Nicolas Ruiz
OECD
OECD Economic Surveys: Estonia 2022
OECD Economic Surveys: Estonia
1. Key Policy Insights
Introduction
Since its independence, Estonia has made tremendous progress towards greater economic prosperity. Per capita income more than tripled over the past 30 years and the gap to the upper half of OECD countries has been continuously narrowed (Figure 1.1). Estonia enjoys solid institutions, political stability, a strong and credible fiscal policy, as well as a robust financial sector. Its major structural strengths are well recognized, including high educational outcomes, a flexible labour market and a business-friendly regulatory framework. Estonia is also a frontrunner in digital governance and innovation, with the highest number of native ICT unicorns per capita in the world. Stable and secure digital services are in fact one of the factors that have allowed Estonia to cushion better than others the sanitary and economic shock from the pandemic, as tools such as electronic health records, digital classrooms and a large range of seamless online services were already in place. A large and timely fiscal response also played a role in stabilising and preserving living standards, and Estonia achieved a solid recovery in 2021.
Figure 1.1. The income gap continues to narrow
GDP per capita
Source: OECD National accounts database.
However, even before Russia’s war against Ukraine, the buoyant recovery started to fuel inflation, reinforcing some pre-existing structural challenges. While Estonia’s labour market has recovered following the peak in unemployment during the pandemic, mounting unmet demand for skilled labour could be a brake on sustainable growth and a driving force of core inflation. Poverty among old-age pensioners and low-paid workers remains a problem, which is likely to have been worsened by the pandemic-induced recession and could be further aggravated by current inflationary pressures. A country that emits large quantities of greenhouse gases per capita, Estonia faces also the challenge of embarking on a fast transition towards low-carbon activities. These various challenges can be addressed with the help of tax reforms, but this should be done without undermining the existing efficiency of the tax mix, which is a key aspect of the country’s attractiveness for both businesses and talent.
This survey charts the road ahead for Estonia to ensure a resilient and inclusive recovery and fostering the convergence process, with the following key messages:
Sustaining the economic recovery will require new reforms. After the impressive short-term rebound and in a particularly uncertain context, a renewed focus on structural reforms will help Estonia remain on a path of rapid convergence. In particular, reforms should focus on addressing labour shortages and skills mismatches, which can also contribute to tame inflation, while protecting the existing flexibility of the labour market.
Achieving the government´s goal of reducing poverty requires further policy measures in the current context. Social transfers could be better targeted towards people left behind, such as the elderly, while social partners should continue to make work pay better. Spending on health and infrastructure should be made more efficient to deliver better value-for-money.
Reducing carbon emissions will also entail significant policy changes. Electricity generation will need to move further away from oil shale towards diversified sources of renewable energy. There is also scope to substantially reduce the use of fossil fuels in the transport sector and to increase the heating efficiency of buildings across Estonia. Achieving this will require a comprehensive approach that combines several policy tools such as carbon pricing, public investment and private-investment incentives. Attention needs to be paid to impacted consumers and workers.
The post-pandemic recovery was strong
In 2020, Estonia has withstood the pandemic shock better than its peers, with a contraction of GDP of 2.7%, one of the softest in Europe (Figure 1.2). While at the beginning of 2021 the fast escalation of infections forced the government to re-introduce containment measures, those were comparatively mild (Figure 1.3). Roughly, 75% of the economy were unaffected or affected only partially through supply chains. As a result, GDP grew by 8.2% in 2021, and activity surpassed its pre-pandemic level. At the beginning of 2022, despite a third wave of infections surpassing the peak of the second, no new significant containment measures were imposed and now contaminations are low. This can be partly traced back to a successful vaccination campaign, including the current booster shots programme. Still, a large share of unvaccinated people remain, warranting further efforts to accelerate vaccination.
Figure 1.2. The economy bounced back strongly before the war, pushing up inflation
Source: OECD Analytical Database.
Investments also surged in 2021, but more than 80% of this came from a major car manufacturer’s software investments in its new Estonian subsidiary, that by now have been withdrawn. While the impact of these specific investments on GDP is small, as they are also recorded as imports/exports of services, they have nonetheless amplified temporarily the current account deficit in a context where several factors were already depressing it. With tourists’ spending still very low, the exports of services remain subdued, in particular passenger services. At the same time, the overall strength of Estonia’s recovery compared to its main partners has boosted imports, led notably by wood and mineral fuels, more than exports for the most part of 2021.
Still, the growth of goods’ exports in value terms is robust, notably in the ICT sector, albeit this is partly explained by the acceleration of export prices. In the manufacturing industry, output has reached the record levels of early 2019. Household consumption remains solid based on growing wages, the drawing down of savings accumulated during the pandemic and the disbursement of funds following the reform of the second pillar of the pension system. However, extraordinarily high inflation will hamper consumption, as wages will grow less than prices. Still, households’ deposit stock remains EUR 2 billion higher than before the pandemic, equal roughly to 14% of annual private consumption. Given how uneven the distribution of household deposits is, as households with large deposits before the pandemic tend to have increased their deposits more than less wealthy households, most of these savings may not further fuel growth through consumption.
Figure 1.3. COVID-19 cases are low
Source: Refinitiv and Oxford University.
Owing to the war in Ukraine, economic sentiment has deteriorated recently, after continuous improvement through 2021 (Figure 1.4). Industrial enterprises’ production and expectations for the services sector are still well above pre-pandemic levels but declining, while consumer confidence is now below pre-pandemic levels due to rapid inflation (Figure 1.2, Panel C), which reached almost 20% in April 2022, the fastest in the euro area at that time. Upward price pressures come mainly from the cost of food (14.6% vs 13.8% in March) and energy (38% vs 44% in March). Producer prices also rose by 31.8% over the year to April.
Figure 1.4. Economic sentiment is declining
Source: Eurostat, European Commission Business and consumer surveys.
Nonetheless, the labour market remains solid, after a sharp downturn during the pandemic (Figure 1.5). With a very flexible labour market, employment dropped markedly in 2020 while the rise in unemployment, peaking at almost 9% at the end of 2020, has been among the highest in the EU, despite the job retention scheme put in place with generous eligibility conditions. Unemployment is now continuously declining and the employment rate is also recovering but is still below its pre-crisis level. Given the integration in the labour market of many low- skilled workers in recent years, the crisis could leave permanent scars on vulnerable groups. For instance, many workers who lost their jobs in the tourism sector have not yet re-integrated the labour market. Demand for new employees has increased significantly in recent months in both industry and services, but supply has not kept pace and almost one third of entrepreneurs are seeing labour shortage as their biggest obstacle to development (Swedbank, 2021).
Figure 1.5. The labour market is recovering from a marked drop
Source: OECD Labour Force Statistics.
The recovery is thus exposing some imbalances already characterized in previous Surveys (OECD, 2019c and 2017c). The lack of suitable labour despite the unemployment rate being substantially higher than before the pandemic underscores skill mismatches as well as changing demographics, while pandemic-induced stricter border controls and administrative hurdles have led to a lack of foreign workers to help filling the gap. Going forward, strengthening up-skilling and re-skilling programmes in line with employers’ needs will be key to address labour shortages, while allowing low-skilled and displaced workers to strengthen their links with the labour market and benefit from the recovery.
Moreover, as currently envisaged by the government work immigration programmes should be made temporarily less restrictive, notably for non-EU immigration, for instance by temporarily extending short-term employment registration above one year. The influx of refugees from Ukraine may help in that regard. So far, Estonia has received 30,000 refugees from Ukraine, two thirds of whom are adults, and 10,000 of whom are expected to join the labour market. While this will increase unemployment in the short term, in the medium-term it will help fill gaps in various sectors such as construction and transportation, if the war is prolonged and not all of the refugees are able to return home. While immigration may influence relative wages for individual categories of workers, at the aggregate level OECD evidence tends to show that pressures on real wages are limited and vanish after a few years, and thus will not have a significant impact on Estonia’s convergence in living standards (Jean et al., 2010).
Labour shortages are putting more upward pressure on wages (Figure 1.6, Panel A). In 2020, with the help of the job retention and wage compensation schemes put in place, wage growth eased to 1%. However, in 2021 average nominal wages grew by 7.1%, a pace faster than the pre-crisis trend level. With Estonian incomes lower than the OECD average, higher wages in real terms will be a welcome development once inflation stabilizes. In 2021, part of wage growth was driven by improved productivity, as companies recruited fewer workers than expected despite accelerating growth, signalling also a shift towards higher value-added activities, especially the IT sector, which could in return durably sustain inflation but also increase the quality of exports. On the other hand, the aforementioned labour market imbalances could make wages grow faster than productivity, reducing cost competitiveness and hampering exporting industries in the long run. Compared to its peers, Estonia’s cost competitiveness was decreasing before the pandemic (Figure 1.6, Panel B).
Figure 1.6. Rapid wage growth could re-apply pressure on competitiveness
Source: OECD Analytical Database.
Despite declining export competitiveness, the pandemic and the war in Ukraine, export performance is holding up well (Figure 1.7). This can be partly traced back to the fact that Estonia‘s main trading partners, Latvia, Sweden and Finland, suffered mild contractions of roughly the same magnitude as Estonia during the pandemic. Moreover, economic links with Russia have weakened considerably since 2014, and Russia’s share in Estonian exports is now below 2%. Nonetheless, the trade deficit widened in 2020 after years of surplus. With restrictions on mobility and economic activity less severe than elsewhere, the hit to domestic demand was smaller than to exports. However, as the Baltics and the Northern parts of Europe remain the core-trading region for Estonia, accounting notably for half of its exports, the war in Ukraine could represent a long-lasting source of uncertainty for a small and open economy like Estonia. However, the constant gains of market share in exports of high-quality and complex services could contribute to lowering this risk by diversifying trade.
Figure 1.7. Export performance remains robust
Source: OECD Analytical Database; UN Comtrade; and Stat.ee.
Looking forward, growth is expected to slow in 2022, as inflation will far surpass nominal wage growth, hampering household purchasing power. In 2023, as price pressures on energy will remain elevated, notably due to the recently announced embargo on Russian oil imports, nominal wage growth will still remain below inflation and GDP is projected to grow at 1.8% (Table 1.1). Export opportunities are expected to shrink, which, together with reduced confidence, will weaken investment. Unemployment is also expected to increase, as a large number of refugees are entering the country and not all of them are likely to find jobs immediately. However, the use of savings accumulated during the pandemic and in individual pension funds will keep private consumption afloat. Moreover, an expected pick-up in EU fund absorption will underpin activity, with rising spending as the previous 2014-2020 financing cycle approaches its end in 2023 and through the EU Recovery and Resilience Facility. Those funds will have a sizeable impact on the growth potential of Estonia and its resilience, by consolidating and deepening further digitalisation, and initiating investments to accelerate the green transition. The implementation of these funds is also expected to pervade across all sectors of the economy.
