payments and income tax are ringfenced by five regional classifications (onshore, shallow water, deep-water, non-associated natural gas, Chicontepec Paleochannel); regional ringfencing (vs licence/asset-level) may delay profit-based government revenue and reduce cap impacts.

### Project example and FARI modeling results (stylized shallow water offshore field)
- Project parameters:
  - Project size: 500 MMBbl
  - Project life: 18 Years
- Project costs (total):
  - Exploration costs: $MM 400; $/Bbl 0.80
  - Development costs excluding drilling: $MM 2,208; $/Bbl 4.42
  - Drilling costs: $MM 1,977; $/Bbl 3.96
  - Operating costs: $MM 3,998; $/Bbl 8.00
  - Decommissioning costs: $MM 558; $/Bbl 1.12
  - Total costs: $MM 9,141; $/Bbl 18.3
- Oil Price used in analysis: $60/Bbl (constant)
- Pre-tax IRR: 35.0 percent
- Modeling framework: FAD’s FARI modeling framework; costs reflect levels reported in a PEMEX investor presentation (November 2018); analysis is project-level and does not model regional consolidated ringfencing benefits.

### Project fiscal results (real 2019 terms, Table 4)
- Pre-Tax project IRR: 35.0%
- Post-tax IRR on total funds:
  - Entitlement Regime: 6.8%
  - Entitlement Regime - Proposed Reforms: 18.0%
  - Production Sharing Regime: 18.5%
- Post-tax IRR on equity:
  - Entitlement Regime: 9.0%
  - Entitlement Regime - Proposed Reforms: 28.9%
  - Production Sharing Regime: 29.6%
- Pre-tax NCF undiscounted: 20,844 (same across regimes)
- Post-tax investor NCF undiscounted:
  - Entitlement Regime: 1,944
  - Entitlement Regime - Proposed Reforms: 5,751
  - Production Sharing Regime: 5,839
- Government Revenue undiscounted:
  - Entitlement Regime: 18,779
  - Entitlement Regime - Proposed Reforms: 14,972
  - Production Sharing Regime: 14,884
- AETR undiscounted:
  - Entitlement Regime: 90.1%
  - Entitlement Regime - Proposed Reforms: 71.8%
  - Production Sharing Regime: 71.4%
- Pre-tax NCF (10% discount): 5,267 (same across regimes)
- Post-tax investor NCF (10% discount):
  - Entitlement Regime: -350
  - Entitlement Regime - Proposed Reforms: 967
  - Production Sharing Regime: 1,024
- Government revenue (10% discount):
  - Entitlement Regime: 5,510
  - Entitlement Regime - Proposed Reforms: 4,193
  - Production Sharing Regime: 4,136
- AETR (10% discount):
  - Entitlement Regime: 104.6%
  - Entitlement Regime - Proposed Reforms: 79.6%
  - Production Sharing Regime: 78.5%
- Interpretation:
  - Under the entitlement regime with a 12.5 percent cost cap, the example project is unviable (AETR well over 100 percent; investor IRR 6.8 percent).
  - Proposed SHCP reforms (raising cap and lowering profit-sharing to 54 percent) reduce regressivity and improve viability (discounted AETR 79.6 percent; investor IRR 18.0 percent).
  - PSC regime produces similar investor outcomes (AETR 78.5 percent; post-tax IRR 18.5 percent).

### Breakeven price, progressivity, and international comparison
- Breakeven price (minimum price to meet investor after-tax hurdle of 12.5 percent real):
  - Entitlement regime breakeven price: USD76.7/barrel.
  - Entitlement regime with increased cap and lower profit-sharing and PSCs display breakeven prices more in line with market trends.
- Progressivity (AETR behavior across outcomes):
  - Increasing cost cap and reducing profit-sharing reduces regressivity, but without substantive progressive components the AETR falls as profitability increases.
  - PSCs can include progressive components (e.g., tiers linked to profitability) that counteract regressive components as profitability rises.
  - Wider range of projects could become commercially viable under the reformed entitlement regime and the PSC regime.
- International comparison:
  - Entitlement regime places a significantly higher burden on projects than other countries in the sample.
  - PSC and reformed entitlement regimes place Mexico better in line with the sample in terms of neutrality while maintaining a comparable government share of revenue.
  - Representative breakeven and AETR figures shown in staff charts include: 76.7, 43.9, 43.8, 43.6, 42.0, 40.1, 39.6, 36.5, 36.0 (as presented in source charts).

### Policy implications and recommended focus areas (PEMEX fiscal regime)
- Short-term:
  - SHCP’s proposal to increase cost cap and lower profit-sharing rate will reduce regressivity for onshore and shallow water projects and may facilitate investment on commercial terms (reducing minimum government share of revenue to 27.5 percent in the relevant proposed reform calculation).
  - These reforms increase returns to PEMEX and may release funds for investment.
- Medium/long-term:
  - Regime still lacks sufficient progressive instruments to share upside; migration of entitlement assets to newer, more balanced contractual regimes would be beneficial although with some revenue trade-offs.
- If cost caps are increased, mitigate cost-inflation risk via:
  - careful screening by CNH of PEMEX’s projects, budgets and work plans;
  - regular high-quality cost and fiscal audits (required by CNH and SAT);
  - competitive, transparent procurement procedures for subcontractor services.
- Additional mechanism:
  - Use farmouts and private participation through migration processes to provide cost oversight and incentivize cost containment.
- Further analysis recommended:
  - Assess specific signed contract terms (variation in bids and contract mechanics, e.g., Ek-Balam).
  - Consider impact of regional consolidated ringfencing on investor returns and government revenue timing.
  - Analyze licence contracts and alignment with PSC cost recovery limits.

