Moldova

Public Finance Review

Fiscal Policy for Resilience and Growth

October 2022

1
Contents
Abbreviations and Acronyms ....................................................................................................................... 3
Overview....................................................................................................................................................... 4
Chapter 1: Current Trends in Public Finance ............................................................................................... 9
Chapter 2: Efficiency and Equity of Public Spending................................................................................. 16
Spending on Health, Education, and Social Protection......................................................................... 18
Health ..................................................................................................................................................... 19
Education ................................................................................................................................................ 23
Social Protection..................................................................................................................................... 25
Pensions .............................................................................................................................................. 26
Social Assistance ................................................................................................................................. 29
Chapter 3: Efficiency, equity and simplicity of the tax system ................................................................. 33
Value-Added Tax .................................................................................................................................... 36
Personal Income Tax .............................................................................................................................. 41
Corporate Income Tax ............................................................................................................................ 44
Annex 1: References................................................................................................................................... 49
Annex 2: A Taxonomy of Budget Rigidities ............................................................................................... 51
Annex 3: Tax System in Moldova ............................................................................................................... 52

This report was prepared by a World Bank team led by Stefano Curto. The team comprised Natasha Rovo, Marcel
Chistruga, Mikhail Matytsin, Tom Bundervoet, Kristina Noelle Vaughan, Daniela Dborkin, Julian Folgar, Roman
Zhukovskyi, Sebastian James, Rajiv Kumar, Martin Holmer, Grzegorz Poniatowski, Lucia Casap, Ilie Volovei, Maria
Ines Badin, Ehab Tawfik and Amshika Amar.

The study benefited from the guidance of Arup Banerji (Country Director), Inguna Dobraja (Country Manager),
Jasmin Chakeri (Practice Manager), and Karlis Smits, Baher El-Hifnawi, and Caryn Bredenkamp (Program Leaders).

The World Bank would like to thank the officials of the Ministry of Finance for their participation and support
during the preparation of this report.

2
Abbreviations and Acronyms

APRA Cold Season Benefit

B40 Bottom 40 Percent
CEM Country Economic Memorandum
CIS Commonwealth of Independent States
CIT Corporate Income Tax
CNAM Compania Naţională de Asigurări în Medicină
ECA Europe and Central Asia
EU European Union
FEZ Free Economic Zone
GDP Gross Domestic Product
GMI Guaranteed Minimum Income
HBS Household-Budget Survey
IMF International Monetary Fund
MOF Ministry of Finance
MDL Moldovan Lei
METR Marginal Effective Tax Rate
NATO North Atlantic Treaty Organization
NBM National Bank of Moldova
NBS National Bureau of Statistics
NCD Noncommunicable Disease
NFRLD National Fund for Regional and Local Development
NIR Notional Ideal Revenue
OECD Organisation of Economic Cooperation and Development
OOP Out of Pocket
PAYG Pay-As-You-Go
PFR Public Finance Review
PHC Primary Health Care
PISA Programme for International Student Assessment
PIT Personal Income Tax
PPG Publicly Guaranteed Debt
SM Subsistence Minimum
SME Small and Medium Enterprise
SOE State-Owned Enterprise
SSC Social Security Contribution
T60 Top 60 Percent
UHC Universal Health Coverage
US$ United States Dollar
VAT Valued-Added Tax
WHO World Health Organization
VTTL VAT Total Tax Liability
yoy Year-on-Year

3
Overview
1. Public finance in Moldova has followed three distinct periods over the past two decades. The
first period, from 2000 to the global financial crisis, witnessed a growing public sector that reached more
than 40 percent of GDP. This period was also characterized by strong GDP growth and revenue collection,
leading to rapid debt reduction. The second period, after the 2009 global financial crisis until the COVID-
19 pandemic, witnessed a reduction in the size of public sector (just above 30 percent of GDP), but also
sustained albeit small fiscal deficit. Fiscal consolidation efforts were put in place following the 2014
banking crisis, which led to an improvement of the fiscal balance in the following years. However, with
the policy changes on public wages and the pension system introduced in 2018, fiscal deficit was set to
widen again in subsequent years. The final period is one of consecutive crises: the 2020 COVID-19
pandemic and the extreme drought, the 2021 gas crisis, and the 2022 war in Ukraine. The escape clause
to the country’s fiscal rule in case of an emergency was activated in 2020 in response to the pandemic,
leading to larger deficit.

2. Despite the above differences, the past twenty years in public finance have been characterized
also by some common traits: procyclical and inefficient spending and the revenue collection skewed
towards indirect taxation. Moldova’s fiscal policy has been generally expansionary and has maintained a
structural deficit,1 which increased in the occurrence of crises, as in 2008, 2014 and, most recently, in
2020. Moreover, persisting procyclicality has partially prevented the build-up of resilience, amplifying
macroeconomic volatility, depressing investment, and resulting in boom-and-bust cycles that have
reduced potential growth and affected the poor the most. While the procyclical stance has been relatively
lower than in its peer countries, Moldova’s structural characteristics—typical of a small economy,
including a narrow production base, overreliance of imports, and vulnerability to exogenous shocks
especially natural disaster—call for a much stronger countercyclical stance to mitigate the socio-economic
impact of shocks when they arise. The overall revenue structure remains highly skewed toward indirect
taxes, in stark contrast to revenue collection in Europe. Other European countries collect more than
double income taxes (as a percentage of GDP) than in Moldova.

3. Fiscal policy has also adapted slowly to demographic changes, including a declining and aging
population as well as a large migration. The population has dwindled by almost one-quarter since its
independence and is expected to shrink by another one-third by 2060. Large-scale migration of the
working-age population, combined with decreasing fertility rates, has accelerated the pace of aging. An
aging population, combined with lifestyle habits,2 is having significant implications for health-care
demand, while education spending is higher than most peer countries with a similar level of income, but
with much lower outcomes. The 2016 Public Finance Review (PFR) emphasized that efficiency gains should
be sought, as Moldova could achieve the same results with spending 40 percent less. At 35 percent of
total general government spending, social protection is the biggest spending item and is expected to
increase over time primarily because of the changes made in the retirement age in 2018, and on minimum
pensions in 2020.

4. The war in Ukraine has substantially deteriorated Moldova’s economic outlook and fiscal
position. The impact of the war in Ukraine is expected to further intensify pre-existing price pressures,
which together with depressed economic activity due to mounting uncertainties and trade disruptions,

1With the exception of 2017.

2Against a world average of 74 percent, almost 90 percent of total deaths are caused by noncommunicable diseases, which include cancer,
diabetes mellitus, cardiovascular diseases, digestive diseases, skin diseases, musculoskeletal diseases, and congenital anomalies.

4
are expected to severely affect household purchasing power, private consumption and investment. As a
result, further weakening of Moldova’s fiscal position can be expected, both from a decline in revenues
due to the lower economic activity and from an increase in social spending to mitigate the impact of the
shock. At the same time, Moldova still has a very large unfinished development agenda to support the
country’s economic, social, and structural transformation. Already before the war in Ukraine, the
Parliament had approved the 2022 Budget Law with a deficit of 6.0 percent of GDP to support investments
and the ambitious reform program to transform the economy and create business opportunities and jobs.
Finally, there are also high macro-fiscal risks stemming from the uncertainties over food and gas prices,
the potential contingent liabilities over the settlement of the historical commercial debt between
Moldovagaz and Gazprom, and further weakening of economic activity and/or tightening of financial
conditions. These risks could dampen growth prospects and reduce the willingness of domestic investors
to roll over existing debt.

5. In the current economic juncture and given mounting uncertainties, the authorities will need to
navigate this uncharted territory by striking the balance between short-term objectives of mitigating
the immediate impact of the crisis and supporting long-term inclusive growth in a fiscally sustainable
manner. While the exchange rate will continue to be the first line of defense against external shocks, and
the National Bank of Moldova (NBM)—the central bank—has demonstrated a swift response to the
increase in inflation, the NBM will face policy tradeoffs between controlling inflation and supporting
economic activity, as well as between managing potential pressure on the domestic currency and
maintaining healthy levels of international reserves. Public finance has an essential role to play in
mitigating the short-term impact, supporting medium-term economic recovery, while ensuring sufficient
fiscal space to help maintain momentum on the long-term reform agenda of steering the economy toward
a more inclusive growth model in a fiscally sustainable manner.

Against this background, the design and implementation of a twofold strategy would be necessary:

6. The first element of this strategy would be to retain fiscal flexibility leading to the winter season
with contingency plans in place if a further increase in food and energy prices materializes. Perhaps the
highest risk to Moldova’s economy today is the potential impact of the war in Ukraine on food and energy
prices, which had already increased by almost 20 percent for food and about 400 percent for gas in 2021.3
In such a scenario the impact would be substantial with poor households expected to be the most
affected. Efficiency gains and budgetary reallocations will need to be quickly identified to partially mitigate
the impact on households, while preserving fiscal sustainability. Given the limited room for maneuver by
the central bank and because of an already high fiscal deficit, the potential policy response to these risks
will need to focus on maintaining the same pre-shock deficit and preserving the fiscal framework and
targets of the IMF program. Because of limited feasibility of changes in tax policy in the current
environment and the strong fiscal multiplier of taxes in Moldova, letting revenues play the role of
automatic stabilizers would likely be the most effective policy option.

7. Without efficiency gains in spending and reallocation efforts, it will be hard to preserve the
purchasing power of the households and counteract the poverty increase in case the risks materialize.
Most of the efficiency gains should come from flexible categories, such as goods and services and
subsidies. There is room for improvement in the procurement system, but also important gains could be

3 While the share of the Russian Federation and Ukraine in total imports represents about 25 percent, Moldova imports 100 percent of its gas
from the Russian Federation as well as various inputs for the economy, including fertilizers and pesticides for agriculture. Moreover, Transnistrian
power plant from which Moldova buys about 85 percent of the total energy, relies on imported gas from the Russian Federation to generate
electricity.

5
achieved by prioritizing state aid. However, these efficiency gains will not be sufficient to contain the fiscal
deficit to the pre-shock level. Capital expenditures will have to be reduced to levels similar to the average
of the past few years and, at the same time, it will be important to apply a proper public investment
management system. The space generated by these measures should be channeled to effective and well-
targeted social protection programs, and help refrain from the temptation to resort to using costly and
inefficient generalized subsidies, which have large leakages to the high deciles of the income distribution
that consume more. This is not only important from a social perspective but also from a macroeconomic
point of view, as the decile at the bottom of the distribution has a much stronger propensity to consume
(close to 1) and therefore able to generate a stronger multiplier effect. As demonstrated during last
winter’s gas crisis, the measures on APRA (the Cold Season benefits) introduced by the Government are
expected to have contained the increase in the poverty rate by about 1.5 percentage points, from an
increase of 6.7 percentage points in the absence of mitigation measures to an increase of 5.2 percentage
points after the mitigation measures.4 The recently amendments to the Law on Ajutor Social is a welcome
step in the right direction to strengthen the main means-tested anti-poverty program, which together
with APRA could channel resources effectively.

8. The second element of the strategy is to advance in parallel medium-term fiscal reforms to
improve spending efficiency and create the space necessary to support the long-term growth, while
safeguarding fiscal sustainability. Improving the efficiency of social spending will be critical in achieving
both social objectives and value for money. Reforms to improve the effectiveness, efficiency, and equity
of public spending should not be put on hold but should be advanced to fully take advantage of the global
economy when it recovers. Striking the right balance between short- and long-term objectives will likely
require improving the intra-sectoral efficiency of spending to create the fiscal space for increased
spending on human development, social assistance, and infrastructure within a tighter budget.5 Efficiency
gains might come from improvements in public procurement through the implementation of digital tools
and procedures, and the strict monitoring of contractual terms. This will be particularly important in the
health sector, but not only there. In the medium-term, network optimization and financing mechanisms,
in particular for higher education and hospital sector rationalization, remain a priority.6 Moreover, a
comprehensive review of the framework for public employment compensation, performance incentives,
adequacy of skills, and career management will be essential to provide effective and efficient service
delivery. It is also important to continue the efforts to improve the social protection system to better
target social assistance spending, keeping fiscal cost of pensions manageable. This will require continued
reform efforts both in improving social assistance, namely in the direction of strengthening the means-
tested anti-poverty program, the Ajutor Social, and further tightening the link between the contributions
and benefits in pensions. Finally, addressing the inefficiencies in capital spending is key to closing the large
infrastructure gap in railways, energy, and water and sanitation. Recent reforms through the
establishment of the National Fund for Regional and Local Development (NFRLD) are a step in the right
direction.7 Strengthening capital investment, as part of a broader “green” fiscal policy approach, will also
help build resilience to climate change.

4 In addition to APRA, the package of measures included a compensation to the tariff of natural gas differentiated by consumption: MDL 4.3 for
the first 50 cubic meters and MDL 3.2 above the first 50 cubic meters. The package also included a compensation of 80 percent to the heating
tariffs up to MDL 815 per month. While the compensation to the tariff of natural gas have been important to support the household, they were
less effective than APRA given their leakages to high income deciles.
5 Cournède, B., A. Pina and A. Goujard (2013), “How Much Scope For Growth And Equity-Friendly Fiscal Consolidation?”, OECD Economics
Department Policy Notes, No. 20 July 2013.
6
Public Finance Review 2016.
7 For the newly established framework, but also for relevant for other public investments to improve investment efficiency, efforts are needed

for: (i) improving preliminary screening and project appraisal mechanisms; (ii) improving the selection of new projects and ensure continuity of
funding for ongoing projects through better prioritization and budgeting processes; and (iii) strengthening the monitoring of project
implementation for cost efficiency and timely delivery of public services.

6
9. Meanwhile, a gradual shift toward domestic revenue mobilization, underpinned by a more
progressive and less distortive taxation system, is necessary to support resilience and rebalance
growth.8 A more equitable system is also necessary to strengthen the social contract and build consensus
around the reform program. On the households’ side, both personal income tax (PIT) and value-added tax
(VAT) can be improved along the principles of tax simplicity, efficiency, equity, and progressivity, while
also increasing revenues. To inject progressivity in the flat PIT system, a number of personal deductions
and exemptions have been included over time, making the system inefficient and cumbersome to
administer. In the short-term, it is important to rationalize and optimize tax expenditures following
international best practices. In the medium-term, it will be critical to review the overall PIT system more
comprehensively, including looking at the possibility of shifting to a progressive PIT system with income
brackets and only a limited number of deductions. As for VAT, the only viable short-term option is
currently to focus on rationalization and optimization of tax expenditures, with a particular emphasis on
minimizing the impact on households, particularly the poor. On corporate taxation, the current corporate
income tax (CIT) system has a wide range of tax facilities and exemptions from the standard applied rate,
leading to overall deficiencies in collections and increasing reliance on large-size taxpayers. In the short-
term, tax incentives should be amended to ensure that they are relevant to Moldova’s development
strategy, effective in providing the right incentives to achieve the stated legislative objectives, do not
generate unintended distortions across and within sectors, and are efficient in using scarce public
resources. In the medium-term, a holistic approach is required to increase CIT collection and broaden the
tax base, while supporting investment and job creation.

