## 1. Macroeconomic Adjustment

## Source details

**Canonical URL:** [1. Macroeconomic Adjustment](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16172.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16172.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16172.pdf.json)

---

### A. The Macroeconomic Adjustment
- Real GDP in Latvia fell by more than 20 percent from its peak in 2007Q3 and remained below the pre-crisis peak for nearly 8 years; per-capita GDP surpassed its pre-crisis level in 2014Q2.
- Labor market mirrored GDP trends with a sharp increase in the unemployment rate and in long-term unemployed.
- Adjustment characterized by internal devaluation through wage restraint and productivity gains:
  - Unit labor costs (ULC) declined by close to 25 percent in one year—from the peak at the end of 2008 to the end of 2009.
  - From 2010 to 2013, ULCs remained low through relative wage restraint and strong growth in labor productivity.
- Productivity gains occurred despite a sharp fall in credit amid financial sector stress:
  - Growth in credit to non-financial corporates was negative and credit growth remained restrained due to deleveraging by a largely foreign owned banking sector.
- Key analytical questions:
  - Were productivity gains driven by sectoral/size composition shifts, firm-level efficiency improvements, catch-up by less productive firms, x-efficiency (same output with a smaller workforce), or differential gains for export-oriented firms?

### B. Sector and Firm Level Data
- Data sources and measurement:
  - Sector-level data from Latvia’s Central Statistical Bureau aggregated into 6 sectors: agriculture and mining, manufacturing, construction, trade, market services, and basic services (public administration, education and health services and other administrative and support services).
  - Firm-level data from Orbis covers about 70 percent of firms (in terms of persons employed) over 2010–2014; after dropping observations with missing key variables, there are over ten-thousand firm-year observations for 2010 onwards.
  - Orbis data are nominal; real values obtained using industry level value added and investment deflators from Eurostat (de-trended using Christiano-Fitzgerald filter at 2 years).
- Firm-level TFP estimation:
  - Real capital stock derived from book value of tangible fixed assets and depreciation using the perpetual inventory method.
  - Value added based TFP calculated using two factor inputs – real capital stock and number of employees.
  - Index number variant of TFP used based on labor shares estimated at the two-digit NACE industry level.
  - Other production function approaches (OLS and control function methods) yield broadly similar results.

### C. Was There a Sectoral Shift?
- Industry shares (2010–2014):
  - Value added and employment shares largely stable; some increase in market services and construction and offsetting declines in agriculture, trade and basic services.
  - Construction share of value added increased by close to 2 percent; market services by around 0.5 percent.
  - Employment shares: construction increased by 1.5 percent; market services increased by close to 3 percent.
  - Manufacturing share stable at around 13 percent of value added and employment.
- Labor productivity (cumulative changes, 2010–2014):
  - Construction: 25 percent
  - Agriculture: 20 percent
  - Basic services: 16 percent
  - Trade: 12 percent
  - Manufacturing and market services: largely flat, but positive
- Total factor productivity (TFP) (cumulative changes, 2010–2014):
  - Manufacturing: 20 percent
  - Construction: 19 percent
  - Trade: 16 percent
  - Market services: 12 percent
  - Agriculture and basic services: largely flat
- Contribution of sectoral composition:
  - Aggregate labor productivity increased by roughly 10 percent between 2010 and 2014.
  - Counterfactual with 2010 sectoral composition yields growth only marginally lower; contribution of sectoral composition to labor productivity growth estimated to be a little over 5 percent.
  - Counterfactual for TFP: sectoral composition explains around 16 percent of the estimated TFP growth.