*Source: 1mexea2019002 — IMF staff analysis excerpt.*

### 1. Expressway Network ________________________________________________________________ 5

### 1. Expressway Network

### A. Background and motivation
- Low productivity growth is a critical challenge for Mexico; upgrading basic public infrastructure, and in particular road infrastructure, raises productivity among firms including small and micro firms.
- Infrastructure quality has declined since around 2013 according to World Economic Forum indicators.
- Public gross fixed capital formation:
  - declined to 3 percent of GDP in 2018,
  - from a peak of 6 percent in 2009,
  - and compared to an average of 4.2 percent since 2000.
- Comparable investment ratios in 2018:
  - emerging market peers: 5.3 percent,
  - regional peers: 3.4 percent.
- Transport and roads specific indicators:
  - 26 percent of firms in Mexico consider transportation—both quality and access—a major constraint (World Bank enterprise survey 2010),
  - 23 percent in Latin America and the Caribbean,
  - 20 percent globally.
- Current spending trends particularly fall short of roads investment needs (Global Infrastructure Hub).

### B. Data, construction of travel-time measures, and sample
- Firm-level panel:
  - Mexican Economic Census covering the universe of Mexican non-agricultural firms with fixed establishments, sample period 1993–2013 in five-year intervals.
  - After cleaning and focusing on manufacturing firms: 398,382 firm-year observations.
- Travel-time measure construction (Box 1):
  - Step 1: Produce cross-sectional travel times at the end of the sample using the 2014 digital GIS road network (INEGI) + average truck speeds from SCT on selected federal and state roads.
    - For federal and state roads without speed information, predict average truck speed by regression on road characteristics.
    - For non-federal/non-state roads with limited information, assume average truck speed = 90 percent of the statutory speed limit.
    - Compute time distance on each road section = distance / speed; apply Dijkstra’s optimal route algorithm to compute shortest time distances between all Mexican localities and the largest metropolitan areas.
  - Step 2: Reconstruct historical networks for each 5-year period 1993–2013 by geocoding expressway construction/improvement (SCT Informe de Labores 2003–2013 + physical maps for 1989, 1995, 2000).
    - Start from 2014 network, eliminate roads built after the period under consideration, attribute a speed penalty to roads improved after the period under consideration.
    - Hold average truck speed constant and vary only length and quality of the federal road network.
    - Reapply optimal route algorithm to compute travel times for each period.
- Identification relies on time variation in travel times driven solely by road construction and improvement; precise firm and road coordinates are used to exclude firms within radii around new/improved roads (radius ranging from 3 to 30 km) to address endogeneity.

### C. Econometric specification and identification
- Baseline specification (Equation 1 and 2):
  - ln(Prod_ft) = α_f + α_l + α_stsec_y + β1 ln(TTimeM_ft) + γ X_ft + ε_ft
  - Extended specification includes interaction terms β2 ln(TTimeM_ft) X_ft.
  - ln(Prod_ft) = log(value added per worker).
  - ln(TTimeM_ft) = log of average travel time between a firm’s locality and the M largest metropolitan areas (baseline M = 20).
  - Controls X_ft include: logged firm age, multi-establishment dummy, share of social security contributions in wage bill (formality), small firm dummy (<10 employees).
  - Fixed effects: firm (α_f), locality (α_l), state-industry-year (α_stsec_y, 4-digit industry).
- Identification strategy:
  - Exclude firms located within radii (3 km, 5 km, 10 km, 20 km, 30 km) around newly built or improved roads to mitigate bias from roads being built where high-growth firms are located.

### D. Key empirical results
- Baseline finding (textual interpretation):
  - For baseline M = 20, the variable is highly significant with a coefficient of about 1.3: "if road construction reduces the average travel time from a firm’s locality to the 20 largest metropolitan areas in the country by 1 percent, the productivity of that firm goes up by 1.3 percent."
- Table 1 coefficients and statistics (preserving reported values):
  - TTime20, Logged coefficients (with clustered standard errors in brackets):
    - All Firms: -1.267*** [0.449]
    - Drop Firms in 3km Radius: -1.213** [0.495]
    - Drop Firms in 5km Radius: -1.236*** [0.407]
    - Drop Firms in 10km Radius: -1.217** [0.478]
    - Drop Firms in 20km Radius: -1.208* [0.618]
    - Drop Firms in 30km Radius: -1.159*** [0.392]
  - Control variable coefficients (selected):
    - Age, Logged: 0.241*** [0.023] (All Firms) and similar across samples (0.240***, 0.238***, 0.233***, 0.238***, 0.237***).
    - Multi-Establishment Firm Dummy: 0.009 [0.020] (All Firms) and small variations across samples (not significant).
    - Formality: 0.008*** [0.001] (All Firms) and similar across samples (0.008***, 0.008***, 0.008***, 0.007***, 0.007***).
    - Small and Micro Firm Dummy (<10 Workers): 0.037 [0.036] (All Firms) and not significant across samples.
  - Sample sizes and goodness of fit:
    - Observations: 398,382 (All Firms); 364,792 (3km radius drop); 352,124 (5km); 286,799 (10km); 220,156 (20km); 180,519 (30km).
    - R-squared: 0.767 (All Firms); 0.770 (3km); 0.772 (5km); 0.779 (10km); 0.779 (20km); 0.786 (30km).
  - Significance legend: *** p<0.01, ** p<0.05, * p<0.1.

