## 1geoea2025002-source-pdf

## Source details

**Canonical URL:** [1geoea2025002-source-pdf](https://www.imf.org/-/media/files/publications/cr/2025/english/1geoea2025002-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2025/english/1geoea2025002-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2025/english/1geoea2025002-source-pdf.pdf.json)

---

### Net Financial Account — Integrated Policy Framework (FXI) and Exchange Rate Policy
- Purpose and main conclusion:
  - IPF offers guidance on how Georgia did and should respond to swings in international and domestic demand for the lari.
  - Prudent monetary and macro‑prudential policy by the NBG has limited the need for further policy responses to these shocks.
  - Shallow FX markets create potential for beneficial government FX intervention (FXI), but low reserves and institutional conditions imply risks and long-run costs likely outweighed benefits.
- Rationale for exchange rate flexibility under inflation targeting:
  - Let the exchange rate be the primary adjustment mechanism because it (i) helps stabilize the economy for some macroeconomic shocks; (ii) adjusts automatically and quickly compared to alternatives like domestic wage cuts; (iii) frees monetary policy to focus solely on inflation.
- When FX intervention is potentially justified (IPF conditions):
  - FXI warranted only when one or more of: shallow FX markets; large private sector currency mismatches; poorly anchored inflation expectations.
  - Interventions should be temporary and not substitute for warranted macroeconomic adjustments.
- Shallow FX markets — evidence and implications:
  - FX market size: "roughly USD 8 billion per month."
  - Market growth: more than triple its size prior to Russia’s invasion of Ukraine in 2022, but still relatively small by international standards.
  - Market characteristics: high and volatile bid‑ask spreads and UIP premia; strong negative relationship between capital inflows/outflows and UIP premia.
  - Model estimate (Chen and others, 2023): market depth in line with other emerging markets but shallowness remains a key friction.
  - Forecast error variance decomposition (small open economy SVAR): investor “risk appetite” shocks account for around two‑thirds at a one‑year horizon, falling to about a third at the three‑year horizon.
- Currency mismatch — extent and sectoral composition:
  - Aggregate private FX debt to GDP: "around 40 percent of GDP", down from pandemic peak of "64 percent".
  - Sectoral shares of GDP: Firms "23 percent"; Households "10 percent".
  - Firm cross‑border debt: "4.4 percent" of GDP.
  - Implication: bulk of FX exposure is domestic and concentrated in firms and households.
- Banking sector exposure and hedging:
  - Bank FX debt to GDP is high compared to comparators but not in top 50 globally by FX debt to GDP.
  - Banks hold high levels of FX assets; net open FX position of domestic banks is very small.
  - Regulatory limits: stress tests, additional capital buffer for currency risk exposure, limit to a "20 percent" open FX position (most banks target "0 percent").
- Household FX exposure and mitigation:
  - "66 percent" of household deposits are in FX.
  - NBG FX mortgage payment‑to‑income ratio limits: "20-30 percent of income" depending on income level.
  - Ongoing stress testing and macroprudential policies to discourage dollarization.
- DSGE / QIPF model findings and FXI role:
  - Model estimated on quarterly Georgian data 2000-2023; shallow FX markets make sterilized FXI have real effects.
  - Risk appetite shocks were a substantial driver of real exchange rate movements since 2015.
  - In shallow markets, a 5 percent depreciation risk appetite shock can be alleviated with FXI, producing smaller depreciation, less contraction in activity, lower inflation, lower interest rates, and smaller domestic spreads; FXI modeled as temporary with reserves run down and rebuilt later.
  - Model limitations: difficulty distinguishing fundamental from sentiment shocks in real time; costs not captured include impeding FX market development.
- Institutional considerations and evaluation of NBG FX interventions 2020–2024:
  - NBG interventions: sold dollars in 2020, purchased in 2022, sold around end‑2024; proxy indicates extensive FXI in 2020-2024.
  - UIP premia ranged from a high of "22.3 percent" in April 2021 to a record low of "-9.9 percent" in April 2023.
  - QIPF counterfactual without FXI suggests greater depreciation in 2020-21 and higher core inflation and interest rates, peaking at 11 and 13.4 percent respectively — 1.5 and 2.4 percentage points higher than realized.
  - Q4 2024 episode: NBG sold USD700 million across September and October 2024; model suggests this offset roughly half of the estimated risk sentiment shock around the election.
  - Reserves impacts: intervention in 2020 contributed to mid-2022 reserves at 80 percent of the ARA metric; Q4 2024 intervention lowered reserves by 15 percent, leaving buffers at 80 percent of the ARA metric at end-2024.
  - Shortcomings: hard-to-identify shock composition, low reserves, limited communication, potential perceptions of targeting the exchange rate risking credibility.
- Conclusions and policy recommendations:
  - Temporary, prudent, targeted FXI to partially offset the largest non‑fundamental shocks can at times be warranted, but implementation must be transparent and cautious.
  - Preparatory steps:
    - Implement all relevant IMF Safeguards Assessment recommendations, including moving to a collegial decision‑making model.
    - Have significantly higher reserve buffers before contemplating FXI.
    - Develop a clear and transparent framework for (i) assessing conditions for FXI, (ii) defining implementation modalities and prudent risk management, and (iii) effectively communicating the approach to the public.

### Informality — definitions, measurement, distribution, and policy implications
- Definitions and margins:
  - Firm-level informality: enterprises operating outside legal/regulatory framework (unregistered, untaxed, noncompliant).
  - Worker-level informality: individuals employed without formal contracts or social protection; includes subsistence farmers.
  - Extensive margin: entire firm informal; intensive margin: informal practices within formal firms.
- Causes and typology:
  - Drivers: burdensome regulations, weak contract enforcement, limited access to finance, high taxes and compliance costs, rigid labor markets, limited formal job creation, high payroll taxes.
  - Three perspectives: constrained entrepreneurs; unfair competitors; economic survivors — latter two argued more relevant for Georgia.
- Measurement approaches and comparative evidence:
  - Direct methods: surveys; indirect methods: models (MIMIC).
  - WBES: "just 12 percent of firms in 2023 said that informal competition was a major or severe constraint."
  - World Bank MIMIC model: puts Georgia’s informal economy at "2 percent of GDP" — among the highest in the world (methodological caveats noted).
  - Historical estimates: 1998 more than half employed informally (37 percent excluding agriculture); World Bank MIMIC estimates "around two-thirds of Georgia’s economy was informal from 1993 to 2008."
  - Reforms and decline: electronic tax filing helped reduce employment informality by "8 percentage points between 2004 and 2011"; Yao (2024) finds informality declined by over "10 percentage points since 2008", reaching around "20 percent of GDP."
  - Caution: MIMIC model limitations and calibration issues for Georgia.
- Distribution and worker characteristics:
  - Sector pockets: food manufacturing (30 percent of firms cited informality as a major constraint); hotel services (19 percent).
  - Non-agricultural informal worker profile: concentrated in construction, manufacturing, retail; more likely middle-aged, male, urban; 27 percent report tertiary education; informal workers earn approximately "13 percent" less than formal employees.
  - Subsistence farming: approximately "400,000" Georgians engaged; including them raises informality from 27.6 percent of non‑agricultural workers to "52 percent" of total workforce.
- Policy implications and recommendations:
  - Limited gains from further regulatory easing—Georgia performs well on World Bank Business Ready indicators.
  - Targeted enforcement justified in sectors like construction; AML/CFT tools recommended for cash‑intensive sectors.
  - Address subsistence farming via agricultural productivity improvements and land reform; incomplete land registration impedes transfer and consolidation.
  - Role of growth: economic growth and urbanization likely reduce informality over time.
- Conclusion — key takeaways:
  - Aggregate informality is not unusually high relative to peers; credible measures indicate relatively low levels consistent with reforms.
  - Remaining informality concentrated in construction, food manufacturing, hotel services, and subsistence farming.
  - Recommended actions: sector-specific measures, targeted enforcement, land reform, and sustained growth.