10. A stronger countercyclical fiscal stance, including through a revision of the fiscal rule, is needed
in the medium-term to favor a more stable and growth-friendly environment and ensure prudent fiscal
management. Moldova is vulnerable to external shocks, particularly from natural disasters, and exposed
to policy and political volatility, a lack of access to capital markets, and high reliance on external financing
on top of demographic challenges. A stronger countercyclical stance is paramount. The fiscal rule
introduced in Moldova in 2014 and suspended in 2020, does not promote countercyclical fiscal policy.
Going forward, it is essential to strengthen the fiscal rule by introducing a cyclically adjusted component.
To ensure simplicity both to help management and communication it is recommended to explore an
expenditure rule that would constrain spending in booms/good times. This could be achieved by having a
ceiling to current expenditure growth not above GDP potential. The introduction of an adequate fiscal
rule should be accompanied by the strengthening of institutional arrangements, such as fiscal councils,
monitoring mechanisms, and enforcement mechanisms. Spending composition also matters. A more rigid
spending structure reduces the room for a countercyclical response during crises. The development of the
domestic debt market could also be an important aspect to support the country’s developmental needs,
help reduce debt roll-over and interest rate risks, and overdependence on official assistance.

11. Finally, it will be important to strengthen the fiscal relationship framework between the central
government and local governments, public agencies and state-owned enterprises (SOEs) to minimize
contingent liability risks. It will be important to introduce a transparent framework with formulas based
on needs, developmental impact, performance and the capacity for intergovernmental transfers,
including to local governments and SOEs. This incentive framework is essential to support an effective and
efficient use of fiscal resources by local governments, public agencies and SOEs.

8 As background for this report, the team has developed two models to replicate the PIT and CIT systems in Moldova, using the HBS and a sample
from the administrative data, respectively. The models can be used not only to replicate the current systems but also to simulate policy scenarios.

7
 Table 2: Summary of policy recommendations and their impact on efficiency, equity, and fiscal balance

Area Action Timeline Expected Impact

Efficiency Equity Fiscal
Health Strategic purchasing, procurement, health-care ST
providers agreements.
Redesigning/rationalizing hospital system and MT
redirecting resources to primary health care
Increase excise taxes on tobacco, alcoholic and ST
sugar-sweetened beverages (health tax)
Education School network optimization and financing MT
mechanisms/ new funding formula
Public employment compensation policy, MT
performance incentives, adequacy of skills and
career management
Pensions Further tightening the link between the MT
contributions and benefits/privileged pensions
Social Redirecting support from generalized ST
Protection subsidies/price controls to more effective and
efficient means tested programs (e.g., cold season
benefit, APRA)
Further strengthening Ajutor Social and other social ST
assistance programs while providing adequate
resources
Personal Optimization/rationalization of PIT personal ST
Income Tax deductions and exemptions
Introducing a progressive system with income MT
brackets while removing the majority of personal
deductions, exemptions and allowances.
Value-Added Optimization/rationalization of tax expenditures ST
Tax while avoiding adverse impact on the poor
Corporate Introducing a progressive system with preferred
Income Tax rate to SMEs while removing the preferred tax rates
by sectors (e.g., agriculture)
Tax Incentives Optimization/Rationalizing of tax incentives and ST
on switch from pure exemptions to tax credits
Investments
Transfers to Introduce a transparent framework with formula MT
Subnational based on needs, developmental impact, and
Governments performance, while considering agencies capacity
and SOEs
Fiscal Rules Introduce cyclically adjusted fiscal rules, simple to MT
manage and communicate, to reduce procyclicality
of spending and budgetary rigidities while ensuring
debt sustainability

Expected Impact NEGATIVE NEUTRAL POSITIVE

LARGE MEDIUM SMALL SMALL MEDIUM LARGE

8
Chapter 1: Current Trends in Public Finance
12. Moldova’s public finance can be divided over the past two decades into three distinct periods.
The first period, from 2000 to the global financial crisis, witnessed a growing public sector that reached
more than 40 percent of GDP. This period was also characterized by strong GDP growth and revenue
collection, leading to rapid debt reduction. The second period, after the 2009 global financial crisis until
the COVID-19 pandemic, witnessed a reduction in the size of public sector (just above 30 percent of GDP),
but also sustained albeit small fiscal deficit. Fiscal consolidation efforts were put in place following the
2014 banking crisis, which led to an improvement of the fiscal balance in 2017 and 2018. However, with
the policy changes on public wages and the pension system introduced in 2018, fiscal deficits were set to
widen again in subsequent years. The final period is one of consecutive crises: the 2020 COVID-19
pandemic and the extreme drought, the 2021 gas crisis, and the 2022 war in Ukraine. The escape clause
to the country’s fiscal rule in case of an emergency was activated in 2020 in response to the pandemic.
Spending increased by over 5 percentage points of GDP in 2020, while revenues were impacted by a weak
economic activity. As a result, the fiscal deficit reached a 10-year record of 5.3 percent of GDP in 2020,
compared with the average of only 1.5 percent between 2010 and 2019.

Figure 2. Revenue, spending, public debt (% of GDP) Figure 3. Primary and overall fiscal balance (% of GDP)

Source: MoF, World Bank staff calculations. Source: MoF, World Bank staff calculations.
Note: Starting with 2010 a new GDP methodology was applied.

13. Strong growth helped stabilize public and publicly guaranteed (PPG) debt at around 35 percent
of GDP over the past decade. PPG debt has been on a downward trend since its peak of about 40 percent
of GDP in 2016, with the exception of an increase in 2020 due to the pandemic-induced increase in
borrowing and the denominator effect from the economic contraction. The strong rebound in economic
activity in 2021 has contributed to a decline in PPG debt to about 31 percent of GDP. Two-thirds of PPG
debt is external and held mainly by multilateral and bilateral development partners; it is mostly medium-
and long-term debt and on concessional terms. One-third is domestic debt, of which 30 percent is long-
term debt securities (government securities with maturity longer than one year). Other domestic
marketable debt is mainly short-term that needs to be rolled over on an annual basis and held by the
banking system. In 2022, the external debt sustainability was assessed as low risk of debt distress,
however the overall risk of debt distress was reclassified to moderate. Moldova’s debt trajectory is
particularly vulnerable to weaker GDP growth, but public debt carrying capacity is assessed as strong.

14. Moldova’s fiscal policy has been generally expansionary and supportive to the economy. The
cyclically adjusted primary balance is an important indicator to assess the structural fiscal stance, by
disentangling the cyclical changes from the discretionary actions of the Government. Fiscal policy is

9
considered contractionary when the cyclically adjusted primary balance increases, and expansionary
when the cyclically adjusted primary balance decreases. Except for 2017, the cyclically adjusted primary
balance has maintained a structural deficit, which increased in the occurrence of crises, as in 2008, 2014
and, most recently, in 2020, signaling an expansionary stance of fiscal policy. Moreover, the fiscal impulse,
which measures the discretionary actions of the Government, has prevailed over the automatic features
of the fiscal framework.9 As shown in the figures below, fiscal impulses (in red) were often positive when
the economy was already booming, and negative when the economy was in a downturn, with limited but
notable exceptions (2007, 2016, 2017, and 2020). This denotes a procyclical behavior of fiscal policy.
Automatic stabilizers (in green) should by nature react to the business cycle, with revenues decreasing
and some spending increasing in times of crisis, and this has been generally the case in Moldova.

Figure 4. Cyclically adjusted primary balance Figure 5. Fiscal impulse, automatic stabilization,
(% of potential output) output gap (% of GDP)

Source: IMF WEO; World Bank staff calculations.

Note: Methodology based on Abdih, Y., Lopez-Murphey, P., Roitman, A., and R. Sahay (2010), "The Cyclicality of Fiscal Policy in the Middle
East and Central Asia: Is the Current Crisis Different", IMF Working Paper.

15. Persisting procyclicality, albeit relatively lower than in peer countries, has partially prevented
the build-up of resilience. While fiscal policy has followed a common global trend, becoming less
procyclical in the past decade, Moldova was not able to markedly change its procyclical bias. Similar to its
peers, Moldova has displayed a procyclical fiscal policy. While the correlation between the cyclical
components of its public spending and GDP growth (see Box below) is almost half of its peers’ average
(0.3 vs 0.6 percent), Moldova’s structural characteristics typical of a small economy, including a narrow
production base, overreliance of imports, and vulnerability to exogenous shocks, call for a much stronger
countercyclical stance and require to build fiscal resilience to mitigate the socio-economic impact. Since
2000, Moldova has witnessed, on average, one major natural disaster every three years, with the 2020
drought being just one of the latest in a long and increasingly frequent series of weather-related events
that are being exacerbated by climate change. Moreover, the energy crisis which started already in 2021
has highlighted Moldova’s untenable energy balance, characterized by expensive, volatile and
undiversified energy sources and imports from one single supplier. Moldova’s procyclical fiscal stance
tends to amplify macroeconomic volatility, depressing investment and resulting in boom-and-bust cycles
that reduce potential growth and affect the poor. As mentioned earlier, the main driver of the procyclical
fiscal stance has been discretionary spending that outweighs the effect of automatic stabilizers.

9
To assess whether the Government responds proactively to the business cycle, the automatic features built (such as a fall in tax revenues or a
rise in unemployment benefits when the economy is in crisis), known as automatic stabilizers are distinguished from the discretionary actions,
namely the fiscal impulse. The fiscal impulse is defined as the change in the cyclically adjusted primary balance. When it is negative (positive), it
means the cyclically adjusted primary balance has decreased (increased), including if it goes from a deficit to a larger (smaller) deficit, reflecting
an expansionary (contractionary) fiscal position caused by fiscal policy with relatively less (more) tax revenue and relatively more (less) spending.

10
16. The sustained spending procyclicality in Moldova is mostly driven by capital spending and goods
and services. Moldova’s wage bill and social benefits—the two biggest spending items in the budget—
have followed countercyclical dynamics throughout the entire period but have become more procyclical
over time. Indeed, Moldova’s wage bill registered a statistically significant negative correlation of -0.39
(countercyclical), while the correlation of social benefits was -0.29 (statistically insignificant) and acyclical.
While the countercyclical wage bill is a positive aspect, it should be noted that the wage bill is permanent
(rigid) in nature and therefore does not necessary act as a countercyclical measure in the medium to long
run, as social benefits (with the expiation of pensions) would do. At 0.49, public investment displayed
procyclical behavior. Given the constraints on Moldova’s absorptive capacity, the procyclicality of public
investment reduces its effectiveness in terms of fiscal multipliers.

Figure 6. Fiscal policy procyclicality, 2000–19 Figure 7. Spending procyclicality, 2000–19

1.2
Correlation between cyclical component of Gov spending

1.0

Procyclical
0.8
0.6
0.4
0.2
and real GDP

0.0
Countercyclical

-0.2
-0.4
-0.6
-0.8
-1.0
Finland

Ukraine
Canada

Austria

Italy
Belgium

Spain

Greece
Croatia
Switzerland

Australia

Moldova
USA
Turkey

UK

N. Macedonia

Bulgaria
Albania

Poland
New Zealand
Kosovo
Netherlands

Belarus

Romania
Serbia
France
Denmark

Germany

Norway

Russia

Hungary
Montenegro
Sweden

Bosnia & Her.

E&D Europe Advanced Economies

Note: A positive correlation implies procyclical government spending policy, as the country is expanding public spending during good times and
cutting it in bad times. Conversely, a negative correlation means countercyclical spending policy.

Figure 8. Cyclicality of main spending items Figure 9. Spending procyclicality pre- and post-2009
0.8 0.64*
Correlation between cyclical component of

0.51 0.49
0.6
0.49*
0.4
real spending and GDP

0.45 0.21
0.37
0.31 0.10
0.2
0.12
0.0 -0.15
-0.13 -0.23
-0.2
-0.26
-0.4 -0.28 -0.29
-0.39*
-0.6
Total Wages Social Gs&Ss Interest Aq. Non
Spending Benefits Fin
2000-2019 2000-2009 2010-2019 Assets

Source: World Bank staff calculations based on IMF WEO database.

Note: Countries below the 45-degree line are those that have become less procyclical (or more countercyclical) between these two periods.

11
 Fiscal Procyclicality
The fiscal stance over the business cycle is key to building fiscal resilience. While the fiscal stance over the
business cycle is important for all countries, it is even more important for small countries with large agriculture
sector, such as Moldova, that are exposed to very large GDP swings and are in need for a forceful countercyclical
action. It is also important for debt sustainability, as fiscal multipliers vary depending on where the economy is in
the business cycle. It is also important for inflation and the current account balance, as the additional contribution
to aggregate demand would be confronted with the output gap and the capacity of the private sector response
to an increase in demand.

Figure a. Moldova’s potential and real GDP Figure b. Moldova’s GDP cycle (HP filter)
35 1.5

Millions
Millions

1.0
30

0.5
25
0.0
20
-0.5

15
-1.0

10 -1.5
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019

2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
Source: IMF WEO, World Bank staff calculations.

A fiscal policy is defined as procyclical (countercyclical) if its expansionary (contractionary) during booms and/or
contractionary (expansionary) during crises. If there is no significant stance, we can define fiscal policy as
“acyclical.” Fiscal policy should therefore actively smooth the business cycle by lowering taxes and increasing
expenditure in bad times, thereby increasing aggregate demand, while reducing expenditure and increasing
savings in good times (Halland and Bleaney, 2011). While automatic stabilizers such as taxes and employment
insurance and other social protection measures play an important role, discretionary measures are key for the
fiscal policy stance over the business cycle. The introduction of stabilization funds cyclically adjusted fiscal rules,
and reforms of budget institutions are some possible options to achieving a more countercyclical fiscal policy
(Braun, 2003). A simple way to evaluate procyclicality is to compare the cyclical component of each different fiscal
instruments (revenue and spending) with the cyclical component of real GDP throughout time.

17. A positive feature of Moldova’s fiscal situation is the low level of budget rigidities that helps
short-term budgetary reallocations in the case of new priorities emerging and/or shocks materializing.
Moldova is among the countries with the lowest level of budget rigidity in Europe. However, budget
rigidity increases significantly when accounting for social security and health fund budgets and transfers
to local governments, the main source of financing at the local level. When compared with structural, as
well as aspirational, peer countries, Moldova’s spending is more flexible, with rigid items—wages, interest
payments and social benefits (mostly related to social security but also including social assistance
programs)—amounting to about 60 percent of the overall general government spending (see Box below
for methodology). Peer countries with a more rigid spending composition tend to also show a higher share
of social spending (Lithuania, the Slovak Republic, Slovenia, Serbia and North Macedonia, accounting for
amounts to 40 percent of total expenditure, much higher than the 34 percent in Moldova) or a higher
wage bill share (Lithuania, Estonia, Latvia and Romania, at around 30 percent of total spending, also much
higher than that observed in Moldova). While the recent decisions to increase the minimum wage and
pensions in 2022 have increased rigidity in the budget, it still remains low compared to other countries.

12
 Figure 10. General government expenditure rigidities, international benchmarking (% of total expenditure),
2019

Source: GFS-IMF.

18. When revenue earmarking is also added to the analysis, flexible fiscal space is reduced to only
12 percent of total spending. By using the more granular data provided by BOOST dataset,10 high rigid
items reach almost 74 percent of total spending, mainly driven by social benefits, the health fund, and
transfers to local governments. When spending funded by earmarked resources is also considered, the
estimated budget rigidity in this case increases to 84.5 percent. This is mostly explained by capital
expenditure, which is financed from earmarked sources. Moreover, some budget items cannot be
changed in the short-term. These medium-term rigid items, related to subsidies to state-owned
enterprises (SOEs), the Road Fund, and the private sector, amount to another 3.4 percent of total
spending, further reducing the capacity for reallocation in response to shocks or emerging priorities, with
flexible fiscal space reduced to only 12 percent.