### D. What Drove Firm Level Differences?
- Methodology:
  - Difference-in-means approach ranks firms by TFP growth between 2010 and 2013, divides them into three buckets, and compares averages of top and bottom buckets (results presented as percent of the average firm value in 2010). Statistical significance assessed at the 5 percent level.
- Catch-up by less productive firms:
  - Firms that increased productivity the most between 2010–2013 had on average 30 percent lower TFP in 2010 relative to firms that saw a relatively smaller increase in productivity; difference statistically significant.
  - The gap in initial labor productivity was over 60 percent (less productive firms being much lower initially); difference statistically significant.
  - Interpretation: a “catch-up” pattern—relatively low productivity firms increased productivity faster and moved closer to the frontier.
- Performance metrics in 2010 (firms that later increased productivity the most had markedly weaker metrics; differences statistically significant):
  - Return on equity: less than one-twentieth (compared to firms that saw the least increase in TFP).
  - Return on assets: just above one-tenth.
  - Profit margin: less than a quarter.
  - Note: analysis subject to survivorship bias as it is based on firms that survived and had data in both 2010 and 2013.
- Role of firm characteristics:
  - Larger firms (value added and employees) improved TFP relatively more, but differences were not statistically significant.
  - Firms with greater capital intensity improved productivity relatively more.
  - Firms with lower average cost of employees improved productivity relatively more; average labor cost used as a proxy for worker skill level, suggesting firms employing relatively lower skilled workers saw greater TFP improvements.

### Financial constraints and TFP
- Financial ratios and credit conditions:
  - Firms with greater TFP growth on average had significantly worse current ratio, liquidity ratio and solvency ratio and greater gearing.
  - Credit conditions were extremely tight, with overall credit to the corporate sector declining over the entire period; firms may have been cut off from the credit market irrespective of their financial ratios.
  - It is surprising financially constrained firms were able to make investments to improve TFP, unless there was considerable slack and sub-optimal use of resources in such firms.

### Gains in x-efficiency and labor use
- Significant gains in “x-efficiency”: firms maintained output with a smaller workforce.
- Difference-in-means results:
  - Firms that increased productivity relatively more increased employment less (holds for 2010–2013 and 2010–2014).
  - Ranking firms by new employment generated: firms that increased employment relatively more saw lower gains in TFP compared with firms that did not aggressively increase the size of their labor force.
- Interpretation: firms that improved TFP the most better utilized available resources and did not expand hiring aggressively.

### Export orientation: did export-oriented firms benefit?
- Trade-theory expectation: internal devaluation should benefit export-oriented firms and reallocate resources toward them.
- Empirical findings (manufacturing, 2-digit NACE export-share proxy):
  - Value added share of manufacturing firms with least exposure to the non-domestic market increased, while those of medium and high export orientation decreased.
  - For employment: the share of the most export oriented firms decreased marginally; the share of firms with medium export orientation increased.
  - Labor productivity increased for low export oriented manufacturing firms and decreased for medium and high export orientation firms.
  - TFP increased for all categories of manufacturing firms, irrespective of export orientation.
- Caveats: analysis uses two-digit industry export shares as a proxy and does not capture within-sector heterogeneity or exports via intermediaries.
- Difference-in-means using export orientation: firms that increased productivity the most were relatively more export oriented in 2010–2013 and 2010–2014, but the difference was not statistically significant. No evidence of a shift in labor resources toward more export oriented firms.

### Key quantitative findings (selected)
- Economic growth since 1995 has averaged 4.1 percent per year.
- Simple growth accounting: bulk of growth driven by TFP growth of about 4 percent (higher than other Baltics and CEE average).
- Post-crisis period: TFP grew at an average of 3.2 percent while growth averaged around 3.6 percent.
- Real total factor productivity (in logs) increased by around 10 percent for an average firm between 2010 and 2014 (controlling for sectoral composition).
- TFP dynamics by initial productivity group:
  - Low productivity group TFP increased by over 35 percent between 2010 and 2014.
  - High productivity group TFP declined marginally by around 3 percent.
  - In 2010 the average firm in the low productivity group was only half as productive as the high productivity group; by 2014 the difference was less than one-fourth.
- Latvia’s labor productivity is only a third of the EU15 average; labor in manufacturing only a forth as productive.
- Potential reform benefits (selected):
  - Estimates suggest better property rights and upgrading the legal system has the potential to improve Latvia’s TFP by close to 20 percent.
  - Authorities aim to double R&D spending from current 0.6 percent of GDP to 1.5 percent by 2020.