### Emigration, Remittances, Return Migration, and Policy Options
- Emigration trends and demographics:
  - Population stabilized around "3.7 million" in the past decade after decline 1994–2014 due to migration.
  - Gross legal migration to Russia declined from about "25,000" in 1997 to fewer than "5,000" in 2023; total to Russia 1997–2023 over "240,000" people (7 percent of current population).
  - Gross legal migration to OECD countries increased from "3,500" in 1995 to "33,000" in 2022; total to OECD 1995–2022 around "340,000" people (10 percent of population).
  - Over "580 thousand" Georgians lived abroad in 2024 ("580,084").
  - Emigration intentions: "9 percent" intended to emigrate permanently (Caucasus Barometer, 2024); "15 percent" in Gallup World Poll (2024); "45 percent" intended temporary relocation (Caucasus Barometer, 2024).
- Demographic impacts:
  - Working-age population declined "5 percent" from 2014 to 2023.
  - Old-age dependency ratio increased; Georgia’s old-age dependency high compared to CCA and developing-economy peers.
  - Return migration: in the past decade returning migrants amounted to two thirds of those emigrating.
  - Migrant remittances averaged over "12 percent of GDP."
- Labor market implications:
  - Emigration contributed to labor shortages; EU Business Climate Survey (2024) reports two thirds of businesses cite high emigration as main factor depleting labor pool.
  - If unchanged, working-age population trend would have reduced annual output by "2 percent."
  - OECD destination outcomes (2020/21): nearly "20 percent" of Georgian labor force was unemployed and over "40 percent" of the working-age population inactive; in Georgia unemployment averaged around "20 percent" and nearly half of working-age population outside labor force.
  - Emigrant education in OECD (2020/21): "30 (31) percent" low (high) educational attainment; share of high-educated emigrants to OECD fell from "38 percent" in 2015/16 to "31 percent" in 2020/21.
  - Comparative: emigrants tend to be older, lower skilled, more female, with higher unemployment/inactivity relative to peers.
- Remittances and macro significance:
  - Net money transfers in 2024: "$3 billon", or "9 percent" of GDP (up from "$1.2 billion" in 2012).
  - Major country contributors 2024: USA "$560 million"; Italy "$550 million"; Russia "$500 million".
  - EU share in 2024: "$1.3 billion".
  - Household-level metrics (2023): "6-7 percent" of households report receiving remittances; average remittances per receiving household near "GEL 900 (USD 330)" per month; remittances make up "40-45 percent" of monthly income for recipients; receiving households earn "25-30 percent" more income; receiving households have "20 percent" lower probability of employment.
- Return migration and immigration trends:
  - During 2020 pandemic, returns exceeded departures; excluding pandemic, average returnees two thirds of emigrants since 2012.
  - Post‑Ukraine invasion record influx of over "150,000" immigrants per year in 2022-23, majority from Russia.
  - Current immigrants working in Georgia: "42,000" (Ministry of Labor); potentially double when including shadow economy.
- Policy options and recommendations:
  - Strengthen migration strategy implementation and monitoring (2021–2030 strategy; 2024 action plan approved December 2023; SCMI progress assessments stalled; last migration profile 2021).
  - Improve vocational education and career development: VET has "4,000" beneficiaries per year; VET graduate employment rate below "70 percent"; spending on unemployment and active labor market programs is less than "0.01 percent of GDP" versus peers' average "0.5 percent of GDP".
  - Mobilize diaspora investment and financial literacy: remittance-receiving households spend "60 percent" more on education.
  - Expand reintegration and entrepreneurial support for returnees (examples: Korea model; Georgia facilitated return of "11,000" migrants during 2016-20 under IOM program; "4,000" types of reintegration assistance benefiting "3,400" returnees).
  - Develop immigration policies and international partnerships (examples: Global Skill Partnerships; bilateral agreement with Germany signed in December 2023 for agriculture and logistics).
- Conclusion:
  - Emigration has mixed effects: human capital loss vs. benefits from remittances and return migration.
  - Policy priorities: invest in education, VET, financial literacy, diaspora mobilization, reintegration programs, and targeted immigration partnerships.

### Large Taxpayers Office (LTO) — Reintroduction, Impact, and Cost-Effectiveness
- Data and empirical framework:
  - Dataset covers "1.45 million active taxpayers from 2017 to 2024".
  - Sample final: "25,292 taxpayer-year observations, 4,215 firms, and 190 and 235 LTO taxpayers assigned in 2021 and 2024, respectively."
  - Estimation: weighted difference-in-differences with CBPS weights; controls include logs of VAT turnover, exports, net property value, and number of employees; POSTt = 1 for 2022–24.
- Aggregate revenue impact:
  - Reintroduction of LTO in 2021 boosted tax revenue by "approximately 0.4 percent of GDP"; expanding LTO to include 2024 cohort increased impact by "about 0.7 percent of GDP" (range "0.4 to 0.7 percent of GDP").
- Per‑taxpayer assessment increases (average treatment effects):
  - Final assessments per taxpayer annually: increase of "1.4 million GEL".
  - By tax type (per taxpayer per year): VAT "666 thousand GEL"; WHD "558 thousand GEL"; DPT "175 thousand GEL".
  - Total initial assessments average effect "1.3 million GEL per LTO taxpayer per year"; final assessments "1.4 million GEL".
- GDP contributions by tax type (total effect for all LTO taxpayers):
  - VAT: "0.21 percent of GDP".
  - WHD: "0.17 percent of GDP".
  - DPT: "0.05 percent of GDP".
- Operating costs and cost-effectiveness:
  - LTO operating costs: "7.7 million GEL".
  - Staff: "47 employees"; average salary "4,572 GEL"; total staff cost approximately "2.6 million GEL".
  - Operating cost represents "2.9 percent" of total LTO effect.
- Time dynamics and cohort expansion:
  - Effect on total final tax assessments increased from "1.1 million GEL per taxpayer in 2022 to 1.9 million GEL in 2024" for the first cohort.
  - When both cohorts considered, total final assessments rose from "1.4 to 1.5 million GEL per taxpayer per year", and initial assessments from "1.3 to 1.5 million GEL".
  - Observed "anticipatory compliance" near the LTO threshold and business-to-business spillovers.
- Persistence, avoidance, and spillovers:
  - Taxpayers who remained in LTO showed consistently large and growing increases, major increase in 2024.
  - Taxpayers entering LTO in 2024 showed large and immediate compliance gains.
  - Taxpayers exiting LTO exhibited weaker or negative trends (DPT and WHD), with VAT compliance more resilient.
  - Firms exiting LTO may have caused a revenue loss of "about 0.1 percent of GDP."
- Enforcement targeting and compliance indicators:
  - LTO assignment increases accrued tax paid: coefficient "0.869***" (SE 0.111) in accrual tax paid.
  - LTO reduces probability of submitting additional DPT and VAT corrections and lowers probability of audits ("Audit (binary)": -0.053* (0.022)).
  - LTO increases likelihood of receiving tax collection letters: "0.108***" (0.025).
- Sectoral heterogeneity:
  - Stronger responses in "Information technology and communication, transportation and storage, and utilities."
  - Weaker or negative responses in "Wholesale and retail trade, finance, construction, manufacturing."
- Detailed table estimates (selected, first cohort final assessments):
  - fDPT: "0.175***" (SE 0.026)
  - fVAT: "0.666***" (SE 0.091)
  - fWHD: "0.558***" (SE 0.031)
  - fTOT: "1.388***" (SE 0.103)
  - Aggregate Total Effect (Percent of GDP) — Final: fDPT "0.054", fVAT "0.205", fWHD "0.172", fTOT "0.428".
- Additional conditional findings:
  - Baseline InLTO unconditional final totals: fDPT "0.295***", fVAT "1.122***", fWHD "0.637***", fTOT "2.023***".
  - Conditional interactions show varying amplification or attenuation when receiving under-payment letters, tax collection letters, or being audited.
- Conclusion and policy recommendations:
  - Targeting large taxpayers through the LTO significantly strengthened revenue collection, yielding gains of "0.4 to 0.7 percent of GDP."
  - Recommended priorities:
    - Maintain a risk-based approach and strong enforcement.
    - Deliver tailored taxpayer services and enhanced communication.
    - Ensure robust governance and continuous monitoring to address strategic avoidance.
    - Strengthen CRM systems and expand LTO coverage to include consolidated economic groups.
    - Carefully manage threshold criteria to sustain and maximize compliance gains.

