## REMITTANCES AND HOUSEHOLDS’ BEHAVIOR: THE CASE OF GUATEMALA

## Source details

**Canonical URL:** [REMITTANCES AND HOUSEHOLDS’ BEHAVIOR: THE CASE OF GUATEMALA](https://www.imf.org/-/media/files/publications/cr/2018/cr18155-guatemalasi.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2018/cr18155-guatemalasi.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2018/cr18155-guatemalasi.pdf.json)

---

### Introduction and key context
- Remittances are a structural feature of the Guatemalan economy:
  - In 2017, remittance flows accounted for over 11 percent of GDP.
  - Remittances benefitted over 1½ million Guatemalan households (over 6 million Guatemalans).
  - Dependence highest in San Marcos, Huehuetenango, and Quetzaltenango.
- Research questions:
  - Do remittance-receiving households spend more (or less) at the margin on food, consumer durables, and education relative to non-receiving households?
  - Do remittances affect households’ labor supply decisions?

### Data and sample
- Data source: 2014 National Household Survey of Living Conditions (ENCOVI).
- Sample coverage:
  - 11,498 urban and rural households in total.
  - 10,389 receive international remittances; 1,109 receive no remittances.
- Selected summary household characteristics (exact reported values):
  - Household head is indigenous: receive remittances 0.31; receive no remittances 0.36; total 0.35.
  - Mean age of household head (years): receive remittances 49.77; receive no remittances 45.84; total 46.19.
  - Area (1=urban): receive remittances 0.48; receive no remittances 0.55; total 0.54.
  - Mean monthly per capita income (incl. remittances) in quetzals: receive remittances 1,032; receive no remittances 1,128; total 1,120.
  - Remittances as percent of total per capita income (incl. remittances): receive remittances 46.5; receive no remittances 10.00; total 4.14.
  - Observations: receive remittances 10,389; receive no remittances 1,109; total 11,498.
- National migration/remittance aggregates (exact series shown):
  - Total population (2017): 16,923,483.
  - Population residing abroad (2017): 2,374,405.
  - Emigration rate (percent) (2017): 14.0.
  - Remittances (USD, thousands) (2017): 8,192,213.1.
  - Remittance flow per immigrant (USD) (2017): 3,450.2.
  - Annual average growth (series): 2.5%-3.9% (earlier), 5.0%, 2.7%, 4.7%, 9.4%, 4.0%, 10.9% (latest entries shown).

### Households’ spending patterns and marginal budget shares
- Survey-average budget shares (exact values):
  - Food: receive remittances 0.43; receive no remittances 0.46; total 0.46.
  - Durable consumer goods: receive remittances 0.04; receive no remittances 0.03; total 0.03.
  - Housing: receive remittances 0.14; receive no remittances 0.14; total 0.14.
  - Education: receive remittances 0.06; receive no remittances 0.04; total 0.05.
  - Health: receive remittances 0.01; receive no remittances 0.01; total 0.01.
  - Other goods: receive remittances 0.32; receive no remittances 0.31; total 0.31.
  - Observations used for expenditure-share calculations: receive remittances 1,109; receive no remittances 10,389; total 11,498.
- Marginal Budget Shares (MBS) — mean response to one-dollar increase (exact values from Table 4):
  - Food: receive remittances 0.35; receive no remittances 0.40.
  - Durable consumer goods: receive remittances 0.06; receive no remittances 0.04.
  - Education: receive remittances 0.08; receive no remittances 0.04.
- Empirical approach:
  - Expenditure modeled using a Working-Leser specification in expenditure-share form with additivity (∑ Ci/EXP = 1).
  - MBS derived from the model capture the response of budget shares to a one-dollar increase in household expenditure conditional on household characteristics Zj.
- Main findings on spending:
  - At the mean, remittance-receiving households allocate each additional dollar:
    - Less to food (MBS 0.35 vs 0.40 for non-recipients).
    - More to education (MBS 0.08 vs 0.04).
    - More to durable goods at the margin (MBS 0.06 vs 0.04).
  - The education-favoring marginal pattern for remittance recipients holds across income quintiles.
- Interpretation:
  - Spending bias toward education interpreted as remittances favoring human capital accumulation.
  - Higher marginal spending on durables may reflect alleviation of liquidity constraints or preferences for status goods.
  - Because recipients have higher per capita income, regression controls were applied to address potential confounding (see Annex I).

### Implications for growth: education, labor supply, and domestic returns
- Two mechanisms determine whether remittance-financed education yields domestic growth:
  - Whether remittances induce disincentives to work (labor supply effect).
  - Whether additional education funded by remittances is rewarded in the domestic labor market.
- Labor supply empirical specification:
  - Conventional specification estimated (equation (3)) with individual socioeconomic and locational controls Zj.
  - Age groups analyzed: 15−25, 26−40, 41−65.

### Labor supply estimation results
- Main empirical results:
  - No evidence of remittance-induced work disincentives.
  - Labor supply for members of remittance-receiving households is relatively more elastic, most markedly for age 41−65.
  - Elasticities (reported as percentual responses to a 1 percent wage increase):
    - Remittance-receiving households: a one percent increase in weekly wages leads to a ½ percent increase in weekly hours worked (noted as 0.5 percent).
    - Non-remittance households: a one percent increase in weekly wages leads to a 0.2 percent increase in weekly hours worked.
- Weekly salary coefficients from labor supply regressions (selected exact coefficients and significance):
  - Households receiving remittances — Weekly salary coefficients:
    - Age 15-25: 0.357***.
    - Age 26-40: 0.324***.
    - Age 41-65: 0.463***.
  - Households receiving no remittances — Weekly salary coefficients:
    - Age 15-25: 0.277***.
    - Age 26-40: 0.275***.
    - Age 41-65: 0.219***.
- Labor market context and constraints (exact figures):
  - Unemployment rate: 3.1 percent of the labor force.
  - Informality: 70 percent of total employment (40 percent in Latin America).
  - Share of employed workers receiving salaries below the minimum wage: 54 percent.
- Policy recommendations to improve labor market outcomes:
  - Tackle informality and increase functional flexibility to enhance participation in the domestic labor market.
  - Combat informality via:
    - (i) Increases in social assistance combined with minimum wage and productivity-enhancing measures for workers earning salaries below the minimum wage (e.g. high-return rural roads and irrigation).
    - (ii) Extended health and pensions coverage for the self-employed.
    - (iii) Tighter tax controls for non-compliances.
  - Minimum wage: consider bringing it closer to regional standards by adopting objective criteria (inflation indexation and some appropriation of economy-wide productivity gains).
  - Functional flexibility and training:
    - Expand apprenticeship schemes such as Mi Primer Empleo and vocational training.
    - Transpose ILO Convention 175 into national legislation to secure rights for part-time employees and set remuneration proportional to effective time worked.
  - Education and part-time notes:
    - Out of total employed workers: one fourth have no education, 30 percent have not completed primary school (90 percent of which are illiterate), and only 3 percent have a secondary school degree (ENEI).
    - Part-time employment (< 48 hours/week) is enshrined in the Labor Code and is to be remunerated at the full ordinary daytime weekly wage.

### Macroeconomic analysis: remittances, shocks, and FSGM simulations
- Methodology:
  - Western Hemisphere Module (WHDMOD) of the Flexible System of Global Models (FSGM) used to simulate tail-risk shocks and transmission channels.
- Risk Event #1 — Pronounced falls in remittances (tail-risk):
  - Shock assumption: immediate and permanent decline in remittances by one-third.
  - Quantitative macro impacts (exact reported numbers):
    - Output: lower by 1 percent after five years; lower by 2 percent in the long run (SS).
    - Private consumption: declines by 8 percent after 5 years; declines by 10 percent in the long run.
    - Real exchange rate: depreciates by just over 12 percent on impact; long-run roughly 5 percent below previous level.
  - Transmission and policy response:
    - Loss of remittance income reduces domestic absorption and imports; real exchange rate depreciation raises exports.
    - Headline inflation temporarily rises due to depreciation but core and headline inflation fall below baseline as demand weakens.
    - Monetary authority lowers the nominal policy rate consistent with its mandate to return inflation to target.
- Variant: Transitory risk premium increase and policy trade-offs:
  - Additional assumption: transitory increase in the risk premium of 150 basis points for corporates, households, and the sovereign.
  - Effects: further declines in real private investment, consumption, and imports; additional exchange rate depreciation.
  - Alternative monetary policy (leaning against depreciation) would require significant tightening, further depressing activity and possibly inducing domestic disinflation/deflation if nominal depreciation is prevented.
- Risk Event #2 — Increased deportations:
  - Assumption: one-fifth of illegal immigrants deported over four years (implying a 15-percent drop in total stock of Guatemalan immigrants in the U.S., given 75 percent of immigrants are illegal).
  - Quantitative macro impacts (exact reported numbers):
    - Real GDP: declines by almost 0.2 percent after five years; declines by 0.8 percent in the long run.
  - Notes:
    - Scenario treated as a lower-bound because returning migrants may enter the labor force over time.
    - Policy suggestion: expand or review programs such as “Bienvenido a casa” and “Quedate” to integrate returned workers.

### Food prices and inflation dynamics
- Food inflation as main driver of headline inflation:
  - Food inflation contributed about two thirds to total inflation since 2012 despite having weight of less than one third in the CPI.
  - CPI Division weights (in percent):
    - Food and Nonalcoholic Beverages 28.75; Housing, Rent, Water, Electric/Gas 12.61; Transportation 10.43; Restaurants and Hotels 9.24; Clothing and Footwear 7.41; Miscellaneous/Other Goods and Services 7.16; Recreation & Culture 5.62; Household Furnishings & Equipment 5.43; Communications 5.15; Health 4.22; Education 3.72; Alcoholic Beverages & Tobacco 0.28; Total 100.
  - Recent months: food inflation has eased and appears on a downward trend; cross-country declines in Central America indicate country-specific supply and demand factors have driven high food inflation in Guatemala since at least 2013.
- Univariate and Phillips-curve evidence (sample 2003Q1−2017Q3):
  - Stationarity: headline, core, and food inflation are stationary (ADF rejects unit root at 99 percent).
  - Sum of autoregressive coefficients (k=5):
    - Core inflation: 0.71 (high inertia).
    - Food inflation: 0.26 (lower inertia).
  - Phillips-curve findings:
    - Headline inflation explained by backward and forward-looking inflation, the output gap (large impact), and the policy rate.
    - Core inflation driven by past core inflation (coefficient 0.67), output gap, and terms of trade; wage growth not significant for core.
    - Food inflation reflects demand (remittances and salaries) and supply (rain and international food prices); remittances and lagged international food prices and drought are significant.
  - VAR results:
    - Food inflation shocks die off after about three quarters; core inflation shocks die off after about seven quarters.
    - Positive shocks to international food prices, remittances, and output gap increase food inflation; remittances shock effect vanishes after about five quarters.
- Policy implications:
  - High persistence of core inflation and limited second-round effects from food to core inflation.
  - Wages are an important driver of food inflation.
  - Given a slowing economy, core inflation at the lower end of target, restrictive fiscal policy in 2017, inflation risks to the downside, well-anchored expectations, and core inflation expected to remain below midpoint of the central bank’s target range, there could be scope to provide further monetary support.
  - Central bank action noted: cut policy rate by 25 basis points in November 2017.
- Caveats:
  - Results sensitive to small number of observations and to CPI rebasing to December 2010 (two CPI methodologies over period).