Diverging economic conditions between Estonia and the Euro area, and the resulting accommodative monetary policy conducted by the European Central Bank, could lead to lower-than-appropriate interest rates given inflation and Estonia’s growth potential (Figure 1.8). To mitigate inflationary pressures, public spending should be focused only on assistance to refugees, defence and infrastructure developments that increase energy security. Support for low-income households to mitigate the negative impact of inflation on essential consumption needs should also remain narrowly targeted. The tightening of macro prudential policies could also be considered.
Table 1.1. Macroeconomic indicators and projections
1. Contributions to changes in real GDP, actual amount in the first column.
2. Harmonised index of consumer prices excluding food, energy, alcohol and tobacco.
3. The Maastricht definition of general government debt includes only loans, debt securities, and currency and deposits, with debt at face value rather than market value.
Source: OECD Economic Outlook database (June 2022).
Figure 1.8. Interest rates are lower than appropriate given inflationary pressures
Note: The Taylor rule interest rate is calculated as: i = annual real potential GDP growth + core inflation + 0.5 × output gap + 0.5 × (core inflation – 2).
Source: OECD, Economic Outlook.
The projections are subject to risks skewed to the downside. Additional disruptions to supply chains, more persistent inflation or prolonged weakness in major trading partner growth could all further weaken the outlook. Moreover, the structural scars of the pandemic remain uncertain in scale and could, in fact, be amplified by the current inflation pace. An increase in poverty, from an already high level, could generate social discontent and present adverse effects. The economy could also face further unforeseen shocks not factored in the projections (Table 1.2).
Table 1.2. Large-scale shocks with a negative impact on the outlook
The macroeconomic policy framework remains solid
Public finances are robust
Estonia entered the pandemic with the lowest general government debt ratio of the OECD, at just 8.6% of GDP in 2019, fully matched by fiscal reserves, meaning that net debt was zero. In 2020, the appropriate force of the policy response, which has greatly mitigated the COVID-19 shock, pushed the 2020 fiscal deficit to 5.6% of GDP (Figure 1.9). This was better than the one originally budgeted (6.6%), as the impact of higher spending was partially offset by a lower-than-anticipated decline in overall revenues. In light of the second pandemic wave at the beginning of the year, the government appropriately carried-over and even amplified the support measures introduced in 2020. The 2021 fiscal package, of 6.5% of GDP, supported the economy recovery. The fiscal deficit declined to around 3.3% in 2021, thanks to strong growth and wage increases that occurred during the year. During the first half of 2022, a supplementary budget, of around 3% of GDP, was passed to cope with the effects of the war in Ukraine, notably for covering initial expenses related to integrating refugees, as well as for strengthening energy security, notably establishing gas reserves and investing in liquefied natural gas capacity and strengthening defence.
Figure 1.9. The largest deficit on record helped to cushion the pandemic
Source: OECD Economic Outlook (January 2021).
The COVID-19 crisis thus led to a significant increase in Estonia’s gross public debt by around 9.8 percentage points in 2020, albeit from a very low level. Nonetheless, it remains comparatively very low as a share of GDP (Figure 1.10). From 2024 on, the authorities wish to restore a prudent fiscal stance and reduce the level of debt to its pre-pandemic level. After having triggered the national escape clauses from fiscal rules for 2020–21, in line with EU guidelines, the 2022-25 State Budget Strategy reaffirms the objective of a structural budget deficit of the general government of 0.5 percent of GDP in the medium term. More precisely, the State budgets will be prepared such that the structural budgetary position of the general government is balanced or in surplus, while annual structural deficit up to 0.5% of GDP can be allowed only if potential cumulative structural surplus have been generated during previous fiscal years. Regarding the short-term transition to this stance, the budgetary deficit will start to be reduced from 2022 onwards, with a deficit that must not exceed 2.2% of GDP that year, and then must be reduced by 0.5% of GDP each year. However, if economic growth remains strong, with tax revenues exceeding expectations, the windfall gains should be used to reduce further the fiscal deficit.
Figure 1.10. Gross public debt has increased but remains very low
Note: Using the general government debt Maastricht criterion, with the exception of the OECD average.
Source: OECD EO database.
In the long run, this fiscal strategy must be framed in the context of future budgetary pressures to cope with an ageing population. With comparatively favourable demographics for the forthcoming decade, the fiscal impact of ageing is less pronounced than in other countries. The OECD Long-term Model estimates that the ratio of structural primary revenue to GDP in Estonia would need to rise by 2% between 2022 and 2060 to stabilise the gross debt-to-GDP ratio at 30% (Figure 1.11). A similar reduction of spending, or a combination of revenue increases and spending cuts would likewise stabilise debt as a share of GDP. Broadening the tax base and increasing the efficiency of the health sector could contribute to alleviate those spending pressures.
Figure 1.11. Ageing costs will add to fiscal pressures
Note: In Panel A, “Other primary expenditure” is projected based on the assumption that governments will seek to provide a constant level of public spending per capita in real terms. Under some reasonable assumptions, the evolution of this expenditure category relative to GDP becomes an inverse function of the projected evolution of the population-to-employment ratio, as expenditure (numerator) follows population whereas GDP (denominator) follows employment. The “other factors” component captures anything that affects debt dynamics other than the explicit expenditure components (it mostly reflects the correction of any disequilibrium between the initial structural primary balance and the one that would stabilise the debt ratio). In Panel B, underlying projected growth rates, interest rates, etc., are from the baseline long-term scenario (for further details, see Guillemette, 2021). The debt path in the “With measures to offset ageing-related costs” scenario assumes the primary budget converges to balance in 2030 and then stays at that level. Interest receipts are assumed to remain at 1% of GDP after 2030. The debt path in the “With stabilizing measures” scenario assumes a rise of 2% of the ratio of structural primary revenue to GDP.
Source: OECD Long-term model.
While Estonia’s fiscal framework is designed to conduct neutral fiscal policy, the latter has often been pro-cyclical in recent years (Figure 1.9). Attempts to achieve balanced budget can lead to pro- cyclicality, as it tends to exaggerate growth through higher-than-necessary public expenditures during economic expansion, while expenditure cuts to balance the budget amplify downturns. As highlighted in previous Surveys (OECD, 2019c), fiscal expansion appeared to have fuelled growth in 2018-19, as windfall revenues were spent instead of creating surplus for future deficits. Eelarvenoukogu (“Estonia’s Fiscal Council”) pointed out that recurrent disagreements occurred on the estimation of Estonia’s cyclical position, leading the government to plan the budget on a worst-case scenario basis based on the recurrent assumption of a negative output gap, while a broader set of indicators than GDP led to think otherwise regarding the health of the Estonian economy (Eelarvenoukogu, 2017). Also pointed out by Eelarvenoukogu, such pro-cyclicality is in fact breaching the fiscal rule, which requires planning surpluses after reaching a balanced budget to offset past deficits. Enhancing the statistical apparatus beyond the single GDP basis used to estimate Estonia’s cyclical position, for example by considering a broad set of labour market indicators or incorporating indicators of tax revenues, would help to get a more accurate picture of Estonia’s cyclical position for the drafting of budget plans.
Moreover, and as already pointed out in recent OECD publication (OECD, 2021b), automatic stabilisers lack size and traction, and the government has limited fiscal tools to smooth cyclical fluctuations beyond escape clauses. Such issue is particularly vivid for unemployment insurance, where contributions exhibit pro-cyclicality (Figure 1.12). This pro-cyclicality is partly due to the setting of contributions based on the needs of the state budget balance rather than labour market needs. This implies that the total tax burden on labour is higher during worse economic conditions and lower during better times. As a result, the system may not accumulate reserves to face longer economic downturns. New principles for setting unemployment insurance contributions could be formulated by introducing counter-cyclicality, for example by incorporating the current reserves, and their forecast, of the unemployment insurance funds. As currently discussed, automatic triggers for higher generosity could also be introduced, such as in the United States (Box 1.1).
Figure 1.12. Unemployment insurance contributions are pro-cyclical
Unemployment rate and unemployment insurance (UI) premium, 2002-2019
Source: OECD (2021b) and Government decrees on unemployment insurance premium in 2002-2019.
Box 1.1. Normal and extended unemployment benefits in the United States
In the United States, normal unemployment benefits are normally available for 26 weeks in the United States under the joint federal state Unemployment Compensation (UC) program. Those can be periodically supplemented and extended during episodes of economic distress, through a combination of permanent and temporary legislation:
The federal Extended Benefits (EB) program, permanently authorized, provides up to 20 weeks of additional unemployment compensation for unemployed individuals who lost jobs in states where the level and change in the state unemployment rate is above a specified threshold. The thresholds or triggers are state specific but most commonly are based on an overall unemployment rate of 6.5 percent (for a 13-week extension) or 8.0 percent (for 20 weeks).
Congressional intervention can also temporarily establish, as in 2008 and 2021, temporary programmes of extended UC benefits when a recession is officially declared by the National Bureau of Economic Research (Cahsel, 2008). These programmes extend the time an individual can claim UC, typically ranging from an additional 6 weeks to 63 week, and have expiration dates. Historically, temporary programmes started operating after the recession had officially ended (albeit this was not the case in 2021 with the UC program as part of the CARES act). One of the main reason is that the exact date of the recession is not known until months after that recession has started, and the National Bureau of Economic Research NBER often announces a recession has begun three or more months after what is later determined to be its official start.
Permanent and temporarily extended UC programmes play an important role for macroeconomic stabilization in the United States. During the Great Recession, it is estimated that they closed about two-fifths of the real GDP shortfall caused by the recession (Nicholson and Needels, 2011). In terms of financing, they allow to collect sufficient reserves when the economy grows, as UC programmes revenues rise through increased tax revenues while UC program spending falls because fewer workers are unemployed and receive benefits, preparing for any economic downturn. In addition, evidence tend to demonstrate that extended UC programmes during the Great Recession may have helped promote labour force attachment among recipients (Rothstein, 2011; Farber and Valletta, 2015).