### E. Interpretation and implications
- Magnitude: The baseline estimate implies a near-proportional productivity response to travel-time reductions (text reports ~1.3 percent productivity increase for 1 percent reduction in travel time to the 20 largest metros).
- Robustness: Negative and significant coefficients on ln(TTime) persist when excluding firms within up to a 30 km radius around new/improved roads, supporting identification.
- Policy implication highlighted in the note:
  - Greater government spending on road infrastructure will support efforts to raise productivity and growth over the medium term, benefiting both large firms and Mexico’s large number of small and micro firms.
- Context for fiscal choices:
  - Within the projected physical capital spending envelope, the share of basic non-energy infrastructure investment is unlikely to increase due to:
    - a shift toward Pemex investments starting with the 2019 budget, and
    - the envelope needing to account for large priority projects (Maya train, Trans-Isthmus Railway).

*Source: 1mexea2019002 - 1. Expressway Network*

### 11.      Our identification strategy suggests that the link running from road investment to

### 11.      Our identification strategy suggests that the link running from road investment to

### Identification strategy and causal interpretation
- Regressions 2–6 of Table 1 implement the identification strategy to mitigate reverse causality concerns (new roads built in areas with expected economic expansion).
- Strategy: exclude all firms located within a radius around any newly built or improved roads; radii used in regressions:
  - Regression 2: 3 kilometers
  - Regression 3: 5 kilometers
  - Regression 4: 10 kilometers
  - Regression 5: 20 kilometers
  - Regression 6: 30 kilometers
- Observations decline across regressions from 398,382 in the baseline Regression 1 to 180,519 in Regression 6 (a decline of some 55 percent in the number of observations).
- Main result: the travel time variable is highly significant in all regressions and its coefficient is stable, fluctuating between 1.16 and 1.27.
- Interpretation: stability and significance of the travel time coefficient are taken as reasonably strong evidence of a causal relationship from road investment (via travel time reductions) to firm productivity.

### Robustness to travel time definition and metropolitan area coverage
- The TTimeM variable is redefined by varying the number of metropolitan areas M used to calculate average travel time; Regression 1 is re-run for M = 20 (baseline), 15, 10, 5, and 1.
- In each specification, TTimeM remains highly significant.
- Coefficient magnitude decreases as M falls:
  - Decreasing travel time to Mexico City (M=1) by 1 percent increases firm productivity by about 0.4 percent.
  - Using the 20 largest metropolitan areas (M=20, baseline) a 1 percent decrease in average travel time increases productivity by 1.3 percent.
- Explanation: reducing average travel time across many metro areas requires more road building than reducing time to a single metro area; hence coefficients decline as M falls.

### Firm-size heterogeneity in productivity response
- Interaction terms between TTime20_ft and firm-size dummies (micro: 0 to 3 employees; small: 4-10 employees; medium: 11-50 employees; large: >50 employees) are included in Regression 1 (medium omitted category).
- All interaction terms except the one for large firms are highly significant.
- Predicted impact of a 1 percent decline in travel times to the largest 20 metropolitan areas:
  - Micro firms: productivity increases by some 1.2 percent
  - Small firms: (coefficient significant; plotted in Figure 4)
  - Medium firms: (omitted category baseline)
  - Large firms: productivity increases by some 1.8 percent (interaction for large firms is noted as insignificant in text; light blue in figure)
- Aggregate interpretation: larger firms benefit more from road investment, but micro and small firms also gain significant productivity improvements.

### Public spending trends and fiscal context (Mexico)
- Programmable budgetary spending and composition:
  - Total Programmable Expenditure:
    - Average (2008-2017): 19.6 percent of GDP
    - 2018: 17.3 percent of GDP
    - 2019 (projected): 16.7 percent of GDP
    - 2020 (projected): 16.6 percent of GDP
  - Wages and Salaries: 5.7; 5.2; 5.0; 5.0 (Average; 2018; 2019; 2020)
  - Subsidies & Transfers: 3.3; 2.7; 2.7; 2.7
  - Pensions: 2.9; 3.3; 3.6; 3.7
  - Other Current Expenditure: 2.8; 2.8; 2.7; 2.4
  - Physical Capital: 4.0; 2.6; 2.6; 2.5
  - Other: 0.9; 0.7; 0.2; 0.2
  - Interest: 2.1; 2.6; 2.7; 2.6
- Functional classification (selected items, percent of GDP):
  - Economic Affairs: 6.4; 5.1; 4.6; 4.6 (Average; 2018; 2019; 2020)
  - Housing and Community Services: 1.5; 1.0; 1.0; 0.9
  - Health: 2.5; 2.4; 2.4; 2.4
  - Education: 3.5; 3.0; 2.9; 2.9
  - Social Protection: 3.4; 3.8; 4.2; 4.3
  - Other (incl Stabilization Funds): 0.3; 0.2; 0.1; 0.0
- Recent outturns (January–August 2018 vs January–August 2019, billion pesos and real % growth):
  - Total Programmable Expenditure: 2,609; 2,590; Real % growth: -4.6
  - Physical Capital: 476; 423; Real % growth: -14.6
  - Health: 343; 342; Real % growth: -4.2
  - Education: 441; 434; Real % growth: -5.3
  - Social Protection: 592; 679; Real % growth: 10.3
  - Interest: 389; 424; Real % growth: 4.9