### Selected difference-in-means statistics (Top third vs Bottom third)
- Total factor productivity: Top third 0.87, Bottom third 1.20, Difference -0.33, t-stat -4.51
- Labor productivity: Top third 0.45, Bottom third 1.22, Difference -0.77, t-stat -6.47
- Return on assets: Top third -0.62, Bottom third 2.42, Difference -3.04, t-stat -11.68
- Return on equity: Top third 0.08, Bottom third 1.66, Difference -1.58, t-stat -5.44
- Profit margin: Top third -0.30, Bottom third 2.02, Difference -2.32, t-stat -11.44
- Capital intensity: Top third 1.09, Bottom third 0.89, Difference 0.20, t-stat 5.72
- Long term debt to total assets: Top third 2.95, Bottom third 0.01, Difference 2.94, t-stat 1.01
- Current ratio: Top third 0.78, Bottom third 1.17, Difference -0.39, t-stat -2.12
- Liquidity ratio: Top third 0.69, Bottom third 1.15, Difference -0.46, t-stat -2.52
- Solvency ratio: Top third 0.70, Bottom third 1.18, Difference -0.48, t-stat -5.81
- Gearing: Top third 1.14, Bottom third 0.86, Difference 0.28, t-stat 1.82
- Change in employment (2010–2013): Top third 0.33, Bottom third 1.52, Difference -1.19, t-stat -2.50
- Share of exports: Top third 1.07, Bottom third 1.03, Difference 0.05, t-stat 0.52

### Policy implications and recommended reform areas
- Strengthen business environment and enhance state-owned enterprise (SoE) governance:
  - Address dominance of SoEs in network sectors (electricity, postal services, airlines, railways, telecom) and foster competition (e.g., unbundling in energy markets).
  - Address OECD-identified shortcomings: non-transparent accountability and shareholder power; perception that SoEs operate less effectively than private enterprises.
  - Simplify license and permit systems; remove compulsory chamber membership of professional services; reduce complexity of regulatory procedure.
- Improve insolvency regime effectiveness and efficiency:
  - Recent steps: increase in number of courts; better specialization and allocation of judges; greater income disclosure for insolvency administrators; stricter supervision.
  - Outcomes: length of proceedings and backlog have fallen; clearance rates have risen.
  - Remaining issues: banks report procedures can be lengthy and administrators obstructive; further implementation and refinement needed; strengthen out-of-court settlement and arbitration; combat grey economy to strengthen property rights.
- Upgrade public infrastructure to attract FDI and technology diffusion:
  - Rail connectivity limited, infrastructure outdated and incompatible with European neighbours.
  - Majority of regional and local roads (over 90 percent of the network) are in poor condition; only a quarter of roads (in km) are categorized in good condition.
  - Energy transmission and distribution costs relatively high; links to wider European gas and energy markets still being developed.
- Promote structural transformation, innovation, and R&D:
  - Latvia’s share of high-technology exports: 9.2% in 2014.
  - Current R&D intensity: 0.6 percent of GDP; target 1.5 percent of GDP by 2020 (doubling).
  - Address low commercialization potential of research and weak cooperation between research, education and industry.
- Improve labor resource use and skill matching:
  - Latvia experiences significant skill shortages; demographic pressures (low birth rate, emigration) may worsen shortages.
  - Authorities emphasize reforms focused on science, technology, engineering, and mathematics; policies needed to stem emigration and encourage skilled return migration.