Table 3. Rigidity analysis by economic classification and source of financing. Central government, including
social security and the health fund

2015 (% of total expenditure) By source of financing TOTAL

By economic classification EARMARKED NON-EARMARKED
HIGH 2.8% 70.7% 73.6%
Wages 0.5% 7.7% 8.2%
Interest Payments 0.0% 2.5% 2.5%
Social Benefits 1.3% 31.8% 33.0%
Goods and services - Health fund 0.0% 12.2% 12.2%
Transfers to local goverments 1.0% 16.7% 17.7%
MEDIUM 0.1% 3.4% 3.5%
Subsidies 0.1% 3.4% 3.5%
LOW 11.0% 12.03% 23.0%
Capital Expenditure 10.1% 1.5% 11.6%
Goods and services (other than health fund) 0.8% 8.3% 9.1%
Other expenditure 2.3% 2.3%
TOTAL 13.9% 86.1% 100.0%
Source: World Bank staff calculations based on BOOST data from 2015.

10
The latest comprehensive BOOST dataset is from 2015.

13
 Budget Rigidity
Budget rigidities limit the fiscal space for governments to adjust the size and structure of the public budget in the
short-term. These constraints can be of an institutional, legal or contractual nature. Several budget components
are naturally inflexible, such as wages, pensions and debt service. However, there are many other inflexibilities
that are rooted in the constitution, laws, or decrees that earmark revenues, set minimum spending requirements,
or link spending to the evolution of certain macroeconomic variables, such as inflation, growth, or unemployment.
Budget management is negatively affected by budget rigidities. These limit the reallocation of public spending in
response to changing needs, promoting poor quality fiscal adjustments, generating a bias toward higher spending
and taxation, and introducing distortions into tax policy choices. The scope for countercyclical fiscal policy is
limited and incentives to improve the efficiency of public spending are weakened.
Public expenditure rigidities arise from different sources. While some degree of budget rigidities may be desirable,
a highly inflexible budget can lead to difficulties in fiscal policy management. Spending rigidities are the result of
past commitments and arrangements made by societies regarding the provision of public goods. Indeed, budget
rigidities result from different factors, as they reflect preferences of the societies about the role of the state,
public policy priorities and their means of financing (Cetrangolo, 2010). While they can help protect some
spending categories (such as public investment) from discretionary policy actions, an extremely rigid budget can
limit the scope of action of fiscal policy, as such rigidity may constrain fiscal consolidation strategies when these
are required.
Budgetary rigidities can change over time and across countries, but they are generally related to legislation or
contractual agreements. Although in the long-term public spending could be more flexible, in the short to
medium-term there are many constraints that limit rapid budget adjustments. For instance, debt commitments
or the size of the social security system place limits on the proportion of public resources that can be easily
reassigned. To identify these different sources of budgetary rigidities, Cetrangolo (2010)11 proposes a seven-item
classification (Annex 1). In this taxonomy, rigidities arise from a diversity of sources, such as: the benefit principle
funding (which relates to expenditures that benefit those who previously contributed), rights and guarantees
established by legislation (protecting spending in prioritized areas), arrangements between different levels of
governments (more relevant in federal or decentralized institutional settings), macroeconomic dynamics
(reflecting interest payments, impact of exchange rate, inflation adjustments), and the size and arrangements
within the public sector (wages).
A deeper analysis of rigidities can be made by exploiting BOOST data. This source of information allows for a
better disaggregation to detect additional expenditure items that suits the categories determined by Cetrangolo
(2010). It also allows for the detection of inflexibilities arising not only from the economic classification of
expenditures but also from their source of financing, separating spending items financed by general revenues
from those categories financed by special funds. Using Alier’s taxonomy (2006),12 budget rigidity results from the
interaction of inflexibilities on both revenue and spending sides, and this interaction creates four categories of
public spending: (i) mandatory expenditure financed by an earmarked source; (ii) mandatory expenditure without
an earmarked source; (iii) discretionary spending financed by earmarked sources; and (iv) discretionary spending
without an earmarked source.

19. Inter-governmental transfers, which account for about 20 percent of total expenditure, are an
important category that affects not only the overall budget rigidity but also the consolidated fiscal
sustainability. Spending at the local level weighs on overall spending and is not sustained by adequate
local revenue generation. Local governments in Moldova spend 8.1 percent of GDP, equivalent to one-
quarter of total spending, much higher than in structural peers, but on a par with the average of
aspirational peers. Most transfers are toward the capital city, Chișinău, which in turn accounts for around

11
Cetrangolo, O., J. Jimenez and R. del Castillo (2010). “Rigidities and fiscal space in Latin America: a comparative case study”, Working Paper 97,
United Nations, Economic Commission for Latin America (ECLAC), (Santiago de Chile).
12 Alier, Max (2006). “Measuring Budget Rigidities in Latin America,” IMF Working Paper, forthcoming, International Monetary Fund, Washington, DC.

14
60 percent of Moldova’s GDP. While inter-governmental transfers have declined over time, the ratio of
local spending funded by transfers from the central government (the so-called vertical fiscal imbalance)
remains particularly high, at 75 percent, given the high dependence of local governments on these
transfers. Moldova’s vertical fiscal imbalance is almost double the average of all comparators, with only
Estonia and Lithuania having a higher ratio, above 85 percent of total spending. This is particularly
important in terms of efficiencies, given that Moldova's current administrative-territorial structure is
suboptimal compared with the size of the country and its population (estimated at 2.7 million in 2019).
There are 35 top-tier local administrative jurisdictions (including two cities, 32 raions [districts], and the
Autonomous Territorial Unit, Gagaúzia) and 896 local jurisdictions with a median land area per
municipality of only about 3 square kilometers. Half of the rural bottom-tier municipalities have fewer
than 2,000 residents, and more than one-quarter have fewer than 1,500 residents. The extent of vertical
fiscal imbalances across the local jurisdictions is a function of the level of per capita income, fiscal capacity,
and demographic characteristics, as well as the central government’s fiscal behavior that reflects fiscal
constraints and policy preferences at the national level (IMF, 2014). Apart from Chișinău and Gaugazia,
most municipal governments depend heavily on transfers from the central government, with a limited
capacity to collect compared with their needs.

Figure 11. Subnational government (SNG) spending and vertical fiscal imbalances (VFI) 2019, Moldova and
other countries (% of total general government spending)

Source: IMF GFS.

Figure 12. Subnational government spending and horizontal fiscal imbalances, 2018

Source: BOOST, 2018.

15
Chapter 2: Efficiency and Equity of Public Spending
20. While public spending has grown on a par with nominal GDP and some categories have grown
even much faster, capital spending has declined in real terms over the past 10 years. Total expenditure
decreased from over 45 percent of GDP in 2009 to close to 31.5 percent in 2019, following a decade of
fiscal consolidation. Fiscal consolidation had been driven mostly by reductions in wages and goods and
services, which have grown at just below 3.5 percent per year in real terms. However, wages have declined
in recent years because of the 2018 reform of the public wage bill, which introduced a new unitary pay
system with the objective of making the system simpler and more transparent. However, these categories
have shrunk as a percentage of GDP over the past 10 years, while transfers are the category that has
grown the most, at the same rate as total spending (over 4 percent per year in real terms). Public
investment, on the other hand,13 has stagnated since 2017. As a result, transfers remain the largest
category in public spending (about 34 percent) followed by wages (about 23 percent). In terms of
functional classification, spending in health has almost doubled over the past 10 years, though most of
the increase has been during the past three years because of the COVID-19 pandemic. Social protection,
on the other hand, has grown steadily at a faster pace than total spending (over 4 percent per year in real
terms). Education instead has remained relatively flat in real terms over the past decade. As a result,
spending in health has surpassed spending in education as the second-largest category, while social
protection remains the largest category at about 34 percent of total spending.

Figure 13. Evolution of public spending (economic and functional classification)

a. Evolution in real terms b. Composition

Source: Ministry of Finance. Source: Ministry of Finance.

13
Public investment is defined as non-financial assets after 2017 and capital expenditure before 2017. The two series are not fully comparable.

16
21. When compared with a set of peer countries, Moldova’s public spending was among the lowest
in the region, but about 2 percentage points higher than the level of its income per capita would imply.14
Given the heterogenous policy response to the COVID-19 pandemic and fiscal position before the crisis,
the comparison is based on 2019 data. At that time, social benefits—the largest spending category—was
below its peers’ average. Public wages—the second-largest category—were in line with structural peers
and just below the average of other peer countries. At 10 percent of total spending, capital spending was
slightly higher than the average of its peers, but below its structural peers with similar development
needs, such Georgia, Kosovo, Albania, and Armenia, with 24, 23, 19 and 12 percent of total spending,
respectively.

Figure 14. General government spending (% GDP) and Figure 15. General government spending composition
GDP per capita (constant, PPP), 2019 (% of GDP), 2019

60 45

Thousands
48.3 40
50 45.8
43.6 42.1 35
41.0 40.5 41.6 41.3
40 38.2 38.0 30
34.6 36.0
31.4 32.5
28.5 29.1 29.5 25
30 25.6 20
20 15
10
10
5
0 0
Hungary

Kosovo
Slovak Republic

Czech Republic

Ukraine
Bosnia and Herzegovina

Romania

Albania
Moldova

Serbia

Slovenia

Estonia

Latvia

Lithuania

Georgia

Armenia

Belarus

North Macedonia

Aspirational Neighbors Structural

GG Capital Spending (% GDP) GG Current Spending (% GDP) GDP per capital, constant PPP

Source: IMF GFS.

22. Over the last period, spending has increased rapidly to mitigate the impact of these subsequent
crises on economic activity, on firms and households, and to meet health-care needs, while revenues
have been subdued. In 2020, owing to the fiscal response to the crisis induced by the COVID-19 pandemic,
total spending reached a record high of 37 percent of GDP and remained high at 34 percent also in 2021.
Current spending reached 31.4 percent by end-2020 and stayed high at 29 percent in 2021. In 2021, health
(+35.4 percent, yoy) and social protection (+13 percent, yoy) were the main drivers of the spending
increase, notably after the increase in minimum pensions. Investment increased by 17.6 percent despite
lower execution of capital investments. Revenue collection rebounded strongly (+23.5 percent, yoy) after
the relaxation of restrictive measures. Due to lower execution and higher-than-expected revenue, the
fiscal deficit in 2021 reached about 2 percent of GDP, almost half of the planned deficit. This strong fiscal
performance in 2021, due to the economic rebound, was reversed in 2022 when the budget foresaw a
higher deficit to support the ambitious government’s reform agenda. Already before the war in Ukraine,
Parliament had approved the 2022 Budget Law with a deficit of 6.0 percent of GDP to support investments

14Comparator countries are divided into three groups. Structural peers include North Macedonia, Albania and Kosovo. Neighboring peers include
Belarus, Ukraine and Romania. Aspirational peers include Serbia, Hungary, Slovenia, the Slovak Republic, Bosnia and Herzegovina, the Czech
Republic, Estonia, Latvia, Lithuania, Georgia and Armenia.

17
and the ambitious reform program. The budget foresaw a substantial increase in social spending (social
assistance and pensions), reflecting the recent increase in minimum pensions, the indexation of pensions,
and measures to mitigate the energy price shocks.15

23. In early 2022, spending was repurposed toward supporting the population, in a concerted effort
with monetary policy to deal with the consequences of rising prices. Since the end of 2021, due to a
combination of a post-pandemic demand boom and new supply disruption, as a consequence of the war
in Ukraine, inflation has been on the rise. As of May 2022, inflation in Moldova had reached 27 percent
yoy, mostly driven by food and energy prices. Food, utilities, and transport make up for almost 80 percent
of the consumption basket of the bottom 40 percent (B40), and about 64 percent of overall consumption
of the top 60 percent (T60) (Figure 16). In an effort to support households to cope with increasing energy
prices first, and food prices more recently, the Government has stepped up fiscal support and the share
of spending directed to social protection has increased significantly compared with previous years.

Figure 16. Consumption basket: B40 and T60 Figure 17. Total spending: 2022 vs previous years
(January–April)

Source: Household Budget Survey. Souce: Ministry of Finance.

Spending on Health, Education, and Social Protection

24. Moldova spends almost 70 percent of its total general government spending on social sectors.
General government spending on social protection, which includes pensions and social assistance,
together with spending on education and health, accounted for 21 percent of GDP in 2019, equivalent to
almost 70 percent of total spending. The decline observed over the past few years was mainly driven by
sectors such as education and health, but the trend has been recently reversed by the need to deal with
the health and economic consequences of the COVID-19 pandemic first, and the energy crisis and the
consequences of the war in Ukraine subsequently. Moldova’s social spending is on a par with its peers,
except for spending on education, which remains higher than in both structural and aspirational peers.
However, spending on health and social protection is in line with structural peers, and slightly lower than
spending by aspirational peers. Social protection is the biggest spending item, at 35 percent of the total,
and is expected to increase over time primarily because of the changes made in the retirement age in
2018, and on minimum pensions in 2020. Moldova’s spending on education remains still higher than that

15
The 2022 budget included MDL 1.8 billion (0.7 percent of GDP) to mitigate the impact of increase in gas prices.

18
of both structural and aspirational peers’ allocations to this sector, at 5.8 percent of GDP in Moldova
versus an average of 4.4 and 4.6 percent, respectively, among structural and aspirational peers.

25. The relatively high level of spending on education is confronted to a declining and aging
population as well as outmigration. Moldova’s total fertility rate is currently below the world average
(1.77 vs 2.40), while the dependency ratio is well above the world average (69 vs 55 percent). While the
population has declined by 6 percent since 2014, the school-age population (6–18 years old) in Moldova
has declined by more than 9 percent due to low fertility rates, as well as outmigration. This suggests that
the population in Moldova will both age and decline in the coming years. By 2060, the population is
projected to decline by 29 percent, to 1.2 million, and the share of older people to triple to 30 percent.

26. Efficiency gains should be sought to generate fiscal space to support emerging needs and long-
term trends. An aging population, combined with lifestyle habits, with almost 90 percent of total deaths
caused by noncommunicable diseases16 against a world average of 74 percent, is expected to have
significant implications for health-care demand. Meanwhile, Moldova’s public educational spending
remains high, despite recent reform efforts. At 5.8 percent of GDP in 2019, Moldova’s government
expenditure on education is higher than that of most countries, in particular those with a similar level of
income, but also similar with aspirational peers. On the composition of spending, Moldova spends more
than the averages of its peers on all levels of education, especially early childhood, primary, and secondary
education, which comprise three-quarters of total spending on education combined.

Figure 18. Population by age group Figure 19. Spending by age group

Source: NBS. Source: HBS, 2019.

Health

27. The COVID-19 pandemic has pushed the health system to its limits, complicated by the
extremely low uptake of vaccines within the country. The vaccination rate in Moldova is among the
lowest in ECA, with only 26.4 percent of population fully vaccinated, second only to Bosnia and
Herzegovina at 25.8 percent.17 Similar to most systems in the world, the health-care system of Moldova
was not fully prepared to face the pandemic, being understaffed, among other things. The number of

16 Noncommunicable diseases include cancer, diabetes mellitus, cardiovascular diseases, digestive diseases, skin diseases, musculoskeletal
diseases, and congenital anomalies.
17
Source: https://coronavirus.jhu.edu/vaccines/international

19
nurses and midwives is among the lowest in ECA (3.9 per 1,000 population). In addition to the burden
imposed by the COVID-19 pandemic on the health system, groups of the population suffered from a lack
of access to routine, essential health services due to the restrictions, as well as pressure on the health
system’s capacity. Over the medium-term, an aging population, combined with lifestyle habits, is expected
to place the health system under stress.