### Analysis of spending composition and sustainability
- Post-GFC adjustment and composition change:
  - Budgetary expenditures were cut by almost 2.8 percentage points of GDP between 2016 and 2018; decline in programmable budgetary spending amounted to 3.4 percentage points of GDP.
  - Decomposition suggests decline largely explained by reduction in capital expenditure and, to a lesser extent, falls in wages and salaries and in subsidies and transfers (mostly in education, health, and housing and community services).
  - Spending increases concentrated in pensions and interest payments.
- Projected fiscal adjustment:
  - Ongoing fiscal adjustment is projected to further shrink programmable expenditure by another ¾ percent of GDP from 2018 to 2020.
  - Pension spending expected to continue increasing significantly; non-PEMEX related physical capital spending set to decline further.
  - Cuts of 0.7 percent of GDP envisaged in other current spending (implying a decline by a quarter).
- Sustainability concerns:
  - Legally-mandated expenditures represent 2/3 of the budget; another 20 percent comprises technically discretionary but inflexible expenditures.
  - Continued cuts focus on a small share of effectively discretionary expenditures, including vital capital investments.
  - To achieve a more growth-friendly and inclusive spending mix while meeting medium-term objectives would require reallocating expenditure toward capital spending and education and making efficiency improvements across areas.
  - Social spending is highlighted as an area where efficiency improvements appear feasible.

### Social spending, benchmarking, and SDG costing
- Definition: social spending includes education, health, social protection, housing and community services, and environmental protection.
- Historical and current levels:
  - Social spending increased by 2.5 percent of GDP from 2007 to 2015, reaching a maximum of 12.1 percent of GDP, then shrank to 10.3 percent in 2018.
  - Social protection and education together absorb more than 60 percent of total social spending; health amounts to close to a fourth.
  - Mexico currently has one of the lowest levels of social spending among peers.
- Health sector fragmentation and efficiency opportunities:
  - Administrative costs account for almost 10 percent of total health spending; aligning administrative and insurance costs with the OECD average of 3 percent would generate savings of at least 0.15 percent of GDP.
  - Eliminating beneficiary overlaps and inconsistencies across insurance schemes could yield further fiscal savings (more than 0.1 percent of GDP).
- SDG costing for health (2030 estimates and scenarios; selected values):
  - GDP per capita (2016 or latest): 8,909.5; 8,115.8; 12,623.2; 8,814.9; 10,234.4 (columns correspond to country groupings and scenarios in Table 6)
  - Doctors per 1,000 population (scenarios): 2.0; 1.9; 2.7; 2.2; 2.7
  - Other medical personnel per 1,000 population: 6.4; 6.3; 7.7; 9.6; 7.7
  - Doctor wages (ratio to GDP per capita): 4.1; 4.2; 4.0; 3.5; 4.0
  - Private share (% total spending) (scenarios): 39.5; 47.8; 31.6; 47.8; 31.6
  - Results — Health spending (percent of GDP) across scenarios and years:
    - All: 6.4; 5.9; 7.1; 5.9; 6.4
    - Public: 3.9; 3.1; 4.9; 3.1; 4.4
    - Private: 2.5; 2.8; 2.3; 2.8; 2.0
  - Per capita spending (USD 2018) in scenarios: 570.4; 482.3; 899.5; 523.8; 653.6
  - Conclusion: Mexico could face an increase of more than 0.5 percent of GDP in total health expenditure by 2030 to meet the SDGs; fiscal pressures would be higher if public sector share in health provision increases.

*Source: Excerpt from IMF chapter, “1mexea2019002 - 11.      Our identification strategy suggests that the link running from road investment to”*

### 13.      Enhancing the quality of education spending is not only important for fiscal

### 13.      Enhancing the quality of education spending is not only important for fiscal

### Education spending: drivers, pressures, and efficiency opportunities
- Given projected demographic trends in Mexico and the constitutional mandate to universalize secondary education, the coming years will likely see a large expansion in education services and a sector-wide shift towards secondary and tertiary levels.
- These developments are expected to generate significant structural fiscal pressures, underscoring the need to enhance value for money in education spending.
- The share of current spending in total education expenditure is very high, potentially crowding out investment in:
  - equipment,
  - facilities,
  - information technology, and
  - modern infrastructure necessary to improve education quality and keep pace with evolving labor demands (World Bank 2016).
- Policy directions and efficiency measures:
  - Shift the composition of education expenditures toward capital spending.
  - Improve transparency and accountability in the education payroll.
  - Improve the quality of early-childhood education.
  - Increase access to education in low-coverage regions and for disadvantaged-background children.
  - Recalibrate the mix of salaries and personnel to emulate levels observed in high performing countries—these tend to have lower teacher wages (in percent of GDP) but also smaller classes (lower student to teacher ratios).
  - Increase enrollment rates.
- Additional spending needs to meet SDGs in education are estimated at 0.5 percent of GDP by 2030.