*Source: IMF staff analysis and chapter excerpts from the referenced PDF.*

### 1. Net Financial Account __________________________________________________________________ 4

### 1. Net Financial Account

### Integrated Policy Framework (IPF) — purpose and main conclusion
- The IPF offers guidance on how Georgia did and should respond to swings in international and domestic demand for the lari.  
- Prudent monetary and macro‑prudential policy by the NBG has limited the need for further policy responses to these shocks.  
- Shallow FX markets in Georgia indicate that movements in the exchange rate can sometimes be inefficient—creating the potential for beneficial government FX intervention (FXI).  
- Review of NBG’s major FXI operations since 2020: within the modeling framework interventions benefited the economy in the short run, but low reserves and institutional conditions imply the risks and long run costs of FXI likely outweighed the benefits.

### Rationale for exchange rate flexibility under inflation targeting
- As an inflation‑targeting regime, Georgia should generally let the exchange rate be the primary adjustment mechanism to capital flow shocks because:
  - It helps stabilize the economy for some macroeconomic shocks.
  - It adjusts automatically and quickly compared to alternatives like domestic wage cuts.
  - It frees monetary policy to focus solely on inflation and avoids costly reserve operations.

### When FX intervention is potentially justified (IPF conditions)
- FXI is warranted only when a country exhibits one or more of:
  - Shallow FX markets,
  - Large private sector currency mismatches,
  - Poorly anchored inflation expectations.
- Interventions are justified to temporarily cushion large fluctuations in capital flows (e.g., to reduce temporary exchange rate depreciation) but must be temporary and not substitute for warranted macroeconomic adjustments.

### Shallow FX markets in Georgia — evidence and implications
- Georgia’s FX market size: "roughly USD 8 billion per month."  
- Market growth: more than triple its size prior to Russia’s invasion of Ukraine in 2022, but still relatively small by international standards.  
- Market characteristics:
  - High and volatile bid‑ask spreads and UIP premia.
  - Strong negative relationship between capital inflows/outflows and UIP premia—indicative of sensitivity to investor sentiment.
- Model estimate: market depth (Chen and others, 2023) is in line with other emerging markets, but market shallowness remains a key friction.
- Forecast error variance decomposition for the nominal effective exchange rate (small open economy SVAR):
  - Investor “risk appetite” shocks account for around two‑thirds at a one‑year horizon, falling to about a third at the three‑year horizon.
- Conclusion: shallow FX markets mean exchange rate movements are substantially influenced by swings in investor sentiment that may not reflect fundamentals, supporting a potential limited role for temporary FXI.

### Currency mismatch — extent and sectoral composition
- Aggregate private FX debt to GDP:
  - Currently "around 40 percent of GDP", in line with long‑term average, down from pandemic peak of "64 percent".
- Sectoral breakdown of FX debt (shares of GDP):
  - Firms: "23 percent" of GDP (firms hold large amount via mortgages/auto loans in USD).
  - Households: "10 percent" of GDP.
  - These put Georgian firms "(17th)" and households "(6th)" as some of the highest holders of FX debt in the world (as a share of GDP) for these sectors.
- Cross‑border composition:
  - Very little of this debt is cross border; households borrow effectively exclusively from domestic banks.
  - Firm cross‑border debt: "4.4 percent" of GDP.
- Implication: the bulk of FX exposure is domestic and concentrated in firms and households rather than being large cross‑border sovereign or banking sector external liabilities.

### Banking sector exposure and hedging
- Bank FX debt to GDP is high compared to comparator countries but not in top 50 globally when measured by FX debt to GDP.
- Georgian banks hold high levels of FX assets (loans to households and businesses), which hedges FX debt: the net open FX position of domestic banks is very small.
- Regulatory and supervisory measures limiting bank FX risk:
  - Stress tests with a large depreciation scenario.
  - Additional capital buffer for currency risk exposure.
  - Limit to a "20 percent" open FX position (though most Georgian banks target "0 percent").

### Household FX exposure and mitigation measures
- "66 percent" of household deposits in Georgia are in FX, providing partial insurance against temporary exchange rate movements for depositors.
- NBG limits on FX mortgages: payment‑to‑income ratio limits of "20-30 percent of income" depending on income level.
- NBG conducts stress testing to limit contagion risk from household losses to the financial sector.
- Ongoing macroprudential policies aim to discourage dollarization and reduce household FX exposure over time.

*Source: FOREIGN EXCHANGE INTERVENTION IN GEORGIA: AN APPLICATION OF THE INTEGRATED POLICY FRAMEWORK (IMF staff), July 2, 2025*

### 16. Non-financial corporations in Georgia face higher exposures than households to

### 16. Non-financial corporations in Georgia face higher exposures than households to

### Exposures and hedging capacity
- Firm deposits are highly dollarized: 40 percent.
- Firms hold FX assets representing around 40 percent of their cross-border debt.
- This implies a degree of temporary hedging, although the data do not reveal whether holders of FX assets are the same as holders of FX debt.
- Firms therefore have a reasonable degree of unhedged cross-border exposure but are more likely than households to accurately assess appropriateness for their business model.

### Exchange rate, inflation, and inflation expectations
- Georgia is a small open economy with 53 percent of its CPI basket being either imported goods or ‘mixed’ goods with domestic and foreign varieties.
- Estimated pass-through: a 10 percent depreciation increases inflation by around 0.5 percentage points at a 2-year horizon.
- Inflation dynamics and policy credibility:
  - Inflation peaked at 13.9 percent in 2021 and fell to -0.1 percent in 2024.
  - The NBG raised the policy rate to 11 percent in 2022 in response to high and rising inflation.
  - Market and firm one-year-ahead inflation expectations since 2013 have been stable and converging toward the NBG’s 3 percent target.
  - Empirical local projection tests find no impact of exchange rate shocks on inflation expectations in Georgia in the 2015-2024 period.
- Policy implication: given anchored inflation expectations, there appears to be little justification for using FX intervention (FXI) solely on inflation expectations grounds.