### Structural policies: competition, transport, and investment
- Competition and food prices:
  - Recommendation: adopt the new competition law and establish a competition authority promptly to analyze competition in food and transport and sanction monopolistic conduct.
  - Rationale: relatively higher trade and transport margins in food prices suggest anti-competitive behavior.
- Investments in transport infrastructure:
  - Recommendation: prioritize transport investments to lower food prices and facilitate rural access to national markets.
- Investment, confidence, and uncertainty (period 2000−16 and monthly VAR/VECM M092004−M092017):
  - Long-run averages: investment ratio (percent of real GDP) 16.9; real GDP growth 3.5 percent.
  - 2016 investment ratio: 14.7 percent of GDP (below emerging market/developing economies 33.5 and LAC 21.7).
  - Main results from VAR/VECM:
    - Global factors explain over one third of investment residuals.
    - Confidence innovations account for nearly one sixth.
    - Uncertainty contributes marginally (< 2 percent), likely due to multicollinearity between confidence and uncertainty.
    - Shocks from labor costs and interest rates contribute about 12 and 11 percent to investment residuals, respectively.
- Policy implications:
  - Improve business environment, undertake integral fiscal reform (tax administration, tax policy, spending flexibility) geared to higher investment in physical and human capital.
  - Expand social protection and sustain governance improvements to raise investment and revenue mobilization.

### Governance, anti-corruption, and judicial integrity
- Anti-corruption efforts post-2015 political crisis:
  - Government reasserted control over tax and customs administration, reformed public procurement, and strengthened prosecutorial independence; GPO and CICIG dismantled suspected networks.
  - Since 2014, GPO reportedly dismantled 178 criminal organizations and confiscated assets for over USD28.5 million (Ministerio Publico, 2018).
- Financial intelligence and enforcement indicators (selected exact entries preserved):
  - STRs to IVE: 2011: 50; 2012: 64; 2013: 67; 2014: 56; 2015: 106; 2016: 1150 (as presented).
  - Other table entries preserved exactly as in source.
- Governance-to-growth regression results (pooled OLS, lagged corruption indicators):
  - Main finding: higher governance (less corruption) is associated with higher per-capita real GDP growth across four corruption indicators (selected coefficients reported):
    - ICRG corruption index (one-period lag): 0.22179*** (0.08034).
    - WB control of corruption: 0.70498*** (0.18733).
    - WEF ethics and corruption: 0.71305*** (0.22274).
    - WEF irregular payments & bribes: 0.53168* (0.28125).
  - Sample periods, observations, number of countries and R-squared reported in table.
- Estimated growth gains from reducing corruption:
  - Closing Guatemala’s governance gap with the sample average could raise real per-capita GDP growth between 0.2 and 0.8 percentage points, depending on the indicator.
  - Targeted improvements in trust in politicians, diversion of public funds, and favoritism reduction deliver largest gains.
- Reform avenues and recommendations:
  - Tax administration reforms: standardized customs valuation controls, effective control of VAT through improved taxpayer registration, publication of VAT-refund recipients, retain qualified SAT staff, integrate information systems across agencies, deepen inter-institutional agreements.
  - Public expenditures and procurement: expand transparency, consolidate fiscal reports, enhance Guatecompras transparency, reconcile procurement-auditing regulations, speed up spending execution while preserving governance, adopt sanctions against fraudulent procurement division.
  - Market regulation and business facilitation: ease construction licenses, e-government services for tax and business registration, arbitration in commercial code, reduce firms’ tax filing burden.
  - AML/CFT: enact risk-based supervision of financial institutions and DNFBPs, increase ultimate beneficial owner transparency, ensure timely access to ownership information, implement FATF recommendations.
  - Enforcement and CICIG: sustain collaboration between CICIG and Prosecutor’s Office; CICIG’s mandate up for renewal in September 2019 (noted in text).

### Judicial integrity and asset disclosure reforms
- Judicial integrity findings:
  - Investigation and prosecution increased but convictions have not risen substantially.
  - Need to enhance transparency in appointment of judges and magistrates and strengthen disciplinary regimes.
  - Recommendations:
    - Revise nominations committee appointment procedures and increase scrutiny.
    - Implement transparent merit-based system for judicial career advancement.
    - Publish judicial decisions online.
- Asset disclosure system (Table 4 figures preserved exactly):
  - Annual figures 2014–2017:
    - Received: 2014 31,128; 2015 17,943; 2016 25,948; 2017 18,967.
    - Verified: 269 220 428 282.
    - Submitted to the GPO: 23 16 208 293.
    - Dismissed by the GPO: 6 2 18 0.
    - Administrative sanctions: 31 36 49 45.
  - Recommended reforms:
    - Electronic filing of asset disclosures.
    - Access to banking information for verification.
    - Online publication of asset disclosures of all public officials (requires legal reforms).
- Legal framework gaps:
  - Draft reforms awaiting parliamentary debate include criminal code, Comptroller’s Office law, Public Procurement Law, and law on Judicial Power.
  - Additional needed laws: reform to the Law on Appointment Committees, reform to the Public Probity law, new AML/CFT law.

### Fiscal policy, incidence, and poverty
- Methodology:
  - Static incidence analysis on ENCOVI 2014 using five income concepts: Market income, Net market income, Disposable income, Consumable income, Final income (with monetized in-kind transfers).
- Baseline inequality and fiscal impact (exact reported figures):
  - Market income Gini coefficient: 47.6 percent.
  - Fiscal policy (market income → final income) reduces Gini by 2.8 percentage points; final income Gini just below 44.8 percent after monetized education and health reduce inequality by an additional 1.7 percentage points.
  - Component effects on Gini (percentage-point changes):
    - Direct taxation: net income Gini decreases by 0.7 percentage points.
    - Government transfers: reduce inequality by 0.2 percentage points.
    - Indirect taxes/subsidies: reduce inequality by 0.1 percentage points.
    - Monetized education and health: additional decline of 1.7 percentage points.
- Poverty impacts (percent of total population, exact reported baseline):
  - Market income poverty: 58 percent.
  - Consumable income poverty: about 61 percent (indirect taxes increase poverty).
  - Reform scenario:
    - Reform would not materially change poverty and inequality compared to baseline.
    - Official revenue potential of reform: 1½ percent of GDP.
    - Consumable income poverty marginally increases under the reform due to higher indirect taxes.
- Incidence and distributional findings:
  - Under current system:
    - Approximately 93 percent of total direct taxes is paid by the wealthy and middle class.
    - 60 percent of total indirect taxes is paid by the wealthy and middle class.
    - Almost 90 percent of total direct transfers is received by the poor and vulnerable.
    - Education benefits: 80 percent to poor and vulnerable groups.
    - Health benefits: almost 50 percent to the middle class and wealthy (contributory coverage).
  - Contributory pensions increase measured inequality because low-income workers lack contribution histories.
  - Mi Bono Seguro (conditional cash transfer): budget 0.06 percent of GDP in 2015; coverage less than 20 percent of poor and less than 30 percent of extremely poor in 2015.
- Policy recommendations on fiscal design:
  - Maximize revenue potential while mitigating negative incidence on the poor.
  - Prioritize PIT reform (less negative effect on lower-income households) while recognizing indirect taxes may be necessary to raise revenue.
  - Use additional revenues to finance well-targeted social spending (expand conditional cash transfers) or investment in infrastructure and human capital to boost growth and reduce poverty indirectly.
  - Sequence reforms: expand social assistance (conditional cash transfers) first, accompany with infrastructure investment and measures to expand formality.

### Banking sector structure, vulnerabilities, and stress tests
- Balance-sheet and funding structure (exact reported averages):
  - Loans ≈ half of banks assets; long-term bonds 14 percent; cash and T-bills 13 percent; other assets 18 percent.
  - Deposits ≈ 75 percent of total liabilities.
  - Foreign liabilities represent 30 percent of total liabilities (up from about 10 percent in 2010).
- Concentration and FX exposure:
  - Nearly 90 percent of total assets concentrated in six largest banks (out of 18); top three account for over 65 percent market share.
  - FX loans/total loans: 38.4 percent.
  - 44 percent of FX loans (or 17 percent of total loans) are to un-hedged borrowers.
  - Net FX exposure / capital: 14.2.
- Key soundness indicators (Percent, September 2017):
  - Total capital / RWA (CAR): 14.8.
  - NPLs (gross)/ total loans: 2.6.
  - Provisions/NPLs: 113.6.
  - Deposits/total liabilities: 73.4.
  - Deposits/total loans: 128.5.
  - ROA (after‐tax): 1.3.
  - ROE (after‐tax): 13.9.
  - Liquid assets/total assets: 27.0.
  - Liquid assets/short‐term liabilities: 65.8.
- Solvency stress test (single and combined shocks):
  - System-wide proportional NPL shock to 3.3 percent of total loans (mean + 2 s.d.) and sectoral shocks:
    - CAR would fall by 1.4 percentage points (system-wide) and 1.0 percentage point (sectoral) but remain above 13 percent on average.
  - Interest rate shock: nominal interest rate increase of 2 percentage points → valuation losses on sovereign bond holdings would reduce system-wide CAR by 3 percent; two banks’ CAR fall below 10 percent.
  - FX shock: 12 percent nominal depreciation → direct FX impact marginally improves CAR (+0.3 p.p.); indirect credit-risk reduces CAR by 0.9 p.p.; overall FX shock reduces CAR by 0.6 p.p.
  - Combined shock (system NPL, interest rate, FX): post-shock system CAR 9.9 percent; CAR change −4.9 p.p.; eight individual banks fall below 10 percent regulatory minimum; identified capital shortfalls ≈ 0.4 percent of 2016 GDP. If profits used, recapitalization needs fall to 0.1 percent of GDP and only three banks below 10 percent.
- Reverse stress test results (NPL levels required for CAR < 10):
  - System-wide CAR < 10: implied NPLs to increase to 11 percent of currently performing loans.
  - 1/2 market share with CAR < 10: implied NPLs to increase to 9 percent.
  - 9 banks with CAR < 10: implied NPLs to increase to 16 percent.
- Liquidity stress tests:
  - Short-term deposit-run scenario (10 percent/day demand withdrawals; 3 percent/day time deposits):
    - Post-shock liquid assets/total assets: 19.4.
    - Post-shock liquid assets/short-term liabilities: 47.2.
    - Number of liquid banks after 5 days: 18 (all banks remain liquid).
  - Medium-term FX funding stress (six largest banks; exact run-off/roll-off schedules and haircuts preserved in source):
    - System-level funding gaps at maturities up to one month under assumed run-off/roll-off rates; cash inflows exceed outflows beyond one month.
    - Counterbalancing capacity (maturing investments/loans, interbank loans, sale of securities, reserves) sufficient to cover initial gaps; after using capacity all banks in exercise remain liquid.
  - Policy recommendation: adopt Basel III LCR in U.S. dollars for monitoring to enhance liquidity risk management.

*Source: IMF staff analysis based on ENCOVI 2014, FSGM/WHDMOD simulations, VAR/VECM and stress-test estimations contained in the content unit.*

### References _____________________________________________________________________________ 14

### REMITTANCES AND HOUSEHOLDS’ BEHAVIOR: THE CASE OF GUATEMALA

### Introduction
- Remittances are described as a structural feature of the Guatemalan economy.
- Key statistics and contextual facts:
  - In 2017, remittance flows accounted for over 11 percent of GDP.
  - Remittances benefitted over 1½ million Guatemalan households (over 6 million Guatemalans).
  - Dependence on remittances is highest in the regions of San Marcos, Huehuetenango, and Quetzaltenango.
- Research questions posed:
  - Do remittance-receiving households spend more (or less) at the margin on food, consumer durables, and education relative to non-receiving households?
  - Do remittances affect households’ labor supply decisions?