### Social protection: structure, fragmentation, and targeting
- Definition:
  - Social protection spending comprises social insurance and social assistance programs.
  - Social insurance: protects households from shocks; typically financed by contributions or payroll taxes.
  - Social assistance (social safety net): protects households from poverty; financed by general government revenue.
- Fragmentation and scale:
  - More than 8,000 programs exist at the federal, state and municipal levels, leading to duplication, redundancy, and fragmentation that reduce effectiveness and efficiency.
  - Some programs suffer from significant leakages to higher-income groups or other unintended beneficiaries.
- Targeting and registry improvements:
  - A social census covering about 20 million households (out of a total of 30 million households) was conducted to identify those in need.
  - Recommendation: if possible, use the social census to clean the existing beneficiary database—the Sistema de Información Social Integral (SISI)—and match SISI with the social census to create a single registry of beneficiaries to reduce errors of inclusion and exclusion, beneficiary overlaps, and program duplications.
- Size and composition of social assistance:
  - Administration spent about 1.8 percent of GDP on social assistance programs in 2018.
  - This level is comparable to other Latin American countries and EMDEs but is lower than the OECD average.
  - Transfers and PAM represented about 40 percent of the total social assistance received by an average household in Mexico and mostly benefit households at the bottom of the income distribution.
  - The 2019 and 2020 budgets increased social assistance spending by about 0.5 percent of GDP and changed composition—with a significant increase in social and disability pensions and a modest decrease in transfers.
  - The conditional cash transfer program Prospera has been canceled and its resources redistributed to other priority areas.
  - To reduce exclusion errors, authorities are targeting indigenous groups, elderly, and people with disabilities.
- PAM specifics:
  - The PAM benefits about 8 million individuals; 60 percent of those aged 65+, and 100 percent of those aged 68+.

### Pension coverage and old-age social assistance pressures
- Multiple pension systems cover private sector employees, different categories of civil servants at different government levels, SOEs, public universities and military personnel.
- Poverty and adequacy:
  - Poverty rate among people over 65 is very high, at more than 30 percent, in part due to insufficient benefits from the contributory pension system.
  - The average contribution rate of 6.5 percent in IMSS for the DC scheme is very low and may at best lead to a replacement rate of 26 percent for a full career average earner, the second lowest replacement rate among OECD countries (OECD 2019).
- Suggested adjustments to improve adequacy:
  - Increase the contribution rate—possibly offset in part by reducing the rate of contribution from wages to the housing fund (Infonavit).
  - Increase the effective retirement age by linking the statutory retirement age to gains in life expectancy.
  - Tighten early-retirement schemes.
  - Increase the contribution period required for a full pension in the old public-sector DB scheme.
  - Increase the age limit to get a full pension in the public sector faster (OECD 2015).

### Conclusions on fiscal strategy and spending efficiency
- Low levels of discretionary spending imply that raising revenues will be indispensable in putting public debt on a downward path.
- There is substantial room for tax policy and revenue administration reforms to raise revenues and ensure that public debt remains on a downward path.
- Spending efficiency should be strengthened to make additional room for priority expenditures.
  - The burden of fiscal adjustment in recent years has fallen on a small number of discretionary spending items such as capital spending, health and education, thus hurting inclusive growth.
  - Spending adjustments should concentrate on boosting spending efficiency, including in social spending.
  - Enhancing value for money in social spending is particularly important given additional spending needs to meet the SDGs in education and health.

### PEMEX’S TAXATION REGIME: objectives, recent reforms, and design issues
- Objective of the note:
  - Using the IMF FARI project-level cashflow modeling methodology, evaluate key characteristics of the current tax regime for PEMEX and compare it to announced reform plans and the production sharing regime applicable to recent licensing rounds.
- Short-term implications of reforms:
  - An increase in the cost cap and a reduction in the profit-sharing rate will reduce the overall tax burden for PEMEX and the regressivity of the regime, and by increasing return to PEMEX, may release funds for further investment.
  - Even with increased cost cap and reduced profit-sharing rate, the regime lacks sufficient progressive instruments to allow the government to share in upside from new developments.
  - In the longer term, migration of entitlement assets to newer more balanced contractual regimes would be beneficial.
- Recent reform actions (2019):
  - SHCP announced plans to ease the tax burden by loosening cost deduction caps under PEMEX’s ‘entitlement’ fiscal regime, aligning them with cost recovery limits of production sharing contracts concluded in 2015-18 and PEMEX’s Ek Balam area which transitioned to production sharing.
  - PEMEX business plan presented in July 2019 announced government plans to reduce the profit-sharing rate from 65 percent to 58 and 54 percent in 2020 and 2021, respectively.
- Background on the sector and PEMEX:
  - The oil industry was nationalized in 1938 and PEMEX was founded with exclusive rights until reforms in 2013-14 opened the sector to private companies.
  - Under ‘Round Zero’, SENER granted PEMEX rights over 83 percent of proven and probable reserves and 21 percent of Mexico’s prospective reserves.
  - PEMEX oil production declined from a peak in 2004 of 3.4 mbpd to 1.6 million barrels per day in 2018.
- Design trade-offs for petroleum fiscal regimes:
  - Balance is required between securing appropriate government revenue and maintaining investment incentives given uncertainty in production, prices, and costs.
  - Production-based instruments (e.g., royalties) provide early revenues but are regressive and should be moderate.
  - Progressive profit-based instruments capture more cashflows as profitability rises and help offset regressive instruments.
  - Modern regimes commonly include resource rent taxes, profit-based production sharing, and additional profits tax mechanisms.
- Entitlements and the entitlement fiscal regime:
  - An entitlement grants PEMEX the right to explore and produce hydrocarbons; PEMEX may conclude service contracts with private firms.
  - There are currently 428 entitlement agreements in place between SENER and PEMEX.
  - Principal components of the entitlement fiscal regime (prior to the changes contemplated by SHCP):
    - Profit Sharing Fee of 65 percent of production value less cost deductions, which are subject to an annual cost cap.
    - Hydrocarbons Extraction Fee, essentially a price linked ad-valorem royalty.
    - Corporate Income Tax at 30 percent of taxable income.
  - The entitlement regime also includes annual fixed surface area fees: the Hydrocarbon Exploration Fee, and the Tax on Hydrocarbon Exploration and Extraction Activity.
  - The profit-sharing fee and hydrocarbons fees are deductible expenses for the calculation of corporate income tax.
- Historical assessment (Box 1 summary from 2012 report):
  - The 2012 regime comprised instruments including the Stabilization Fund Duty (DSHFE) 10 percent royalty when oil prices exceeded $31 per barrel; Extraordinary Export Duty 13.1 percent royalty; an ordinary duty, a 71.5 percent income tax subject to a cost cap of $6.50/bbl; and special regimes for Chicontepec and deep water.
  - At an assumed oil price of $115 per barrel and industry benchmark costs of $7.5 per barrel, the 2012 fiscal regime generated average effective tax rates (AETR) in the range 70-80 percent.
  - The AETR was notably higher when assumed costs matched PEMEX’s costs of $17/barrel.
  - Sensitivity analysis showed the ordinary regime could generate AETR of over 100 percent over a $50-80 price range, rendering projects unviable at industry benchmark and PEMEX cost levels.
  - The 2012 report concluded that in the long run the cost cap should be eliminated and the fiscal regime aligned with normal IOC taxation; interim adjustments to the cap could be considered to reduce distortionary effects; and emphasis should be placed on narrowing the gap between PEMEX and industry benchmark costs and strengthening audit controls and independent oversight.