### Financial sector context and credit constraints (selected)
- Commercial banks’ assets accounted for almost 130 percent of GDP at end-2015.
- Alternative corporate funding is limited: total stock market capitalization is close to 4 percent of GDP; outstanding non-financial corporate debt is less than 1 percent of GDP; government securities are close to 5 percent of GDP.
- Out of 27 commercial banks, around half are foreign: 3 are subsidiaries from the Nordic region, and 10 are branches of foreign banks.
- Foreign banks account for around 81 percent of domestic lending.
- Non-resident deposits form more than half of the deposit base in Latvia; NRD banks have around 13 percent share of domestic loan portfolio (as of February 2016) and hold close to 60 percent liquid assets (liquidity ratio).
- NRD deposit outflow example: suffered a 30 percent outflow during 2008.
- Financial sector repair indicators:
  - Pre-crisis peak: 2006 Q3 credit growth reached almost 70 percent.
  - Crisis impact: NPLs peaked at 20 percent.
  - Repair outcomes: NPLs decreased to 6 percent as of end 2015; provisions to NPLs have increased substantially.
  - Banking system capital adequacy ratio at 22.7 percent (CET1 of 19.7 percent) as of end-2015.
  - Banking sector liquidity ratio stood above 65 percent.
  - For banks focused on residents, CET1 was 26 percent.
- Real credit contraction:
  - Real credit has shrunk for over 5 consecutive years despite positive GDP growth.
  - Comparative durations: Estonia ceased real credit contraction after 3 years; Lithuania after 4 years.
  - Bijsterbosch and Dahlhaus (2011): out of 211 recoveries studied, only 12 experienced 3 consecutive years of negative credit growth.
- Supply-side conclusions:
  - Capital and liquidity availability at aggregate level are adequate, but higher regulatory costs and risk model calibration raise lending costs and tighten standards.
  - Legacy issues (collateral recovery, heightened risk aversion, centralized risk management) are significant contributors to persistent tight lending conditions, particularly for SMEs and households lacking collateral or documentable income.
  - Continued reform implementation (courts, insolvency procedures) and balanced supervision are vital to restore sustained credit growth without embedding excessive risk-aversion.

### Structural factors constraining credit access and policy lessons
- Structural constraints:
  - Large size of the gray economy excludes many firms and individuals from bank credit because “official” income is inadequate.
  - Slow process of reclaiming collateral impedes lending to riskier borrowers.
  - Insolvency framework reform has made progress, but implementation lags; administrators and judges could be better trained on insolvency-specific issues.
- Banking sector structure and foreign bank behavior:
  - Countries with more international banking linkages tend to see larger contractions in lending as foreign banks deleverage disproportionately from non-strategic stressed regions.
  - Latvia constitutes 2 percent of the Swedish banks’ assets.
  - Bank-level data show loan growth is much more sensitive to GDP growth in the Baltics than in Sweden, producing highly cyclical subsidiary behavior.
  - Domestic banks have not broadly picked up the slack left by foreign banks; NRD sector’s lending to the domestic economy remains limited at 13 percent of total lending.
- Demand-side and sectoral reallocation:
  - Monetary policy and macro conditions unlikely to be a drag on credit demand given current stance, improving property prices, and strengthening balance sheets.
  - Outstanding credit reduced by 26 percent since end-2010, masking sectoral differences; construction and real estate deleveraged strongly (these two sectors constituted 36 percent of domestic corporate credit at end-2015, compared to 44 percent at end-2010).
  - Excluding real estate and construction, year-on-year credit growth turned positive in the final quarter of 2015.
- Policy lessons and recommended actions to resuscitate credit growth:
  - Facilitate access to credit for SMEs and first-time borrowers where market failures are largest.
  - Government programs such as loan guarantee schemes and subordinated loans by the Single Development Institute (SDI) are welcomed; effective guarantee schemes should be targeted, commercially operated via competent banks, and avoid misallocation.
  - Establishment of a credit bureau (this year) is an important step; bureaus should provide positive and negative data, including land register and court proceeding information.
  - Review benchmarking results of banks’ internal risk models (in cooperation with other competent authorities) to ensure models appropriately capture risk and that crisis legacy does not unduly constrain lending.
  - Institutional improvements: ensure uniformity in insolvency administration quality; strengthen out-of-court settlement and arbitration; combat grey economy.