28. Both government health spending and private out-of-pocket (OOP) spending are particularly
high in Moldova, with implications for the poor and most vulnerable. In 2019, the Government spent
about US$170 per capita on health, representing more than 4 percent of GDP and 13 percent of total
spending. This increased to 4.8 percent of GDP, or 13.6 percent of total spending, in 2020 because of the
increased needs to deal with the pandemic. At the same time, OOP expenditure amounts to 40 percent
of current health expenditure, which is higher than in peer countries and in aspirational countries,
respectively, at 37 and 35 percent (Figures 20 and 21). Pharmaceutical spending is by far the main source
of OOP health spending, at 73 percent,18 and improving drug coverage and the price-setting mechanism
in the market is crucial to achieving universal health coverage.

29. The technical efficiency of public spending is low, however, largely because of inefficiencies in
hospital and specialized outpatient care, which is usually attached to hospitals outside of the capital
city. Oversupply of hospital infrastructure absorbs considerable public resources because much of the
infrastructure is not used optimally. A long-postponed rationalization of Moldova’s hospitals is one reason
for a delay in efficiency gains. Improvements in the efficiency of inpatient care are mostly expected to
come from reducing the length of hospital stays and providing more day-care services. Meanwhile, the
system can expect higher demand to provide care to chronic patients, which can be mostly provided
without requiring an overnight stay. The COVID-19 pandemic has shown that, beyond the number of beds
or hospitals, it is the capacity to treat patients that matters most, with an adequate mix of human
resources, equipment, and enabling environment. The COVID-19 crisis has reinforced the need to redesign
the hospital system, with a strategic plan to guide future investment in order to maximize value for
patients.

30. As a result, the health outcomes for Moldova are poor despite high expenditure. Life expectancy
at birth is 72 years for Moldova, while the average of its peers is 73 years and the aspirational average is
77 years (Figure 22). While Moldova fares relatively well compared with its peers with regard to maternal
health, its child outcomes are the worst. The mortality rate measured as the number of infant deaths is
very high, at 13 per 1,000 live births, almost double the EU average of eight infant deaths (Figure 23).
Around one in six households in Moldova experiences extremely high health spending. Over 10 percent
of the population still lacks coverage, mainly because entitlement is linked to payment of health insurance
contributions. The number of people who work in the informal sector and cannot afford to pay
contributions is significant. For those who are covered by the health insurance fund, financial protection
may be undermined by the limited range of publicly financed outpatient medicines, heavy co-payments
for these medicines, underdeveloped strategic purchasing, and the practice of making informal payments.
As a result of these gaps in coverage, poorer households are at high risk of being uninsured, facing financial
barriers to access, and experiencing exorbitant health spending.19 Financial hardship is heavily
concentrated among the poor, pensioners and people living in rural areas. Outpatient medicines are the
largest single driver of exorbitant OOP payments and their contribution to financial hardship has increased
over time.

18 NBS 2016 https://statistica.gov.md/public/files/publicatii_electronice/acces_servicii_sanatate/Acces_servicii_sanatate_2016.pdf

19
WHO, 2022.

20
31. There are also large inequities across age groups, men and women, and rural and urban areas.
First and foremost, efforts to improve equity and equality of opportunity depend on timely early
childhood development. Poor health in early childhood, including poor nutrition, limits the development
of physical, cognitive, and socioemotional skills. Access to health services for the poorest 20 percent of
Moldovans has deteriorated since 2012. According to a 2016 survey,20 only 16.9 percent of the poorest
benefited from access to health, compared with 28.4 percent of the top quintile. For poorer households,
inpatient care is the second-largest driver of health spending, perhaps linked to informal payments for
hospital care (the use of cash or presents to secure services), which have also increased. Therefore, there
are inequities in the access of the population to assistance services medical. The decline in access was
most evident in primary health care (PHC) and hospital services, which led to a sizable reduction in health
benefits for the poorest Moldovans. The share of people in rural areas who have made appointments to
see a doctor and who spend more than 20 minutes outside the doctor's office waiting for medical
consultation is 2 times higher than in urban areas. Access of women and girls with locomotor disabilities
to health-care facilities and adapted gynecological chairs remains challenging. Women, regardless of age,
need medicines 1.5 times more often than men.

32. During the COVID-19 outbreak, Moldova has used reserve funds to guarantee free hospital
treatment for everyone, regardless of their health insurance status. This measure adopted to deal with
the COVID-19 emergency proves the importance of universal health coverage (UHC) – namely, the access
for everyone to health services without financial hardship. Countries have adopted different approaches
to achieving UHC, with governments intervening to a different extent, including also de-linking
entitlement to all health services from contributions. Ensuring UHC is particularly relevant for a country
where households are faced with large health spending, mainly driven by OOP payments for outpatient
medicines for all income groups, and especially for poorer households. Re-designing the healthcare
insurance scheme would be important to ensure UHC. For example, Moldova has broadened the coverage
including those who seek employment and assistance beneficiaries, a key action to contribute to UHC.
Focusing on the coverage of outpatient medicines may help strengthen financial protection, and
introducing managed entry agreements for single-source expensive and innovative medicines may help
improve efficiency of available resources. Moving toward UHC requires a commitment to steadily increase
and seek efficiencies in public spending on health and higher taxes, including also health taxes, feasible
given the expected reduction in households’ OOP expenditure (Figure 21).

Figure 20. General government health spending vs Figure 21. OOP expenditure: Moldova vs peer
OOP expenditure countries’ average

20
NBS (2016) https://statistica.gov.md/public/files/publicatii_electronice/acces_servicii_sanatate/Acces_servicii_sanatate_2016.pdf

21
 Figure 22. Life expectancy at birth: Moldova vs peers’ Figure 23. General government health spending vs
average infant mortality rate

Source: WDI, 2018.

33. Increasing the efficiency of public spending is a priority to free up fiscal space for health in the
medium-term. Moldova’s allocations to health care are adequate and on a par with countries at a similar
stage of economic development. However, Moldova now needs to prioritize more efficient use of those
public resources. In the short-term, improving strategic purchasing and selectively contracting health-care
providers and diversifying procurement mechanisms, together with introducing managed entry
agreements for single-source expensive and innovative medicines, can help optimize resource use and
increase efficiency. Unless efficiency improves, the Government may not be able to secure the resources
necessary to meet the healthcare needs of the nation. Moldova has been working to improve purchasing
mechanisms for all types of care. The options include age-adjusted capitation in PHC, case-mix funding for
hospitals, and performance-based financial incentives. It is important to empower Compania Naţională
de Asigurări în Medicină (CNAM) to become a strategic purchaser in the healthcare market, and facilitate
selecting and contracting of the most efficient providers and introducing performance-based incentives.
Benchmarking hospital-care providers by the achievement of efficiency and performance indicators would
better inform CNAM’s strategic purchasing decisions. Finally, diversifying procurement mechanisms, and
increasing transparency and efficiency in purchasing medicines and medical devices, including for modern
methods of contraception, would improve cost efficiency, as well as health outcomes.

34. In the medium-term, redesigning the hospital system and redirecting resources toward PHC
services are key to increasing efficiency and quality. Improving the efficiency of hospitals can be
undertaken through rationalization. Savings on unwarranted hospital services can be reallocated to
increase funding for population-based public health services and PHC. Raising the efficiency of hospital
service use by reducing unnecessary hospitalizations and managing noncommunicable diseases (NCDs) on
a PHC level could significantly improve health-care spending, with the savings used to provide more
services to the public. Giving PHC a higher priority would improve the efficiency and cost-effectiveness of
health care, improve its resilience, and give the poorest Moldovans better financial protection.

35. Reducing the prevalence of NCD risk factors with public health interventions, including health
taxes. A reduction of NCD risk factors would contribute toward a substantial improvement of Moldova’s
health status. These interventions include, but are not limited to, increases in excise taxes and prices on
tobacco, alcoholic and sugar-sweetened beverages, enacting and enforcing comprehensive bans on
tobacco advertising, and reducing salt intake through the reformulation of food products.

22
Education

36. There is a strong correlation between education and poverty. Adults with only completed
primary education but no further education are most likely to be poor. The poverty rate among adults
(18+) with only primary education stands at 46.9 percent, compared with 25.1 percent among individuals
with completed secondary education and 7.8 percent among individuals with completed tertiary
education. The correlation between low levels of education and poverty is in part through low levels of
education strongly impeding access to, and returns in, the labor market. Lower labor force participation
rates among working age adults in the B40 are largely accounted for by low education levels. In 2020,
individuals with primary education or less had employment rates of only 2.6 percent, compared with 60.4
percent among tertiary-educated individuals. Low levels of education have also been associated with
lower earnings—individuals with tertiary education reported wages 4.7 times higher than those with
primary education (HBS, 2020). Overall, the gap in labor force participation rates between the B40 and
T60 amounted to 10 percentage points in 2019. Differences in education of working age adults between
the B40 and the T60 explain the biggest part of the gap in labor force participation—42 percent. This is
driven by the low share of adults in the B40 with tertiary education (only 6 percent, compared with 29
percent in the T60), and the strong correlation between higher education and labor force participation.

37. Though Moldova’s spending on education has been declining over time, it remains relatively
high compared with its peers and its outcomes. Education spending accounted for almost 6 percent of
GDP in 2019, and it is higher than the average for its peers (4.5 percent). Out of total spending in 2018, 26
percent was spent on secondary education, followed by early and primary education and higher
education, at 18 and 12 percent, respectively (Figure 24). About 90 percent of the education budget goes
toward current expenditures, and this applies across all different education levels,21 crowding out needed
quality enhancing investments in capital and educational materials. Efficiency in spending is low, leading
to lower learning outcomes—as measured by OECD Programme for International Student Assessment
(PISA) scores—and low education outcomes in primary, secondary and tertiary education, as measured,
for example, by school enrolment rates. The country performed below its peers in the 2018 PISA tests,
and its results have remained unchanged since 2015.

38. Inequities to access to education also persist by socio-economic groups and location. Education
in Moldova is largely public, with universal access to primary and secondary education with a duration of
compulsory education of 12 years, among the highest in peer countries.22 However, large inequities
persist by socio-economic groups and location. The net attendance rate for upper secondary in rural areas
is 64.2 percent compared with 85.4 percent in urban areas. Gross enrolment is low for Moldova at all
levels of education compared with peer countries. For the poorest, enrolment rates for upper secondary
and lower secondary education are lower.

21 Education Statistics database, 2018.

22
The enrolment in private institutions in Moldova at primary education is low at 1.5 percent.

23
 Figure 24. Spending by level of education, as a share of Figure 25. School enrolment and pupil-teacher ratio
total spending in education

Source: WDI.

Figure 26. PISA reading score, 2018 Figure 27. Efficiency frontier: education

Source: OECD. Source: World Bank staff calculations; WDI.

Note: The efficiency scores are calculated based on DEA Analysis, output-oriented model. Inputs are domestic general government education
expenditure (% of GDP). The outcome variable is the PISA score, 2018 in science. The sample includes 34 ECA countries (advanced and emerging
economies).

39. Improving the efficiency of education will be critical for both supporting a new growth model
and value for money. Education network optimization and financing mechanisms, in particular for higher
education, remain a priority. In the past decade, Moldova has made some improvement in the efficiency
of its primary and secondary school system. Since 2014, efforts to optimize the school system through
school mergers have made it more efficient. As a result, the average student-to-teacher ratio for grades
1–12 increased from 10.8:1 in 2012 to 12.4:1 in 2020. However, with a continuous decline in the size of
the school-age population anticipated, there is room for more progress in this area. To address challenges

24
in access to preschool education, a new funding formula should be introduced and implemented. The new
formula of funds allocation to institutions should be based on a per-capita principle, with specific
coefficients to support children from disadvantaged groups of the population. It would need to be
supported by legislative changes that would require local authorities to adequately finance preschool
education institutions and ensure universal access, including for children with disabilities, Roma children,
and children from rural areas. Moreover, a comprehensive review of the framework for public
employment, compensation, performance incentives, adequacy of skills and career management will be
essential in providing effective and efficient service delivery.

Social Protection

40. Pensions and social assistance remain the main drivers of income growth for the poor. In the
years before the pandemic, income growth among the poor was driven primarily by pensions and social
assistance, while incomes of households above the poverty line were boosted by an increase in wages. At
the same time, the downward trend in remittances by more than 30 percent over the past decade has
had a substantial impact on income growth for all income groups except for the poorest households,
highlighting the inherent vulnerability of a development model based on exporting workers. Social
transfers such as pensions and social assistance benefits are an important source of income for the bottom
40 percent of Moldovans. In 2019, social protection programs covered 53 percent of the population and
91 percent of the bottom quintile. Pensions and other social insurance benefits covered 37 percent of the
population and 75 percent of the poorest quintile. The coverage of social assistance transfers was 41
percent of the population and 73 percent of the poorest quintile.

Figure 28. Pensions and social assistance remain the main drivers of income growth for the poor
(Decomposition of income growth, 2019–20)

Source: HBS 2019–20. World Bank staff calculations.

41. However, Moldova’s expenditures on social safety nets are low in regional comparison and
benefits often do not allow the beneficiaries to graduate out of poverty. Coverage of the most
vulnerable, especially those in the poorest income quintile, is limited.23 On social assistance, the eligibility

23Cash-based social protection for children in the Republic of Moldova, UNICEF, 2018; “Strengthening the effectiveness of the social safety net
project: Consultancy for Ajutorul Social evaluation and design”, Oxford Policy Management, 2017; SWOT analysis of Child Protection System,
Government of Moldova, 2014 concluding that “the existing social benefits do not adequately support families with children”.

25
criteria (income threshold) of Ajutor Social remains too restrictive to lift the poor out of poverty. On
pensions, there is a still small but growing number of people who reach retirement age without having
enough of a contributory period to earn an old age pension. Moreover, benefits are relatively low. Due to
concerns surrounding the adequacy of pensions, in 2021, minimum pensions were increased by 62
percent.

Figure 29. Cross-country analysis of expenditures on social protection (% of GDP)

Source: SPeeD database.24

Pensions

42. Households that are dependent on pensions account for the majority of the poor. In 2020, the
poverty rate among pensioners was 36.6 percent compared with 23.0 percent for working age adults and
26.0 percent for children. The high poverty rate among pensioners is reflected in a high poverty rate
among households with pensioners only (38.6 percent) and among households with mixed adults
(working-age and pensioners) (31.4 percent). Similarly, households that reported pensions as their main
source of income reported a poverty rate of 40.5 percent. The high rate of poverty among pensioners and
pension-dependent households is due in part to the relative magnitude of pensions—in 2020, per-capita
income from pensions was, on average, 3.5 times lower than per-capita income from employment (NBS).
The pension system, regulated by Law 156/1998, is based on a Pay-As-You-Go (PAYG) system for all
individuals above the statutory retirement age (63 for men, 59 for women as of 2020).25 The contribution
rate amounts to 6 percent of their income for employees, and 18 percent (reduced from 23% in 2018) of
the payroll for employers in the private sector. Through a series of subsequent reforms, starting in 2016,
the general PAYG retiree replacement rate reached 30 percent for new and 26 percent for current
pensioners of the national average wage. This is double than the pre-reform replacement rate for new
pensioners, while replacement rate for the existing pensioners was expected to slip to 16 percent.26

24 https://tab.worldbank.org/#/site/WBG/views/AggregateSPExpenditureAnalysisexpenditure4c/CrossCountry-latest-year?:iid=1
25 Starting with July 1, 2019, the standard retirement age of 63 years for men is established and, starting with
July 1, 2028, the standard retirement
age of 63 years for women.
26
World Bank. 2019. Moldova Economic Update. Special Focus Note: Unfinished pension reform. World Bank: Washington, DC.