*International Monetary Fund*

### 12.      The cost caps associated with the profit-sharing fee reflect an effort to try and contain

### 12.      The cost caps associated with the profit-sharing fee reflect an effort to try and contain

### Cost caps, profit-sharing fee, and effective royalty
- Prior to the 2014 reform, cost caps were calculated annually in absolute monetary terms (agreed annual portfolio–wide expenditures by PEMEX divided by the number of barrels expected to be produced in the year).
- The 2014 reform introduced caps expressed as a percentage of revenue in the Hydrocarbons Revenue Law:
  - Cost caps range from 12.5 percent to 80 percent depending on location and hydrocarbon type.
- Drivers of the relatively low cost caps:
  - Low operating costs associated with the Cantarell field.
  - Government experience of cost inflation issues in PEMEX.
- The minimum government share of project revenues (the ‘effective royalty rate’) under the entitlement regime results from the Hydrocarbon Extraction Fee combined with the cost cap limit and the profit-sharing fee.
- The cap on cost deductions functions like a royalty by securing up-front revenues once production starts, ensuring a minimum quantity of production revenue is subject to the profit-sharing fee.
- Trade-off: A high minimum government share increases investor-perceived risk because less petroleum is available for cost recovery and payback/recovery periods lengthen.

### Fiscal terms (entitlement regime) — specific parameters
- Production Sharing Fee (percent of value of hydrocarbons): 65 percent
- Cost Deductions:
  - 100 percent expensing of exploration costs, 4 year straight-line depreciation of development costs
- Cap on Cost Deductions (percent of value of hydrocarbons):
  - Onshore Oil: 12.50 percent
  - Offshore Oil Shallow: 12.50 percent
  - Offshore Oil Deep: 60 percent
  - Natural Gas: 80 percent
  - Chicontepec: 60 percent
- Hydrocarbon Extraction Fee (percent of value of hydrocarbons):
  - When oil price is less than $48 per barrel, 7.5 percent.
  - When oil price greater than or equal to $48 per barrel, (12.5 percent*Petroleum price) + 1.5 percent
- Corporate Income Tax:
  - Rate: 30 percent
  - Depreciation: 100 percent immediate expensing of exploration costs, 4 year straight-line depreciation of development costs

### Minimum government revenue (Table 2) — selected figures
- Minimum Royalty (percent of Production Revenue) (base): 7.5 percent across listed regions
- Cost Cap (percent of Production Revenue):
  - Onshore: 12.5
  - Offshore Shallow Water: 12.5
  - Offshore Deepwater: 60
  - Natural Gas: 80
  - Chicontepec Paleochannel: 60
- Profit Sharing Fee (percent of Production Revenue - Costs):
  - Onshore: 65
  - Offshore Shallow Water: 65
  - Offshore Deepwater: 65
  - Natural Gas: 65
  - Chicontepec Paleochannel: 65
  - Proposed Reform (Chicontepec row shows profit sharing 54 percent)
- Effective Royalty Rate (percent) (as computed in Table 2):
  - Onshore: 60.1
  - Offshore Shallow Water: 60.1
  - Offshore Deepwater: 31.6
  - Natural Gas: 19.5
  - Chicontepec Paleochannel: 31.6
  - Proposed Reform (Chicontepec / Paleochannel): 27.5
- Note in source: at current price levels of approximately $60/barrel, the royalty rate would be higher than the minimum rate, at around 9 percent ((0.125*60)+1.5). On July 3, 2019, the Mexican basket price was $59.33 per barrel.
- International comparison: median effective rate in a sample of 56 countries surveyed is approximately 20 percent. Mexico’s effective royalty rate for oil under the entitlement regime is far above international norms.