### Structural and quantitative assessment using an estimated DSGE / QIPF model
- Model and estimation:
  - An estimated DSGE model based on Chen and others (2023) implements the QIPF model.
  - The model allows for shallow FX markets and foreign investor risk appetite shocks.
  - The model is estimated on quarterly Georgian data 2000-2023 using Bayesian likelihood techniques.
- Key modeling insights:
  - Market shallowness generates realistic exchange rate volatility and implies sterilized FX interventions have real effects.
  - Counterfactuals indicate substantially lower exchange rate volatility if Georgia’s FX markets had been deep.
  - Risk appetite shocks were a substantial driver of real exchange rate movements since 2015, notably the 2020 depreciation and 2022 appreciation after the outbreak of war in Ukraine.
- Role of FXI in the model:
  - In shallow markets, a risk appetite shock that would cause a 5 percent depreciation can be alleviated with FXI, producing:
    - smaller depreciation,
    - less contraction in economic activity,
    - lower inflation,
    - lower interest rates, and
    - smaller domestic spreads.
  - FXI is modeled as temporary: reserves are run down initially and rebuilt in following periods.
- Model limitations and risks:
  - Difficulty in real time of distinguishing fundamental from non-fundamental (sentiment) shocks; news shocks could be misidentified as sentiment shocks.
  - Costs not captured by the model include the risk of impeding FX market development if participants expect repeated central bank intervention.

### Institutional considerations for FXI
- Effective FXI requires:
  - strong legal and operational autonomy for the central bank, and
  - clear communication about policy goals.
- Without these, FXI could impair central bank credibility and raise doubts about independence or commitment to inflation targeting.

### Evaluation of NBG FX interventions 2020–2024
- Intervention chronology and scale:
  - The NBG pursued active FXI in 2020-2024, initially selling dollars during the pandemic, then purchasing in 2022 following inflows due to the invasion of Ukraine, and selling dollars at the end of 2024 around Parliamentary elections.
  - A proxy measure indicates extensive FXI in 2020-2024 (series constructed largely from Adler and others (2023) and public NBG information).
  - UIP premia ranged from a high of 22.3 percent in April 2021 to a record low of -9.9 percent in April 2023.
- Model counterfactuals and outcomes:
  - QIPF counterfactual without FXI suggests greater depreciation in 2020-21 and higher core inflation and interest rates, peaking at 11 and 13.4 percent respectively — 1.5 and 2.4 percentage points higher than realized.
  - Conclusion: Georgia likely benefited in the short run from FXI in 2020-2024 when interventions smoothed (but did not halt) exchange rate movements; reserves were prudently rebuilt after initial 2020 intervention when large inflows occurred in 2022.
- Q4 2024 specific episode:
  - Across September and October 2024, the NBG sold USD700 million in response to downward pressure on the lari around the Parliamentary election.
  - The QIPF model suggests this intervention offset roughly half of the estimated risk sentiment shock around the election.
- Risks and shortcomings observed:
  - Shocks in 2020 and 2024 were hard to identify contemporaneously; both contained fundamental and non-fundamental elements.
  - Reserves were low in both episodes; intervention in 2020 contributed to mid-2022 reserves at 80 percent of the ARA metric and the Q4 2024 intervention lowered reserves by 15 percent, leaving buffers at 80 percent of the ARA metric at end-2024.
  - Communication and operational frameworks were insufficient; market communication by the NBG was limited and did not sufficiently clarify alignment with the policy framework.
  - Use of FXI around politically sensitive events can reinforce perceptions that the NBG targets the exchange rate, risking credibility.

### Conclusions and policy recommendations
- Summary judgments:
  - Sound policies have limited the scope for intervention in response to capital flow and domestic currency preference shocks: credible inflation targeting has anchored expectations and macroprudential policy has limited FX mismatches.
  - Georgia has shallow FX markets; temporary fluctuations in investor sentiment can substantially move the exchange rate.
  - Prudent and targeted FXI to partially offset the largest non-fundamental shocks can at times be warranted.
  - Georgia likely benefited in the short run from FXI in 2020-2024, but implementation should have been more transparent and cautious given uncertainty over drivers of exchange rate pressure, low reserve levels, and concerns about NBG independence.
- Preparatory steps to operationalize FXI while committing to inflation targeting and a flexible exchange rate:
  - Strengthen institutional independence by implementing all relevant IMF Safeguards Assessment recommendations, including moving to a collegial decision-making model.
  - Have significantly higher reserve buffers from which to contemplate FXI operations.
  - Develop a clear and transparent framework for:
    - (i) assessing the conditions under which FXI is justified,
    - (ii) defining implementation modalities consistent with a prudent risk management approach, and
    - (iii) effectively communicating the approach to the public.

*Source: IMF staff analysis in the referenced chapter.*

### 2. Informality exists at both the firm and worker level. Firm-level informality refers to

### 2. Informality exists at both the firm and worker level. Firm-level informality refers to

### Definitions and margins
- Firm-level informality: enterprises operating outside the legal and regulatory framework—typically unregistered, untaxed, and noncompliant with labor or business regulations.
- Worker-level informality: individuals employed without formal contracts or access to social protection, including informal workers within both formal and informal firms.
- Includes many subsistence farmers, which is particularly relevant for Georgia.
- Extensive margin: whether a firm is informal in its entirety (e.g., unregistered).
- Intensive margin: informal practices within otherwise formal firms—such as employing workers "off the books."
- Many formally employed individuals may supplement income with informal “side” jobs.

### Causes and typology of informal firms
- Informality often reflects institutional and economic constraints:
  - Barriers to formalization: burdensome regulations, weak contract enforcement, limited access to finance.
  - Incentives to avoid formalization: high taxes and compliance costs.
  - Worker-side drivers: rigid labor markets, limited formal job creation, high payroll taxes.
- Three perspectives on informal firms:
  - (i) Constrained entrepreneurs: held back by barriers such as regulation or lack of finance.
  - (ii) Unfair competitors: avoid taxes and regulation despite capacity to formalize.
  - (iii) Economic survivors: operating informally out of necessity due to low skills and limited opportunities.
- For Georgia, the latter two explanations are argued to be more relevant for most observed informality.

### Measurement approaches and aggregate evidence
- Direct methods: surveys of firms or workers (and occasionally tax audits); estimate level and characterize nature of informality.
- Indirect methods: models that infer informality from input variables associated with informality (causes or symptoms).
- Comparative findings:
  - WBES firm-level data: "just 12 percent of firms in 2023 said that informal competition was a major or severe constraint."
  - Labor market-based estimates: show informality rates that are high but broadly in line with international norms for Georgia's GDP per capita.
  - World Bank MIMIC model: puts Georgia’s informal economy at 2 percent of GDP—among the highest in the world.
- Historical context:
  - 1998: "more than half of all employed persons were working informally, falling to 37 percent when excluding agriculture."
  - As late as 2008: "over half of firms in the WBES reported facing informal competition, and a quarter cited it as a major constraint."
  - World Bank MIMIC model: estimates "around two-thirds of Georgia’s economy was informal from 1993 to 2008."
- Reforms and measured decline:
  - Early 2000s reforms (governance, tax reform, regulatory simplification) likely reduced informality significantly.
  - Akitoby (2018): electronic tax filing "helped reduce employment informality by 8 percentage points between 2004 and 2011."
  - Yao (2024), integrating direct survey data, finds informality in Georgia "has declined by over 10 percentage points since 2008, reaching around 20 percent of GDP."
- Methodological caution on indirect models (MIMIC):
  - Limitations: unrealistic assumptions, lack of external validation, dependence on external calibration.
  - Specific concerns for Georgia: high dollarization, distinct demographic and institutional structure, indicators (currency growth, labor force participation) may behave differently.
  - In World Bank informality database: most variation in MIMIC-based estimates from 1993–2020 is cross-sectional rather than time-series, with heavy influence from initial calibration.