### Data
- Data source: 2014 National Household Survey of Living Conditions (ENCOVI).
- Sample and coverage:
  - 11,498 urban and rural households in total.
  - 10,389 receive international remittances; 1,109 receive no remittances.
- Remittance questions in the survey:
  - (i) does your household receive remittances from people living abroad;
  - (ii) how much remittances did your household receive in the past 12 months?
- Summary household characteristics (selected exact values):
  - Household head is indigenous: receive remittances 0.31; receive no remittances 0.36; total 0.35.
  - Mean age of household head (years): receive remittances 49.77; receive no remittances 45.84; total 46.19.
  - Area (1=urban): receive remittances 0.48; receive no remittances 0.55; total 0.54.
  - Mean monthly per capita income (incl. remittances) in quetzals: receive remittances 1,032; receive no remittances 1,128; total 1,120.
  - Remittances as percent of total per capita income (incl. remittances): receive remittances 46.5; receive no remittances 10.00; total 4.14.
  - Observations: receive remittances 10,389; receive no remittances 1,109; total 11,498.
- National remittance aggregates and migration (exact series shown):
  - Total population (2017): 16,923,483.
  - Population residing abroad (2017): 2,374,405.
  - Emigration rate (percent) (2017): 14.0.
  - Remittances (USD, thousands) (2017): 8,192,213.1.
  - Remittance flow per immigrant (USD) (2017): 3,450.2.
  - Annual average growth (series): 2.5%-3.9% (earlier), 5.0%, 2.7%, 4.7%, 9.4%, 4.0%, 10.9% (latest entries shown).

### Households’ Spending Patterns on Consumption
- Survey-average budget shares (exact values):
  - Food: receive remittances 0.43; receive no remittances 0.46; total 0.46.
  - Durable consumer goods: receive remittances 0.04; receive no remittances 0.03; total 0.03.
  - Housing: receive remittances 0.14; receive no remittances 0.14; total 0.14.
  - Education: receive remittances 0.06; receive no remittances 0.04; total 0.05.
  - Health: receive remittances 0.01; receive no remittances 0.01; total 0.01.
  - Other goods (household services, transport, communications, legal and personal services): receive remittances 0.32; receive no remittances 0.31; total 0.31.
  - Observations: receive remittances 1,109; receive no remittances 10,389; total 11,498 (note: observation counts in this table reflect the breakdown used for expenditure-share calculations).
- Marginal Budget Shares (MBS) by expenditure category — mean responses to one-dollar increase in expenditure (exact values from Table 4):
  - Food: receive remittances 0.35; receive no remittances 0.40.
  - Durable consumer goods: receive remittances 0.06; receive no remittances 0.04.
  - Education: receive remittances 0.08; receive no remittances 0.04.
- Empirical approach:
  - Expenditure modeled using a Working-Leser specification in expenditure-share form (additivity ensured: ∑ Ci/EXP = 1).
  - Marginal budget shares derived from the model capture the response of budget shares to a one-dollar increase in household expenditure conditional on household characteristics Zj.
- Main empirical findings on spending patterns:
  - At the mean, remittance-receiving households spend less on food and more on education, for each additional dollar of income, compared with non-recipient households.
  - Remittance-receiving households spend more on durable goods at the margin than what they would have spent without remittances.
  - The education-favoring marginal pattern for remittance recipients holds across income quintiles when households are partitioned into quintiles based on total annual per capita income including remittances.
- Interpretation notes:
  - The spending bias toward education is interpreted as remittances favoring human capital accumulation.
  - The greater marginal spending on durables by recipients may reflect either the alleviation of liquidity constraints or preferences for status-oriented goods.
  - Because remittance-receiving households also have higher per capita income levels, the authors note a possible confounding effect of income level on the marginal shares and apply controls in regressions (Annex I contains regression results).

### Implications for Growth and Labor Supply (preview of subsequent analysis)
- The paper highlights two mechanisms determining whether remittance-financed education translates into domestic growth:
  - Whether remittances induce disincentives to work (i.e., reduce labor supply) — to be examined in the labor supply section.
  - Whether additional education funded by remittances is adequately rewarded by the domestic labor market (as opposed to foreign labor markets), which affects the growth dividend of remittance-financed human capital.
- The authors frame subsequent analysis to assess remittances’ effects on households’ labor supply decisions and the domestic returns to education.

*Source: IMF staff analysis based on the 2014 ENCOVI Survey and related tables and figures contained in the content unit.*

### 11.      Remittances may reduce labor market effort. Remittance receipt would be expected to

### 11.      Remittances may reduce labor market effort. Remittance receipt would be expected to 

### Theory: channels through which remittances affect labor supply
- Remittances are simple income transfers; recipients may substitute unearned remittance income for labor income because labor income must be earned through the expenditure of effort.
- Moral hazard problems may arise due to asymmetric information—distance separating remitter and recipient makes monitoring difficult—potentially inducing recipients to divert resources to the consumption of leisure (Chami, Fullenkamp, and Jahjah, 2003).

### Empirical specification used to estimate effects on labor supply
- A conventional specification was estimated (equation (3) in the source), where j denotes individuals and Zj denote individual socioeconomic and locational characteristics (family size, number of children, education, geographic region, rural-urban and ethnic divide) allowing for shifts in the intercept and/or slope of the demand functions.

### Estimation results and labor supply elasticities
- Equation (3) was estimated for age groups 15−25, 26−40, and 41−65.
- There is no evidence of remittance-induced work disincentives.
- Labor supply for members of remittance-receiving households is relatively more elastic, most markedly for the 41−65 age group:
  - A one percent increase in weekly wages leads to a ½ percent increase in weekly hours worked for members of remittance-receiving households.
  - For non-remittance-receiving households, a one percent increase in weekly wages leads to a 0.2 percent increase in weekly hours worked.
- Source data for the labor supply regressions: ENCOVI 2014 and Fund staff calculations.

### Key labor market and labor force facts (from the analysis)
- Unemployment rate: 3.1 percent of the labor force.
- Informality: 70 percent of total employment (40 percent in Latin America).
- Share of employed workers receiving salaries below the minimum wage: 54 percent.
- Reasons for non-contribution to Social Security (IGSS) and informality:
  - Legal restrictions (15 percent of workers, mostly self-employed, are not allowed to contribute to IGSS).
  - Inadequate control (12 percent of workers evade IGSS obligations).

### Policy recommendations to improve labor market outcomes
- General assessment:
  - Absent remittance-induced disincentives on households’ labor supply decisions, low education and employment outcomes primarily reflect inadequate labor market policies.
  - Tackling informality and increasing functional flexibility are important to enhance participation in the domestic labor market.
- Combatting informality requires a multi-pronged approach:
  - (i) Increases in social assistance combined with minimum wage and productivity-enhancing measures for workers earning salaries below the minimum wage (e.g. high-return rural roads and irrigation).
  - (ii) Extended health and pensions coverage for the self-employed.
  - (iii) Tighter tax controls for non-compliances.
- Minimum wage policy:
  - Guatemala’s minimum wage is one of the highest in Latin America.
  - Authorities should consider bringing it closer to regional standards by adopting objective criteria for its determination (as per ILO guidelines), typically inflation indexation and some appropriation of the economy-wide productivity gains.
- Functional flexibility and training:
  - Low education hinders hiring and limits mobility out of agriculture and informality.
  - Apprenticeship schemes such as Mi Primer Empleo and vocational training should be expanded.
  - To make part-time work a successful route to formal employment, particularly for women and the youth, the recently ratified ILO Convention 175 should be transposed to national legislation with aims to: (a) secure labor rights for part-time employees and (b) set remunerations in proportion to the effective time worked.
  - Note on education distribution: Out of the total employed workers, one fourth have no education, 30 percent have not completed primary school (90 percent of which are illiterate), and only 3 percent have a secondary school degree (ENEI).
  - Note on part-time employment: Enshrined in the Labor Code as a legitimate modality of employment, part-time employment (less than 48 hours per week) is to be remunerated at the full ordinary daytime weekly wage.

### Annex I — selected regression results (high-level)
- Household expenditure OLS estimates (separate tables for households receiving no remittances and households receiving remittances) report coefficients for categories: Food, Consumer goods, durables, Housing, Education, Health, Other. Examples of reported coefficients (households receiving no remittances):
  - Log total annual per capita household expenditure (γ): Food −0.064***; Consumer goods, durables 0.018***; Housing 0.023***; Education −0.003**; Health 0.004***; Other 0.021***.
  - R2 values: Food 0.335; Consumer goods, durables 0.185; Housing 0.191; Education 0.266; Health 0.032; Other 0.075.
- Labor supply regressions (Table 3) — Weekly hours of work regressions report coefficients and significance by age group and by whether households receive remittances:
  - Weekly salary coefficients for households receiving remittances: Age 15-25 0.357***; Age 26-40 0.324***; Age 41-65 0.463***.
  - Weekly salary coefficients for households receiving no remittances: Age 15-25 0.277***; Age 26-40 0.275***; Age 41-65 0.219***.
  - Other weekly income coefficients for households receiving remittances: Age 15-25 −0.016; Age 26-40 −0.072***; Age 41-65 −0.051.
  - Other weekly income coefficients for households receiving no remittances: Age 15-25 −0.024***; Age 26-40 −0.028***; Age 41-65 −0.024***.
  - Observations: Households receiving remittances — Age 15-25: 135; Age 26-40: 186; Age 41-65: 123. Households receiving no remittance — Age 15-25: 1,372; Age 26-40: 2,680; Age 41-65: 1,832.
  - R-squared values reported: e.g., households receiving remittances — Age 15-25 0.450; Age 26-40 0.342; Age 41-65 0.308. Households receiving no remittance — Age 15-25 0.186; Age 26-40 0.188; Age 41-65 0.151.
- Notes: *, **, and *** denote significance at the 0.1, 0.05, and 0.01 levels. Source of household and labor supply regressions: 2014 ENCOVI Survey. Instituto Nacional de Estadística; IMF staff calculations.

*Source: IMF staff analysis based on ENCOVI 2014 and IMF staff calculations (Annex I).*

### 7.      Food price increases are a main

### cr18155-guatemalasi - 7.      Food price increases are a main

### Food inflation as a main driver of headline inflation
- Food inflation contributed about two thirds to total inflation since 2012, despite having a weight of less than one third in the CPI.
- CPI Division weights (in percent):
  - Food and Nonalcoholic Beverages 28.75
  - Housing, Rent, Water, Electric/Gas 12.61
  - Transportation 10.43
  - Restaurants and Hotels 9.24
  - Clothing and Footwear 7.41
  - Miscellaneous/Other Goods and Services 7.16
  - Recreation & Culture 5.62
  - Household Furnishings & Equipment 5.43
  - Communications 5.15
  - Health 4.22
  - Education 3.72
  - Alcoholic Beverages & Tobacco 0.28
  - Total 100
- Recent months: food inflation has eased and appears on a downward trend; cross-country declines in Central America indicate country-specific supply and demand factors have driven high food inflation in Guatemala since at least 2013.

### C. Macroeconomic evidence — Univariate approach and persistence
- Sample and data:
  - Quarterly seasonally-adjusted series spanning 1Q2003−3Q2017.
  - With the exception of interest rates, all variables are in log first differences.
  - Series seasonally-adjusted using the Census X-13 procedure where needed.
- Stationarity and unit root tests:
  - Over the past 15 years, headline, core, and food inflation proved stationary.
  - Augmented Dickey-Fuller (ADF) tests: the null of unit root can be rejected at 99 percent confidence in headline, core, and food inflation.
  - Variance-ratio tests: variance-ratios of headline, core, and food inflation decrease over time; relative stability of variance ratio indicates some inflation persistence.
- Measures of inertia:
  - Sum of autoregressive coefficients (k=5):
    - Core inflation: 0.71 (evidence of high inertia; large weight on backward-looking component).
    - Food inflation: 0.26 (lower inertia).
  - Additional tabulated univariate evidence (quarterly data, 2003Q1-2017Q3):
    - CPI: 0.86 0.74 0.38 0.26  -3.91***
    - CPI, core: 0.98 0.75 0.48 0.51  -6.11***
    - CPI, food: 0.70 0.60 0.33 0.22  -4.80***
    - Note: *** means the null of unit root is rejected at the 99 percent level.