### Ringfencing
- Payments under the entitlement regime and income tax are ringfenced by five regional classifications: onshore, shallow water, deep-water, non-associated natural gas, and Chicontepec Paleochannel.
- Regional ringfencing (rather than licence/asset-level ringfencing) implications:
  - May reduce the impact of the cost caps and defer government revenue.
  - Without tight licence-level ringfencing, PEMEX can deduct exploration or development costs for each new project against income of producing projects (unless constrained by the cost cap), delaying profit-based government revenue.

### Contractual regimes and alignment with PSCs
- Post-reform legal framework allows multiple contract types: license contracts, production-sharing contracts (PSCs), profit-sharing contracts, and service contracts—each implying different fiscal regimes.
- Some terms are set under Hydrocarbons Revenue Law; others are biddable variables in tenders. Each awarded contract may have a slightly different fiscal regime.
- SHCP is seeking to align PEMEX cost caps with cost recovery limits of recent PSCs:
  - Cost recovery limit under PSCs set at 60 percent for oil and 80 percent for gas (aligns with entitlement deep-water oil and natural gas caps).
  - Ek-Balam contract referenced: cost recovery set at 60 percent.
- Bidding results:
  - Average first tier (government) profit share bid was 54 percent for shallow water oil PSCs—this aligns with the proposed profit-sharing rate in 2021.
  - Bids varied widely; further analysis of specific signed contract terms recommended.

### Project example — stylized shallow water offshore field (project parameters)
- Project size: 500 MMBbl
- Project life: 18 Years
- Project costs (total):
  - Exploration costs: $MM 400; $/Bbl 0.80
  - Development costs excluding drilling: $MM 2,208; $/Bbl 4.42
  - Drilling costs: $MM 1,977; $/Bbl 3.96
  - Operating costs: $MM 3,998; $/Bbl 8.00
  - Decommissioning costs: $MM 558; $/Bbl 1.12
  - Total costs: $MM 9,141; $/Bbl 18.3
- Oil Price used in analysis: $60/Bbl (constant; source notes USD 60 per barrel used based on current trends and expectations)
- Pre-tax IRR: 35.0 percent

### Economic modeling approach and assumptions
- Modeling framework: FAD’s FARI modeling framework with a stylized offshore oil field example (reflecting broad cost structure of prospects such as Ek-Balam).
- Costs reflect levels reported in a PEMEX investor presentation published in November 2018.
- Analysis considers variations in price and cost; further refinement contingent on more detailed project economics (including for natural gas).
- Analysis conducted on a project level; regional consolidated ringfencing benefits (ability to offset new costs against revenue from producing fields) are not modeled.

### Revenue generating capacity — Average Effective Tax Rate (AETR) and project viability
- AETR defined as ratio of government revenue from a profitable project to the project’s pre-tax net cash flows; calculated undiscounted and discounted (discount rate 10 percent).
- Key modeled outcomes (Table 4: Project Fiscal Results; all figures in real 2019 terms):
  - Pre-Tax project IRR: 35.0% (for Entitlement Regime, Proposed Reforms, Production Sharing Regime)
  - Post-tax IRR on total funds:
    - Entitlement Regime: 6.8%
    - Entitlement Regime - Proposed Reforms: 18.0%
    - Production Sharing Regime: 18.5%
  - Post-tax IRR on equity:
    - Entitlement Regime: 9.0%
    - Entitlement Regime - Proposed Reforms: 28.9%
    - Production Sharing Regime: 29.6%
  - Pre-tax NCF undiscounted: 20,844 (same across regimes)
  - Post-tax investor NCF undiscounted:
    - Entitlement Regime: 1,944
    - Entitlement Regime - Proposed Reforms: 5,751
    - Production Sharing Regime: 5,839
  - Government Revenue undiscounted:
    - Entitlement Regime: 18,779
    - Entitlement Regime - Proposed Reforms: 14,972
    - Production Sharing Regime: 14,884
  - AETR undiscounted:
    - Entitlement Regime: 90.1%
    - Entitlement Regime - Proposed Reforms: 71.8%
    - Production Sharing Regime: 71.4%
  - Pre-tax NCF (10% discount): 5,267 (same across regimes)
  - Post-tax investor NCF (10% discount):
    - Entitlement Regime: -350
    - Entitlement Regime - Proposed Reforms: 967
    - Production Sharing Regime: 1,024
  - Government revenue (10% discount):
    - Entitlement Regime: 5,510
    - Entitlement Regime - Proposed Reforms: 4,193
    - Production Sharing Regime: 4,136
  - AETR (10% discount):
    - Entitlement Regime: 104.6%
    - Entitlement Regime - Proposed Reforms: 79.6%
    - Production Sharing Regime: 78.5%
- Interpretation:
  - Under the entitlement regime with a 12.5 percent cost cap, the example project is clearly unviable (AETR well over 100 percent; investor IRR 6.8 percent).
  - Proposed SHCP reforms (increasing cost cap and lowering profit-sharing rate to 54 percent) reduce regressive impact and improve viability:
    - Discounted AETR falls to 79.6 percent.
    - Investor IRR rises to 18.0 percent.
  - PSC regime produces an AETR of 78.5 percent and post-tax IRR of 18.5 percent; slightly lower AETR than the reformed entitlement regime due to cost uplifts that improve investor relief during investment recovery.