26
43. Measures to increase the minimum pension have strengthened support to pensioners. Due to
concerns surrounding the adequacy of pensions, in 2021, minimum pensions were increased by 62
percent, from an indexed MDL 1,233.9 to MDL 2,000 per month. Data from the NBS suggest that, in 2021,
average monthly per capita pensions increased from MDL 449.4 to MDL 490.9. In addition, household
consumption expenditure rebounded in 2021 from the impacts on the pandemic by 8.9 percent, on
average, contributing to declines in poverty. Isolating the impact of the pension increase from the
rebound in general household consumption suggests that the pension increase is likely to have led to a
decline in overall poverty of 1.8 percentage points, and in old-age poverty specifically of 6.5 percentage
points. Working-age poverty declined by 0.5 of a percentage point and child poverty remained unchanged.
As a result of the pension increase, the poverty rate among households with pensioners is expected to
decline only by 8.4 percentage points and the poverty rate among households with mixed adults is
expected to decline by 3.1 percentage points.27 Other household types saw negligible changes in the rate
of poverty. Going forward, increases in minimum pensions may have a broader muted effect on poverty
because of informality and other factors leading to insufficient contributions. This is reflected also in the
growing number of people in the country reaching retirement age without attaining eligibility for old-age
pensions, as will be discussed later on.

44. To reduce the negative impact on the financial balance of the pension system, the Government
subsequently restored the previous retirement age and canceled the double indexation. While
parametric changes of 2016 envisaged a near fiscally neutral pension system (with long-term average
financial balance being close to -0.1 percent GDP), subsequent waves of policy changes undermined this
achievement. After the minimum pension increase, the short-term financial balance was expected to
worsen to about -3.6 percent GDP (Figure 30). In December 2021, the Government restored the original
retirement age that had been temporarily relaxed, and replaced the double indexation with a weighted
indexation (100 percent CPI and 50 percent real GDP), which is expected to improve pensions financial
balance substantially. The long-term average of the deficit of the pensions system is around 1.4 percent
of GDP.

Figure 30. Effect of waves of measures on the financial balance of the pension system of Moldova

Source: PROST simulation, World Bank staff calculations.

27Prior to the pension increase, the poverty rate among households with pensioners only stood at 38.6 percent and the poverty rate among
households with mixed adults (working age and pensioners) stood at 31.4 percent.

27
45. Moreover, there are still several challenges to be addressed for the pension system to become
an effective and inclusive tool for social protection of the elderly. First, the elderly are well covered with
pension benefits, even though the benefits are relatively low. Old-age pensions are the major source of
income for the elderly and cover nearly 63 percent of the poor (and 37 percent of the population in
general). More importantly, further tightening of the link between the contributions and benefits in social
insurance is essential, as there is a still small but growing number of people who reach retirement age
without having paid in enough during the contributory period to earn an old age pension, particularly in
the case of day laborers in the agriculture sector. Simulations show that, over time, this will lead to nearly
half of all people over the retirement age being ineligible for old-age pensions due to a lack of
contributions.

Figure 31. Eligibility of the elderly for the old-age pensions

700,000

600,000

500,000

400,000

300,000

200,000

100,000
2038

2059
2020
2023
2026
2029
2032
2035

2041
2044
2047
2050
2053
2056

2062
2065
2068
2071
2074
2077
2080
2083
2086
2089
2092
2095
2098
People above retirement age 63/63 Old Age pensioners

Source: PROST simulation, World Bank staff calculations.

46. Finally, important inequities in the system remain because of privileged pensions, such as those
for members of the judiciary and the military. For judiciary pensions, although the number of such
pensioners is very low (283), their benefits are 5.4 times higher than the PAYG old-age pension (MDL 9,834
vs MDL 1,834). Meanwhile, military pensions, the bulk of the privileged pensions (over 21,000), are on
average 2.3 times higher than general PAYG pensions (the average military pension is MDL 4,633), and
members of the military retire about 20 years earlier. In addition, the military pension benefit is calculated
based on the income of the last year of service (higher pensionable base), but in the PAYG system the
average income of the whole contributory period is considered (lower pensionable base). In terms of the
replacement rate, new pensioners are granted 60 percent of the average military wage, or 85 percent of
the average wage, and for the simulation period the replacement rate for existing military pensions
averages 39 percent of the average military wage, or 56 percent of the national average wage, much
higher than the general PAYG retiree replacement rate.

47. There are significant differences in the ability to work after retiring among privileged and
general groups of pensioners. Beneficiaries of military pensions on average retire at 43, while the general
male retirement age is 63. While it is usual for military service to have a lower retirement age, in the North
Atlantic Treaty Organization (NATO), the Commonwealth of Independent States (CIS), and other
comparable countries, the retirement age is usually 10 years later. While mechanisms are in place to

28
encourage military personnel to serve longer, the incentive further increases inequity, given that 3
percent accrual per year of service beyond the minimum of 25 years is included.

Social Assistance
48. Social assistance is the major source of income support for the poor (41 percent of the
population and 71.4 percent of the poorest quintile received some sort of social assistance in 2020).28
However, the adequacy29 of social assistance benefits is low (23 percent of Q1 consumption compared
with over 85 percent of social insurance benefits), which results in a relative incidence30 for social
assistance of 16.1 percent. Of overall social assistance spending, which is close to 1.5 percent of GDP,
about 31 percent was leaked to the non-poor in 2020.

49. After major reforms over a decade ago, noncontributory social assistance attained important
results. In 2008, the Ajutor Social program (as well as an add-on Heating Allowance program) was
introduced, targeting social benefits to the poorest and replacing some inefficient entitlement-based
(categorical) benefits. Elimination, in 2012, of the costly, poorly targeted Nominative Compensations and,
in 2017, of categorical benefits from the Fund for Social Support of Population, improved the spending
efficiency and sustainability of the social safety net. A gradual expansion of targeted cash transfers helped
to substantially increase the share of social assistance benefits going to the poorest 20 percent of the
population from 40 to 69 percent by 2020. Targeted social assistance works very well in Moldova, with 83
percent of Ajutor Social benefits accruing to the poorest quintile. However, Ajutor Social is small, covering
only 8.2 percent of the poor in 2020. Other social assistance benefits with far higher coverage (for
instance, categorical benefits) display a worse targeting performance, with 23 percent of the benefits
leaking to the non-poor. Ajutor Social, by contrast, has low leakage and is the most cost-effective program
in terms of reducing poverty in Moldova’s social protection architecture.

Table 4. Performance of various social protection programs in reducing the poverty gap

Coverage of the Share of benefits going

Coverage (%) Benefit-cost ratio
poor (%) to poorest quintile (%)

All social protection 52.7 78.6 49.0 0.71

All social insurance 37.8 61.2 49.9 0.70
Old age pensions 37.6 60.7 50.1 0.70
All social assistance 40.8 61.7 45.5 0.69
Social pensions 7.0 13.3 53.4 0.84
Child benefits 10.2 10.0 14.1 0.50
Ajutor social 3.9 8.2 82.5 0.89
Categorical benefits 29.7 49.2 45.6 0.68
Cold season (APRA) 5.0 10.6 76.2 0.69
Source: ADePt analysis by the World Bank staff on the basis of HBS2020. Benefit-cost ratio is the poverty gap reduction in a unit of local
currency for each 1 unit spent on the social program.

28
The 2021 Household Budget Survey is not yet available.
29 Adequacy is the mean transfer amount received by a group as a share of the total welfare of the beneficiaries in that group. I.e., how meaningful
is the benefit for the recipients
30 Relative incidence is the transfer amount received by a group as a share of total welfare aggregate of the group. I.e., how meaningful is the

benefit for specific income quintile or the poor in general

29
50. However, the gains of these reforms have been eroded over time. Ajutor Social covered 7
percent of the total population in 2017, but its relevance as the main poverty-focused social assistance
program has been gradually reduced by various policy measures (such as accelerated indexation of
pensions, ad-hoc top-ups, increases in public sector wages). Moreover, the number and cost of categorical
benefits and ad-hoc pension top-ups have increased since 2018, diverting budget resources from
expanding and sustaining the adequacy of the means-tested benefits. Preferences toward expanding
benefits with the broader coverage (e.g., increasing the minimum pension) is expected to result in lower
efficiency in poverty reduction. Similarly, existing categorical transfers leak about 27 percent of the total
social assistance spending of about 2 percent of GDP (excluding social pensions) to people in the top three
quintiles. Categorical benefits cover about 30 percent of the population, with about 62 percent of the
bottom quintile households receiving at least one categorical benefit. About 29 percent of categorical
benefits recipients are non-poor. As a result, the share of families with children receiving Ajutor Social
declined from 52 to 34 percent between 2014 and 2020.

51. Leveraging categorial benefits during the recent gas crisis underscores their inefficiency despite
an overall positive impact on poverty. The majority of fiscal resources were devoted to compensating
households from the increased gas prices with gas subsidies on the basis of the consumption volume. The
package of measures included compensation to the tariff of natural gas differentiated by consumption:
MDL 4.3 for the first 50 cubic meters and MDL 3.2 above the first 50 cubic meters. The package also
included compensation of 80 percent to heating tariffs of up to MDL 815 per month. Already before the
war in Ukraine, Parliament had approved the 2022 Budget Law with a deficit of 6.0 percent of GDP to
support investments and the ambitious reform program, including also a substantial increase in social
spending, of which 0.8 percent of GDP to mitigate the impact of increase in gas prices. Additional spending
of about 1.3 percent of GDP is expected to protect firms from the economic impact of the war in Ukraine
and to protect households’ disposable income from increasing energy and food prices. While the overall
package is expected to have reduced poverty by 3 percentage points from an increase of 6.7 percentage
points in the absence of mitigation measures to an increase in poverty of 3.7 percentage points after the
mitigation measures, gas subsidies are a costly and inefficient protection tool, as coverage of the poor is
limited: over 60 percent of the poorest are not connected to the gas network, and therefore were not
able to benefit from subsidies. While expansion of the coverage of the well-targeted cold season benefit
(APRA) by increasing the income threshold (from a 2.2 multiple of the income threshold used for the
Ajutor Social program to a 2.6 multiple) had better outcomes, some categories have nonetheless
benefited more (i.e., pensioners, female-headed households, and urban families) than others (i.e.,
families with more than three children, rural families in the south). Further expansion of Ajutor Social
would ensure better protection of the vulnerable. However, these measures are expected to have
reduced the poverty rate by about 1.5 percentage points from an expected increase of 6.7 percentage
points in the absence of mitigation measures to an increase in the poverty rate of 5.2 percentage points
after the mitigation measures.

Figure 32. Share of the gas subsidy accruing to different income deciles

Source: World Bank staff calculations.

30
52. At the same time, the design of Ajutor Social remains too restrictive to lift the poor out of
poverty. The income threshold per adult equivalent historically is only about a half of the subsistence
minimum (SM), and for children it is even smaller.31 The current design is not conducive for the graduation
of vulnerable families, as an increase in income translates into equal withdrawal of the benefit, leading to
large fluctuations in the number of beneficiaries when income sources fluctuate, sometimes even due to
the indexation. This fluctuation clearly does not indicate any sustainable improvement in the
beneficiaries’ situation. However, the program still provides some support and allows for it to be scaled
up in times of crisis—as happened during the emergency period announced in April 2020 due to the
COVID-19 outbreak.32 The increase in the minimum pension in 2021 has led to about 20 percent of the
beneficiaries leaving the program, despite still significant poverty rates among families with pensioners.
Indirect evidence that the income threshold does not adjust quickly enough and becomes more restrictive
over time is from the increasing share of new applicants being rejected due to their failure to pass the
income test.

Figure 33. Number of applicants to Ajutor Social and their application status

Source: SIAAS database (admin data).

53. Raising means-tested support for children is an effective measure, not only to combat poverty
but also to support long-term growth. The poverty rate among households with five or more members is
38.2 percent, though these account for only 9.7 percent of households. Households with two adults and
three or more children have the highest poverty rate, at 40.1 percent, but only account for 3.6 percent of
households. Similarly, households where there are three or more children have a poverty rate of 41.0
percent but account for only 4.3 percent of households. By contrast, single-person households have a
poverty rate of 31.0 percent and account for 31.0 percent of households. High poverty rates among large
households are largely due to high age-dependency ratios, whereas high poverty rates among single-
person households are due to these households being overwhelmingly comprised of pensioner-only
households (62.1 percent). Not only families with many children are overrepresented among the poorest

31 A child has a 0.75 adult equivalency coefficient. While after the recent indexation in April 2022 (GMI increased to MDL 1,363), the income
threshold temporarily reached almost 60 percent of subsistence minimum for the able-bodied (as of H2 2021), while for children it is still below
50 percent. This gap of about 10 percent of subsistence minimum for children is persistent.
32 The COVID-19 Emergency Response project supported temporary expansion of Ajutor Social coverage achieved by increasing the income

threshold from MDL 1,107 to MDL 1,300. This led to almost 50 percent increase in coverage and 34 percent increase in the benefit amount. The
expansion made use of the shared registry with APRA program. However, newly vulnerable households may not have been able to enroll due to
lockdowns.

31
households, but a failure to cover the basic needs of children negatively affects formation of the human
capital and productive capacity of the country. Children from low-income backgrounds often grow up with
developmental deficits that are difficult, if not impossible, to reverse later in life without intervention. The
existence of these deficits is a key driver of inequality and contributes to the intergenerational poverty.33
Providing households with 60 percent of SM and MDL 500 of additional income disregard per child –
namely, the level of income that is not considered to compute the benefits –would increase the number
of beneficiaries by 49,800 households to a total of 70,500 families with children. The share of the extreme
poor would be reduced from 10.6 to 8.7 percent.

Table 5. Simulation of the scenario with children provided 60 percent of SM and working poor – MDL 500 of
additional income disregard per child
Baseline Simulated outcome
All beneficiaries 49,783 70,506
Two+ adults and a child 4,167 6,858
Two+ adults and two children 7,407 16,142
Two+ adults and three+ children 7,094 12,892
One adult and child/ren 3,076 6,572
Only working age adults 21,137 21,137
Only pensioners 5,132 5,132
Mixed adults, no children 1,770 1,770
Households with children 21,743 42,466
Households w/o children 28,040 28,040
Source: Simulations on the basis of HBS2019, World Bank staff calculations.

54. To be socially and fiscally sustainable, the social protection system needs to better target social
assistance spending while keeping fiscal costs manageable. The recently enacted law on Ajutor Social is
an important step in the right direction in strengthening the main social protection programs to mitigate
the impact of future shocks by further increasing and linking the guaranteed minimum income (GMI) for
children to the SM for children, while containing the costs. This change will support families with children,
which are among the most vulnerable to shocks and were relatively less well protected from the gas crisis
in the winter of 2021. The legislation also increases the income disregard. Changes in income disregard
will also have a component that depends on the number of children in the family. The increased income
disregard may also help further support formalization of envelope wages. Moreover, since higher energy
costs translate into a larger SM, the supported measures will also allow for the incorporation of seasonal
fluctuations in the amount of the basic needs, as the SM is recalculated. However, continued reform
efforts in moving in the direction of strengthening the means-tested anti-poverty program, the Ajutor
Social, as well as finetuning social assistance programs, are central during this period of heighted
uncertainty and in the long term given the country’s vulnerability to shocks.