### Phillips curve regressions — drivers of different inflation measures
- Headline inflation specification:
  - Inflation related to output gap, intrinsic persistence, forward-looking component (one-year-ahead inflation expectations), and the policy rate.
  - Headline inflation is explained by both backward and forward-looking inflation in about equal proportions, by the output gap (large impact), and by the policy rate.
- Core inflation specification:
  - Driven by past core inflation, the output gap, and the terms of trade.
  - Past core inflation has a large and significant coefficient (0.67) pointing to high inertia of core inflation.
  - Including international energy and food prices reduces the output gap coefficient, but past core inflation coefficient remains large.
  - Wage growth does not appear significant for core inflation.
- Food inflation specification(s):
  - Food inflation reflects demand (remittances and salaries) and supply (rain and international food price) factors.
  - Key findings:
    - Food inflation is explained by past food inflation, remittances and lagged international food prices, and drought (lack of rain).
    - In an alternative specification, wage growth appears significant for food inflation with a large coefficient.
    - If salaries are included in the second specification, salaries and past food inflation are insignificant, while remittances and rain are significant at the 90 percent confidence level only.
- Nominal wage inflation specification:
  - Wage growth driven by past and current core inflation and past wage growth.
  - Results: nominal wage growth is partially explained by past or current core inflation and to a little extent by current food inflation; past food inflation becomes insignificant for wage growth.
- Summary of estimation results (quarterly data, 2003Q1-2017Q3):
  - Estimated coefficients generally have expected signs and are significant at the 95 percent confidence in most cases.
  - Inflation expectations: food inflation (or international food and energy prices) does not appear to influence inflation expectations strongly; potential second-round effects from food-price shocks contaminating core inflation should be limited.
  - Wages emerge as an important driver of food inflation.

### VAR approach — interactions and impulse responses
- Model setup:
  - Standard VAR with two lags; exogenous block includes international food prices (price index in U.S. dollars).
  - Endogenous vector Yt = {remit_t, neer_t, ygap_t, p_foodt, p_coret, w_t, i_t}.
  - Cholesky recursive identification; baseline orderings:
    - For food inflation: food_t → remit_t → neer_t → ygap_t → p_foodt
    - For core inflation: remit_t → neer_t → ygap_t → p_foodt → p_coret
- Baseline VAR results:
  - Food inflation:
    - Less persistent than core; shocks to food inflation die off after about three quarters.
    - Positive shocks to international food prices, remittances, and the output gap increase food inflation.
    - Impact from international food prices shock is large but dies off after four quarters.
    - Impact from remittances shock is large and vanishes after about five quarters.
  - Core inflation:
    - Highly persistent; shocks to core inflation die off after seven quarters in baseline specification.
    - Impact from a food inflation shock on core inflation is short lived (effect vanishes after two quarters).
    - Output gap impact on core inflation is more persistent (effect vanishes after five quarters).
- Robustness and alternatives:
  - Results robust to alternative Cholesky orderings (Alternative 1–4 described in source).
  - Alternative 3 and 4 show higher wages have a small persistent impact on core and food inflation.
- Caveats:
  - Number of variables relative to observations makes the system overdetermined; coefficient accuracy may be poor.
  - Two specifications with only 4 variables aim to reduce overdetermination.

### D. Conclusions and policy implications
- Key conclusions:
  - High persistence of core inflation and limited second-round effects from food to core inflation.
  - Food inflation exhibits a low degree of persistence: shocks to food inflation die off after about three quarters, versus seven quarters for core inflation.
  - Non-persistent component of core inflation is dominated by output gap dynamics; international food and energy prices play a lesser role.
  - Wages are an important driver of food inflation.
  - Potential for second-round effects from shocks to food prices on core inflation (via inflation expectations) appears limited.
- Policy-relevant observation:
  - After being on hold for two years, the central bank cut the policy rate by 25 basis points in November 2017.
  - Given a slowing economy (negative output gap), core inflation at the lower end of the target range, restrictive fiscal policy in 2017, inflation risks to the downside, well-anchored inflation expectations, and core inflation expected to remain below the mid-point of the central bank’s target range, there could still be scope to provide further monetary support to the economy.
  - Further easing would be particularly warranted if fiscal policy remains constrained in supporting demand.
- Limitations and caution:
  - Results are sensitive to the relatively small number of observations.
  - A rebasing of the CPI to December 2010 implies working with two distinct CPI methodologies over the analyzed period, which may lead to imprecise results.

*Source: IMF staff calculations and analysis in the chapter "Food price increases are a main" (cr18155-guatemalasi).*

### 29.      On structural policies, promoting competition and investments in transport

### 29.      On structural policies, promoting competition and investments in transport

### Competition and food prices
- Recommendation: Adopt the new competition law and establish a competition authority promptly.
- Rationale:
  - Relatively higher trade and transport margins in food prices compared to other countries in the region point to anti-competitive behavior.
  - The new competition authority should analyze competition in food and transport industries and sanction cases of monopolistic conduct.

### Investments in transport infrastructure
- Recommendation: Prioritize investments in transport infrastructure to lower food prices and facilitate access to national food and other markets for those living in rural areas.
- Rationale:
  - Investments in transport infrastructure will facilitate access to national food and other markets for those living in rural areas.

### Caveat on inflation estimates
- The analysis notes one caveat: the estimates are based on quarter-on-quarter changes in inflation—rather than year-on-year changes, and the latter represent the basis on which Banguat makes monetary policy decisions.

### Investment, confidence and uncertainty in Guatemala: key findings and statistics
- Context and long-run averages:
  - In the period 2000−16, the average investment ratio (percent of real GDP) and real GDP growth stood at 16.9 and 3.5 percent, respectively.
  - At 14.7 percent of GDP in 2016, Guatemala’s investment ratio was less than half the emerging market and developing economies’ (33.5 percent) and less than two thirds the Latin American and The Caribbean’s average (21.7 percent).
- Recent dynamics:
  - Gross fixed capital formation expanded at double-digit growth rates prior to the GFC, but has been underperforming in recent years; the overall investment rate has stagnated at about 14 percent of GDP.
  - The share of public investment in gross fixed capital formation has been narrowing and foreign direct investment (FDI), which had averaged about 4 percent of GDP in the first half of this decade, has been on a declining trend as well.
  - Over the short term, the outlook for investment remains subdued based on current policies.
- Drivers and measurement:
  - The paper examines the role of confidence and uncertainty using a VAR/VECM approach on a dataset comprising monthly series spanning M092004−M092017 (157 observations).
  - Confidence index: constructed from four questions in Banguat’s Encuesta de Expectativas Económicas; sub-indexes are constructed by subtracting negative-answer percentages from positive-answer percentages and adding 100; the confidence index is the average of the sub-indexes, expressed as a percent of the maximum possible average.
  - Uncertainty measure: follows Bachmann et al (2013): Uncertainty = root square [(%opt_t + %pes_t – (%opt_t - %pes_t)^2] and is built from answers to the first question in Banguat’s Economic Expectations Survey.
  - Several series are interpolated from quarterly to monthly frequency for the analysis; all series used are seasonally-adjusted except the uncertainty series and the real interest rate series where X-13 was incompatible.
- Econometric approach:
  - A standard VAR with an exogenous block for global conditions was specified; principal components of {wgdp_t, S&P 500_t, ffr_t, oil_t, vxo_t} summarize global conditions.
  - Identification via Cholesky recursive ordering; two alternative orderings tested (4a and 4b) differing in the causal relationship between GDP and confidence/uncertainty.
  - A VECM was estimated because variables are a combination of I(0) and I(1) and some cointegrating vectors exist.
  - Historical decomposition of the investment residual was used to determine major drivers.
- Main results (period 2005−17 and sub-periods):
  - Global factors explain over one third of the investment residuals; their innovations trended upwards since the end of the GFC.
  - The contribution of confidence accounted for nearly one sixth of the investment residuals; confidence innovations gradually increased after the GFC, peaking in early 2013, declining through 4Q2015, rose rapidly with the advent of the Morales administration in early 2016, and declined over the last couple of years amidst a politically-charged environment.
  - Uncertainty contributed marginally to investment residuals, at less than 2 percent; this likely reflects multicollinearity between the indicators of confidence and uncertainty used in the analysis (about -55 percent). Alternative VECM specifications using confidence and uncertainty separately resulted in contributions of 9.7 and 9.3 percent, respectively.
  - Shocks from labor costs and interest rates to investment residuals contribute about 12 and 11 percent, respectively; the contribution of GDP has been more muted in comparison.
  - Results are broadly robust to changes in Cholesky ordering and other specification variations.
- Tabled/period figures summarized (seasonally adjusted contributions to investment residuals, percent):
  - 2005−17: Global factors contributed the largest share; confidence larger than interest rates, labor costs or GDP; uncertainty marginal (see detailed decomposition in the source).
  - 2008−10 and 2016−17 sub-period decompositions are reported in the source tables and figures.

### Policy implications and recommendations
- The relatively high sensitivity of private investment to global conditions and investors’ confidence highlights the need for investment-enhancing policies.
- Suggested actions (multi-front, coordinated effort):
  - Improve business environment conditions.
  - Undertake an integral fiscal reform encompassing continued tax administration efforts, tax policy changes, and spending flexibility, all geared at higher investments in physical and human capital.
  - Expand social protection by creating more formal jobs in the economy.
  - Sustain efforts to improve governance, which is critical to durably raise investment and support efforts at revenue mobilization.

*International Monetary Fund — Guatemala country report content unit*

### 2.      A VAR with a combination of I(0) and I(1) series was estimated; thereafter a VECM was

### 2.      A VAR with a combination of I(0) and I(1) series was estimated; thereafter a VECM was

### VAR/VECM specification and data choice
- A VAR with a combination of I(0) and I(1) series was estimated; thereafter a VECM was applied since no cointegration was found.
- The VAR was initially run with quarterly data but degrees of freedom were low; averaging monthly data to obtain quarterly indicators lost variability information. For these reasons, the analysis switched to monthly data.
- Note: "It is usually suggested that if monthly data are used, the VAR, at least in its initial formulation, should have twelve lags. When the model was run with 12 lags, the most reliable test of lag length (Hanna Quinn) still suggest two lags."

### Diagnostic test results and residual properties
- Residual correlation matrix:
  - No particular high positive or high negative correlation between two variables was highlighted.
  - Highest correlation reported: confidence index (ci) and uncertainty (unc) at -54 percent (seasonally adjusted) and -49 percent (non-seasonally adjusted).
- Polynomial roots:
  - The VAR does not have polynomial roots outside the unit circle when seasonally adjusted.
  - Using non-seasonally adjusted series (NSAS) with 12 lags yielded 3 polynomial roots outside the unit circle, precluding convergence of the model towards equilibrium.
- Residual autocorrelation:
  - Autocorrelation of residuals (up to 12 lags) within +/- 2 standard deviations does not suggest autocorrelated errors for seasonally adjusted series.
  - Correlogram counts:
    - Non-Seasonally Adjusted Series: 12 out 36 beyond 2 s.d.
    - Seasonally Adjusted Series: 5 out 36 beyond 2 s.d.
- Lagrange Multiplier (LM) test for autocorrelation:
  - For NSAS: It does not reject Ho (no autocorrelation) only for the first lag at 95 percent confidence level.
  - For SAS: It does not reject Ho (no autocorrelation) for the first three lags.
- Lag length tests (initial number of lags = 12):
  - Non-Seasonally Adjusted Series: 12 lags (AIC), 12 lags (HQ), and 2 lag (SC).
  - Seasonally Adjusted Series: 4 lags (AIC), 2 lags (HQ), and 1 lag (SC).
  - When the VAR was estimated with 4 and 12 lags, the lag length tests (SC, AIC, and HQ) suggested in the case of the former that 2 lags would be appropriate, while in the case of the latter the most reliable test (HQ) still suggested 2 lags.