Table 1.3. Past OECD recommendations on macroeconomic conditions
The financial sector is sound but the housing sector should be monitored
The financial sector cushioned the shock well. Benefiting from higher precautionary savings, banks did not show any sign of deteriorating solvency or liquidity. Banks’ capital-to-asset-ratios remain among the highest in the OECD, while the share of non-performing loans is at an all-time low (Figure 1.13). Some of the measures taken during the first wave of the pandemic to ease financing conditions, such as lowering the systemic risk buffer, are also gradually withdrawn in light of the recovery. A recent Central Bank’s stress test confirmed banks’ low exposure to sectors severely impacted by the pandemic. The exercise suggests that even in the case of an adverse scenario, non-performing loans would rise by a much smaller magnitude than during the global financial crisis.
Figure 1.13. The banking sector displays healthy ratios
Source: IMF, Financial Soundness Indicators database.
However, a large share of the loan portfolios of the banks operating in Estonia is composed of loans to real estate and construction companies. These were significantly affected by the crisis in 2020 but have since recovered and the real estate market, while remaining affordable, is now accelerating again (Figure 1.14). While the Russian invasion of Ukraine initially made households in Estonia cautious about borrowing, demand for loans recovered quite quickly. However, there is no sign of overheating according to the authorities, and the Central Bank considers price dynamics to be in line with fundamentals. That being said, the expected gradual withdrawal of accumulated deposits, particularly by wealthy households, matched with overall easy financing conditions, could pose a risk. Developments in the housing market should be carefully monitored and standard macro-prudential instruments, e.g. loan-to-value and debt-to-income ratios, should be promptly adjusted should overheating signs increase. In that regard, the planned clarification and tightening of eligibility criteria for housing loans backed by state guarantees are welcome.
Figure 1.14. House prices have regained momentum but affordability remains
Source: OECD Analytical house prices indicators database; and Statistics Estonia.
Box 1.2. Quantifying the fiscal impact of selected reforms
The following estimates quantify the fiscal impact of selected medium-term reforms. They are taken from a variety of sources, including costs observed in other countries, and hence serve only an illustrative purpose. Some of the measures involve one-off (e.g. the rollout of broadband), while others involve continuous disbursement of public funds (e.g. VET spending). The negative second-round effects of these reforms on GDP in the case of tax increases is not considered.
Table 1.4. Illustrative fiscal impact of selected reforms
1. Assuming that rollout costs are proportionate with the size of the area and based on EUR 14 billion estimated costs for rural area coverage.
2. Based on the assumption of 700M submitted to the 14% rate in 2021.
3. Based on calculations from the Ministry of Finance of Estonia.
4. Based on the assumption that the real estate tax will be in the same magnitude (as a ratio to GDP) as the average OECD country.
Source: OECD calculations
Box 1.3. Quantifying selected structural reforms
The following estimates roughly quantify the cumulative GDP impact of reform scenarios after 10 years and are illustrative. The estimates, subjects to data availability and statistical significance, are based on empirical modelling of the relationship between the reform measure and total factor productivity, capital deepening and the employment rate.
Table 1.5. Illustrative GDP impact of selected recommendations
Source: OECD calculations based on Égert (2018).
Structural challenges to get the recovery right
Moving from job retention to job creation and reallocation
As most countries, in response to the COVID-19 crisis Estonia took active measures to preserve jobs and support incomes. It introduced a job retention scheme, covering 70% of usual wages for firms experiencing a temporary decline in business activity under broad eligibility conditions. This welcome scheme was in place intermittently up to May 2021, when it was discontinued. Preliminary evidence on the effectiveness of such schemes during the first six months of the COVID-19 crisis suggests that their role was significant in limiting job losses and averting a surge in unemployment, albeit there are large uncertainties by how much, while evidence about how these schemes could have potentially hampered job creation by locking workers into firms with structural difficulties is more limited (OECD, 2021a). After the pandemic, policies conducive to job creation and reallocation will be central to a strong recovery and to address the long-standing issue of skilled labour scarcity. Moreover, the pandemic has also demonstrated the need for swift labour reallocation while in parallel, the acceleration of structural changes in labour markets, including the green transition (see Chapter 2), requires effective reskilling policies to help businesses, start-ups and workers to create opportunities and facilitate transition to occupations and sectors with higher growth potential.
Strengthen engagement in vocational education and training
Prior to the pandemic, evidence demonstrated that high-performing VET systems provide a very effective means of integrating learners into the labour market and opening pathways for further learning and personal growth (OECD, 2020b). During lockdown measures, the reliance on vital services such as healthcare, many of which relying on vocational education, has shown the importance of vocational education and training (VET). In Estonia, students are little involved in VET, as well as businesses, even when it comes to ICT training (Figure 1.15).
Figure 1.15. Students and businesses are little involved in VET
Note: Panel B data refer to businesses with 10 or more employees that provided any type of training to develop the ICT related skills of their employees within the last 12 months.
Source: OECD Education at a glance 2020; and OECD ICT Access and Usage by Businesses Database.
Increasing the participation of employers and trade unions, at both local and national levels, and combining school-based with work-based programmes, would make VET more attractive to students. This would give students the opportunity to assimilate relevant skills while being in the workplace, in contrast to existing practices that involve little exposure to the workplace (Figure 1.16). It is also crucial that VET programmes provide students with the skills needed for tomorrow. Occupations involving routine tasks are being transformed, restructured or are disappearing entirely due to increasing levels of automation. Hence, VET will need to focus more on those occupations demanding higher levels of autonomy, planning, teamwork, communication and customer service skills that are more able to resist automation.
Figure 1.16. Work experience during vocational education lacks availability
Distribution of 25-34 year-olds with vocational upper secondary or post-secondary non-tertiary education without work experience while studying, 2016
Source: OECD Education at a glance 2020.
Promoting pathways from VET to higher levels of education would also be important to support students in developing skills that provide value in the workplace. Estonian students tend to be relatively less inclined to engage in programmes that can lead to tertiary education (Figure 1.17). Moving between programme types, including into higher education, would signal that VET programmes can open the door to further learning and self-development. Providing prospects for higher education would also encourage vocational students to complete their education. Although the completion rates of students in vocational upper secondary programmes, tend to be lower than in general ones, vocational students are more likely to complete their qualification when the programme provides access to tertiary education than when it does not (OECD, 2020b).
Figure 1.17. The engagement in vocational programmes leading to tertiary education could be improved
Completion rate of vocational programme with access to tertiary education, 2020
Source: OECD EAG 2020.
Advance the use of artificial intelligence to connect people with jobs
Estonia is already equipped with tools to analyse labour market trends, thanks to its jobs and skills forecasting system (Tööjõu-ja oskuste vajaduse seire-ja prognoosisüsteem, OSKA) and through its Unemployment Insurance Fund (Töötukassa), which provide recommendations to stakeholders in education, training and support measures (CEDEFOP, 2020). However, such tools could be further improved (OECD, 2021a). Additional use of Artificial intelligence (AI) can help to connect people with jobs, taking into account the skills of jobseekers. Learning algorithms can also spot emerging patterns and speed up the reabsorption of displaced workers into industries requiring similar skillsets. AI can process large pools of jobseekers more quickly than workers in public employment services.
The tailoring of active labour market policies (ALMPs) could also be improved with AI. The pandemic has increased the number of job seekers in Estonia, and preliminary evidence suggest that their composition has changed, as some groups have been more affected than others (OECD, 2021a). As a result, ALMPs’ traditional profiling tools, either digital or through human assessment, can become less accurate. AI algorithms can allow for rapid and consistent adjustment of individual profiling, meaning that services can be adapted swiftly and with higher accuracy. The planned incorporation of the new 2021 population census in OSKA offers an opportunity to revise and upgrade the algorithms based on a set of fresh data to improve recommendations. Scenario buildings could also be developed to understand and anticipate better structural changes. For example, the transportation and logistics sector in OSKA could take further into account, and elaborate different scenarios on, the impact of driverless cars and smart transportation systems for skill demand.
Improve the match between ALMPs and people with skills obstacles
While participation in active labour market policy programme is high among persons with health obstacles, thanks to past reforms, it is comparatively lower for those with low skills (OECD, 2021b and Figure 1.18). Further efforts should be made to refer and convince the low-skilled to participate in training measures, not least because the potential positive effects of their participation are likely to be higher than for the high-skilled (Leetmaa and al., 2015).
Figure 1.18. ALMPs participation is comparatively low for those with skill obstacles
Share of working-age people with weak attachment to the labour market in Estonia receiving ALMPs by their labour market obstacles, 2017-2019
Source: OECD (2021b).
While Estonia public employment services (EUIF) already use guidelines to detect training needs and encourage low-skilled persons to enrol in training, more can be done to fully implement these guidelines (OECD, 2021b). In addition, training programmes could be adapted to fit better the needs of the low-skilled, minimising discouraging aspects and preventing dropping out. The latter could be done by providing counselling and mentoring during participation in training programme, discussing the challenges met and trying to address them. Furthermore, the outreach strategy to low skilled people should be revised to promote up- and re-skilling directly and personally, as raising general awareness might not be sufficient. Specific training programmes for this group could also be created, as has been done successfully in Germany (Box 1.4).
Box 1.4. Specific training programmes for the low-skilled in Germany.
In 2016, the German Ministry of Employment and Social Affairs and the social partners in the public employment supervisory board jointly initiated the second chance initiative Zukunftsstarter (“future starter”), which aims at increasing qualification levels among young people. In particular, it targets 25- 35 year-olds who have no vocational degree (which is a requirement for working in many jobs, including in many technical, administrative and service occupations), or have been working in a job that does not require such a degree for at least four years. The programme is open to employed people as well as people who do not work and sponsors training, usually for two years, permitting to obtain a vocational degree.
The objective of Zukunftsstarter was to provide training to 120 000 unskilled young adults between 2016 and 2020. Ultimately, this will contribute to alleviating shortages of skilled labour, improving people’s employment prospects and lowering unemployment. A similar second change initiative, Spätzünder gesucht (“late starters wanted”), ran between 2013 and 2015 as Zukunftsstarter’s predecessor. The results of Spätzünder gesucht are encouraging, pointing to the high potential of this type of initiative. Following the start of the programme in 2013, entries into degree-oriented training programmes increased markedly, in total by 19%.