33 Orazio Attanasio, Sarah Cattan & Costas Meghir. Early Childhood Development, Human Capital and Poverty.
https://www.nber.org/papers/w29362

32
Chapter 3: Efficiency, equity and simplicity of the tax system
55. Revenue has grown on a par with nominal GDP over the past 10 years, with taxes slowly
reorienting toward income taxes, although these are still below 15 percent of total revenue. General
government revenues have declined steadily since the global financial crisis, when they reached 39
percent of GDP. Over the past 10 years, total revenues followed the same trend of nominal GDP, with the
largest increase coming from corporate income tax (CIT), which has grown twice as fast as GDP. Also,
personal income tax (PIT), social contributions (social and health insurance) and value-added tax (VAT)
have displayed positive performances in real terms, growing at just above the GDP growth rate. On the
other hand, taxes on foreign trade, property tax, and non-tax revenue have displayed a weak performance
throughout the past decade, despite recent gains. As a result, the overall revenue structure has slowly
reoriented toward direct taxes. But VAT still represents the lion’s share of revenue (about 32 percent),
followed by social contributions (about 29 percent).

Figure 34. Revenue evolution in real terms Figure 35. Revenue structure

Source: MoF.

56. The overall revenue structure remains highly skewed toward indirect taxes, in stark contrast to
revenue collection in Europe. Income taxes (as a percentage of GDP) collect on average more than double
in other European countries than in Moldova. On the other hand, VAT collection in Moldova is 2
percentage points higher than the collection in Europe. While Moldova’s collections are higher than the
level of income would imply, this performance is primarily driven by social contributions and VAT. Income
tax and, to lesser extent, tax on foreign trade are below what countries with similar per capita incomes
are collecting. The potential revenue loss, as measured by the tax gap, is around 7 percent of GDP from
VAT collections only. This gap is a combination of a policy gap and a compliance gap, which will be
discussed later in the report.

33
 Figure 36. Tax capacity and performance, 2019

Source: IMF GFS, World Bank staff calculations.

57. When we benchmark Moldova’s revenue with specific peer countries, these results are more
nuanced, with Moldova having revenue lower than the average for peer countries but in line with its
structural peers.34 Given the heterogenous impact of the COVID-19 pandemic on the economy and
revenue collection, the comparison is based on 2019 data. At that time, general government revenues for
Moldova as a percentage of GDP were around 30 percent, about 7 percentage points lower than the
average for peer countries but in line with structural peers. In 2019, the average revenue from income,
profits and capital taxes for peer countries was 6.6 percent of GDP, while for Moldova it was 4.5 percent
of GDP, 2.1 percentage points lower than the average for peer countries. However, the average revenue
from income, profits and capital taxes is close to its structural peers. Similarly, revenues from social
contributions (8.8 percent of GDP) are also below its peers’ average (10.1 percent of GDP) by 1.3
percentage points.

34Comparator countries are divided into three groups. Structural peers include North Macedonia, Albania and Kosovo. Neighboring peers include
Belarus, Ukraine and Romania. Aspirational peers include Serbia, Hungary, Slovenia, the Slovak Republic, Bosnia and Herzegovina, the Czech
Republic, Estonia, Latvia, Lithuania, Georgia and Armenia.

34
 Figure 37: General government revenues (% of Figure 38: GDP per capita, Moldova and peers, 2019
GDP), Moldova and peers, 2019
60 45 Kosovo 26.8

Thousands

Structural
40 Albania 27.4
50 48.4
43.7 44.2 North Macedonia 30.4
43.3 44.1 35
40.7 40.8 39.5 Romania 31.7

Neig hbors
40 38.3 37.4 30
34.8 Ukraine 39.5
30.0 31.7 25
30.4 Belarus 44.1
30 27.4 26.8
25.8 25.3 20 Armenia 25.3
North Macedonia 30.4
20 15
Lithuania 34.8
10
10 Latvia 37.4
5

Aspi rational
Estonia 38.3
0 0 Slovak Republic 40.7
Czech Republic
Slovak Republic

Ukraine
Belarus
Moldova

Serbia

Slovenia

Bosnia and Herzegovina

Estonia

Latvia

Lithuania

Georgia

Armenia

Romania

North Macedonia

Albania
Hungary

Kosovo
Czech Republic 40.8
Bosnia and Herzegovina 43.3
Hungary 43.7
Slovenia 44.2
Serbia 48.4
Moldova 30.0
0 5 10 15 20 25 30 35 40 45 50
Aspirational Neighbors Structural Taxes on income, profits, & capital gains Taxes on payroll & workforce Taxes on property
GG Revenues (% GDP) GDP per capita, constant PPP Taxes on goods & services Taxes on int trade & transactions Other taxes n.e.c.
Social contributions Other revenue Grants revenue

Source: GFS and WEO IMF.

58. While Moldova’s revenue acts moderately well as an automatic stabilizer, income taxes are not
as sensitive to the GDP cycle as in its peers and advanced economies.35 The correlation coefficient
between cyclical components of overall revenue and GDP reached 0.72 in the entire period, which is in
line with regional peers and advanced economies. VAT correlation stood at +0.83, while taxation on
international trade also recorded a high correlation of +0.71, being the two most procyclical sources of
revenue. However, income taxes, particularly PIT, have reduced their procyclicality in the past decade.
This can be attributed to several policy factors, such as the flat tax rate, high informality and a multitude
of deductions.

Figure 39. Revenue buoyancy

1.0
Correlation between cyclical component of

0.8

0.6
Gov revenue and real GDP

0.4

0.2

0.0

-0.2

-0.4

-0.6

-0.8
Total VAT Social PIT CIT Intern. Excise Grants
Revenue Insurance Trade
2000-2019 2000-2008 2010-2019

Source: IMF GFS; World Bank staff calculations.

35
This study focuses on revenue collection dynamics and not on tax rate modifications over time.

35
Value-Added Tax

59. The VAT tax exemptions, which are directly impacting the policy gap, have strongly narrowed
the tax base and are undermining overall revenue collections efforts. Moldova has four VAT tax rates,
ranging from zero percent on exports and transport services, up to the standard 20 percent rate. A
reduced rate of 8 percent is applicable on several goods, including primarily on bread and bakery products,
milk and dairy products, certain categories of medicines, electrical energy, heating and hot water
production, and natural and liquefied gas. A reduced rate of 12 percent is applicable for accommodation
and catering services. Additionally, there is a wide range of VAT exempted goods and services, including
cars, selected financial activities, health and education services, and most of the fast-growing e-commerce
transactions.

60. With multiple VAT rates, the VAT policy gap in Moldova is relatively high compared with its peer
countries, where there are less reduced rates and exemptions. Moldova could have potentially collected
around MDL 17 billion (7 percent of GDP36) more at full tax compliance, at a standard rate, and if all the
exemptions and reduced rates were abolished, which is 42.8 percent of notional ideal revenue (NIR).
While the policy gap in Moldova is close to the EU median of 44.5 percent of NIR,37 high-income countries
have higher shares of non-taxable final consumption. The policy gap in Moldova’s peer countries (Bulgaria,
Romania, and Ukraine) is substantially lower, pointing to the far less prevalent application of reduced
rates and exemptions in these countries. Except for the energy sector, around 5.5 percent of the forgone
VAT revenue (around MDL 2 billion, or 1 percent of GDP) from the application of the 8 percent reduced
rate is due to food and agricultural products, and another 2.7 percent is related to vehicles (around 0.5
percent of GDP). High informality that accompanies the hospitality sector resulted in forgone revenues of
less than 0.2 percent of GDP. In addition, with the unobserved economy varying at around 30 percent of
Gross Value Added in the past five years, the additional revenues that could be generated are even higher.

Figure 40. Policy gap (% of NIR), 2019 Figure 41. VAT compliance gap (% of VTTL and MDL
billion)

Source: World Bank staff calculations. Source: World Bank staff calculations.

36MoF calculations using a different methodology identified tax expenditures amounting to 2.5 percent of GDP for 2022.
37Notional ideal revenue (NIR) is defined as the upper limit of VAT revenue (i.e., the revenue levied at a uniform rate in the environment of
perfect tax compliance).

36
61. The compliance gap for VAT in Moldova followed a downward sloping trend between 2016 and
2019 and seems relatively low compared with the scale of the undeclared economy. Between these
years, the VAT gap (as percent of the VTTL)38 fell by about 6.5 percentage points, as the collections from
VAT increased faster than the liabilities. The difference between actual VAT revenue collections and
collectable VAT revenue under the assumption of full compliance of the current laws was around 11
percent (MDL 2.4 billion) in 2019, down by 6.5 percentage points from the peak in 2016. This
measurement is on a par with some peer countries in the region. However, a high VAT c-efficiency rate39
(around 60 percent) would indicate that the compliance gap could be much higher, implying a more
targeted focus on tax administration.

Figure 42. VAT compliance gap in Moldova (% of the Figure 43. VAT compliance gap (% of the VTTL)
VTTL and MDL billion)

Source: World Bank staff calculations based on WEO database. Source: European Commission (2020), Dubrovskiy (2017), World Bank
staff calculations. .

62. The VAT gap analysis revealed reform areas with the potential to increase collections and level
the distortions in the economy. The VAT gap can be further decomposed into the exemption gap (13.2
percent) and the rate gap (29.6 percent), which captures the loss in VAT liability due to the
implementation of exemptions, and the loss in liability due to application of reduced rates. Reduced rates
for the agriculture sector and vehicles are particularly distortionary, as the reduced rate of 8 percent in
the agriculture sector is not applicable to food processing industries, which operate at the standard 20
percent rate, cannot deduct the VAT.

63. Additional challenges to VAT collections stem from the structure of the consumption basket
and the impact on consumers, particularly the poor. Given the composition of the consumption basket,
and the existence of exempt, reduced or zero-rated items, the effective VAT rate faced by households is
generally lower. The weighted average VAT rate faced by households is 9.4 percent, which is at the lower
end of the distribution compared with EU countries. As a result, the overall tax expenditures are heavily
tilted toward VAT.

38 The VAT gap is estimated collectively for all economic activities. The estimation follows a top-down consumption-side approach, which relies
on estimating the expected VAT liability (VTTL) by modelling tax rules, applying them for the aggregate tax base available in national accounts’
supply and use tables (SUT), and then by comparing it with the actual receipts. The formula used to estimate the VTTL consists of five main sub-
aggregates: household final consumption (HHC), government consumption (GOV); intermediate consumption (IC); gross fixed capital formation
(GFCF); and other, largely country-specific, net adjustments
39 The C-efficiency ratio is the most commonly used indicator for evaluating the revenue. performance and overall efficiency of the VAT system.

It is the ratio of actual revenues to theoretical revenues from a perfectly enforced tax levied at a uniform rate on all consumption.

37
 Figure 44. Spending breakdown of households

Source: World Bank staff calculations based on HBS data.

64. The only viable option in the short-term would be the rationalization/optimization of tax
expenditures, while protecting the poor and most vulnerable. For this reason, two scenarios were
considered to assess the impact on the population and on revenue collections. In the first scenario (AGR
VAT20), the VAT rate for food and non-alcoholic drinks was increased to 20 percent, with all the rest of
the rates remaining the same as in the current law. The second scenario (VAT_exemp_0_to_4) examines
the introduction of a 4 percent VAT on all products currently exempt from VAT.

Figure 45. Relative change in final income, % Figure 46. Relative change in final income across
different types of households, %

Figure 47. Relative change in poverty headcount at Figure 48. Impact on the budget, compared with
consumable income level, % the baseline, %

Source: World Bank staff simulations, based on 2019 HBS. Source: World Bank staff simulations, based on 2019 HBS.

38
65. Overall, increasing VAT would lead to a moderate decrease in the welfare of the population,
but would result in a substantial increase in public revenues. With 46 percent of the consumption basket
composed of food products, the simulations show that increasing VAT for food products to the standard
rate of 20 percent would lead to a relatively homogenous decrease in final income across all income
deciles, averaging -1 percent compared with the current situation. Nevertheless, with lower deciles
consuming more food and beverages, the impact of the decrease on the first two deciles would be 0.2 of
a percentage point greater than for the remaining deciles. As a result, the poverty rate would increase by
1.7 percentage points. In terms of the structure of households, households with no children (increase of
3.2 percentage points), followed by pensioners (2.9 percentage points) and working age adults (2.2
percentage points) will be the most affected. In the case of the second scenario, the decrease of final
income will be less prominent, and the increase in the poverty rate will total 1.4 percentage points, with
similar, but lower in magnitude, impacts on deciles and types of households. The increase in collections
in the case of the AGR VAT20 scenario is around 0.5 percent of 2019 GDP, while for the second scenario
is around 0.2 percent of GDP.

66. In the medium-term, the authorities have the option of slowly converging with the EU’s rates.
For this, another two scenarios were examined. In the first one, an increase in the VAT rate to 20 percent
(All VAT20) for all goods and services is envisaged, leaving the zero rated and exempt goods and services
untouched as in the current law. In the second scenario, the standard VAT remains at 20 percent, while
VAT reduced rates for all food and non-food products and services (i.e., 8 percent), including hotel and
restaurants (i.e., 12 percent), would increase to 15 percent (reduced VAT15), leaving the zero rated and
exempt goods and services untouched as in the current law. Bringing the VAT rate to 20 percent would
decrease the final income of the population by 1.6 percent, while increasing the reduced VAT rate to 15
percent would result in a decrease of 0.9 percent. In the case of the first scenario, the overall impact on
the deciles is largely homogeneous. However, deciles 1 to 7 will experience a marginally larger decrease
in their final incomes. In the case of bringing the reduced rates to 15 percent, the third decile would be
the most affected. However, the impact is again homogeneously spread across all deciles. In terms of
households, in the case of higher VAT for all products, pensioners and households with mixed adults and
no children would be the most affected, and their final incomes would decrease by 2.2 and 1.7 percent,
respectively. The same pattern is followed when the reduced VAT rates are increased, but by a lower
magnitude. In both scenarios, poverty levels would increase by 2.2 percentage points for the All VAT20
and 1.9 percentage points for the reduced VAT15 scenario. Both scenarios would generate considerable
collections to the budget from indirect taxes. In the case of increasing VAT to 20 percent, indirect taxes
would increase by 0.8 percent of 2019 GDP, and in the case of the second scenario the increase would
amount to 0.4 percent of GDP.

Figure 49. Relative change in final income, % Figure 50. Relative change in final income across
different types of households, %

39
 Figure 51. Relative change in poverty headcount at Figure 52. Impact on the budget collections,
consumable income level, % compared with the baseline, %

Source: World Bank staff simulations, based on 2019 HBS. Source: World Bank staff simulations, based on 2019 HBS.

67. The analysis shows some negative impact on the welfare of the population in all the policy
options. Spending on food and non-alcoholic drinks and housing and utilities is particularly high among
the bottom 40 of the population. Therefore, in the short term, the optimization of VAT tax expenditures
will have to go in parallel with elaborating mitigation mechanisms for the poorest layers of the population.
As a first step, authorities could rationalize the list of VAT tax expenditures that are non-compliant with
the EU Directives.40 In the longer term, changes to the VAT rate system are warranted and could be
accompanied by the overall process of convergence with the EU. The simulations do not account for the
second-round effects that would have an impact on the economy, and consequently on collections and
poverty outcomes. The simulation also did not consider any mitigation measures through social protection
programs that could be considered.