### Cointegration and VECM implementation
- Johansen Cointegration Test (trace) used to determine number of cointegrating equations.
- An unconstrained VECM with the appropriate number of cointegrating equations and lags was run (the number of lags in a VECM equals the number of lags when the model is run in levels minus one).
- Johansen trace test results (Table 1):
  - NSAS: 4 cointegrating equations.
  - SAS: 2 cointegrating equations.
- Because no cointegration was found overall in initial checks, a VECM was applied where appropriate.

### Seasonally Adjusted Series (SAS) versus Non-Seasonally Adjusted Series (NSAS)
- Test results for NSAS are generally weaker than for SAS.
  - Example: in correlogram, NSAS had one third of autocorrelations above 2 standard deviations, while SAS had one sixth.
- NSAS with 12 lags produced 3 polynomial roots outside the unit circle, preventing convergence.
- Residual plots:
  - NSAS residuals: randomly dispersed around horizontal axis; show some seasonality in a few cases.
  - SAS residuals: randomly dispersed around horizontal axis.

### Treatment of exogenous variables
- Exogenous block cannot be combined with endogenous variables to estimate overall impact on the VAR system because domestic endogenous variables cannot affect global exogenous variables by definition.
- Impulse estimators are used to calculate the impact of changes in exogenous variables:
  - A change in an exogenous variable in time t is transmitted to t+1, t+2 ... t+N-1 (N is the size of the sample).
  - Impact in each period is measured by impulse responses; overall impact of changes for each exogenous variable is the summation of N partial aggregations.

### Historical decomposition and contributions of structural innovations
- If the estimated innovation in the period had the same sign as the residual of the factor being explained, the contribution was acknowledged; otherwise it was omitted with a zero value.
- This procedure produced periods whose components for a specific period were all either positive (including zeros) or negative (including zeros) innovations.
- The absolute value was then applied and figures corresponding to innovations of each factor were added over time to avoid cancellation.
- Summations for each factor in each sub-period were then expressed as percent of the totals.
- This approach is an approximation because historical decomposition of the residual in a given period does not yield innovations that are all positive or negative.

### Selected model results (Table 1 summary, Period: 2005 M9 - 2017 M9; 157 Monthly Observations)
- Residual correlation:
  - NSAS: highest correlation between ci and unc @ -49 percent.
  - SAS: highest correlation between ci and unc @ -54 percent.
- Number of polynomial roots outside unit circle:
  - NSAS: 30 (as reported in table).
- Plot of residuals:
  - NSAS: Randomly dispersed; show some seasonality in a few cases.
  - SAS: Randomly dispersed.
- Autocorrelation +/- 2 Standard Deviations (Correlograms):
  - NSAS: 12 out 36 beyond 2 s.d.
  - SAS: 5 out 36 beyond 2 s.d.
- LM Test - Autocorrelation:
  - NSAS: It does not reject Ho (no autocorrelation) only for the first lag @ 95 percent confidence level.
  - SAS: It does not reject Ho (no autocorrelation) for the first three lags.
- Lag Length Tests (AIC, SC, HQ) - Initial Number of Lags = 12:
  - NSAS: 12 lags (AIC), 12 lags (HQ), and 2 lag (SC).
  - SAS: 4 lags (AIC), 2 lags (HQ), and 1 lag (SC).
- Johansen Cointegration Tests (trace):
  - NSAS: 4 cointegrating equations.
  - SAS: 2 cointegrating equations.

### Alternative Cholesky orderings — contributions to Investment Residuals (seasonally adjusted figures)
- Annex II presents percentage contributions to investment residuals under different Cholesky orderings for sub-periods and samples (examples given):
  - For ordering "glotpmulcgdpinvciunc" — Total = 100.0 with sample rows such as:
    - 2008-10: 35.3 17.6 12.5 0.8 22.4 8.4 3.0 100.0
    - 2016-17: 51.7 1.7 7.7 1.3 4.4 32.3 0.9 100.0
    - 2005-17: 37.0 10.6 11.7 1.1 2.7 15.5 1.4 100.0
  - Other Cholesky orderings and corresponding percentage rows are reported similarly in Annex II.
- Source for these percentage decompositions: IMF Staff Estimations.

### Governance and growth — empirical context and channels
- Anti-corruption efforts since the 2015 political crisis: government reasserted control over tax and customs administration, reformed public procurement practices, enhanced independence of general attorney’s office; GPO and CICIG have dismantled networks suspected of illegal activities.
- Since 2014, GPO has dismantled 178 criminal organizations operating at the national level and confiscated assets for over USD28.5 million (Ministerio Publico, 2018).
- Financial Intelligence Unit (IVE) initiatives:
  - Increased number of STRs and financial intelligence reports disseminated to AML/CFT stakeholders.
  - Increased onsite inspections and capacity building.
- Financial intelligence and enforcement metrics (selected entries from Table 1 in the governance section):
  - STRs to IVE: 2011: 50; 2012: 64; 2013: 67; 2014: 56; 2015: 106; 2016: 1150 (as presented).
  - IVE requirements: 75 175 258   13733   20337   12867 (table formatting preserved as in source).
  - STRs used for filing of legal reports: 142 165 239 395 183 (as presented).
  - Ratio STR reported over used: 0.28 0.35 0.42 0.37 0.16 (as presented).
  - Prosecutions for ML offences: 8 7 10 48 41 51 11 12 (as presented).
  - Convictions for the crime of ML as a stand-alone offence: 0 0 1 7 15 15 (as presented).
  - ML cases filed: 128 214 164 313 441 (as presented).
  - Note: Table entries preserved exactly as in source; acronyms: IVE = Financial Intelligence Unit, ML = money laundering, STR = suspicious transactions reports.
- Public perceptions:
  - Anti-corruption efforts in response to 2015 crisis were appreciated by the public; perceptions that corruption had increased during 2016 were the second lowest in the Latin America region (Global Corruption Barometer, Transparency International, 2017).
- Channels through which weak governance undermines growth:
  - Investment: weak institutions increase cost of doing business and reduce investor risk appetite; Guatemala’s gross investment rate stagnated at 15 percent of GDP after the Global Financial Crisis.
  - Human capital: weak governance reduces revenue mobilization and provision of public goods (health and education).
  - Productivity: weak institutions act as barrier to entry, hinder employment growth and innovation, and slow public investment execution.
- Cross-country empirical approach:
  - Panel regressions (pooled OLS) for 89 countries; regressions use one-period lagged corruption index to mitigate reverse causality.
  - Four corruption indicators used in separate regressions: (i) ICRG corruption index; (ii) World Bank’s control of corruption (CCI) index; (iii) World Economic Forum (WEF)’s ethics and corruption index; and (iv) WEF’s irregular payments and bribes index.

*Source: Staff Estimations.*

### Annex I). Following Barro (2001, 2015) and Contreras and Pinto (2015), we include the following

### Annex I). Following Barro (2001, 2015) and Contreras and Pinto (2015), we include the following control variables

### Regression results: governance and per-capita real GDP growth
- Main finding: Regression results suggest a significant and positive relationship between governance and GDP growth (Table 2).
- Interpretation:
  - For all four corruption indicators used, a higher score indicates lower corruption. The coefficients of the one-period lags of the corruption indicators are all positive and significant, suggesting that lower corruption would raise per capita GDP growth.
  - Control variables: signs broadly as expected. The coefficients are significantly positive for log GDP and negative for the square of log GDP, implying a likely non-linear relationship between growth rate and level (for poorest countries marginal effect of GDP level on growth is positive; for richest countries it becomes strongly negative).
  - Longer life expectancy and higher fertility rate increase population size and therefore tend to lower GDP growth per capita.
  - Years of schooling are in general positively correlated to per-capita GDP growth, though insignificant results may reflect poor quality of schooling data.
  - Government consumption ratio is negative and strongly correlated to per-capita GDP growth.
  - Higher inflation is estimated to have negative correlation with GDP growth.
  - Trade openness is estimated to be positively related to growth, as is improvement in terms of trade.

- Table 2: Guatemala: Impact of Corruption on Per-Capita Real GDP Growth (selected coefficient estimates and statistics)
  - Use of corruption indicator columns: ICRG corruption index | WB control of corruption | WEF ethics and corruption | WEF irregular payments & bribes
  - Corruption indicator (one-period lag):
    - ICRG corruption index: 0.22179***
      - (0.08034)
    - WB control of corruption: 0.70498***
      - (0.18733)
    - WEF ethics and corruption: 0.71305***
      - (0.22274)
    - WEF irregular payments & bribes: 0.53168*
      - (0.28125)
  - Log(Initial GDP per capita):
    - 5.46109*** (0. (1.45362))
    - 7.82421*** (1.82539)
    - 7.65690** (3.53075)
    - 6.28859 (5.04501)
  - Log(Initial GDP per capita) squared:
    - -0.34681*** (0.07552)
    - -0.46960*** (0.09917)
    - -0.46732** (0.19474)
    - -0.35053 (0.28039)
  - 1/(Life expectancy):
    - 115.98194 (99.78348)
    - 226.90330* (121.82109)
    - 68.113445 (349.06584)
    - 53.90564 (576.78935)
  - Log(Fertility rate):
    - -1.95831*** (0.42826)
    - -1.22302** (0.52395)
    - 0.55976 (0.92985)
    - 1.98328 (1.25006)
  - Years of schooling (male):
    - 0.00326 (0.01516)
    - -0.01171 (0.01846)
    - 0.00088 (0.02651)
    - -0.00754 (0.02721)
  - Years of schooling (female):
    - -0.00080 (0.01332)
    - 0.01854 (0.01561)
    - 0.02341 (0.02344)
    - 0.02903 (0.02635)
  - Government consumption/GDP:
    - -0.14848*** (0.02661)
    - -0.21483*** (0.03389)
    - -0.31429*** (0.05977)
    - -0.35046*** (0.08205)
  - Inflation rate:
    - -0.01529*** (0.00388)
    - -0.05046*** (0.01749)
    - -0.05350*** (0.01712)
    - -0.07845*** (0.01412)
  - Trade openness:
    - 0.00869*** (0.00200)
    - 0.00714*** (0.00209)
    - 0.00876*** (0.00292)
    - 0.00903** (0.00353)
  - (Terms of trade shock)*(Trade openness):
    - 0.00038** (0.00019)
    - 0.00047** (0.00021)
    - 0.00083** (0.00033)
    - 0.00110* (0.00057)
  - Constant:
    - -15.22457** (7.69232)
    - -27.43257*** (8.70208)
    - -28.46918* (15.16164)
    - -25.40244 (19.83483)
  - Sample period:
    - 1984-2016 | 1996-2016 | 2007-2016 | 2007-2016
  - Number of observations:
    - 1,876 | 1,228 | 499 | 298
  - Number of countries:
    - 84 | 84 | 78 | 73
  - R-squared:
    - 0.08591 | 0.10137 | 0.20895 | 0.26347
  - Notes: Robust standard errors in parentheses. * p<0.1, ** p<0.05, *** p<0.01.
  - Source: IMF staff estimates.