### What informality looks like in Georgia (distribution and worker characteristics)
- Sectoral pockets with higher informality:
  - Food manufacturing: 30 percent of firms cited informality as a major constraint.
  - Hotel services: 19 percent of firms cited informality as a major constraint.
- Non-agricultural informal worker profile (GEOSTAT Labor Force Survey):
  - Concentration in construction, manufacturing, and retail.
  - Informal employees are more likely to be middle-aged, male, and urban-based.
  - Often employed by micro-enterprises and tend to expect job duration less than a year.
  - Educational attainment: 27 percent of informal workers report having completed tertiary education.
  - Wage gap: Mincer-type regression shows informal workers earn approximately 13 percent less than formal employees (based on reported net earnings).
- Subsistence farming:
  - Approximately 400,000 Georgians engage in subsistence farming.
  - Including them raises overall informality from 27.6 percent of non-agricultural workers to 52 percent of the total workforce.
  - Policy note: subsistence farmers may alternatively be classified as outside the labor force or as unemployed if actively seeking formal employment; their causes and consequences differ and warrant distinct policy approaches.

### Policy implications and recommendations
- Limited gains from further regulatory easing:
  - Georgia performs well on World Bank’s Business Ready indicators (business entry and location) and on regulatory framework and operational efficiency.
  - Research suggests policies reducing formalization costs often have limited or no impact.
- Targeted enforcement for specific sectors:
  - Increased enforcement (detection and penalization) can encourage formalization; may be justified in sectors like construction.
  - AML/CFT tools recommended for cash-intensive and lightly regulated sectors: risk-based regulation and supervision, customer due diligence, beneficial ownership transparency.
  - Caveat: Georgia has a relatively efficient tax system (e.g., scoring higher than many European countries on VAT C-efficiency), so additional tax revenue gains from formalization may be limited.
- Addressing subsistence farming:
  - Policy response should focus on improving agricultural productivity and supporting structural transformation.
  - Constraint: incomplete land registration impedes transfer and consolidation of land.
  - Continued progress on land reform (accelerated over the past two years) will facilitate selling land and transitioning households into formal, higher-paying employment.
- Role of economic growth:
  - Growth likely reduces informality over time by expanding opportunities in larger, more efficient businesses.
  - Rising urban incomes can pull workers out of subsistence farming and shift consumer preferences toward formal enterprises.

### Conclusion — key takeaways
- Best available evidence: Georgia's aggregate informality is not unusually high relative to peers.
- Most credible measures (particularly firm-level surveys) indicate relatively low levels of informality, consistent with the timing of reforms.
- Older MIMIC-based high estimates should be viewed with skepticism; improved models incorporating direct data converge with direct measures.
- Remaining informality is concentrated in construction, food manufacturing, hotel services (including short-term rentals), and subsistence farming.
- Sector-specific measures, targeted enforcement, land reform, and sustained economic growth are recommended to further reduce informality.

*Prepared by Will Abel (MCD). — Excerpt from the IMF Georgia country chapter on informality.*

### 2. Large scale emigration has contributed to a declining and aging population. Between

### 1geoea2025002-source-pdf - 2. Large scale emigration has contributed to a declining and aging population. Between

### Trends of emigration
- Georgia’s population declined steadily between 1994 and 2014 primarily due to outward migration; the population has stabilized at around 3.7 million in the past decade.
- Destination and flow facts:
  - Gross legal migration from Georgia to Russia declined from about 25,000 people in 1997 to fewer than 5,000 in 2023.
  - From 1997 to 2023, migration to Russia totaled over 240,000 people, representing 7 percent of Georgia’s current population.
  - Gross legal migration of Georgians into OECD countries increased from 3,500 people in 1995 to 33,000 in 2022.
  - Between 1995 and 2022, migration to OECD countries totaled around 340,000 people, or 10 percent of Georgia’s population.
  - Over 580 thousand Georgians lived abroad in 2024 (figure reference: 580,084).
- Destination composition and recipient-country patterns:
  - Germany, Italy, Spain, and France received the majority of emigrants to Western Europe.
  - Türkiye, Poland, Greece, and the United States also received around a third of emigrants to OECD/non-Western Europe destinations.
  - Patterns by country: Poland mostly received men aged 30-44 on a short-term employment basis; Greece and Italy received more women aged 45-64 on long-term and family/humanitarian or employment reasons; Germany attracted the largest share of youth (aged 0-29) and students.
- Emigration intentions and permit reasons:
  - Between 2010-23, half of all first resident permits issued by European countries to Georgian nationals were for employment reasons.
  - 9 percent of Georgians expressed intention to emigrate permanently in the 2024 Caucasus Barometer Survey; 15 percent in the 2024 Gallup World Poll.
  - 45 percent indicated intention to relocate temporarily (Caucasus Barometer Survey, 2024).

### Demographic impacts
- Age structure and population composition:
  - Persistent emigration—especially among young adults—combined with declining birth rate and low fertility below replacement level has increased the median age and decreased the working-age population.
  - The working-age population experienced a 5 percent decline from 2014 to 2023.
  - The old-age dependency ratio has increased in the past decade as the working-age population shrank and the elderly population rose; Georgia’s old-age dependency is particularly high compared to CCA and developing-economy peers.
- Return migration and remittances:
  - In the past decade, returning migrants amounted to two thirds of those emigrating.
  - Migrant remittances averaged over 12 percent of GDP.
- Potential fiscal pressures:
  - Aging and a higher old-age dependency ratio will likely strain social welfare and pension systems, though the latest fiscal risks statement (MoF, 2024) estimates the long-term impact of aging on health and pension spending as manageable.

### Labor market implications
- Labor shortages and business constraints:
  - Emigration likely contributed to labor shortages; the EU Business Climate Survey (2024) reports that two thirds of surveyed businesses cited high emigration rates as the most important factor depleting the local labor pool and hampering expansion.
  - If policies and preferences remained unchanged, the working-age population trend would have resulted in a corresponding annual reduction in labor force participation and supply, and a drag on annual output by 2 percent.
- Emigrant labor market outcomes and skills composition:
  - In OECD destination countries (2020/21), nearly 20 percent of the Georgian labor force was unemployed and over 40 percent of the Georgian working-age population was inactive (not in the labor force).
  - In Georgia during the same period, the unemployment rate averaged around 20 percent and nearly half of the working-age population was outside the labor force.
  - Educational attainment of emigrants in OECD countries (2020/21):
    - 30 (31) percent of working-age Georgian emigrants had low (high) educational attainment.
    - Within Georgia’s local labor market the share of individuals with high educational attainment remained around 40 percent, while those with low educational attainment decreased from 12 percent to 4 percent during the same period.
  - Employment rates varied by destination: employment highest in Poland for both sexes; unemployment highest in France (especially for females) and Greece (especially for males).
- Comparative assessment:
  - Compared to regional peers, Georgian emigrants to OECD countries have tended to be older, lower skilled, more female, and to show higher unemployment and inactivity rates; Georgia has sent a greater number and proportion of legal emigrants to OECD and European countries since 1995 than CCA peers.
  - Emigration from Georgia has not produced excessive brain drain relative to peers—emigrants are, on average, older and less educated than non-migrants, and the share of high-educated emigrants to OECD fell from 38 percent in 2015/16 to 31 percent in 2020/21.

### Policy implications and options
- Objectives highlighted:
  - Improve domestic labor market conditions to develop and retain talent.
  - Promote return migration to expand the local labor pool and capture human capital gains.
  - Incentivize the use of remittances to enhance physical and human capital (education and healthcare investments).
  - Develop policies and partnerships to attract foreign talent.
- Rationale:
  - Policies should aim to mitigate adverse effects of workforce contraction and population aging while leveraging remittances and return migration benefits (skills, experience, investments) to enhance productivity and innovation.