Source: from OECD (2021b)
Table 1.6. Past OECD recommendations on VET and ALMPs
Tackling a persistent poverty challenge
Despite important progress over the last two decades, poverty remains a challenge in Estonia. While overall inequality is below the OECD average and absolute poverty is on the path of complete eradication after years of sustained reduction, relative poverty remains high by OECD and EU standards (Figure 1.19). It is also particularly acute for some specific populations and regions (Figure 1.20). With significant inflationary pressures, this challenge could become more acute. While 2020 figures show a slight decrease by 0.1 points of the relative poverty rate compared to 2019, still one fifth of Estonians are living at risk of poverty, making the government´s goal of reducing this risk to 15% by 2023 (in terms of 60% of median income) particularly important.
Figure 1.19. Severe material deprivation has been continuously reduced but relative poverty stays high
1. Data refer to the total population and are based on equivalised household disposable income, i.e. income after taxes and transfers adjusted for household size. The poverty threshold is set at 60% of median disposable income in each country. The income-based poverty rates exclude lump-sum payments, which are frequent in the retirement schemes of some countries (e.g. Australia, Switzerland).
2. The absolute poverty rate is the share of persons with an equivalised annual disposable income lower than the national definition of the absolute poverty threshold.
Source: OECD Income Distribution Database (IDD), http://stats.oecd.org/Index.aspx?DataSetCode=IDD; and Statistics Estonia.
Figure 1.20. Poverty prevalence is especially high for some populations and regions
Source : Statistics Estonia, Estonian Social Survey (EU SILC) 2021.
Transfer targeting and in-work benefits
The transfer system could be made more effective in reducing poverty (Figure 1.21). In particular, social transfers could be better targeted towards people living in poverty. Before the pandemic, almost similar amounts of cash transfers were received by the 20% richest individuals and by the 20% poorest (Figure 1.22, Panel A). Moreover, guaranteed minimum-income benefits (including housing assistance) cover only half of the relative poverty threshold (Figure 1.22, Panel B). This coverage is the highest among Baltic states, but is lagging behind some Nordic countries such as Finland. Increasing further the rate of subsistence benefits would thus help to reduce poverty. Estonia could also make better use of its advance in digital technology to increase the take-up of benefits (Võrk and al., 2016). This could involve, for example, creating a unique identifier linked to a financial account to supplement the normal social program registers.
Figure 1.21. There is room to increase the effectiveness of the transfers system for the poorest
Reduction in relative poverty by the transfers system Percentage points reduction in relative poverty at 60% of median income, 2020 or latest
Source: OECD Income Distribution Database.
Figure 1.22. Cash transfers are not tightly targeted and lack coverage
Source: Calculations based on OECD Income Distribution Database; and OECD Tax-Benefit Models, via http://www.oecd.org/els/soc/benefits-and-wages.htm
Increasing benefit generosity might lower work incentives of low-wage earners. Social reforms should therefore seek to encourage the labour-market reintegration of beneficiaries. Moreover, in the current context of labour shortages, in-work benefits would tend to increase labour supply (OECD, 2005). To do so, in-work benefits, which currently do not exist in Estonia, could be introduced. Making work pay more would be particularly important in Estonia, where working does not necessarily generate above-poverty income. One out of ten persons in employment (waged or self-employed people) lives in a household whose total equivalised disposable income is under the poverty line, a rate higher than the EU average and the highest in the Baltics (Figure 1.23).
Figure 1.23. In-work poverty is high
In-work poverty rates for employed persons aged 18 or above (2020 or latest year available)
Source: Eurostat.
Tackling old-age poverty
Estonia has a three-pillar pension system: a mandatory state pension as the first pillar, mandatory private accounts as the second pillar and voluntary savings accounts as the third pillar. The levels of pensions are indexed annually to consumer prices (with a weight of 20%) and to the increase of the pension part of the social tax revenues (with a weight of 80%). Additionally, the Pension Insurance Act requires the government to evaluate the impact of pension increases on financial and social sustainability, and propose changes to the indexation if necessary. Under this functioning, the pension system is sustainable. However, and as highlighted in previous Surveys (OECD, 2019c), old age poverty remains particularly acute in Estonia, questioning pension adequacy. In 2020, 40% of the persons aged 65 or over were at risk of poverty. While annual increases are guaranteed, wages and salaries grew faster, leading to a continuous increase in the poverty rate among the elderly over the last decade: in 2011, one-fourth of pensioners were at risk of poverty.
The recent reform, allowing the withdrawal of funds from the second pension pillar, is likely to aggravate this trend, by diminishing further pension adequacy. The foreseen basic pension increase in 2023, which is expected to reduce old age poverty by 1.6 points according to the authorities, is welcome in this regard. However, by establishing the average pension at 654 euros in 2023, and given the rapid wage growth in the economy envisaged over the next two years, old age poverty will remain problematic. Increasing basic pension further, at a pace closer to curent economic developments, would contribute to put old age poverty on a downward trend.
Further discussion about the minimum wage among social partners
In the past, Estonia has tended to favour wage-setting policy, with the emphasis on minimum wages, to tackle poverty. With a low coverage of collective bargaining (3.9% of companies and 19% of employees were covered in 2015), and where only two sectors (healthcare and transports) have minimum wage agreements in place, the main wage-setting policy in Estonia is the statutory national-level minimum wage, which is agreed by the social partners and enforced by government decree. According to the authorities, the labour market has managed to absorb increases in the minimum wage for many years and low wage employees were mostly able to find employment. Recent empirical evidence also shows that minimum wage appears to be effective in lowering wage inequality and poverty in Estonia (Ferraro and al., 2016).
With the minimum wage at 584 euros in 2020 and with almost 16% of all workers receiving it or less, social partners can play a key role to keep all workers out of poverty. The recent agreement between unions and the Employers’ Confederations to raise the minimum wage by 12% in 2022, to 654 euros, is welcome in this regard after last year during which it has not been re-evaluated. Social partners should continue discussions to keep minimum wage increases at least in line with national average wage growth, and eventually raise it above the poverty line, as it currently remains 10% lower (Figure 1.24).
Figure 1.24. The minimum wage is lower than in other countries
Source: OECD Employment and Labour Market Statistics; and Statistics Estonia.
Raising the minimum wage further would also entail the additional advantages of strengthening mechanically social protection, as in Estonia the minimum wage is linked to different social transfer payments. For example, the minimum wage is the basis for: the minimum sum of unemployment insurance (50% of the minimum wage in the previous year); the rate of parental benefit when there are no earnings from the previous year (100% of the minimum wage in the previous year); the minimum sum of parental benefit when previous earnings are lower than the national minimum wage; and the minimum sum of sickness or care benefit. However, the minimum wage enters into the calculation of the minimum social tax contributions paid by employers, and thus a raise in the former could lead to an increase in the latter, while the labour tax wedge is already high in international standards for workers with low earnings (Figure 1.25). The 2018 reform has been effective at lowering its level but kept it nonetheless above the OECD average. Reducing the social security contributions of low wage earners while keeping their level of benefits entitlement, especially employees’ contributions which account for most of the wedge, would contribute to reduce poverty by strengthening employment and support take-home pay directly, notably for young workers.
Figure 1.25. The tax wedge for low wage earners could be further reduced
Labour tax wedge, % of total labour compensation, single person at 67% of average earnings without children, 2021
Source: OECD Taxing Wages - Comparative tables.
It must be noted that introducing in-work benefits and raising the minimum wage further are two policy instruments that are not substitutes but that could in fact complement each other. For instance, an in-work benefit that would top up wages in such a way as to make net wages equal to the minimum wage is not feasible, as it could be partially captured by the employers to reduce their labour costs and delivering less than otherwise expected gains for the employees. However, the financing of these income gains for the poorest differs between the two scenarios. Introducing in-work benefits could be financed by strengthening the tax base (see below).
Close the gender pay-gap
Another wage-setting mechanism to tackle poverty and promote fairness is equal pay regulation. In Estonia, the basis for ensuring equal pay for different groups are the Gender Equality Act and the Equal Treatment Act, in force respectively since 2004 and 2009. However, some pay differences remain and women are particularly affected. While Estonian women have high employment rates and outperform men in the education system (OECD, 2017d), the gender pay gap is high (Figure 1.26). Progress was made in recent years, with the gender wage gap declining from 21% in 2017 to 15% in 2021 (preliminary data), Nonetheless, Estonia’s wage gap remains one of the highest in the OECD. Pay differences is directly related to poverty (McKnight and al., 2016), and the gender pay gap has consequences for some specific groups such as one-parent families, a group often at higher risk of poverty and where women are overly represented (Rense, 2020).
Figure 1.26. The gender pay-gap remains high
2020 or latest available
Note: The gender wage gap is defined as the difference between male and female median wages divided by the male median wage.
Source: OECD Employment and Labour Market Statistics.
Closing the gender pay gap is a priority in the Government`s Action Plan for 2021-2023. This includes new amendments to the Gender Equality Act, to be presented in early 2022, after the ones proposed in 2019 were dropped due to the expiry of the Parliament’s mandate. The reform intended to provide the Labour Inspectorate with a right to exercise administrative and state supervision over the implementation of equal pay for women and men in the public sector organisations. It also sought to specify some already existing requirements regarding data collection and the information provided to employees or their representatives.
Those were steps in the right direction that should be reintroduced and enhanced. For instance, in those past amendments private sector companies would have been inspected only by volunteering. Involving employee representatives in the private sector would help in preventing pay discrimination. In Lithuania for example, at the request of employee representatives employers with more than 20 employees are required since 2017 to provide average wages by professional groups and gender at least once a year. Providing average wages would not conflict with the rule of wage confidentiality in the private sector, which is important in Estonia. More generally, the combination of legislative, administrative and awareness-raising measures is crucial in continuing to tackling persistent discrimination against women in wage setting, and reducing poverty.
Enhancing the labour-market integration of ethnic non-Estonians through targeted language support
Other groups in Estonia are also facing labour market challenges, and in particular ethnic non-Estonians. Their integration is still lagging behind, both in terms of employment and wages. These differences lead to a higher prevalence of poverty among ethnic non-Estonians: 25.3% versus 20.7% for the whole population in 2020. As many of the ethnic non-Estonians reside in the north eastern region dominated by large industrial enterprises, higher unemployment and poverty rates partly reflect weaknesses in regional economic policy. Language skills are also a key factor. As the preservation of the Estonian language, spoken by only one million individuals in the world, is of critical importance, efforts should be pursued in the availability of language training availability for all residents, but also in fostering the engagement of ethnic non-Estonians in training, in particular to facilitate the integration of Ukrainian refugees into the labour market.