### Estimated growth gains from reducing corruption
- Regression estimates suggest that, if Guatemala were to close the governance gap with the average country in the sample, its real per-capita GDP growth could be between 0.2 and 0.8 percentage points higher relative to baseline, depending on the corruption indicator considered.
- More granular findings: improving trust in politicians, tackling diversion of public funds and reducing favoritism in decisions of government officials can deliver the largest growth gains.
- Figure summary (textual):
  - Aggregate Governance Indicators (in pc GDP growth terms) show gains up to 0.8 (ICRG corruption, Irregular payments and bribes, WB Control of Corruption, WEF Ethics and corruption).
  - Granular Governance Indicators (in pc GDP growth terms) with notable items: Transparency of policymaking, Judicial independence, Ethical behavior of firms, Favoritism in decisions of gov. officials, Diversion of public funds, Public trust in politicians — all with positive estimated effects (indicators taken from WEF Institutions pillar; all coefficients significant at varying levels).

### Reform avenues — sustaining reform momentum through a multi-pronged strategy
- Cross-cutting observation: International experience shows an effective anti-corruption strategy requires perseverance, strong political ownership, and sustained civil society support.

- Tax revenues — tax administration reforms
  - Background: Unveiled cases of tax graft in 2015 eroded SAT credibility, weakening tax compliance and leading to collection losses.
  - Measures adopted in 2015: new SAT law aimed at strengthening integrity and accountability of SAT staff.
  - Ongoing efforts:
    - New standardized, transparent processes (strengthened control of goods and their valuation in customs; control of special regimes; effective control of VAT through improved taxpayer registration).
    - Publication of the list of VAT-refund recipients to avoid undue reimbursements.
    - Retain qualified SAT staff and fully integrate information systems to facilitate flow of information between SAT, National Civil Police, IGSS, Ministry of Labor and Commercial Register.
    - Deepen inter-institutional agreements involving SAT and private operators and border authorities; harmonize tax and custom processes with bordering countries.
  - Figure 4: Customs prices and tax collection time series and imports composition by product in 2015 (fuel products, intermediate goods, consumption goods, capital goods) — sources: Superintendencia de Administración Tributaria and Staff calculations.

- Public expenditures — transparency and quality of spending
  - Visible efforts to promote transparency should be expanded to release resources for investment in physical and human capital.
  - Positive initiatives:
    - Third National Action Plan of Open Government 2016-2018 jointly elaborated with CSOs and the Open Budget exercise.
  - Further scope:
    - Consolidate and expand coverage of fiscal reports.
    - Enhance transparency in financial statements of local governments.
    - Complete the public sector personnel census.

- Public investment — procurement reforms
  - Reforms of the Procurement Law in the past two years have facilitated greater oversight and contributed to tackling corruption (forbid government contracts with state officials and sponsors of political parties; adoption of e-outlet “Guatecompras”; introduction of reverse auctions).
  - Remaining steps to increase transparency and governance in procurement:
    - Further increase transparency of “Guatecompras”: disclose documents at all stages, publish municipal procurement activity, provide certification and technical assistance from Ministry of Finance on large procurement contracts, and make portal fully operational with technical assistance to companies, executing units, supervisors.
    - Clarify scope of concurrent regulations and reconcile legal interpretations by procuring and auditing agencies. Public Prosecutor’s unit at the CGC to facilitate pre-screening of cases to avoid unnecessary escalation from administrative to criminal sanctions. Adopt new regulations to increase transparency of fideicomisos.
    - Speed up spending execution without diluting governance focus: allow streamlined procedure for MINFIN to approve payments on pending contracts; allow auditors from the Comptroller General to participate as observers in public bids; define procedures outside of the penal system to handle procedural errors in contracting.
    - Introduce sanctions to combat fraudulent dividing up of procurement transactions and remove tax incentives to entities breaching procurement regulations.
    - The newly created vice-ministry for transparency and procurement will oversee and coordinate entities involved in procurement with focus on fiscal transparency.
  - Footnote: fraudulent dividing up of procurement transactions defined as division of a large procurement operation into smaller transactions to bypass bidding requirements.

- Market regulation — business facilitation reforms
  - Government implementing comprehensive business facilitation reform to deter corruption: easing procedures for construction licenses, providing e-government administrative services on tax reporting (declaraguate) and business registration (minegocio.gt).
  - Commercial code now introduces arbitration for business disputes as alternative to civil suit.
  - Note: SAT exploring possibility to consolidate firms’ quarterly tax filings into the annual tax report, saving companies close to 40 hours in tax filing per year (240 hours currently). Online registration would allow registering companies in Commercial Register, SAT, IGSS and Ministry of Labor.

- Anti-Money Laundering (AML) and Countering the Financing of Terrorism (CFT)
  - Rationale: Proceeds of corruption are usually laundered; an effective AML framework contributes to prosecuting and deterring corruption.
  - Under FATF standard, AML framework must require:
    - risk based supervision of financial institutions and designated non-financial businesses and professions (DNFBPs);
    - transparency of the ultimate beneficial owner of corporate vehicles;
    - effective and operational agencies specialized in preventing and combatting money laundering.
  - Recent steps: intensified on-site inspections by Guatemala’s Financial Intelligence Unit (IVE) and enhanced customer due diligence have helped uncover laundering of proceeds of corruption.
  - Recommended further measures:
    - Prompt submission to parliament of a bill (drafted with support from international organizations) to enact risk-based supervision of financial institutions and DNFBPs, ensure effective enhanced customer due diligence of politically exposed persons, and suspicious transaction reporting.
    - Increase transparency on the ultimate beneficial owner of legal persons: bearer shares eliminated and converted to nominative shares; IVE’s 2014 guidelines require reporting entities to identify shareholders with participation greater than 10 percent.
    - Additional measures: obligation for reporting entities to refrain from or terminate business relationships when ultimate beneficial owners cannot be identified (FATF Recommendation #10); ensure authorities have timely access to accurate information on ultimate beneficial owners.

- Enforcement — CICIG and collaboration with General Prosecutor’s Office
  - Background: Government requested UN assistance in 2003 to create mechanism to support investigation and prosecution of organized crime; CICIG commenced operations in 2007. CICIG’s mandate extended five times; up for renewal in September 2019.
  - CICIG role and structure:
    - Independent UN body created in 2006; led by a Commissioner (recruited internationally, grade of Under Secretary General); staff recruited internationally.
    - Mandate: investigate illicit security forces and clandestine organizations; collaborate with Government in disarticulation of clandestine organizations and prosecution; make policy recommendations; authorized to make administrative complaints against public officials and act as interested third party in disciplinary procedures; provide technical assistance to government agencies.
    - Special Prosecutor’s Office against Impunity (FECI) within Prosecutor General’s Office created to work with CICIG; cases decided by High Risk Tribunals (Decree number 21-2009).
  - Effectiveness:
    - Collaboration between General Prosecutor’s office and CICIG effective at promoting independent investigation of corruption; increased prosecutorial capacities and independence; strengthened legal framework through CICIG-promoted reforms; uncovering of several high-profile cases.
  - Examples of high-profile cases uncovered:
    - Tax Fraud: La Linea case (2015) involving expedited customs clearance/dismissal of customs duties in exchange for payouts; investigations leading to one of the largest fines on record for company attempting to bribe tax officials.
    - Illicit Government Contracts: criminal networks controlling public institutions to assign/pay contracts, alleged financing of political campaigns; irregular awards in provision of goods and services (medicines, medical equipment, promotional events) by selected companies lacking capacity.
    - Judiciary: bribery charges against judges and magistrates for illegal decisions in exchange for payouts or promotions.
    - Financial sector: bank executives acquiring undervalued properties through irregular auctions with funds transferred to paper companies’ current accounts.

*Source: IMF staff estimates (extracted from the provided content).*

### 21.      However, for anti-corruption efforts to be effective, judicial integrity must be

### 21.      However, for anti-corruption efforts to be effective, judicial integrity must be

### Judicial integrity: findings and recommendations
- Finding: Investigation and prosecution of acts of corruption has increased, but the number of convictions has not seen a significant impact (Table 3).
- Finding: Recent convictions and investigations against members of the judiciary highlight the need to enhance judicial integrity.
- Possible causes identified:
  - Lack of transparency in the appointment of judges and magistrates.
  - Insufficient disciplinary regime that does not cover all magistrates (Box 2).
- Recommendations to strengthen judicial integrity:
  - (i) Transparency and accountability throughout the selection process of judges and magistrates, beginning with a review of the appointment procedures of the nominations committee, and deeper scrutiny of the members.
  - (ii) Effective implementation of a transparent merit-based system for the advancement of judges’ careers.
  - Recommendation: Online publication of all judicial decisions as a step towards reinforcing legal certainty.

### Box 2 — Appointment procedure (description)
- Constitutional/legal framework:
  - According to article 217 of the Constitution, magistrates are appointed from a list compiled by a “nominations Committee”.
  - The committee gathers one representative from each Law University in the country, an equal number of representatives from the Bar Association, and an equal number of Supreme Court magistrates.
  - The Law on the Nominations Committees: the committee drafts the profile of the candidates and a scoring table for each requirement.
  - Vacancy announcement and candidates’ names must be published once selected.
  - After verification and interviews, candidates are scored and the commission compiles a list based on those scores; the list is then submitted to Congress for subsequent appointments.
- Transparency gaps:
  - There is no requirement that the selection process or the voting must be made public.
  - Candidates’ scores are not binding for the final appointment decision.
- Recommendation: Increased transparency in this selection process could improve judicial integrity and subject commission members to higher scrutiny.

### Table 3 (Prosecution and Conviction of Acts of Corruption) — figures as reported
- Header/context: Table 3. Guatemala: Prosecution and Conviction of Acts of Corruption
- Reported figures (as presented):
  - 2010   2017
  - Avg. 
  - 2010-17
  - Totals 
  - 2010-17
  - Indictments71111.088
  - Convictions325.141
  - Indictments041.19
  - Convictions020.54
  - Indictments0112.621
  - Convictions020.43
  - Indictments020.43
  - Convictions000.11
- Source: Office of the General Prosecutor.
- Categories listed below the table: Passive bribery; Active bribery; Misappropriation; Trading in influence.

### Judicial effectiveness: reforms to improve efficiency
- Finding: Higher judicial effectiveness is needed alongside improved integrity.
- Proposed production-based business model reforms:
  - (i) Reorganization of courts (separating technical from managerial functions).
  - (ii) Simplification of administrative procedures, online access to services, proactive case management, improved budgetary mechanisms, and performance accountability (Esposito, Lanau and Pompe, 2014).

### Asset disclosures: usage and reform needs
- Finding: Asset disclosures (Public Probity Law) have been used by the General Prosecutor and the CICIG in investigations of corruption and illicit enrichment, with good results; emblematic cases and ongoing investigations show asset disclosures are an important source of information (Table 4).
- Implementation to date:
  - Since 2014, the Comptroller General’s Office has received, verified and sanctioned officials administratively for delayed disclosure.
  - The Comptroller General’s Office has submitted a total of 540 cases for criminal investigation by the PGO, including breaches to the Public Probity Law and delays or unexplained patrimonial increases.
- Recommended reforms for an effective system:
  - (i) Implementation of electronic filing of asset disclosures to relieve resources currently used to digitalize forms.
  - (ii) Access to banking information for the purpose of verification.
  - (iii) Online publication of asset disclosures of all public officials to increase transparency and accountability.
  - Note: These require legal reforms.

### Table 4 (Asset Disclosures and Administrative Sanctions) — annual figures
- Header/context: Table 4. Guatemala: Asset Disclosures and Administrative Sanctions
- Annual figures (2014–2017) as reported:
  - 2014 2015 2016 2017
  - Received 31,128     17,943     25,948     18,967
  - Verified 269 220 428 282
  - Submitted to the GPO 23 16 208 293
  - Dismissed by the GPO 6 2 18 0
  - Administrative sanctions 31 36 49 45
- Source: Office of the General Prosecutor.