*Source: IMF staff note summarizing emigration trends, labor market impacts, and policy options for Georgia.*

### 16. However, emigration has also led to significant remittance inflows. In 2024, net money

### Emigration, Remittances, Return Migration, and Policy Options

### Remittance inflows and macroeconomic significance
- Net money transfers to Georgia in 2024 amounted to $3 billon, or 9 percent of the country’s GDP.
- This represents a more than doubling since 2012 (from $1.2 billion).
- Major country contributors in 2024:
  - USA: $560 million
  - Italy: $550 million
  - Russia: $500 million
- The EU’s share (mainly from Italy, Germany, Greece, Spain, France, and Poland) totaled $1.3 billion in 2024.
- Remittance inflows as a share of GDP are comparable to the average of CCA peers but significantly higher than Eastern and Central European and developing countries.

### Remittances as household income and foreign currency source
- Remittances are a key source of foreign currency and are comparable to export revenues, tourism receipts, and foreign direct investment.
- Remittances are countercyclical and provided a financial buffer during downturns (e.g., 2020).
- Household-level metrics (household income and expenditure survey data, 2023):
  - Around 6-7 percent of households report receiving remittances.
  - Average remittances per receiving household near GEL 900 (USD 330) per month in 2023.
  - Remittances make up around 40-45 percent of monthly income for receiving households.
  - Receiving households earn 25-30 percent more income than those receiving no remittances.
  - Remittance-receiving households have a 20 percent lower probability of employment, indicating reduced labor force participation incentives.

### Return migration and immigration trends
- Return migration has been significant and can bring human capital gains from skills and experience acquired abroad.
- During the pandemic in 2020, the number of Georgians returning home exceeded those leaving Georgia.
- Excluding the pandemic, the average number of Georgian returnees has been two thirds of those emigrating since 2012.
- Following Russia’s invasion of Ukraine, Georgia saw a record influx of over 150,000 immigrants per year in 2022-23, with a majority coming from Russia.
- A significant share of immigrants were ethnic Georgians returning home along with Russian immigrants, likely contributing to high economic activity.

### Policy options to mitigate challenges and enhance opportunities
- Strengthen migration strategy implementation and monitoring:
  - Georgia has a migration strategy for 2021-2030 implemented through annual action plans; the latest action plan for 2024 was approved in December 2023.
  - Progress and impact assessments by the State Commission on Migration Issues (SCMI) have stalled; the latest migration profile document was published in 2021.
- Improve vocational education and career development to retain talent:
  - Vocational education and training (VET) has only 4,000 beneficiaries per year.
  - VET challenges: access, quality, business partnerships, inadequate funding, limited coordination capacity, and employment rate among VET graduates below 70 percent.
  - Georgia’s State Employment Promotion Program (administered by SEPA) provides vocational and soft skills training, internships, employment in public works, wage subsidies, career counseling, job fairs and portals, and facilitates cross-border agreements on circular migration.
  - Georgia’s spending on unemployment and active labor market programs is less than 0.01 percent of GDP versus peers’ average 0.5 percent of GDP; higher funding could strengthen SEPA’s capacity.
- Promote financial literacy and mobilize the diaspora for productive investment:
  - Remittance-receiving households spend 60 percent more on education (household income and expenditure survey, 2023).
  - Mobilize diaspora investment in local businesses and infrastructure through matching programs (examples cited in source: Mexico’s Tres por Uno; El Salvador's EduRemesa).
- Expand reintegration and entrepreneurial support for returnees:
  - Korea’s diaspora management model includes relocation grants, job placement services, VET, subsidies for starting businesses, tax exemptions and reductions, and R&D support in high tech industries.
  - Georgia facilitated the return of 11,000 migrants during 2016-20 under IOM’s Assisted Voluntary Return and Reintegration program.
  - 4,000 different types of reintegration assistance were provided, benefiting 3,400 returnees (SCMI, 2021).
  - Such programs could be expanded and evaluated more frequently.
- Develop immigration policies and international partnerships to attract foreign talent:
  - According to the Ministry of Labor, 42,000 immigrants currently work in Georgia, and potentially double that number if including the shadow economy.
  - Legalizing work through registry and work permits is essential.
  - Establish partnerships such as Global Skill Partnerships (example: Germany) and bilateral agreements to promote temporary legal employment (example: bilateral agreement with Germany signed in December 2023 for agriculture and logistics).

### Conclusion: balancing costs and benefits
- Emigration’s economic impacts are multifaceted: human capital loss vs. benefits from remittances and return migration that boost incomes, foreign currency, and skills.
- Policy priorities identified:
  - Invest in education, vocational training, and career development to retain talent.
  - Promote financial literacy and mobilize the diaspora to leverage remittances for physical and human capital investments.
  - Encourage return migration through reintegration programs and entrepreneurial support.
  - Attract foreign talent through tailored immigration policies and international partnerships.

*Source: National Bank of Georgia; PMC Research Center (2022); GEOSTAT; State Commission on Migration Issues (SCMI); IOM; Ministry of Labor; Fund staff calculations.*

### 6. This analysis assesses the revenue impact and behavioral changes following the LTO’s

### 6. This analysis assesses the revenue impact and behavioral changes following the LTO’s re-establishment

### Data and empirical framework
- Dataset:
  - Covers "1.45 million active taxpayers from 2017 to 2024".
  - Includes information on "LTO selection criteria, tax assessments (DPT, VAT, WHD), taxpayer characteristics (sector, establishments, founders), and enforcement actions."
  - Nominal values deflated to 2021 using the CPI.
  - Sample restricted to firm taxpayers with positive VAT turnover, focusing on the top 5 percent of the VAT turnover distribution.
  - Final dataset after filters: "25,292 taxpayer-year observations, 4,215 firms, and 190 and 235 LTO taxpayers assigned in 2021 and 2024, respectively."
- Estimation approach:
  - Weighted difference-in-differences comparing LTO-assigned taxpayers (treatment) with those never assigned in 2021 or 2024 (control).
  - Probability of LTO assignment estimated using selection criteria: "VAT turnover, exports, net property value, and employees."
  - Covariate Balancing Propensity Score (CBPS) weights applied; common support restrictions imposed by trimming extreme VAT turnover values.
  - Model includes taxpayer and year fixed effects and controls: "the logarithm of VAT turnover, the logarithm of exports, the logarithm of net property value, and the logarithm of the number of employees."
  - Treatment indicator 퐼푛퐿푇푈푖 = POSTt × LTUi, with POSTt = 1 for years after LTO establishment (2022–24) and 0 before; LTO introduced in September 2021, 2022–24 considered post-LTO period.

### Key quantitative findings: revenue yields and costs
- Aggregate revenue impact:
  - Reintroduction of the LTO in 2021 boosted tax revenue by "approximately 0.4 percent of GDP", with impact increasing over time.
  - Expanding the LTO to include the 2024 cohort increased impact by "about 0.7 percent of GDP" when considered (range reported across discussion: "0.4 to 0.7 percent of GDP").
- Per-taxpayer assessment increases (average treatment effects):
  - "LTO resulted in a substantial increase of 1.4 million GEL in final tax assessments per taxpayer annually."
  - VAT: average rise of "666 thousand GEL per taxpayer per year."
  - WHD: "558 thousand GEL" per taxpayer per year.
  - DPT: "175 thousand GEL" per taxpayer per year.
  - Total initial tax assessments: estimated average effect of "1.3 million GEL per LTO taxpayer per year", compared to "1.4 million GEL for final assessments."
- GDP contributions by tax type (total estimated effect for all LTO taxpayers):
  - VAT accounted for "0.21 percent of GDP".
  - WHD contributed "0.17 percent of GDP".
  - DPT contributed "0.05 percent of GDP".
- Operating costs and cost-effectiveness:
  - LTO operating costs estimated at "7.7 million GEL".
  - Staff: "47 employees", average salary "4,572 GEL", total staff cost approximately "2.6 million GEL".
  - Operating cost "represents a mere 2.9 percent of the total LTO effect."