Table 1.7. Past OECD recommendations on poverty and inclusiveness
Modernizing the tax system without sacrificing efficiency
Estonia’s tax system has many strengths, including easy compliance and simple administration. Both business and individual taxpayers widely use the possibility to declare taxes on-line via the e-Tax Board. A simple tax code makes it easy to state the taxable income: corporate profits are taxed only when they are distributed as dividends, and personal income is taxed at a flat single rate aligned with the corporate income tax rate, thus pre-empting the possibility of arbitrage.
Options for tax reform
Total tax revenue is close to the OECD average (Figure 1.27, Panel A). Estonia collects almost half of its total tax revenues from goods and services (Figure 1.27, Panel B), the highest share in the OECD after Chile. By contrast, income taxes (personal + corporate) are among the lowest in the OECD, both as a share of GDP and as a share of total revenues. This reflects the minimal progressivity of the personal income tax and the taxation of corporate income only when it is distributed (Box 1.5). Property taxes contribute very little to revenues, as only land (not buildings) is taxed. This section discusses various options for tax reform, which either could have a neutral impact on government revenue (i.e. limited to changes in the tax mix) or increase total government revenue. Tax reform should wait until the sanitary situation is fully restored and the economic situation has normalized, but analysis and debate can proceed immediately.
One goal of such tax reform could be to repay the public debt accumulated during the pandemic to finance health measures and protect workers and businesses, rather than leave this debt to future generations. Tax reform could also help the country deal with future challenges, such as financing the large investments required to mitigate climate change, prepare for demographic ageing, reduce poverty and make economic growth more inclusive. In the post-pandemic context, a more than proportional contribution of high-income taxpayers to these various objectives could be well justified.
Figure 1.27. Tax revenues are tilted toward indirect taxation
Source: OECD Tax revenues statistics.
Box 1.5. Specificities of Estonia’s corporate tax system
In 2000, Estonia established a unique system, later adopted by Latvia, of corporate taxation, by abolishing its conventional corporate income tax and replacing it by a tax on corporate distributions. This distribution tax is levied on dividends and on certain other corporate expenses, which could be seen as hidden dividends (e.g. fringe benefits not subject to personal income tax, loans to participators). The distribution tax is payable by all Estonian resident companies and by Estonian permanent establishments of foreign companies. The main tax rate on distributions is currently 20% of the gross distribution. In addition, in 2018 a lower rate of 14% has been introduced, with the goal of rewarding companies with a more regular dividend distribution and to establish a countermeasure to prevent one of the most common indirect outflows of profits from the company, i.e. lending to the parent company or another company in the group.
The Estonian distribution tax comes close to a source-based cash flow tax, which is the most optimal corporate tax setting according to the economic theory literature (Meade, 1978). In fact, and if the Estonian distribution tax had allowed a deduction for the firm’s revenue from new share issues, making it a tax on net distributions, it would be optimal in the sense that it would not induce distortions regarding the firm’s investment and financing distribution. Stated otherwise, the debt-equity choice becomes unrelated to the tax environment. For mature corporations, which can satisfy all of their need for equity through retained earnings, a dividend tax like the Estonian distribution tax is optimal.
However, having a zero corporate tax rate on retained profits does not make Estonia, which has joined at the end of 2021 the OECD-G20 two-pillar solution to address the tax challenges of the digital economy, a low-tax jurisdiction. In fact, while such system reduces the cost of capital when investment is financed out of retained profits, a distribution tax does increase the entry cost of capital when foreign direct investors consider establishing themselves in Estonia. According to some estimates, the marginal cost of capital for direct investment into Estonia appears in fact to be roughly at the same level as the cost of capital for domestic investment in Finland, which has a more traditional corporate tax system (Kari and Ylä-Liedenpohja, 2005). However, because it exempts retained earnings, the base of the Estonian distribution tax is narrower than that of a conventional corporate income tax.
Property taxation
Property taxation is deemed the least distortionary among all fiscal revenues (Arnold et al., 2011). As recommended in previous Surveys (OECD, 2019c and 2017c), increasing property taxes would not only bring additional, non-distortive revenues, but would also remove existing inefficiencies, as it contributed to the previous housing boom, and would also help to ease the current mounting tension on the housing market. The additional revenues generated could also contribute to the renovation and insulation of buildings, where the needs are significant in Estonia (see Chapter 2). The new land evaluation, to be carried out in 2022, represents an opportunity to also evaluate the housing stocks and business properties and then expand the property tax base beyond land.
Personal income tax progressivity
Personal income tax in Estonia is often viewed as being a flat tax but, due to basic exemptions and allowances, it exhibits a progressive schedule, with two tax brackets at 0% and 20%. In fact, the 2018 reform, which increased the basic income tax allowance from EUR 170 to 500 per month, and the abolition for 2022 of the income tax exemption for mortgage payments interests, both amplified the progressivity of the income tax. Nonetheless, progressivity remains very low by international standards (Figure 1.28). Moreover, progressivity is hampered by a specific provision of the personal income tax code, the so-called “sliding scale”, which makes the basic income tax allowance decreasing linearly in the 1200-2100 euros income range. This rule triggers local regressivity of the personal income tax, where incomes in this range are marginally taxed at a rate above 20% while income above this range are taxed at the 20% flat rate. The sliding scale should instead be maintained fixed as income increases, as is the case in other countries such as France and Spain, to fully align the aim of the tax schedule with effective marginal rates and restoring progressivity along the income scale.
Figure 1.28. Income tax progressivity is low
Ratio between the top marginal tax rates and the first non-zero rate, 2021
Source: OECD Tax database.
The personal income tax schedule in Estonia has undoubtedly played a role in propelling the Estonian economy after the independence with its simplicity, which reduced tax evasion and improved economic efficiency through lower tax distortions (Barrios and al., 2020), and the degree of progressivity of a personal income tax, and thus of redistribution, remains first and foremost a question of social preferences (Limberg, 2020). However, the introduction of a third tax bracket for top incomes would have the potential to generate large tax revenues – for instance with a tax rate of 25% slightly above the flat rate of 20% (Box 1.2). Albeit income concentration at the top is relatively contained in Estonia, the richest 10% of households own 35% of the national income before taxes (Figure 1.29). The creation of a third tax bracket would make richer earners contribute somewhat more to the various challenges faced by the country. In terms of efficiency, recent evidence shows that progressivity tends to induce small efficiency costs (Gerber and al., 2018). Moreover, it could strengthen the counter-cyclicality of fiscal policy, providing a buffer against output volatility: by moving taxpayers into lower income brackets and reducing their average tax rate during recession, while doing the reverse during expansions, disposable income is made more stable and thus consumption and investment less volatile (Alessandrini, 2021).
Figure 1.29. Income concentration at the top is contained
Pre-tax national income share held by the richest 10% of the population, 2020
Source: World Inequality Database, https://wid.world/fr/country/estonie/.
Evaluate and monitor the recent corporate tax reform
The corporate income tax rate cut of 2018 from 20% to 14% for companies that pay dividends for three consecutive years has, as anticipated, increased revenue in the short-term, by encouraging the companies to distribute profits. Two years after the reform, dividends under the 14% regime have soared by 77% in 2020 and 40% during the first half of 2021 (for all companies excluding banks and state-owned companies). However, by freeing-up past retained profits such boost will be temporary by nature, and in the long-run corporate tax revenues will decrease below their pre-reform level. With an already favourable system for investment and entrepreneurship, and comparatively low corporate tax-revenues, the cost and benefits of the lower tax rate should be evaluated and monitored. In particular, according to the authorities this new tax regime has added complexity to an otherwise simple and efficient system. Moreover, this lower tax rate conflicts with the OECD-G20 two-pillar solution, as it includes a 15% global minimum corporate income tax rate to which 340 foreign subsidiaries and four multinational Estonian firms are foreseen to be subjected to by the tax authorities.
Moreover, because the corporate tax in Estonia is paid at the point of distribution and is therefore similar to a final withholding tax, no further tax is payable at the shareholder level on distributed income, except in the case of private shareholders where an additional 7% income tax is applied. This new lower tax regime could thus potentially widen an already large gap on overall dividends taxation, where Estonia stands at almost 22 percentage points below the OECD average (Figure 1.31).
Figure 1.30. The taxation of dividends is very low
Overall tax rate on dividends (personal plus corporate income taxes), 2022
Source: OECD Tax database Table II4.
The alignment of the personal income tax rate with the corporate income tax avoids arbitrage, discarding the possibility for individuals to requalify their labour income into capital income and vice versa. That being said, strengthening the progressivity of the personal income tax may induce some distortions for individuals eligible to a new top personal income tax bracket. For example, after the 1993 Finnish tax reform, which established a dual income tax with a lower rate on capital income, there were significant shifts of labour income to capital income among the self-employed (Pirttilä and Selin, 2011). However, in Finland the taxation of dividends and the top income tax rate remain far apart, with Finland having one of the highest top personal income tax rate in the world (Kaisa and Matikka, 2017). Thus, those insights may not be directly applicable to Estonia. Moreover, the shift from personal to corporate income tax in the presence of differential in rates can be mitigated by limiting the opportunities for tax shifting if the costs of combatting it are not too high (Piketty and al., 2014), notably by creating a strict nomenclature that distinguishes between capital and labour income.
Raising health spending efficiency
Despite its strong fiscal position, Estonia needs nonetheless to conduct a prudent budget policy to be ready for future challenges. Spending pressures are mounting because of population ageing, putting additional stress on a healthcare system already stretched by the pandemic. Raising the efficiency of the health system is therefore crucial.
Better health outcomes can be achieved with similar expenditures
Overall, the health status of Estonians is converging to the OECD average. Although there is room for improvement, many health outcomes have improved significantly over the last decades. Gains in life expectancy have been almost twice the OECD average over the last fifteen years while infant mortality has been cut by more than half over the same period (Figure 1.31, Panel A and C). Healthy life years, which indicate the share of remaining years free of disability, have also increased, leading Estonians to live longer but also healthier (Figure 1.31, Panel B). However, Estonia is still lagging behind in amenable mortality (Figure 1.31, Panel D).
Figure 1.31. Health outcomes continue to improve
Source: Worldbank WDI database; and Eurostat.