### Legal and institutional framework: gaps and needed laws
- Finding: An updated legal framework could enhance anti-corruption efforts; a suite of draft laws is with Parliament awaiting debate and approval.
- Draft reforms included in the package:
  - Reform to the criminal code (including the crime of illicit political party financing).
  - Reform to the General Comptroller’s Office law (to include open government reforms and preventive auditing mechanisms).
  - Reform to the Public Procurement Law.
  - Reform to the law on the Judicial Power (to separate administrative and jurisdictional functions).
- Additional laws identified as needed but not yet in the package:
  - Reform to the Law on Appointment Committees.
  - Reform to the Public Probity law.
  - The new AML/CFT law.

### Contextual fiscal note (linked content): August 2016 tax reform summary
- Finding: Official estimates indicate the August 2016 tax reform proposal would have raised tax revenues by 1½ percent of GDP.
- Main components of the reform proposal:
  - Increase in corporate income tax (CIT) rates both under the net income and gross sales regimes (Table 1).
  - Increase in personal income tax (PIT) rates and introduction of two additional brackets to the scale (Table 2); re-establishment of the tax credit for VAT on personal purchases for wage earners; introduction of new deductible expenses for medical, private education, and contributions to private pension plans.
  - Higher excise taxes on fuel products by 3 quetzales (USD 40 cents) per gallon of gasoline, diesel, jet fuel and kerosene (earmarked to finance the road maintenance fund, COVIAL).
  - Higher excise tax on cement from 1.50 quetzales (USD 20 cents) to 5 quetzales (USD 70 cents) per bag.
  - Higher royalties from 1 to 10 percent for gold, silver and platinum, and to 3 percent for other metallic minerals; part of the royalties to be distributed to municipal governments where mining operations are located.
  - Other tax administration measures aimed at widening the tax base.

*International Monetary Fund. Guatemala: selected sections on judicial integrity, asset disclosures, and related fiscal context.*

### 7. Given the nature of the analysis, based on household microdata, this study examines

### 7. Given the nature of the analysis, based on household microdata, this study examines

### Data and Methodology
- Household microdata source: ENCOVI 2014 (latest available).
- Analysis type: static incidence analysis based on five core income concepts (following Lustig (2018)):
  - Market income: factor income, private transfers, imputed rent, self-consumption.
  - Net market income: market income minus direct personal income taxes (simulated) and social security contributions.
  - Disposable income: net market income plus direct government transfers.
  - Consumable income: disposable income plus indirect subsidies minus indirect taxes (simulated from household consumption, current rates and exemptions). For fuel taxes, indirect effects from increased prices of other goods using fuels in production/transportation were estimated.
  - Final income: consumable income plus monetized in-kind government transfers (education and health). Education values imputed using national average per-level expenditure for public-school attendees in ENCOVI; health values imputed for users of health centers/public hospitals and IGSS affiliates using Ministry of Public Finance functional classification and IGSS program spending.
- Indirect effect methodology note: matching ENCOVI 2014 consumption categories with Social Accounting Matrix (SAM) disaggregation and multipliers from the cost-push model by Jellema and Inchauste, 2018; SAM estimated by Escobar (2015) using National Accounts and ENCOVI 2011.
- Key measure: difference between Gini coefficients for relevant income definitions before and after reforms.
- Static framework: no behavioral feedbacks or dynamic taxpayer adjustments included.

### Main Results — Fiscal Policy Impact on Income Inequality
- Baseline inequality:
  - Market income Gini coefficient: 47.6 percent.
- Effect of policy components on Gini (percentage-point changes):
  - Direct taxation: net income Gini decreases by 0.7 percentage points.
  - Government transfers: reduce inequality by 0.2 percentage points.
  - Indirect taxes/subsidies: reduce inequality by 0.1 percentage points.
  - Monetized education and health (in-kind transfers): additional decline of 1.7 percentage points, bringing Gini to just below 44.8 percent for final income.
  - Overall fiscal policy effect on inequality: 2.8 percentage points (market income to final income).
- Comparative statement: Guatemala is among the least redistributive countries in Latin America given the small overall fiscal effect.
- Reform scenario impact on inequality:
  - Gini coefficients under the reform are broadly the same as baseline.
  - Proposed reform to indirect taxes on fuel products and cement would slightly improve redistribution: reduction of about 0.1 percentage points in the Gini index for consumable income.
- Marginal contributions (difference in Gini with and without measure):
  - Direct taxes and government transfers (especially in-kind education and health) are equalizing.
  - Contributory pensions increase inequality because low-income workers often lack sufficient contribution histories.
  - The conditional cash transfer program Mi Bono Seguro is more equalizing than the non-contributive pension Aporte al Adulto Mayor; other conditional transfer programs are neutral.
  - Indirect subsidies are neutral; indirect taxes are slightly more equalizing under the reform due to higher fuel excises.
  - Primary education has the largest equalizing marginal contribution among social spending items; in-kind university education and social security health services worsen inequality because they are accessed more by higher-income households.
- Incidence by socio-economic group:
  - Under the current system, approximately 93 percent of total direct taxes is paid by the wealthy and the middle class, and 60 percent of total indirect taxes is paid by the wealthy and the middle class.
  - Almost 90 percent of total direct transfers is received by the poor and vulnerable.
  - More than 50 percent of indirect subsidies are received by the middle class and wealthy.
  - Education benefits: 80 percent to poor and vulnerable groups.
  - Health benefits: almost 50 percent to the middle class and wealthy (incidence of contributory health coverage).
- Reform effect on tax burden concentration:
  - The tax reform would increase the share of total direct taxes paid by the top quintile by 0.5 percent of total payments.
  - The reform would increase the top quintile’s share of total indirect taxes by 5.1 percent of total payments, proportionally reducing the burden on lower quintiles.

### Fiscal Policy Impact on Poverty
- Baseline poverty and fiscal components (percent of total population):
  - Market income poverty: 58 percent.
  - Net market income poverty (after direct taxes): marginally higher than market income, indicating direct taxation increases the share below the poverty line.
  - Disposable income (after direct transfers): direct transfers lower poverty but do not fully offset the poverty-increasing effect of direct taxation.
  - Consumable income (after indirect taxes/subsidies): indirect taxes increase poverty to about 61 percent of the total population.
- Final income poverty: not evaluated due to difficulty quantifying minimum income necessary for basic health and education services.
- Reform scenario effect on poverty:
  - Net market and disposable income poverty under the reform are the same as baseline (reform’s PIT changes affect only the top 20 percent).
  - Consumable income poverty would marginally increase under the reform because higher indirect taxes reduce purchasing power across all groups, including the poor.
- Drivers and mitigation:
  - The poverty-increasing effect of direct taxation mainly stems from social security contributions (paid by low-income workers) and, to a lesser extent, PIT.
  - Non-contributory pension Mi Adulto Mayor and conditional cash transfer Mi Bono Seguro effectively reduce poverty but have very small budgets and limited coverage:
    - Mi Bono Seguro budget: 0.06 percent of GDP in 2015.
    - Mi Bono Seguro coverage: less than 20 percent of poor and less than 30 percent of extremely poor population in 2015.
  - Indirect taxation has a negative effect on poverty, marginally higher under the reform due to higher fuel and cement excises.
- Recommendation summarized: expanding cash transfers—especially conditional cash transfers that enhance human capital—would help mitigate the poverty-increasing effects of fiscal policy.

### Conclusions and Policy Recommendations
- Overall reform impact:
  - The fiscal reform would not have materially impacted poverty and inequality compared to the current system.
  - Revenue potential of the reform: 1½ percent of GDP per official estimates.
  - Revenues could be used to: (a) provide direct income assistance to poor households (decreasing poverty and inequality), or (b) step up investment to boost growth, labor demand, productivity, and potentially lower food prices—indirectly reducing extreme poverty.
- Design principles for a well-designed tax reform:
  - Maximize income potential while mitigating negative incidence on poor households.
  - Direct taxation (PIT) has milder negative effects on lower-income households than indirect taxation; PIT is a privileged candidate for reform given very low PIT rates in Guatemala.
  - However, indirect taxes typically yield more revenue; given Guatemala’s significant tax revenue gap, reforms may need to involve indirect taxes as well.
  - The redistributive outcome depends on how additional revenues are spent: using higher indirect tax revenue to finance well-targeted social spending could mitigate or offset negative effects on the poor.
- Policy sequencing and priorities:
  - Use granular incidence and concentration estimates to design fiscal measures that maximize revenue and progressivity while minimizing adverse effects on poor households.
  - Prioritize an initial expansion of social assistance programs—particularly conditional cash transfers to enhance human capital.
  - Complement social assistance with increased infrastructure investment and measures to expand formality to initiate a virtuous cycle of growth, higher labor demand, and poverty reduction.

### Selected Banking Sector Vulnerability Note (contextual observation)
- Banking system indicators (as of September 2017):
  - CAR for the banking system: 15 percent.
  - Minimum regulatory CAR: 10 percent.
  - Impaired loans: on average less than 3 percent of total loans (higher for selected small banks).
  - System is liquid with healthy profits, though rising non-performing loans could erode profitability.

*Source: Authors’ estimates and analysis as presented in the provided IMF chapter.*

### 2.      Banks are mainly oriented towards lending activities and deposits are the main source

### 2.      Banks are mainly oriented towards lending activities and deposits are the main source

### Balance-sheet composition and funding structure
- On average, loans represent about half of banks assets; the rest comprising of long-term bonds (14 percent), cash and T-bills (13 percent), and other assets (18 percent).
- Deposits account for almost 75 percent of total liabilities across the system, with little difference among banks.
- The deposit to loans ratio is well above 100 for the system as a whole as well as individual banks, with only two exceptions.
- Foreign liabilities represent 30 percent of total liabilities and have increased from about 10 percent in 2010, exposing the system to rollover risks.

### Concentration and FX-induced vulnerability
- Nearly 90 percent of total assets are concentrated in the six largest banks (out of 18); the top three account for over 65 percent of market share.
- About 40 percent of banks loans are denominated in foreign currency.
- 44 percent of foreign-currency loans (or 17 percent of total loans) are extended to un-hedged borrowers, exposing the system to FX risk through credit risk.
- About one third of bank liabilities are denominated in foreign currency.
- The system overall has positive net FX exposure, partially mitigating depreciation risk.

### Key banking soundness indicators (Percent, September 2017)
- Total capital / RWA (CAR): 14.8
- NPLs (gross)/ total loans: 2.6
- Provisions/NPLs: 113.6
- FX loans/total loans: 38.4
- Deposits/total liabilities: 73.4
- Deposits/total loans: 128.5
- ROA (after‐tax): 1.3
- ROE (after‐tax): 13.9
- Liquid assets/total assets: 27.0
- Liquid assets/short‐term liabilities: 65.8
- Net FX exposure / capital: 14.2

### Solvency stress test: coverage and single-shock impacts
- The top-down solvency stress test includes: (i) credit risk (aggregate NPL shock and sectoral shocks); (ii) market risk (interest and exchange rate risk); (iii) contagion risk through interbank exposure; and (iv) reverse tests.
- Credit risk shock: system-wide proportional increase in NPLs to 3.3 percent of total loans (the mean value plus two standard deviations of NPLs over the last decade); sectoral shocks of 6½ percent in trade, manufacturing and non-bank financial sectors (together almost 40 percent of total loans).
  - On average the CAR for the system would fall by 1.4 percentage points (system-wide shock) and 1 percentage point (sectoral shock), but would remain above 13 percent on average and above the minimum 10 percent requirement for individual banks.
- Interest rate risk shock: assumes a nominal interest rate increase of 2 percentage point.
  - Flow impact from interest-sensitive gaps and stock impact from bond repricing included.
  - Valuation losses on sovereign bond holdings would reduce the system-wide CAR by 3 percent; two banks’ CAR would fall below the minimum 10 percent.
  - Mitigant: current accounting standards require some securities to be valued at acquisition cost.
- FX risk shock: assumes a 12 percent nominal depreciation of the bilateral exchange rate with the US dollar.
  - Direct FX impact marginally improves capitalization due to net long FX position (+0.3 percentage points to post-shock CAR).
  - Indirect FX impact (through credit risk) reduces CAR by 0.9 percentage points.
  - Overall FX shock reduces CAR by 0.6 percentage points; system CAR remains above regulatory minimum.