### Time dynamics and cohort expansion
- Effects strengthened over time:
  - Estimated effect on total final tax assessments increased from "1.1 million GEL per taxpayer in 2022 to 1.9 million GEL in 2024" for the first cohort.
  - When both cohorts considered, "Total final tax assessments when both cohorts are considered rose from 1.4 to 1.5 million GEL per taxpayer per year, and initial assessments from 1.3 to 1.5 million GEL."
- Spillovers and anticipatory responses:
  - Observed "anticipatory compliance" by taxpayers near the LTO threshold.
  - Business-to-business spillovers where "non-LTO firms aligned their practices to maintain business with LTO taxpayers."
  - Some firms "strategically adjusted their behavior either to enter or avoid the LTO", magnifying positive impact when more LTOs in 2024.

### Behavioral changes and persistence
- Persistence and avoidance:
  - Taxpayers who remained in the LTO: "consistently large and growing increases in tax assessments", major increase in 2024.
  - Taxpayers entering LTO in 2024: "large and immediate compliance gains."
  - Taxpayers exiting LTO: "weaker or even negative trends, especially for DPT and WHD", while "VAT compliance remained resilient."
  - Behavioral adjustments by firms exiting LTO may have led to a revenue loss of "about 0.1 percent of GDP."
- Enforcement targeting and compliance indicators:
  - LTO assignment increases accrued tax paid and likelihood of receiving tax collection letters.
  - LTO reduces probability of submitting additional DPT and VAT corrections and lowers probability of audits, suggesting improved voluntary compliance and lower audit needs.
  - LTO enforcement effectively targeted at higher-risk taxpayers.

### Sectoral heterogeneity
- Stronger LTO responses in sectors with formal records:
  - "Information technology and communication, transportation and storage, and utilities" respond more strongly to LTO oversight.
- Weaker or negative responses in sectors with limited documentation:
  - "Wholesale and retail trade, finance, construction, manufacturing" show weaker or even negative responses, consistent with persistent evasion models.

### Additional supporting evidence
- Descriptive profile of LTO taxpayers:
  - LTO taxpayers: VAT turnover "16 times higher" on average than non-LTO taxpayers; represent "44 percent of total VAT turnover" while comprising only "5.3 percent of taxpayers."
  - Final tax assessments "10 times higher" and represent "26 percent of total assessments."
  - LTO taxpayers have fewer late declarations.
- Anecdotal survey and risk-flagging:
  - A 2023 survey found "98 percent of large taxpayers were satisfied with the LTO’s services" and "0 percent reported that their assigned officer improved their understanding and compliance."
  - LTO identified "67 high-risk companies" flagging over "259 million GEL in potential tax losses."

### Conclusion and policy recommendations
- Summary conclusion:
  - Targeting large taxpayers through the LTO "has significantly strengthened revenue collection in Georgia, yielding gains of 0.4 to 0.7 percent of GDP."
- Recommended policy priorities:
  - Maintain a risk-based approach and strong enforcement.
  - Deliver tailored taxpayer services and enhanced communication.
  - Ensure robust governance and continuous monitoring to address strategic avoidance.
  - Strengthen CRM systems.
  - Enhance enforcement while preserving service orientation.
  - Expand LTO coverage to include consolidated economic groups.
  - Carefully manage threshold criteria to sustain and maximize compliance gains.

*Source: 1geoea2025002-source-pdf - 6.*

### 2021. Dotted black lines indicate the lower bound and upper bound of the percentile common support. For Common Support

### 1geoea2025002-source-pdf - 2021. Dotted black lines indicate the lower bound and upper bound of the percentile common support. For Common Support

### Methodology and Common Support Definitions
- Common support [02.5;99.9]: dropped taxpayers below the 2.5th percentile and above the 99.9th percentile of the VAT turnover distribution in either the treated or control group.
- Common support [1;99.9]: dropped taxpayers below the 1st percentile and above the 99.9th percentile of the VAT turnover distribution in either the treated or control group.
- Common support [min; max]: dropped taxpayers below the minimum and above the maximum of the VAT turnover distribution in either the treated or control group.
- Weights constructed by applying the CBPS to estimate the probability of LTO assignment based on VAT turnover, exports, net property value, and number of employees.
- Standard errors are clustered at the taxpayer level.

### LTO First Cohort — Average Effects on Tax Assessments (Table I.2)
- Outcome categories: fDPT (final distributed profit tax), fVAT (final value-added tax), fWHD (final withholding tax), fTOT (final total); and iDPT, iVAT, iWHD, iTOT for initial assessments.
- Estimated InLTO (average effect per taxpayer per year) — Final Assessments:
  - fDPT: 0.175*** (SE 0.026)
  - fVAT: 0.666*** (SE 0.091)
  - fWHD: 0.558*** (SE 0.031)
  - fTOT: 1.388*** (SE 0.103)
- Estimated InLTO — Initial Assessments:
  - iDPT: 0.171*** (SE 0.028)
  - iVAT: 0.589*** (SE 0.090)
  - iWHD: 0.576*** (SE 0.133)
  - iTOT: 1.313*** (SE 0.158)
- Observations and sample counts:
  - Final assessments observations: 20744, 21047, 21041, 21047 (columns 1–4)
  - Initial assessments observations: 15192, 20421, 19566, 20876 (columns 5–8)
  - # of LTO (final): 165, 176, 176, 176
  - # of LTO (initial): 151, 170, 172, 175
  - # of Non-LTO (final): 3440, 3481, 3478, 3481
  - # of Non-LTO (initial): 2238, 3306, 3124, 3424
- Aggregate impacts:
  - Total Effect (million GEL) — Final: 33.237, 126.479, 106.016, 263.787 (fDPT, fVAT, fWHD, fTOT)
  - Total Effect (million GEL) — Initial: 32.418, 111.916, 109.476, 249.453 (iDPT, iVAT, iWHD, iTOT)
  - Total Effect (Percent of GDP) — Final: 0.054, 0.205, 0.172, 0.428
  - Total Effect (Percent of GDP) — Initial: 0.053, 0.182, 0.178, 0.405
- Pre-treatment balance and significance (2017–21 comparisons):
  - Pre-treatment Unweighted Diff. — Final: 0.517, 0.886, 2.335, 3.707
  - P-Value Unweighted Diff. — Final: 0.001, 0.114, 0.000, 0.000
  - Pre-treatment Weighted Diff. — Final: 0.034, -0.514, 0.784, 0.297
  - P-Value Weighted Diff. — Final: 0.340, 0.159, 0.287, 0.719
  - Std. Unweighted Difference — Final: 0.272, 0.139, 0.990, 0.502
  - Std. Weighted Difference — Final: 0.047, -0.146, 0.201, 0.054

### LTO First Cohort — Effects Over Time (Table I.3)
- Interaction estimates InLTO X Year:
  - InLTO X Year=2022 — Final: 0.101** (fDPT), 0.602*** (fVAT), 0.424*** (fWHD), 1.125*** (fTOT)
    - SEs: (0.038), (0.133), (0.045), (0.149)
  - InLTO X Year=2023 — Final: 0.033, 0.629***, 0.443***, 1.103***
    - SEs: (0.036), (0.127), (0.043), (0.143)
  - InLTO X Year=2024 — Final: 0.386***, 0.764***, 0.803***, 1.931***
    - SEs: (0.036), (0.129), (0.043), (0.145)
- Year comparison tests (reported as Effects 2022=2023, 2022=2024, 2023=2024) — Final (examples):
  - Effects 2022=2023: 0.138, 0.869, 0.726, 0.902
  - Effects 2022=2024: 0.000, 0.320, 0.000, 0.000
  - Effects 2023=2024: 0.000, 0.385, 0.000, 0.000
- Aggregate and balance statistics largely mirror Table I.2:
  - Total Effect (million GEL) — Final: 32.879, 126.328, 105.768, 263.412
  - Total Effect (Percent of GDP) — Final: 0.051, 0.205, 0.169, 0.423
  - Pre-treatment Unweighted Diff.: 0.517, 0.886, 2.335, 3.707; P-values and weighted diffs as in Table I.2.