Although Estonia’s spending on health per capita is relatively low in international comparison, it is among the highest compared to the Baltic and CEE economies. However, this does not necessarily translate into significantly better health outcomes (Figure 1.32). Some OECD countries such as Turkey, Costa Rica, Greece or Chile exhibit higher life expectancy than Estonia, despite similar or lower health spending per capita. While such figures do not control for additional differences in life-style and social factors, which undoubtedly play an important role (Giorno and Londáková, 2017), this suggests nonetheless that at current spending levels, Estonia health care system has room to increase efficiency. OECD evidence suggests that there is no single best health system: each model has its strengths and weaknesses (Joumard and al., 2010). Estonia’s model is characterized by a mandatory public system of healthcare, with almost universal coverage and a single insurer. When reforming the system after independence in the 1990’s, a multi-insurer model was ruled out on the grounds that given Estonia’s small population, effective competition between insurers would be impossible. General health insurance covers about 95% of the population, 45% of whom are non-contributors (i.e. not working). The state makes contributions on behalf of some of them, notably the unemployed and parents on parental leave. Overall, the Estonian system performs well on most efficiency metrics (Habicht and al., 2018), but some pockets of inefficiency remain.
Figure 1.32. There is room to improve efficiency in health care
Source: OECD Health expenditure and financing database; and Worldbank WDI database.
Promote the use of generic drugs and biosimilars
Pharmaceuticals currently fail to provide good value as the penetration of generic drugs remains low (Figure 1.33). There is no mandatory generic substitution in pharmacies in Estonia. The regulations stipulate that doctors prescribe pharmaceuticals by their International Non-proprietary Name (INN). If prescribing by brand name, the doctor has to justify this in the patient’s medical record (e.g. the patient refuses generic, or the cheapest option is not available). Customer awareness on generic drugs needs to be strengthened, to further encourage their take-up, notably by providing better information regarding the equal quality of the original product and the substitute. Generics should also be made the standard for reimbursement of every prescription, with their substitution mandatory. Moreover, while recent the promotion of biosimilar treatments could be explored, as significant savings for the health care system could be achieved (Box 1.6).
Figure 1.33. The share of generics in the total pharmaceutical market is low
Share of generics in the total pharmaceutical market, 2020 (or nearest year)
Source: OECD Health statistics.
Box 1.6. Current and future savings from the use of biosimilars
In parallel with generic drugs, biosimilar treatments, which are biologic treatments designed to work in almost the same way as treatments already approved by sanitary authorities, could realise significant savings for health care systems. For example, between 2016 and 2020 eight key biologics have lost patent protection. Analysis of data available for five European countries (France, Germany, Italy, Spain and the United Kingdom) and the United States suggests that a 20% reduction in price per treatment-day across these eight products could result in cumulative savings exceeding EUR 50 billion in 2020 (IMS Institute for Healthcare Informatics, 2016). In 2015, following the introduction of biosimilar competition in one of the most often used classes of biologics – erythropoietins (EPOs) – the observed price reduction varied from 39% in France to 55% in Germany (IMS Institute for Healthcare Informatics, 2016).
The European Union approved the first biosimilar in 2006 and is the leader in the number of approved products, with 79 as of 2021. Yet biosimilars’ use shows wide variation in the European Union. Even the first biosimilar still has little or no uptake in some countries (e.g. Greece, Ireland and the Slovak Republic), while in Poland it is used in almost all relevant therapies (Ekman and Cornes, 2016). The United States adopted the legislative framework for licensing biosimilars in 2010, but the first biosimilar was approved only in March 2015 (Belloni et al., 2016).
Policies to increase uptake of generics can also be applied to biosimilars. For example, physicians and patients often worry that biosimilars will compromise quality of treatment (IMS Institute for Healthcare Informatics, 2016). Thus, regulators should communicate their knowledge more actively and, most importantly, strive to take clear positions on interchangeability between biologics and biosimilars. In Norway and Denmark, where physicians are at the heart of decision making, uptake of biosimilars was rapid and sustained. Similarly, biosimilar competition is strong in Germany, where insurance funds invested in communication with physicians on the subject and subsequently introduced prescribing quotas for biosimilars (IMS Institute for Healthcare Informatics, 2016).
Source: OECD (2017), Tackling Wasteful Spending on Health.
Expand the number of generalists
In terms of human resources, pressure is mounting on the health workforce, with consequences on the quality and efficiency of care provided (OECD, 2018a). As in other Baltics countries, Estonian health workers are ageing: almost 50% of doctors are older than 55 years (Figure 1.34, Panel A), and among family doctors this share is close to 60% (NIHD, 2018). Furthermore, in 2015, 24% of physicians and 16% of nurses working in health institutions were older than 65 years, i.e. past retirement age (NIHD, 2018). Moreover, the lack of foreign trained doctors is also affecting the number of practitioners (Figure 1.34, Panel B). Combined with the unbalanced mix of doctors providing basic and special primary care, with significantly fewer generalists than specialists (23% of doctors are generalists in Estonia compared with 30% in the OECD on average, (OECD, 2019)), ageing is creating barriers to the performance of the health system.
Figure 1.34. The health workforce is under ageing pressure
Source: OECD Health statistics.
Generalists plays a key role for a well-performing health system. The ageing of general practitioners from an already low share can be an issue for life-long training and the skill mix. As older general practitioners are less likely to undertake retraining to update their practices, they are less trusted to play a gate-keeping role, which will bear on efficiency (Moore and al., 2009). Albeit few specialist organizations (e.g. the professional associations of family medicine, cardiologists and surgeons) have instituted systems for regular recertification, for which the health care professionals must undergo continuous medical education and present proof of professional activities performed, there is no statutory relicensing or reaccreditation for general practitioners in Estonia, as is required by medical regulation in the United States and which turns effective in maintaining clinical knowledge (Vandergrift and al., 2017). This should be implemented. There is also a need to retain doctors as well as to create incentives to attract more Estonian students to enrol in health faculties, and also doctors from abroad, to ensure sufficient supply in the medium to long-term. Additionally, although the admission quota to the medical faculty has been gradually increased, from 100 in 2000 to 185 in 2018, it still falls below the level that would cover the estimated future needs, which the Ministry of Social Affairs estimates at 200 admissions (Habicht and al., 2018), and should then be adjusted consequently.
Streamline the secondary use of health data
As a front-runner in digital technologies, Estonia has advanced in digital innovation in health and provides many health tools and services online. These include electronic health records, national image archiving, ePrescriptions, eReferrals, eAmbulance and eConsultations. Most recently, Estonia has introduced a central system for booking appointments, beginning in specialist ambulatory care but with plans to expand to GPs and dental care. The current data exchange system allows linkages between various eServices in both the public and private sectors, and more than 96 % of Estonians hold an ID card that allows for digital authentication and access to eServices via the Internet. Additionally, Estonia has used electronic billing data since the late 1990s, ahead of many other EU Member States. However, and despite those remarkable achievements, Estonia reports as lagging behind in term of preparedness for secondary use of health data (Figure 1.35, Panel A). Technical readiness is high but there are needs in data governance and capacity to put data to work (Figure 1.35, Panel B).
Figure 1.35. Data governance is hampering the secondary use of health data
Source: Oderkirk, J. (2017), "Readiness of electronic health record systems to contribute to national health information and research", OECD Health Working Papers, No. 99, OECD Publishing, Paris, https://doi.org/10.1787/9e296bf3-en; and THE OECD 2019 OPEN USEFUL REUSABLE DATA (OURDATA) INDEX.
Secondary use of health data has the potential to remodel services around patient needs, which could be conducive to sizeable gains for the productivity of the health care system and health outcomes in general (Meystre and al., 2017). Secondary use is generally cheap, so greater use of existing data to generate knowledge and improve services will often turn to be a highly cost-effective way of improving health outcomes (OECD, 2019b). This may not necessarily imply cost savings for any country, as the use of data and ICT can, for example, uncover unmet need and make new models of care delivery and digital services more accessible, increasing demand and thus aggregate expenditure. However, in the case of Estonia this is unlikely to be the case given its already advanced capacity in digital services. As a result, secondary use of data should be stepped-up, by implementing a legal framework that enables these data to be securely extracted and used for secondary purposes, while tackling technical challenges by encouraging common approaches to data terminology and exchange standards, as implemented recently in Sweden (Box 1.7). Some recent and on-going initiatives, such as the New Generation Health Information System Project, which aims at implementing new principles of health data governance and processing, go in the right direction and should be pursued further.
Box 1.7. Recent advances on health data interoperability in Sweden
In Sweden, successive digital health strategies have been in place since 2006 (Swedish Ministry of Health and Social Affairs, SALAR, 2016). The latest strategy was jointly developed by the national government and the Swedish Association of Local Authorities and Regions (SALAR), and was endorsed in 2016 to guide the digital transformation through to 2025. The goal of the strategy is to “make it easier for people to achieve good and equal health and welfare, and to develop and strengthen their own resources for increased independence and participation in the life of society” (SALAR, 2016). The Swedish strategy comprises three pillars:
Increasing digital information exchange, both between different public authorities and with citizens, while safeguarding privacy and data security
Advancing semantic interoperability of data in the health system
Ensuring technical interoperability of ICT systems.
This new strategy emphasizes the use of standards for health data to ensure that codes, concepts, terms and structures used are valid and usable in the work of responsible entities to enable the exchange of information that is needed to guarantee quality and security. The goal is to structure data in such a way that the services can analyse and draw conclusions about the outcomes of actions over time, comparing different actors and different processes and forms of treatment. The Government also works to provide the responsible entities with national support. Municipalities and county councils cooperate among themselves and with relevant central government actors in implementing common concepts, terms and classifications or structures in their services.
Currently, Sweden reports near-universal use of electronic medical records in primary care and hospitals, the second highest rate of availability and linkage of key health datasets for secondary use and the use of routine data for monitoring medicine use and expenditure (OECD, 2019b).
Promote healthy lifestyles
Finally, health promotion and disease prevention have a major role to play in reducing avoidable health care costs while delivering better health outcomes. Estimates show that almost half of all deaths in Estonia (above the EU average of 39 %), can be attributed to behavioural risk factors, including dietary risks, tobacco smoking, alcohol consumption and low physical activity (OECD/European Observatory on Health Systems and Policies, 2019). Estonia has implemented several policies to reduce smoking, contributing to a decline in the prevalence of this risky behaviour. In addition to increased taxes on cigarettes, Estonia has adopted advertising bans for tobacco and alcohol products, including banning advertisements at point of sale. Prevention programmes and investment in treatments and counselling services were also stepped-up. Reducing alcohol consumption has been a public priority over the last decade, albeit the former government reduced excise taxes on alcohol from July 2019. Moreover, despite climbing overweight and obesity rates, no significant action has been taken on the issue (OECD/European Observatory on Health Systems and Policies, 2019). The renewal of the Public health Act, planned for 2022, offers an opportunity to strengthen the legal basis for health promotion and protection, while existing services should be complemented and made further available.