### Combined solvency shock and contagion
- Combined shock (system-wide NPL increase, interest rate shock, FX shock):
  - Post-shock system CAR (percent): 9.9
  - CAR change (pct points): −4.9
  - Aggregate CAR would fall by almost 5 percentage points to just below the 10 percent regulatory minimum.
  - Eight individual banks would fall short of the minimum regulatory CAR and would require some recapitalization.
  - Identified capital shortfalls would amount to 0.4 percent of 2016 GDP.
  - If profits are used for defense, only three banks would have CAR below 10 percent and total recapitalization needs would fall to 0.1 percent of GDP.
- Contagion through domestic interbank exposures: limited. Interbank lending in Guatemala is very thin and there is no second-round effect following the combined macro-shock.

### Reverse stress test results
- NPL increases required for CAR to fall below 10 percent:
  - System-wide CAR < 10: implied NPLs would need to increase to 11 percent of currently performing loans.
  - 1/2 market share with CAR < 10: implied NPLs would need to increase to 9 percent.
  - 9 banks with CAR < 10: implied NPLs would need to increase to 16 percent.

### Short-term liquidity stress test
- Scenario: 10 percent per day withdrawal of demand deposits and 3 percent per day withdrawal of time deposits across all banks.
- Post-shock liquid assets/total assets: 19.4
- Post-shock liquid assets/short-term liabilities: 47.2
- Number of liquid banks after 5 days: 18
- Result: Although liquid assets fall sharply, all banks remain liquid after 5 days with no need for outside liquidity support, largely due to maturing assets being rolled off and converted into cash inflows.

### Medium-term liquidity stress test on foreign currency flows (six largest banks, end-September 2017; FX converted at 7.3 Quetzales/USD)
- Purpose: gauge resilience to severe and protracted loss of foreign funding (including Correspondent Banking Relationships interruption).
- Assumptions on run-off rates (selected exact values preserved):
  - Retail funding: sight deposits — Stable: 8% (1 to 7 days), 4% (8 to 15 days), 4% (16 to 30 days), 3.5% (31 to 90 days), 0% (91 to 180 days), 0% (More than 180 days); Unstable: 12%, 6%, 6%, 2.5%, 0%, 0%.
  - Retail funding: savings deposits — Stable: 8%, 4%, 4%, 4%, 0%, 0%; Unstable: 12%, 6%, 6%, 3%, 0%, 0%.
  - Retail funding: term deposits — 35%, 35%, 35%, 30%, 25%, 20%.
  - Other deposits — 35%, 35%, 35%, 30%, 25%, 20%.
  - Secured wholesale funding from other financial institutions — 25% across all buckets.
  - Unsecured wholesale funding from other financial institutions — 75%, 75%, 75%, 30%, 30%, 20%.
  - Outflows from derivatives — 100% across all buckets.
  - Other obligations — 100% across all buckets.
  - Undrawn volume of committed credit/liquidity lines — 10%, 5%, 5%, 10%, 20%, 0%.
- Roll-off rates (inflows usable):
  - Securities in trading book, available for sale, held to maturity, inflows from derivatives — 100% across all buckets.
  - Loans maturing — 50% across all buckets.
  - Other inflows — 80% across all buckets.
- Haircuts:
  - Cash items: 0.0%
  - Securities in trading book: 10.0%
  - Securities available for sale: 10.0%
  - Securities held to maturity: 10.0%
- Maturity distribution findings:
  - 70 percent of assets mature beyond six months.
  - FX liabilities exceed assets at almost all maturities below six months (except 8-15 days).
  - Share of assets to liabilities averages about 80 percent across maturities up to six months, rising to 116 percent beyond six months.
  - Wholesale funding share increases with maturity: about 3 percent at the one-week horizon to over 90 percent at horizons beyond 6 months.
  - Highly liquid cash items and securities available for sale account for over 90 and 40 percent of total assets at the shortest horizons; loans and other assets dominate at longer maturities.

### Medium-term liquidity test results and implications
- System-level funding gaps emerge at maturities of up to one month (0–30 days) under assumed run-off/roll-off rates; cash inflows exceed outflows beyond one month with substantial excess liquidity beyond six months.
- Individual-bank outcomes:
  - Most banks experience liquidity shortfalls in the first three buckets and surpluses afterwards.
  - One bank stays liquid at all maturities.
  - One bank presents a funding gap even beyond six months.
- Counterbalancing capacity (sequentially accessed: inflows from maturing investments/lending, cash and interbank loans, sale of securities with haircuts, required reserves) is sufficient to cover initial funding gaps:
  - At one-week maturity the system-wide funding needs can be covered using cash items.
  - Securities and reserves would be accessed to cover needs up to one month.
  - After using counterbalancing capacity, all banks in the exercise remain liquid.
- Policy recommendation from the test:
  - Adopt the Basel III LCR in U.S. dollars for monitoring purposes to enhance banks’ management of liquidity risk.

*Source: IMF staff analysis as presented in the chapter "2.      Banks are mainly oriented towards lending activities and deposits are the main source."*

### 1.      Remittances have become an increasingly important source of income to Guatemalans

### 1.      Remittances have become an increasingly important source of income to Guatemalans

### Overview
- Remittances have grown in importance since the early 2000s.
- In 2017, remittance flows accounted for over 11 percent of GDP and benefited more than 1.5 million households.
- Anti-immigration measures in the U.S.—including termination of TPS and CAM programs, likely termination of DACA, and claims to end the Diversity Visa Lottery—raise tail-risk scenarios for Guatemala by making it more difficult for immigrants to come to or stay in the U.S.
- Current immigration context: about 75 percent of Guatemalan immigrants in the U.S. are illegal.

### Methodology
- Analysis uses the Western Hemisphere Module (WHDMOD) of the Flexible System of Global Models (FSGM).
- FSGM: a system of annual, multi-region, general equilibrium models developed by RES; monetary and fiscal policy governed by endogenous rules; WHDMOD calibrated to Western Hemisphere economies; international linkages modeled in aggregate.
- Simulations explore transmission channels and net economic effects of potential shocks.

### Risk Event #1: Pronounced Falls in Remittances (tail-risk scenario)
- Shock assumption:
  - Immediate and permanent decline in remittances by one-third due to more stringent U.S. policies on money transfers by illegal immigrants.
  - Rationale: Controls on remittances are difficult to enforce; with 75 percent of immigrants illegal, a one-third fall is assumed under the tail-risk scenario.
- Macro impacts (from FSGM simulation):
  - Output is lower by 1 percent after five years and 2 percent in the long run (steady state, “SS”).
  - Private consumption declines by 8 percent after 5 years and 10 percent in the long run.
  - On impact, the real exchange rate depreciates by just over 12 percent and then gradually adjusts to a new long-run equilibrium roughly 5 percent below its previous level.
  - The nominal exchange rate depreciates by a similar extent to the real exchange rate.
  - Private consumption and investment fall owing to the loss of remittance income; real exports increase due to real exchange rate depreciation.
  - Headline inflation temporarily rises due to the sizeable exchange rate depreciation, but negative demand pressures quickly dominate and both headline and core inflation fall below the baseline.
  - Consistent with its mandate and the decline in core inflation, the monetary authority lowers the nominal policy rate to dampen the impact and return inflation to its target rate.
- Economic adjustment mechanism:
  - Permanent loss of real income requires reductions in domestic absorption and imports, and increased exports to cover imports previously financed with remittances.
  - Real exchange rate depreciation (with transitory overshooting) facilitates this rebalancing.

### Variant: Risk Premium Increase and Monetary Policy Deliberations
- Additional assumption:
  - Transitory increase in the risk premium of 150 basis points faced by corporates, households, and the sovereign (linked to sharp currency depreciation and higher debt-service costs for firms with unhedged USD liabilities).
  - The increase in the risk premium is assumed transitory; long-run equilibrium remains unchanged as firms reduce currency mismatches.
- Short-run effects:
  - Real private investment, consumption, and imports drop further; additional exchange rate depreciation occurs reflecting further easing in monetary policy.
- Alternative policy reaction—leaning against the exchange rate depreciation:
  - Monetary policy attempting to mitigate nominal exchange rate depreciation (supplementing FX intervention with policy rate increases) would further depress domestic activity.
  - Proportional to the nominal depreciation avoided, the risk premium decreases in response to the central bank’s reaction, but eliminating nominal depreciation via uncovered interest parity requires significant tightening of monetary policy, inducing additional excess supply and a larger decline in inflation.
  - If the nominal exchange rate is constrained while a fundamental shock requires a long-term real exchange rate change, the real exchange rate will adjust via domestic disinflation or deflation, requiring additional domestic weakness.
  - Effectiveness of intervening depends on resilience: foreign-currency indebted firms need natural hedges or use hedging contracts to benefit from flexible nominal exchange rate policy.

### Risk Event #2: Increased Deportations
- Assumptions:
  - One-fifth of the illegal immigrants are deported over four years.
  - Returning migrants are not absorbed into the domestic labor market (no positive benefits on Guatemala’s potential output); the main effect comes through loss of remittance income.
  - Note: The model does not distinguish between legal and illegal immigrants; assuming 20 percent of illegal immigrants deported effectively means a 15-percent drop in the total stock of Guatemalan immigrants in the U.S. (since 75 percent are illegal).
- Macro impacts:
  - Real GDP declines by almost 0.2 percent after five years and 0.8 percent in the long run.
  - Gradual drop in remittances depresses disposable income and private consumption; per-capita consumption drop is even larger.
  - Private firms lower investment due to reduced domestic demand.
  - Gradual real effective exchange rate depreciation helps adjust the economy by dampening import demand and raising demand for exports.
- Lower-bound caveat and labor market implications:
  - It is likely migrants will enter the labor force to some extent over time, making this deportation scenario a lower-bound on possible macroeconomic outcomes.
  - To mitigate adverse impacts, national strategies for prevention and attention to migrants should be equipped to face the surge in the labor force.
  - Programs such as “Bienvenido a casa” and “Quedate” should be expanded or reviewed to help integrate returned workers.

### Key quantitative items (as presented)
- 2017: remittances > 11 percent of GDP
- Benefited households: more than 1.5 million
- Share illegal among Guatemalan immigrants in U.S.: about 75 percent
- Assumed remittance fall under tail-risk: one-third
- Assumed deportations: one-fifth of illegal immigrants over four years (implying a 15-percent drop in total stock of Guatemalan immigrants in the U.S.)
- Risk premium transitory increase: 150 basis points
- Output impacts (Remittance shock): -1 percent after 5 years; -2 percent in long run (SS)
- Private consumption impacts (Remittance shock): -8 percent after 5 years; -10 percent in long run
- Real exchange rate: depreciates just over 12 percent on impact; long-run roughly 5 percent below previous level
- Real GDP impacts (Deportations): -0.2 percent after 5 years; -0.8 percent in long run

*Source: FSGM simulators; International Monetary Fund.*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2018/cr18155-guatemalasi.pdf_