### Government revenue profile and timing
- Under the entitlement regime government captures significant revenues from commencement of production (due largely to royalty and minimum production share/profit sharing fee).
- Raising the cost cap to 60 percent and lowering the profit-sharing rate to 54 percent reduces the heavy early-year government take and provides relief to investors during investment recovery.
- Production sharing (PSC) with cost uplifts similarly provides additional early-year relief to investors through more generous cost recovery.

### Policy implications and recommended focus areas (as presented in text)
- The current effective royalty rates for onshore and shallow water oil operations are highly regressive and may discourage investment (especially for marginal or less profitable projects).
- SHCP’s proposal to increase the cost cap and lower the profit-sharing rate would reduce the regressive impact for onshore and shallow water offshore oil projects and may facilitate investment on commercial terms (reducing minimum government share of revenue to 27.5 percent in the relevant proposed reform calculation).
- Aligning PEMEX cost caps with cost recovery limits in PSCs (60 percent oil, 80 percent gas) reduces divergence across regimes and is the apparent focus of SHCP’s recent announcement.
- Further analysis recommended:
  - Assess specific signed contract terms (variation in bids and contract mechanics, e.g., Ek-Balam).
  - Consider the impact of regional consolidated ringfencing treatment on investor returns and government revenue timing.
  - Analyze licence contracts and their potential alignment with PSC cost recovery limits.

*Italic: Source — Excerpt from IMF staff analysis in the provided content unit.*

### 33.      The analysis also compared the relative burden that the different options would put

### 33.      The analysis also compared the relative burden that the different options would put

### Breakeven price and investor hurdle
- Key indicator: the “breakeven price” or the minimum price required to meet the minimum after-tax rate of return required by the investor (assumed in the model to be 12.5 percent in real terms).
- Entitlement regime:
  - Breakeven price is USD76.7/barrel.
  - Driven by its highly regressive nature, the breakeven price is well above current price levels.
- Entitlement regime with increased cost cap and lower profit-sharing rate:
  - Breakeven prices move more in line with current market trends and expectations (breakeven price figures illustrated but not restated beyond comparative description).
- Production sharing regime:
  - Displays breakeven prices more in line with current market trends and expectations.
- Source for breakeven price calculations: Staff calculations (Figure 5).

### Progressivity (AETR behavior across project outcomes)
- Method:
  - AETR variation obtained by varying oil prices and unit costs to produce a range of project pre-tax IRRs.
- Findings:
  - Increasing the cost cap and reducing the profit-sharing rate reduces regressivity of the entitlement regime, but without substantive progressive components the AETR falls as profitability increases.
  - Under the PSC regime:
    - The AETR initially falls as profitability increases due to regressive components.
    - Progressive components counteract that effect once profitability increases enough to trigger higher tiers of the production sharing mechanism.
  - Implication:
    - A wider range of projects could be developed commercially under the reformed entitlement regime and the PSC regime.
    - Government take from an individual project may be lower under the PSC than the entitlement regime at lower project profitability levels, but such projects might not be developed under the current entitlement regime (resulting in no government revenue).
- Source for progressivity charts: Staff calculations (Figure 6).

### International comparison of fiscal regimes
- Sample includes: Angola, Norway, Indonesia, Colombia, Brazil, China (as indicated in figures).
- Findings:
  - Entitlement regimes place a significantly higher burden on projects than other countries in the sample.
  - The PSC and the reformed entitlement regime place Mexico better in line with the sample in terms of neutrality while maintaining a comparable government share of revenue.
  - In terms of progressivity:
    - The PSC regime places Mexico in line with other regimes that link production sharing to profitability indicators (e.g., Angolan rate-of-return linked production sharing) or those with additional profit tax mechanisms (e.g., Norwegian Special Petroleum Tax).
- Representative breakeven and AETR figures shown in Figures 7 and 8 (Staff calculations). Specific breakeven prices cited in figures include: 76.7, 43.9, 43.8, 43.6, 42.0, 40.1, 39.6, 36.5, 36.0 (as presented in the source charts).

### Observations and policy recommendations
- Short-term effects of reforms:
  - Increasing the cost cap and reducing the profit-sharing rate will reduce regressivity.
  - These reforms would increase the return to PEMEX, improving its ability to undertake new onshore and shallow water oil projects commercially and increase available cashflow for additional investment.
- Longer-term considerations:
  - Even with increased cost cap and reduced profit-sharing rate, the regime lacks sufficient progressive instruments to allow the government to share in upside from new developments.
  - Migration of entitlement assets to newer, more balanced contractual regimes would be beneficial in the longer term, although this would come with some revenue loss to the government.
- If cost caps are increased, mechanisms to mitigate the risk of cost inflation should include:
  - (i) careful screening by CNH of PEMEX’s projects, budgets and work plans;
  - (ii) regular high-quality cost and fiscal audits (which should be required by CNH and SAT, the Tax Administration Service);
  - (iii) competitive, transparent procurement procedures for subcontractor services.
- Additional mechanism:
  - Through the migration process, private sector participation through farmouts can provide a mechanism for cost oversight and incentivize cost containment.

*Source: Staff calculations and analysis as presented in the provided IMF chapter excerpt.*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2019/1mexea2019002.pdf_