Delivering on physical and digital infrastructures
For a small and open economy like Estonia, high-quality infrastructures are a key factor underpinning long-term growth. Infrastructures have been upgraded significantly over the last two decades, benefiting from large amounts of public investment and EU funds. However, the World Economic Forum Global Competitiveness Index shows that the quality of infrastructure is perceived as low, across CEE and OECD countries (Figure 1.36, Panel A). This can be partly explained by road and railroad quality (Figure 1.36, Panel B and OECD, 2019c).
Figure 1.36. The perceived quality of infrastructure remains low
Source: Worldbank Global competitiveness index 4.0 dataset.
In the past, evidence have shown that some EU-funded projects have been designed too large and have not fully delivered (National Audit Office of Estonia, 2013). Now, in the wake of the recovery, and as a substantial part of the new EU-funds will be devoted to continuing upgrading infrastructures, improving the selection, monitoring and decision making on how infrastructure projects are selected, as well as establishing tools to assess the value-for-money and socio-economic impacts of planned investment (e.g. on regional development, environment and safety), will be key.
Raise the efficiency of infrastructure selection and management
Currently, Estonia lacks a formal policy to ensure project selection and an assessment of the performance of each project (OECD, 2017c). Implementing such policy, based on systematic cost-benefit analysis, would help to ensure efficient selection, value-for-money, and to manage risks throughout the operational phase of the projects. Similar shortfalls have been identified recently by the International Monetary Fund, specifically on project selection, which has rated the effectiveness of the selection in Estonia as low (IMF, 2019). Two of the concerns it identifies are: i) the lack of a “comprehensive pipeline” from which projects can be selected, based on consistent, published criteria; and ii) the lack of review of nationally funded projects by an independent agency. The IMF rated these as high priority areas for reform, together with the need to adopt a standard appraisal methodology. Moreover, in contrast to European Union-funded projects, appraisals of nationally funded projects are not published, impeding scrutiny by stakeholders and thus reducing accountability.
Expand data-driven decision-making and go digital on infrastructures
Sound infrastructure policy should be based on data. Putting in place systems that ensure a systematic collection of relevant data and institutional responsibility for analysis, which are currently lacking in Estonia (OECD, 2017c), could contribute to evaluate and monitor the projects’ performance and to base future decision and delivery modalities and contracts on comparable data and information. Establishing a central, systematic and formal collection of information on financial and non-financial performance of infrastructure, that makes it possible to compare various forms of infrastructure delivery models, would contribute to deliver value-for-money for the forthcoming infrastructure projects in Estonia and support the recommendation-making process of an independent infrastructure agency. The information collected should also be disclosed to the public in an accessible format and in a timely fashion, to raise transparency and confidence among the stakeholders.
Beyond physical infrastructures, digital infrastructures are now an essential pillar of any comprehensive infrastructure strategy, a requirement that the pandemic has exacerbated. Estonia has solid and stable digital infrastructures, notably physical access to high-speed broadband. Final deployment are currently completed for middle-mile network of fibre-optic cables, to cover 98% of all residential buildings, companies and public buildings, as well as last-mile network for 100 000 additional addresses. However, those coverage efforts should be complemented by the strengthening of broadband speed, currently lagging behind (Figure 1.37). While the broadband network weathered well the pandemic, coping with higher-than-usual demand without important degradation in performance, the need for high-speed connection has turned crucial. Investment should be increased on broadband speed, with the goal of reaching the EU target of the gigabit society, where at least 100 Mbps speed should be made available to all Estonians. Along the same line, after some delays the awards of 5G bands should be expedited to meet the country’s ambitious goal to deploy 5G in major cities by 2023 and along transport corridors by 2025.
Figure 1.37. The download speed of fixed broadband is lagging behind
Speedtest, 2021
Sources: Speedtest (Ookla), https://www.speedtest.net/global-index.
The future success of service delivery and value-for-money from infrastructures depends on physical ones becoming increasingly interlinked with and operated using digital technologies. The ability to interlink such systems offers opportunities for improving existing services and creating completely new types of services. For example, the integration of physical–digital systems shows the ways that smart cities of the future will make use of advanced technologies within and between their various component systems to improve their effectiveness and efficiency (OECD, 2020a). For logistics, where the perceived quality is low in Estonia (Figure 1.38), the recent experiment conducted jointly by Estonia, Latvia, Lithuania and Poland on electronic consignment notes (e-CMR), a digital technology to ease logistics, is a case in point on digital technologies becoming part of a physical process, and shows how some weaknesses in the latter can be solved by the former (Box 1.8).
Figure 1.38. The quality of logistics infrastructure is lagging behind
Overall logistics performance index, 2018
Source: World Bank WDI.
Box 1.8. The 2020 Baltics/Poland electronic consignment notes (e-CMR) experiment
Between the Baltics and Poland, thousands of shipments are made daily, accompanied by information, documentation, licenses, and other information relevant to both cargo owners, transport companies as well as public authorities. Rules for such documentation are covered by various regulations and among the most important ones, bound to the CMR Convention (the United Nations Convention for the carriage of goods). The CMR document deriving from that Convention is mandatory for international shipments and holds key information about the goods, the transporting and receiving parties. Until recently, the CMR notes were only used in a form of paper, making it time-consuming as well as costly to process them. Moreover, during the pandemic CMR notes were implying unnecessary human contacts in cargo collections and delivery activities.
In 2020 Estonia, together with Latvia, Lithuania and Poland, tested an electronic consignment note (e- CMR) prototype in order to achieve cross-border exchange of e-CMR information by setting up an indexing scheme between countries. The countries used real-time freight information to reduce paperwork in the road transport sector and unjustified downtime during the roadside checks. With the prototype, it was no longer necessary to stop the vehicle, as the load can be controlled by just entering the truck and trailer registration number to the system. Tests were successful and the idea of the e-CMR indexing scheme proved its benefits to all parties in the supply chain. Estonia plans to go further together with neighbouring countries by creating a Pan-European interoperable system for the exchange of electronic freight information, a step towards an EU Digital Single Market.
This type of solution has the potential to significantly improve the performance and efficiency of the transport sector and infrastructures. As border handling remains a significant bottleneck for import and export processing times in several countries (World Bank, 2020), digital solutions like e-CMR eliminates middlemen and middleware and paves new ways of handling tangibles commodities and managing physical infrastructures.
Continuing the progress on corruption and money laundering
Corruption harms the business climate, distorts competition and diverts public resources, and generates mistrust in institutions and corrodes the social fabric. Estonia ranks the highest in the Baltics in the Transparency International’s Corruption Perceptions Index (Figure 1.39, Panel A and Transparency International, 2020), and the effective control of corruption (Figure 1.39, Panel B). In fact, Estonia has improved its ranking considerably since the early 2000s (Figure 1.39, Panel C), and scores at or above the OECD average for each sector-based sub-components (Figure 1.39, Panel D). Estonia is also fully compliant with the standards set by the Global Forum on Transparency and Exchange of Information for Tax Purposes (OECD, 2018b).
Figure 1.39. Perceived corruption is declining
Note: Panel B shows the point estimate and the margin of error. Panel D shows sector-based subcomponents of the “Control of Corruption” indicator by the Varieties of Democracy Project.
Source: Panel A: Transparency International; Panels B & C: World Bank, Worldwide Governance Indicators; Panel D: Varieties of Democracy Project, V-Dem Dataset v11.
In 2013, Estonia set an objective in its 2013-2020 anti-corruption strategy to reach the group of countries with the lowest perceived levels of corruption in the Transparency International’s Corruption Perceptions Index (the value of the index must be higher than 64 points). This objective was achieved and in 2020 Estonia adopted a new anti-corruption plan until 2025. The goals of the previous anti-corruption strategy were to promote awareness of corruption, increase the transparency of decision-making and activities, develop the investigative capacity of investigative bodies, and prevent corruption that threatens security. The objectives of the new plan remain broadly similar, but with now a specific focus on short-term activities for tackling these objectives, including raising the awareness of companies on business corruption, which is currently low and where efforts need to be deployed. Currently, more than half of managers in Estonia have encountered at least one corruption practice within their own business sector, while 48% believe bribes are justified when competitors bribe (Johannsen and al., 2016). Raising awareness and motivating the fair conduct of business should then be a priority, with the State playing a more active role in explaining the nature of private corruption and conflicts of interest on the business environment, and promoting the adoption of specific ethical codes of conduct.
Close some remaining legislative gaps
Estonia has also made further progress on strengthening its anti-money laundering and counter-terrorism financing framework. Those steps were necessary in light of the Danske Bank money laundering scandal and Versobank’s regulatory breaches. Estonia has transposed the fifth anti-money laundering EU directive into national law in 2020. In 2021, an institutional reform established the Financial Inspections Unit (FIU) as an autonomous institution with its own budget, expanded to strengthen its strategic analytical abilities. The FIU also now partakes in the anti-money laundering Baltic Working Group and seeks to enhance its cooperation with Baltic and Nordic colleagues. Overall, all those steps to fight corruption and money-laundering goes in the right direction. However, some of the gaps identified in the previous Survey remain (OECD, 2019c). Estonia’s legislative framework could be improved further, including in clarifying corruption-related offences in the Criminal Code, allowing surveillance activities to counter corruption and adopting legislation guaranteeing confidentiality to private sector whistle-blowers, which is currently not regulated. Also, to meet the requirements of the OECD Anti-Bribery Convention, Estonia should amend its legislation to waive the statute of limitations following a mutual legal assistance request, expand the scope of its false accounting offences and increase the corresponding sanctions.
Use data linkages to automatize inspection
Taking seize of Estonia’s digital capacities, fostering the availability and interoperability of data could make the handling of corruption and money laundering activities more efficient. For example, by linking together data available in different state information systems, while maintaining compliance with privacy regulation, it would be possible to identify dubious transactions (e.g. by linking together data from the Public Procurement Register, the Commercial Register, and the balance record information system to increase the transparency of transactions). In the healthcare sector, the Covid-19 pandemic has created, as in many countries, the opportunity for corruption to flourish as massive resources have been rushed to address the health crisis, while procurement oversight and enforcement efforts have diminished by the constraints of the crisis. Increasing inspection activities in the health sector through new digital tools, such as machine learning, would help to expose corruption. More generally, and given its already existing digital infrastructure, Estonia has the capacity at few costs to develop further digital effective system of auditing and accountability. Such avenues to achieve a more effective grasp on corruption are envisaged in the new anti-corruption plan and should be implemented.