68. Large gains could be generated in the fast-growing e-commerce sector and introducing VAT on
cars. The existing tax code does not have the necessary provisions to fully cover e-commerce.41 Important
gains could be achieved by an obligation on suppliers of digital goods and services based outside Moldova
that supply those goods and services for consumption in Moldova to account for VAT on those sales.
Important steps to register the suppliers of digital goods and services were made and reducing the
threshold for VAT exempted imported goods through consignments, but this could be further fortified by
applying international good practices. Revenue mobilization could be strengthened by introducing VAT
on cars. Moldova’s cars are taxed at acquisition through an excise based on fuel, engine capacity, and age.
Passenger cars are, however, VAT exempted. Estimates suggest that foregone revenue from this VAT
exemption could have reached up to 0.5 percent of GDP in 201642. Reforming the existing excise tax and
introducing a complementary VAT would: i) increase revenue, ii) skew the revenue from the car taxation
system at acquisition towards more expensive cars, and iii) provide the possibility to reduce the import of
old environmentally unfriendly cars. Additionally, recent calculations done by MoF identifies major tax
expenditures in crops, horticulture, and livestock production, and in supply of education and health
services.

40 For instance, goods and services delivered to the International Free Port of Giurgiulesti and the International Free Port of Marculesti; electricity,
heat, and hot water for residential sector; hydraulic turbines with capacity of not more than 1,000 kW; long-term tangible assets directly used to
produce goods, provide services; tractors for agricultural operations; and secondary raw materials, including paper and cardboard, rubber resin,
plastic, glass (broken glass), ferrous and non-ferrous metals waste and residue, industrial waste containing metals or alloys.
41 Note on Indirect Taxation of E-Commerce in Moldova – Practical Aspects of Collecting VAT from Suppliers of E-Services Based Outside Moldova,

2019.
42
Moldova Economic Update – Special topic: Strengthening revenue mobilization in Moldova: Car taxation, 2016.

40
69. Tax expenditures related to commodities that are currently zero-rated in Moldova’s VAT system
appear to be less problematic. Zero-rating is mostly applied to electricity consumed by residential
sector43, and exports and re-exportation of goods produced or transiting through Moldova. The latter is
consistent with practices in VAT systems across the world. Nonetheless, special care should always be
paid to relief of VAT and excise on goods destined for exportation, as these goods have been known to
sometimes make their way back into the domestic economy through fraudulent practices. This
phenomenon is well documented in developing economies with weaker tax and customs administrations.
This could also be the case given the size of the tax expenditures associated with these zero-rated goods
in Moldova.

Personal Income Tax

70. Moldova’s PIT collections are constrained by structural inefficiencies, primarily related to a
skewed distribution of employment income. Moldova has a flat PIT tax rate of 12 percent, and the system
relies on many exemptions and deductions to inject progressivity, at the cost of its simplicity. Multiple
exceptions, difficulties in administration and high informal employment (on average, around 23 percent
in the past three years) are holding back PIT collections. More than 70 percent of employees earn less
than the average salary in the economy. At the high end of the distribution, individuals earning 10 times
more than the average salary are around 1 percent. As highlighted in both 2019 Country Economic
Memorandum (CEM) and 2016 PFR, labor taxation in Moldova does not encourage formal employment
and compliance.

Figure 53. Estimated structure of PIT taxpayers Figure 54. Distribution of salaries, 2021

Source: World Bank staff calculations. Source: World Bank staff calculations.

71. The progressivity of Moldova’s PIT system relies on hard-to-administer deductions44 that
benefit mostly the middle-income deciles45. While there is not minimum threshold in Moldova for the 12
percent flat tax rate, the number of deductions available to taxpayers have an impact on the income that
is taxable (MDL 25,000). While all deciles with the exception of the top first benefit from the deductions,
the 6th and 7th deciles are those that benefit the most. More disaggregated data show that the removal of

43 MoF estimates the amount of zero-rated supplies of goods and services, like electricity to residential consumers at around 700 million MDL.
44 There are six deductions that are applied in Moldova PIT system, namely: individual deductions that is converging to the existing minimum and
has the biggest impact on the progressivity, increased individual, deductions for the partner/spouse, increased deduction for the partner/spouse,
and deductions for dependents and increased deductions for dependents.
45 The simulations do not include other forms of revenues, like pensions, subsidies and social security payments, and scholarships, as they are

treated as non-taxable income and not captured in the dataset. Nevertheless, in the Tax Expenditure Review, MoF identifies these three
categories as the largest tax expenditures amounting to 1.7 billion MDL.

41
all deductions would be harmful for overall poverty rates, with an estimated increase in poverty of around
1.6 percentage points. The most affected by the removal of the deductions would be households with two
adults with more than one child. The total revenue loss from the existing deductions amount to around
72 percent of the current level of PIT collections. However, it is key for injecting progressivity into the
system, but it comes at a cost of cumbersome and hard-to-administer procedures.

72. A progressive PIT system based on multiple brackets and the removal of most of the deductions
could be one option. A progressive PIT with the first bracket at the zero rate is the most efficient direct
tax in terms of redistribution effects among the direct taxes, without the need for cumbersome
deductions. In this respect, a review was conducted of the tax system of peer countries to identify country
examples—Albania and Kosovo were selected as countries with tax systems that could provide some
insights of potential policy changes. Based on a model that uses household budget survey (HBS) data
adjusted to address the inherent bias of HBS coverage of households in the top two deciles,46 a simulation
was undertaken where the progressive PIT systems of Albania and Kosovo were applied to Moldova,
simulating the distributional effects and welfare of the population. The progressive PIT system in Albania
has three brackets with a cut-off threshold of MDL 48,000 for the bracket that has a tax rate of zero (the
bracket emulates Albanian system by applying the same ratio of the threshold to the minimum salary).
Kosovo, on the other hand, has four brackets with a cut-off threshold of MDL 16,800 for the bracket that
has a tax rate of zero.
Table 6. Policy scenarios
Moldova PIT_Albania PIT_Kosovo
Baseline
(2021)
Deductions
Individual 25,200
Increased individual 30,000
Threshold for individual deduction 360,000
Partner 18,900
Increased partner 18,900
Dependent 4,500
Increased dependent 18,900
Personal income tax rates
Standard PIT rate 12.0%
Separate PIT for farmers yes No no
PIT rate for farmers 7.0%
Brackets Rate cut-off rate cut-off
(annual) (annual)

1st 0.0% 48,000 0.0% 16,800

nd
2 12.0% 235,000 8.00% 58,800
3rd 23.0% no 12.0% 100,800
4th 16.0% no

73. The simulations show a differentiated impact across deciles and underlines the sensitivity of
the choice of tax rates and income thresholds. The scenario with the PIT system from Albania applied to

46
The high-income correction is based on the methodology described in Ruiz, N. and N. Woloszko (2016).

42
Moldova has a strong positive impact on the final income of all deciles at the expenses of revenue losses
(30 percent), while the scenario with the PIT system from Kosovo applied to Moldova has a negative
impact on the population but with an increase in revenue collection (15 percent). At the same time, the
upper deciles (particularly 8th to 10th) will gain more in terms of final income from the application of PIT
Albania scenario, while the effect on lower deciles is more neutral, as their income falls under the first
bracket (less than MDL 48,000) and is taxed at zero rate, which is comparable with the current situation
including all the deductions. When compared with the current situation, the overall poverty rate will
marginally decrease by 0.1 of a percentage point, with households with two adults and three children
most affected, with a poverty rate increase of 0.3 of a percentage point, while households with only
working age adults would benefit the most. Their poverty rate would decrease by 0.4 of a percentage
point. The impact of the Kosovo system on poverty and final income is negative. In this scenario, the first
cut-off threshold below which the income is not taxed is 34 percent lower than in the current Moldova
scenario. As a result, some of the taxpayers that are not currently paying any PIT in Moldova would pay
taxes at 8 percent in the case that the PIT system in Kosovo were applied. This would lead to an increase
in the poverty headcount for the PIT Kosovo scenario of 0.3 of a percentage point.

Figure 55. Percentage change in final income with Figure 56. Percentage change in final income with
respect to the 2021 baseline, by deciles respect to the 2021 baseline, by type of households

Figure 56. Percentage change in final income with Figure 58. Relative change in poverty headcount at
respect to the 2021 baseline, by locality consumable income level, %

43
 Figure 59. Relative change in poverty headcount at Figure 60. Impact on revenue and budget, compared
consumable income level, % with the baseline, %

Source: World Bank staff calculations. Source: World Bank staff calculations.

74. There is indeed room to improve PIT system along the principles of progressivity and simplicity,
while also increasing revenues. A high labor tax wedge undermines a country’s competitiveness and
creates incentives to under-report wages and employment. PIT collection is weaker than in peer countries
due to multiple exemptions, low incomes, large informality, and poor management. In the short-term, it
is important to review the PIT expenditures by consolidating and optimizing tax expenditures considering
international best practice, and weigh the importance and efficiency of the existing system. There is space
to improve the progressivity of full-time contracts and move away from the regressive taxation of self-
employed, patent holders, and liberal professions, etc. In the medium-term, it will be critical to look at
the overall PIT system to address more comprehensively the challenges of striking the right balance
between the principles of tax simplicity, efficiency, equity and progressivity. The number of personal
deductions and exemptions has made the system inefficient and cumbersome to inject progressivity. As
pointed out in the 2022 Tax Expenditure Review by MoF, there large amounts of tax expenditures,
particularly social security payments that also need a thorough attention. The examples of Albania and
Kosovo show how other countries have dealt with these tax principles. Calibrating the income thresholds
and applicable rates of Kosovo, which are at the moment more stringent, to Moldova’s demographics
could optimize tradeoffs between progressivity and revenue generation without the need for
cumbersome deductions.

Corporate Income Tax

75. Moldova’s CIT system has a wide range of tax facilities and exemptions from the standard
applied rate, leading to overall deficiencies in collections and increasing reliance on large-size taxpayers.
Under the current legislation, the standard CIT rate is set at 12 percent, while agricultural enterprises are
subject to a 7 percent CIT rate and SMEs that are not registered as VAT payers and correspond to specific
criteria may opt for a special CIT regime of 4 percent. Large benefits are offered to residents of the seven
free enterprise zones (FEZs) in Moldova, with 34 subzones, which had 213 companies employing over
16,500 employees. FEZ residents have a 50 percent exemption from the tax rate of 12 percent for the
income obtained from the export of goods or services, but also a 25 percent exemption from the tax rate
of 12 percent for income obtained from activities other than the export of goods and services. The IT park
companies pay a reduced rate of 7 percent on their sales revenue. There are also exemptions on the
income of banks and micro-lending institutions for loans and credits for terms exceeding three years (and
a reduction in tax for terms between two and three years). As a result, several sectors (agriculture, mining,

44
energy and gas, construction, IT, FEZs and finance)47 are not contributing in line with their share in the
economy.

76. Alternative options can be explored to improve the current CIT to increase the efficiency of
revenue collection, including by the removal of some of the existing income tax expenditures. The
model (the World Bank’s tax analyzer framework) uses a sample of 1,167 VEN taxpayer’s returns (income
returns), with the total population of the taxpayers amounting to 35,526 in 2019. Most of the sectors of
the economy and legal forms are represented in the sample. Around 94 percent of all entities in the
sample are SMEs. This is close to 98 percent—the share of SMEs in the administrative data—making the
sample largely representative for the whole population of active business entities in Moldova. In terms of
income, the sample is dominated by six taxpayers with gross income above MDL 500 million and four
above MDL 1,000 million. The model was used to simulate three policy reforms, namely: (i) the removal
of tax expenditures; (ii) the introduction of a higher uniform flat rate at 15 percent; and (iii) the
introduction of a flat rate of 15 percent only for large entities, and a rate of 12 percent to other firms
including farmers.

Table 7. Increase of revenue from CIT reform options

Flat CIT rate of 15% A flat CIT rate of 15% only to
Firm size Number of entities large entities
0-9m 909 27.2% 1.8%
9m-25m 142 27.5% 1.9%
25m-50m 56 25.2% 0.0%
50m-200m 43 25.1% 25.1%
200m-500m 11 24.9% 24.9%
500m-1b 2 25.2% 25.2%
>=1b 4 25.0% 25.0%
ALL sample 1167 25.4% 19.7%
Source: World Bank staff calculations.

77. The introduction of a 15 percent income tax rate will increase CIT collections by about 25
percent. Most of the increase in CIT collections (around 76 percent) would come from large companies
with a gross income of more than MDL 50 million. The distribution of CIT collections would not change
significantly, with the share of SMEs marginally increasing by 0.3 of a percentage point. Only 9 percent of
the total increase would be generated by additional collections from “very small” businesses with a
turnover of less than MDL 9 million, which would experience a 27 percent increase in CIT liability. To this
end, introducing a 15 percent CIT rate only for large entities and 12 percent to other firms including
farmers would lead to an overall increase in additional CIT collections of around 20 percent, with a
marginal increase for small businesses. Small businesses would also register an increase caused by
increasing the tax rate for farmers from 7 to 12 percent. Ninety-eight percent of the increase would be
from large businesses with more than MDL 50 million in annual gross income.

78. Reform of the CIT regime would require a holistic approach to increase collections and broaden
the tax base. These results reveal the overreliance on large taxpayers for CIT and the need to broaden the
tax base. Based on the existing sample, raising the collections from CIT by bringing SMEs to the standard
rate and eliminating the multiple existing tax rates would not improve overall efficiency significantly.
Simulations show that, due to a narrow tax base, even if the CIT rate is increased to 15 percent in absolute

47To note, the MoF identifies tax exemption of income from international grants for research and development, and from special funds and
grants approved by the government as one of the biggest (590 million MDL).

45
terms, CIT collections would generate around 25 percent of additional revenues compared with the
current law. The removal of tax expenditures is a potentially important source of additional CIT collections
that could strengthen overall CIT efficiency. In the short-term, to increase CIT collections to levels
comparable to peer countries in the region, the simplest technical option would be to increase the CIT
rate to 15 percent and tax more large enterprises with a gross income of more than MDL 50 million. In
the medium-term, however, the CIT system could be designed so that SMEs are taxed at 12 percent and
large companies at 15 percent. This kind of reform would have to be accompanied by a review of all tax
expenditures, particularly in the agriculture sector and SMEs. In parallel, other supportive schemes, more
targeted and transparent, should be developed and tailored for specific objectives.

79. Broadening the tax base for CIT collections would require tackling the undeclared economy.
According to the fiscal service of Moldova, the budget loses around MDL 20 billion annually because of
undeclared revenues, unofficially payment of salaries and tax fraud. Companies in the agriculture sector,
as well as SMEs, particularly holders of entrepreneur “patents”, are major elements of the undeclared
economy. Patent holders operate only on cash basis and are a major source of undeclared revenues. The
elimination of patents and encouraging cashless transactions would result in positive gains in terms of the
mobilization of CIT collections and the economy overall.

80. Moreover, some rebalancing of the tax system may be warranted to reduce distortions across
sectors and firms.48 Overall, the tax system in Moldova is conducive to investment and is on a par with
the countries in the region. Marginal effective tax rates (METRs) in Moldova are closer to the statutory
rate (12 percent) for those sectors that do not benefit from a special tax regime (16.5 percent in
manufacturing and 20.4 percent in services). The slight difference between METRs and the statutory rate
is mainly driven by: (i) lower tax depreciation compared with the economic depreciation; (ii) property and
transfer taxes; (iii) low inflation in the analyzed period; (iv) asset structure; (v) the debt-to-asset ratio; and
(vi) tax on debt and equity.