### LTO Second Cohort — Average Effects on Tax Assessments (Table I.4)
- Estimated InLTO — Final Assessments:
  - fDPT: 0.258*** (SE 0.023)
  - fVAT: 0.629*** (SE 0.078)
  - fWHD: 0.631*** (SE 0.029)
  - fTOT: 1.505*** (SE 0.089)
- Estimated InLTO — Initial Assessments:
  - iDPT: 0.256*** (SE 0.026)
  - iVAT: 0.650*** (SE 0.076)
  - iWHD: 0.651*** (SE 0.106)
  - iTOT: 1.542*** (SE 0.131)
- Observations and sample counts:
  - Final assessments observations: 23139, 23473, 23465, 23473
  - Initial assessments observations: 16772, 22722, 21755, 23247
  - # of LTO (final): 245, 259, 259, 259
  - # of LTO (initial): 223, 251, 251, 256
  - # of Non-LTO (final): 3551, 3593, 3590, 3593
  - # of Non-LTO (initial): 2264, 3386, 3197, 3517
- Aggregate impacts:
  - Total Effect (million GEL) — Final: 70.983, 172.992, 173.626, 413.827
  - Total Effect (million GEL) — Initial: 70.343, 178.797, 178.922, 424.047
  - Total Effect (Percent of GDP) — Final: 0.115, 0.281, 0.282, 0.672
  - Total Effect (Percent of GDP) — Initial: 0.114, 0.290, 0.290, 0.688
- Pre-treatment balance:
  - Pre-treatment Unweighted Diff. — Final: 0.483, 0.712, 2.037, 3.206
  - P-Value Unweighted Diff. — Final: 0.001, 0.093, 0.000, 0.000
  - Pre-treatment Weighted Diff. — Final: 0.047, -0.479, 1.210, 0.768
  - P-Value Weighted Diff. — Final: 0.116, 0.118, 0.108, 0.357
  - Std. Unweighted Difference — Final: 0.247, 0.127, 0.863, 0.474
  - Std. Weighted Difference — Final: 0.065, -0.139, 0.302, 0.139

### Triple-Difference: Second Cohort, Yearly Effects and Threshold Changes (Table I.5)
- Interaction categories include Drop Out, Drop In, and Remain LTO, each interacted with Year=2022, Year=2023, Year=2024.
- Selected interaction estimates — Final Assessments (examples):
  - InLTO X Drop Out X Year=2022: -0.020, 0.752***, 0.130, 0.861***
  - InLTO X Drop In X Year=2024: 1.005***, 0.731***, 1.281***, 3.007***
  - InLTO X Remain LTO X Year=2024: 0.932***, 0.982***, 1.665***, 3.478***
  - InLTO X Remain LTO X Year=2022: 0.240***, 0.614**, 0.683***, 1.523***
- Observations and sample counts mirror Table I.4:
  - Final assessments observations: 23139, 23473, 23465, 23473
  - Initial assessments observations: 16772, 22722, 21755, 23247
  - # of LTO (final): 245, 259, 259, 259; # of Non-LTO (final): 3551, 3593, 3590, 3593
- Pre-treatment balance statistics replicate prior tables:
  - Pre-treatment Unweighted Diff.: 0.483, 0.712, 2.037, 3.206
  - P-Value Unweighted Diff.: 0.001, 0.093, 0.000, 0.000
  - Pre-treatment Weighted Diff.: 0.047, -0.479, 1.210, 0.768
  - P-Value Weighted Diff.: 0.116, 0.118, 0.108, 0.357
  - Std. Unweighted Difference: 0.247, 0.127, 0.863, 0.474
  - Std. Weighted Difference: 0.065, -0.139, 0.302, 0.139

### Channels: Compliance and Enforcement Effects (Table I.6)
- Outcome variables and estimated InLTO (first cohort) — point estimates (SEs):
  - Accrual tax paid: 0.869*** (0.111)
  - Add. DPT (binary): -0.126*** (0.028)
  - Add. VAT (binary): -0.054* (0.026)
  - Add. WHD (binary): 0.013 (0.031)
  - Und-pay. Let. (underpayment letter, binary): -0.019 (0.028)
  - Tax Coll. Let. (tax collection letter, binary): 0.108*** (0.025)
  - Late dec. (late declaration, binary): 0.003 (0.017)
  - Late pay. (late payment, binary): -0.004 (0.025)
  - Audit (binary): -0.053* (0.022)
- Observations: 21047 for all columns; # of LTO: 176; # of Non-LTO: 3481.
- Notes on variables:
  - Columns 2–4: binary indicators equal to 1 if the taxpayer makes additional DPT, VAT, and WHD assessments.
  - Columns 5–9: binary indicators equal to 1 if the taxpayer receives an underpayment letter, a tax collection letter, makes a late declaration, makes a late payment, or is audited.

### Conditional Effects: Final Assessments Conditional on Enforcement Measures (Table I.7)
- Models estimate fDPT, fVAT, fWHD, fTOT conditional on three enforcement states: Receive under-payment letter; Receive tax collection letter; Be audited.
- Baseline InLTO (unconditional) — Final:
  - 0.295*** (fDPT), 1.122*** (fVAT), 0.637*** (fWHD), 2.023*** (fTOT) — SEs: (0.036), (0.125), (0.042), (0.140)
- Interaction InLTO X Cond (example coefficients and significance):
  - Conditional on Receive under-payment letter:
    - InLTOXCond: -0.259***, -1.025***, -0.189**, -1.439*** (SEs: (0.054), (0.188), (0.064), (0.212))
  - Conditional on Receive tax collection letter:
    - InLTOXCond: -0.109, -0.404, -0.254**, -0.754** (SEs: (0.069), (0.248), (0.084), (0.280))
  - Conditional on Be audited:
    - InLTOXCond: -0.003, 1.524***, -0.344***, 1.182*** (SEs: (0.075), (0.269), (0.091), (0.304))
- Additional conditional terms reported:
  - Cond, CondXPost, LTOXCond with associated coefficients and SEs (see table for exact values).
- Observations and sample counts:
  - Observations: 20744, 21047, 21041, 21047 across columns.
  - # of LTO: 165, 176, 176, 176; # of Non-LTO: 3440, 3481, 3478, 3481.
- Pre-treatment balance statistics repeat values reported in earlier tables:
  - Pre-treatment Unweighted Diff.: 0.517, 0.886, 2.335, 3.707
  - P-Value Unweighted Diff.: 0.001, 0.114, 0.000, 0.000
  - Pre-treatment Weighted Diff.: 0.034, -0.514, 0.784, 0.297
  - P-Value Weighted Diff.: 0.340, 0.159, 0.287, 0.719
  - Std. Unweighted Difference: 0.272, 0.139, 0.990, 0.502
  - Std. Weighted Difference: 0.047, -0.146, 0.201, 0.054

*International Monetary Fund. Georgia: LTO assignment and tax assessment tables and notes (extracted content).*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2025/english/1geoea2025002-source-pdf.pdf_