Table 1.8. Past OECD recommendations on money laundering
References
Alessandrini D. (2021), “Progressive Taxation and Economic Stability”, Scandinavian Journal of Economics 123(2), 422–452, 2021.
Arnold J., B. Brys, C. Heady, A. Johansson, C. Schwellnus and L. Vartia (2011), “Tax Policy for Economic Recovery and Growth”, the Economic Journal, 121 (February), F59–F80.
Barrios S., V. Ivaškaitė-Tamošiūnė, A. Maftei, E. Narazani & J. Varga (2020), “Progressive Tax Reforms in Flat Tax Countries”, Eastern European Economics, 58:2, 83-107.
Belloni, A., D. Morgan and V. Paris (2016), “Pharmaceutical Expenditure and Policies: Past Trends and Future Challenges”, OECD Health Working Papers, No. 87, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jm0q1f4cdq7-en.
Cashell, B. (2008), “What is a Recession and Who Decided When It Started?”, CRS Report R40052.
Cedefop (2020), “Strengthening skills anticipation and matching in Estonia: capitalising on OSKA’s potential to realise national ambitions”, Publications Office of the European Union, Luxembourg, http://dx.doi.org/10.2801/12542.
Égert, B. (2018), “The quantification of structural reforms”, OECD Economics Department Working Papers No. 1482. https://doi.org/10.1787/18151973.
Eelarvenoukogu (2017), “The new look of Estonian fiscal policy”, https://eelarvenoukogu.ee/news/the-new-look-of-estonian-fiscal-policy.
Ekman, N. and P. Cornes (2016), “Reducing Health Care Costs and Building Trust in Biosimilar Medicines”, Generics and Biosimilars Initiative Journal, Vol. 5, 17 June, pp. 84-88.
Farber, H. and R. Valletta (2015), “Do Extended Unemployment Benefits Lengthen Unemployment Spells? Evidence from Recent Cycles in the U.S. Labor Market.” The Journal of Human Resources 50 (4): 873–909.
Ferraro, S., Meriküll, J., Stæhr, K. (2016), (Minimum wages and the wage distribution in Estonia”, Working paper series / Eesti Pank. Eesti Pank, Tallinn.
Gerber C., A. Klemm, L. Liu, and V. Mylonas (2018), “Income Tax Progressivity: Trends and Implications”, IMF Working Paper No. 18/246.
Giorno, C. and K. Londáková (2017), “Improving the efficiency and outcomes of the Slovak health-care system”, OECD Economics Department Working Papers, No. 1404, OECD Publishing, Paris, http://dx.doi.org/10.1787/f2b496cd-en.
Guillemette, Y. (2021), “The long game: fiscal outlooks to 2060 underline need for structural reform”, OECD Economics Department Working Paper, forthcoming.
Habicht T, Reinap M, Kasekamp K, Sikkut R, Laura Aaben L, van Ginneken, Estonia: Health system review. Health Systems in Transition, 2018; 20(1): 1 – 193.
IMF (2019), “Republic of Estonia: Public Investment Management Assessment”, Technical Assistance Report, IMF Country Report No. 19/152, International Monetary Fund, Washington, D.C., https://www.imf.org/en/Publications/CR/Issues/2019/06/03/Republic-of-Estonia-Technical-AssistanceReport-Public-Investment-Management-Assessment-46963.
IMS Institute for Healthcare Informatics (2016), “Delivering on the Potential of Biosimilar Medicines: The Role of Functioning Competitive Markets”, March, Parsippany, United States.
Jean, S., O. Causa, M. Jimenez, and I . Wanner (2010), “Migration and labour market outcomes in OECD outcomes”, OECD Journal: Economic Studies, Volume 2010 . http://dx.doi.org/10.1787/9789264281318-en.
Johannsen L., Pedersen K., Vadi M., Reino A., Sööt M. (2016), “Private-to-Private Corruption: a survey on Danish and Estonian business”, Aarhus University, Tartu University and Ministry of Justice of Estonia, https://www.korruptsioon.ee/sites/www.korruptsioon.ee/files/elfinder/dokumendid/private-to-private_corruption_final_report_2.pdf.
Joumard, I., C. André and C. Nicq (2010), “Health Care Systems: Efficiency and Institutions”, OECD Economics Department Working Papers, No. 769, OECD Publishing, Paris, http://dx.doi.org/10.1787/5kmfp51f5f9t-en.
Kaisa K. and T. Matikka (2017), “Revenue-maximizing top earned income tax rate in the presence of income-shifting”, Nordic Tax Journal 2017:1, 100-107.
Kari, S. and J. Ylä-Liedenpohja (2005). “Cost of capital for cross-border investment: The fallacy of Estonia as a tax haven”, Baltic Journal of Economics 5, Autumn/Winter 2004/2005, 28-43.
Leetmaa, R. et al. (2015), Counterfactual Impact Evaluation (CIE) of Estonian Adult Vocational Training Activity, Praxis.
Limberg, J. (2020), “What’s fair? Preferences for tax progressivity in the wake of the financial crisis”, Journal of Public Policy, Volume 40, Issue 2, June 2020, pp. 171 – 193.
McKnight, A., Stewart, K., Himmelweit, S. M., Palillo, M., (2016), Low pay and in-work poverty: preventative measures and preventative approaches. Luxembourg: Publications Office of the European Union.
Meade, J. (1978). The Structure and Reform of Direct Taxation. London: Allen and Unwin.
Meystre, S. et al. (2017), “Clinical Data Reuse or Secondary Use: Current Status and Potential Future Progress I Introduction”, Yearbook of Medical Informatics, pp. 38-52, http://dx.doi.org/10.15265/IY-2017-007.
Moore Jr, D. E., J. S. Green, and H. A. Gallis (2009), “Achieving Desired Results and Improved Outcomes: Integrating Planning and Assessment throughout Learning Activities”, Journal of Continuing Education in the HealthProfessions 29 (1): 1–15.
National Audit Office of Estonia (2013), Overview of the use and preservation of state assets in 2012–2013 - Summary of Problems in the Development and Economy of Estonia by the National Audit Office, Tallinn, 2013.
Nicholson, W. and K. Needels (2011), “The EUC08 Program in Theoretical and Historical Perspective.” Princeton, NJ: Mathematica Policy Research.
NIHD (2018), Health Statistics Database. Tallinn, National Institute for Health Development.
Swedish Ministry of Health and Social Affairs, SALAR (2016), Vision for eHealth 2025.
OECD (2021a), OECD Employment Outlook 2021: Navigating the COVID-19 Crisis and Recovery, OECD Publishing, Paris, https://doi.org/10.1787/5a700c4b-en.
OECD (2021b), Improving the Provision of Active Labour Market Policies in Estonia, Connecting People with Jobs, OECD Publishing, Paris, https://doi.org/10.1787/31f72c5b-en.
OECD (2020a), Smart Cities and Inclusive Growth, OECD Publishing, Paris.
OECD (2020b), Education at a Glance 2020: OECD Indicators, OECD Publishing, Paris, https://doi.org/10.1787/69096873-en.
OECD (2019a), Health at a Glance 2019: OECD Indicators, OECD Publishing, Paris, https://doi.org/10.1787/4dd50c09-en.
OECD (2019b), Health in the 21st Century: Putting Data to Work for Stronger Health Systems, OECD Health Policy Studies, OECD Publishing, Paris, https://doi.org/10.1787/e3b23f8e-en.
OECD (2019c), OECD Economic Surveys: Estonia 2019, OECD Publishing, Paris, https://doi.org/10.1787/f221b253-en.
OECD (2018a), OECD Economic Surveys: Czech Republic 2018, OECD Publishing, Paris. https://doi.org/10.1787/eco_surveys-cze-2018-en.
OECD (2018b), Global Forum on Transparency and Exchange of Information for Tax Purposes: Estonia 2018 (Second Round): Peer Review Report on the Exchange of Information on Request, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264291034-en.
OECD (2017a), Health at a Glance 2017: OECD Indicators, OECD Publishing, Paris, http://dx.doi.org/10.1787/health_glance-2017-en.
OECD (2017b), Tackling Wasteful Spending on Health, OECD Publishing, Paris. http://dx.doi.org/10.1787/9789264266414-en.
OECD (2017c), Getting Infrastructure Right: A framework for better governance, OECD Publishing, Paris. http://dx.doi.org/10.1787/9789264272453-en.
OECD (2017d), OECD Economic Surveys: Estonia 2017, OECD Publishing, Paris. http://dx.doi.org/10.1787/eco_surveys-est-2017-en.
OECD (2017e), The Pursuit of Gender Equality: An Uphill Battle, OECD Publishing, Paris. http://dx.doi.org/10.1787/9789264281318-en.
OECD (2005), OECD Employment Outlook 2005, https://doi.org/10.1787/empl_outlook-2005-en.
OECD/European Observatory on Health Systems and Policies (2019), Estonia: Country Health Profile 2019, State of Health in the EU, OECD Publishing, Paris/European Observatory on Health Systems and Policies, Brussels.
Piketty, T., Saez, E. and Stantcheva, S. (2014), “Optimal taxation of top labor incomes: A tale of three elasticities”, American Economic Journal: Economic policy, 6(1): 230-271.
Pirttilä, J., and H. Selin (2011), “Income Shifting within a Dual Income Tax System: Evidence from the Finnish Tax Reform of 1993”, Scandinavian Journal of Economics, 113(1): 120–144.
Ravallion, M. (2012), "Why Don't We See Poverty Convergence?" American Economic Review, 102 (1): 504-23.
Rothstein, J. (2011), “Unemployment Insurance and Job Search in the Great Recession.” Brookings Papers on Economic Activity. Washington, DC: Brookings Institution.
Rense N. (2020), “Directions of thought for single parents in the EU”, Community, Work & Family, Taylor & Francis Group.
Swedbank (2021), Economic Outlook August 2021, Swedbank Macro Research, https://www.swedbank-research.com/pvs9qb/english/swedbank_economic_outlook/2021/q2/seo_aug_2021_eng_final.pdf.
Transparency International (2020), “Corruption Perception Index”, transparency.org. Available at: https://www.transparency.org/en/cpi/2020/.
Vandergrift, J.L., Gray, B.M. and Weng, W. (2018), “Do State Continuing Medical Education Requirements for Physicians Improve Clinical Knowledge?”, Health Service Research, 53: 1682-1701.
Võrk, A., A. Paulus and C. Leppik (2016), EUROMOD Country Report: Estonia 2011-2016, Colchester: University of Essex.