Figure 61. Comparison of Moldova's METRs with other selected countries in the region

Source: World Bank staff calculations based on WEO database.

48 METRs show some variations across sectors (manufacturing vs services). This reflects the differences in the actual asset mix/structure of a given
sector, given the differences of economic and tax depreciation across assets (for example, buildings depreciate far slower than heavy machinery).
Thus, even when the METRs for each separate asset class are the same in different sectors, the fact that sectors use different mixes of assets
causes the METRs to vary. METR calculations use an asset basket for each sector based on Canada data, due to the lack of actual data available
for Moldova. This is not particularly issue as asset structures of similar industries are not likely to vary much across countries, i.e., it is less likely
that hotels in Moldova have a major difference in the composition of buildings, machinery and land in the asset structure compared with Canada.
Lastly, METR for inventory is the result of the FIFO accounting for inventory in Moldova, whereby assets that are bought first are treated as sold
first. This means that any changes in the value of inventory due to inflation results in higher taxation and higher METRs

46
81. METRs for sectors with special tax regimes in Moldova are significantly lower than the statutory
CIT tax rate of 12 percent. FEZs (8 percent), the agriculture sector (11 percent), and small businesses in
manufacturing benefit the most (8 percent) from the current system. The main driver of this low rate is
the low effective CIT rate (4 percent for small businesses and 6 percent in FEZs and agriculture sector).
This lower tax rate more than compensates for the allowed reductions in taxable income. Significant tax
advantages exist for businesses in FEZs. While lower tax rates and abundant tax incentives for investments
made in FEZs are being debated, it is always a risk that these rate differentials potentially undermine
revenue, with possibly limited impact on growth and competitiveness. These sectors usually do not
promote better jobs, technology adoption, and reinvestment of profits in capital goods. The METR for the
manufacturing sector is lower than the one for services, mainly because of a higher proportion of
equipment (52 percent) in its asset mix. Equipment enjoys a lower METR (14 percent) than buildings due
to its tax depreciation (five years), being closer to its economic depreciation (five years). In contrast,
buildings have a higher METR (22.5 percent), as their tax depreciation (20 years) is much lower than their
economic depreciation (10 years). The services sector has a higher METR due to a higher proportion of
buildings (68 percent) in its asset mix, but this also implies that capital allowances do not reduce the
impact of the other taxes on capital.

Table 8: Marginal effective tax rates on capital in Moldova

Manufacturing Services Agriculture Manufacturing - Small business -

FEZ manufacturing
METR Equipment 14% 14% 7% 7% 4%
METR Buildings 22% 22% 12% 12% 8%
METR Land 18% 18% 26% 26% 29%
METR Inventory 13% 13% 7% 7% 5%
METR Total 17% 20% 11% 8% 6%

Source: World Bank staff calculations.

82. Business investments are highly affected by higher Social Contributions /payroll tax wedge.
Business incentives to invest decrease if we consider all factors of production (capital and labor). A higher
tax wedge pushes up the overall METRs on production, including in agriculture and FEZs. The calculations
of the overall METRs on production show that METR for production is much higher than METR on capital
due to a higher rate of SSC (24 percent). In the case of special tax regimes, the higher tax wedge pushes
up the overall METRs on production, in particular for FEZs.

Table 9: METRs and SSC/payroll tax wedge in Moldova

Manufacturing Services Agriculture FEZs Small

business
METR (Capital) 17% 20% 11% 8% 6%
METR (Production) 26% 28% 23% 21% 20%
Capital Share 0.5 0.5 0.5 0.5 0.5
Payroll Tax Wedge 36% 36% 36% 36% 36%
- PIT 12% 12% 12% 12% 12%
- SSC 24% 24% 24% 24% 24%

Source: World Bank staff calculations

47
83. In addition, the existing tax code allows companies to make various deductions that total
around 9 percent of paid CIT. The removal of all tax expenditures would result in an increase of CIT
collections of around 9 percent, with collections from SMEs increasing the most. Collections from SMEs
would increase by more than 40 percent compared with the current law. Nevertheless, in absolute terms,
most of the increase (around 63 percent) from the removal of tax expenditures would come from large
companies with a gross income of more than MDL 50 million. From the list49 of exemptions, the removal
of income deductions, such as revenues obtained from state securities, donations, selling of fixed assets,
and others, represent around 76 percent of total possible CIT collection gains. The first three sectors
account for 86 percent of revenue losses from deductions from business income. The financial sector is
by far the biggest beneficiary from income deductions, followed by the production of beverages and
agriculture. The second-largest yield would generate the removal of expenditure adjustments (around 21
percent), such as family-related spending, insurance, losses incurred in the event of exchange or sale of
assets, the provision of services between interdependent persons, expenses related to the obtaining of
income that is exempt, etc. The impact from the removal of other tax expenditures is negligible.

Table 62. Simulation of removing CIT expenditures Figure 63. Revenue loss from the income adjustments
and deductions by sector
Elimination of
all tax
expenditures
Number of (change relative
Firm size entities to current law)
0-9m 909 17.4%
9m-25m 142 9.8%
25m-50m 56 16.8%
50m-200m 43 9.7%
200m-500m 11 1.2%
500m-1b 2 4.1%
>=1b 4 8.9%
ALL sample 1,167 8.7%
Source: World Bank staff calculations. Source: World Bank staff calculations.

84. Tax expenditures should be reviewed to ensure that they are relevant, effective, and efficient.
While some tax expenditures are born out of good policy intentions, companies activating in FEZs,
particularly those that are not exporting, and IT sector should be scrutinized. From a practical perspective,
providing a complete tax exemption for businesses that invest above a certain level causes issues related
to the measurement of the level of investment, and these are often complex and a burden to the tax
administration. In turn, tax credits could be offered instead of pure exemptions, where taxpayers would
need to submit a tax return to obtain the tax reduction, thus becoming known to the tax authorities. The
same could be implemented in the IT sector, where a low CIT rate is not likely to be among the top
priorities for software development. Many companies in the IT parks are not developing software, but
rather manage and/or maintain databases and provide customer support for foreign companies where
the salaries are higher. If what the authorities are looking to encourage is research and technological
development, they could do so in a less expensive and more efficient fashion by providing tax incentives
for research and development, expenses that are already fully deductible under Moldova’s tax code.

49Total 'fiscal facility' tax credit [VEN130]; exempt taxable income [VEN0902]; adjustment to business expenditure [VEN030]; personal and family
exemptions [VEN0701]; exempt gifts [VEN050]; exempt no-documentation [VEN060]; adjustment to business income [VEN020]

48
Annex 1: References
Abdih, Y., Lopez-Murphey, P., Roitman, A., and R. Sahay. 2010. "The Cyclicality of Fiscal Policy in the Middle
East and Central Asia: Is the Current Crisis Different", IMF Working Paper.
Braun M. dnd L. Di Gresia. 2003. “Towards Effective Social Insurance in Latin America: The Importance of
Countercyclical Fiscal Policy”, Annual Meetings of the Inter-American Development Bank and Inter-
American Investment Corporation Milan, Italy.
Frankel, J., C. A. Végh, and G. Vuletin. 2013. “On Graduation from Fiscal Procyclicality.” Journal of
Development Economics 100: 32-47.
Carneiro, Francisco, and Leonardo Garrido. 2015. "New evidence on the cyclicality of fiscal policy." World
Bank Policy Research Working Paper 7293.
Galeano L., A. Izquierdo, J. Puig, C. Vegh, G. Vuletin. 2021. “Can automatic government spending be
procyclical?” NBER, Working paper 28521.
Guerguil, Martine, Pierre Mandon, and René Tapsoba. 2017. "Flexible fiscal rules and countercyclical fiscal
policy." Journal of Macroeconomics 52: 189-220.
Halland H. and M. Bleaney. 2011. “Explaining the procyclicality of fiscal policy in developing countries”,
CREDIT Research Paper, No. 11/09, The University of Nottingham, Centre for Research in Economic
Development and International Trade (CREDIT), Nottingham
Herrera, Santiago; Kouame, Wilfred A.; Mandon, Pierre. 2019. “Why Some Countries Can Escape the Fiscal
Pro-Cyclicality Trap and Others Cannot?” Policy Research Working Paper, no. WPS 8963; Policy Research
Working Paper; No. 8963. World Bank, Washington, DC.
IMF. 2014. Fragmentation and Vertical Fiscal Imbalances: Lessons from Moldova. IMF Working Paper
14/233.
IMF. 2019. "A Strategy for IMF Engagement on Social Spending." IMF Policy Paper.
IMF. 2019. “Improving the Efficiency and Equity of Public Education Spending: The Case of Moldova.” IMF
Working Paper No. 2019/042
Manasse P. 2006. “Procyclical Fiscal Policy: Shocks, Rules, and Institutions—A View From MARS”, IMF
Working paper, Fiscal Affairs Department, WP/06/27.
Muñoz, Ercio, and Eduardo Olaberria. 2019. "Are Budget Rigidities a Source of Fiscal Distress and a
Constraint for Fiscal Consolidation?" World Bank Policy Research Working Paper 8957.
Republic of Moldova. Ministry of Finance. 2022. “Tax Expenditure Assessment Report”.
Ruiz, N. and N. Woloszko. 2016. "What do household surveys suggest about the top 1% incomes and
inequality in OECD countries?", OECD Economics Department Working Papers, No. 1265, OECD Publishing,
Paris, https://doi.org/10.1787/5jrs556f36zt-en
Schick, Allen. 2008. "Budgeting for fiscal space." In Asian OECD Senior Budget Officials Meeting, pp. 1-10.
Serven, Luis. 1998. “Macroeconomic Uncertainty and Private Investment in LDCs: an Empirical
Investigation.” Working Paper No. 2035 (Washington, DC: The World Bank).
Vegh, Carlos; Lederman, Daniel; Bennett, Federico R. 2017. “Leaning Against the Wind: Fiscal Policy in
Latin America and the Caribbean in a Historical Perspective.” LAC Semiannual Report; April 2017. World
Bank, Washington, DC.

49
World Bank. 2013. “Capital Expenditures: Making Public Investment Work for Competitiveness and
Inclusive Growth in Moldova.” Public Expenditure Review, Report 76310 (Washington, DC: World Bank,
2013).

World Bank. 2016. “Republic of Moldova Public Finance Review: Towards More Efficient and More
Sustainable Public Finances.” Washington, DC. World Bank.

World Bank. 2016. “Moldova Economic Update. Special topic: Strengthening revenue mobilization in
Moldova: Car taxation” World Bank: Washington, DC.

World Bank. 2019. “Moldova Economic Update. Fall 2019.” Special Focus Note: Unfinished Pension
Reform. World Bank: Washington, DC.

World Bank. 2019. “Moldova: Rekindling Economic Dynamism. Country Economic Memorandum.” World
Bank, Washington, DC.

World Bank. 2022. “Moldova Policy Notes 2021: Sectoral Recommendations.” World Bank, Washington,
DC.

World Health Organization. 2020. “Can people afford to pay for health care? New evidence on financial
protection in the Republic of Moldova.” World Health Organization. Regional Office for Europe.

50
Annex 2: A Taxonomy of Budget Rigidities
Annex Table 2.1: A taxonomy of budget rigidities
Origin Type of Expenditure Description
1. Benefit principle of funding Social security benefits Expenditures that benefit those who
previously contributed
Government entities’ expenses
financed by their own revenues Revenue generated by entities to fund
their operating costs
Infrastructure maintenance costs
financed by user charges Costs paid by users of public
infrastructure/facilities
2. Rights and guarantees established Meritorious goods Earmarking of certain revenues for
in the constitution or regulations specific sectors
Guarantees of assured service provision
and access to it Guarantee of minimum spending levels
in specific sectors
Priority policy objectives
3. Fiscal federalism/decentralization Revenue-sharing system Transfers to cover vertical and
Earmarked or conditional transfers horizontal imbalances in fiscally
decentralized structures

Finance decentralized service delivery

(by lower levels of governments)
4. Macroeconomic dynamics Interest payments on debt Spending is defined by exogenous
factors outside the control of current
Clauses for civil-service wage policymakers such as inflation, GDP
adjustments or benefit adjustments growth, the interest rate, or the
exchange rate.
5. Extraordinary income Natural resource exploitation Proceeds from non-renewable revenue
sources allocated according to
Privatization proceeds derivation criteria (i.e., compensated
for externalities to producer areas) or to
wealth/stabilization funds to promote
intergenerational fairness.

Once-for-all proceeds are allocated to

predefined uses.
6. Resource disputes within the Allocations to judiciary or legislative Ensure adequate funding to ensure
public sector branches autonomy and proper functioning of
the branches of government.
Politically inflexible spending /
compensation of employees Civil-service unions that ensure
adequate wages for civil service
7. Policy commitments Tax expenditures Incentives to promote sectors

Subsidies Social programs that have a permanent

delivery
Social programs
Multi-year implementation of
Investment investment projects

Future maintenance cost of

infrastructure

51
Annex 3: Tax System in Moldova
Type of Tax Rates Remarks
Income Taxes
Corporate Income Tax 12% FEZ: Export-oriented businesses are subject to CIT of 6%. Others are
subject to CIT rate of 9%. Moreover, export-oriented businesses enjoy
tax holiday for 3 to 5 years (which can be extended) if additional
investment of US$ 1 to 5 million is made in fixed assets.
Income Tax on Small 4% Small and medium companies that are not registered as VAT payers
Businesses (i.e., if the total turnover within the last 12 consecutive months is
below MDL 1.2 million) and correspond to specific criteria may opt for
a special CIT regime of 4% on their aggregated income determined for
accounting purposes, except for certain types of income (e.g.,
subsidies, dividends, exchange rate differences).
Personal Income Tax 12% 6% in case of individuals who carry out activities in the field of
purchases of phytotechnic and / or horticulture products and / or
objects of the vegetable kingdom (Chapter 10 of the Tax Code)
7% in case of farmers
Treatment of interest income 12%
Treatment of Dividend income 6%

Treatment of Capital Gains 12%

Tax Depreciation Straight line

- Buildings 5%
- Construction 8%
- Roads, certain 12.5%
equipment
- Industrial equipment 20%
- Car, computer, 30%
furniture
Indirect and Property Taxes
VAT Standard: 20% Reduced rate: applies to local supplies of bread and bakery products;
Reduced: 8%, milk and dairy products; transport and distribution of natural gases
12% services; biofuels used for electricity, heating, and hot water
Exports & FEZ: 0% production; and specific phytotechnical, horticultural, and
zootechnical products.
Land Tax 0.1 – 0.3%

Property Tax 0.05 – 0.4%

Customs Duty Various rates
Other Taxes
Single tax for IT park residents 7% TITLE X of the Tax Code
Private tax 1% the tax rate is established annually by budget law (Law No. 258/2020)
Wealth tax 0.8% TITLE VI1 of the Tax Code
Non-Tax parameters
Interest rate (lending, 2014) 21.2%
Inflation rate 1.49%
International interest rate 1% LIBOR rate for 3 months securities

52
Dividend payout ratio 50% 50% of the profits are reinvested while 50% are distributed. This has
relevance when calculating the tax rate on equity.
Te=γTd+(1-γ)Tc
where Te is the weighted average tax rate on equity, Td is the “personal”
tax rate on dividends (and/or the tax rate on dividend distributions), Tc
the accrual equivalent tax rate on capital gains, and γ is the dividend
payout ratio.
Debt-Asset Ratio 0.4

53