*Source: IMF staff analysis in “1. Macroeconomic Adjustment” and related sections of the IMF staff report (May 17, 2016).*

### 1. Macroeconomic Adjustment ___________________________________________________________ 4

### 1. Macroeconomic Adjustment

### A. The Macroeconomic Adjustment
- Real GDP in Latvia fell by more than 20 percent from its peak in 2007Q3 and remained below the pre-crisis peak for nearly 8 years; per-capita GDP surpassed its pre-crisis level in 2014Q2.
- Labor market mirrored GDP trends with a sharp increase in the unemployment rate and in long-term unemployed.
- The adjustment was characterized by internal devaluation through wage restraint and productivity gains:
  - Unit labor costs (ULC) declined by close to 25 percent in one year—from the peak at the end of 2008 to the end of 2009.
  - From 2010 to 2013, ULCs remained low through a combination of relative wage restraint and strong growth in labor productivity.
- Productivity gains occurred despite a sharp fall in credit amid stress in the financial sector:
  - Growth in credit to non-financial corporates was negative and credit growth remained restrained due to deleveraging by a largely foreign owned banking sector.
- Key questions examined:
  - Were productivity gains driven by sectoral/size composition shifts, broad-based firm-level efficiency improvements, catch-up by less productive firms, x-efficiency (same output with a smaller workforce), or differential gains for export-oriented firms?

### B. Sector and Firm Level Data
- Data sources and coverage:
  - Sector-level data from Latvia’s Central Statistical Bureau aggregated into 6 sectors: agriculture and mining, manufacturing, construction, trade, market services, and basic services (public administration, education and health services and other administrative and support services).
  - Firm-level data from Orbis covers about 70 percent of firms (in terms of persons employed) over 2010–2014; after dropping observations with missing key variables, there are over ten-thousand firm-year observations for 2010 onwards.
  - Orbis data are nominal; real values obtained using industry level value added and investment deflators from Eurostat (de-trended using Christiano-Fitzgerald filter at 2 years).
- Firm-level TFP estimation:
  - Real capital stock derived from book value of tangible fixed assets and depreciation using the perpetual inventory method.
  - Value added based TFP calculated using two factor inputs – real capital stock and number of employees.
  - Index number variant of TFP used based on labor shares estimated at the two-digit NACE industry level.
  - Other production function approaches (OLS and control function methods) yield broadly similar results.

### C. Was There a Sectoral Shift?
- Industry shares:
  - Value added and employment shares were largely stable between 2010 and 2014, with some increase in market services and construction and offsetting declines in agriculture, trade and basic services.
  - Construction share of value added increased by close to 2 percent; market services by around 0.5 percent.
  - Employment shares: construction increased by 1.5 percent; market services increased by close to 3 percent.
  - Manufacturing share was largely stable at around 13 percent of value added and employment.
- Labor productivity (cumulative changes, 2010–2014):
  - Construction: 25 percent
  - Agriculture: 20 percent
  - Basic services: 16 percent
  - Trade: 12 percent
  - Manufacturing and market services: largely flat, but positive
- Total factor productivity (TFP) (cumulative changes, 2010–2014):
  - Manufacturing: 20 percent
  - Construction: 19 percent
  - Trade: 16 percent
  - Market services: 12 percent
  - Agriculture and basic services: largely flat
- Contribution of sectoral composition:
  - Aggregate labor productivity increased by roughly 10 percent between 2010 and 2014.
  - Counterfactual with 2010 sectoral composition yields growth only marginally lower; contribution of sectoral composition to labor productivity growth estimated to be a little over 5 percent.
  - Counterfactual for TFP: sectoral composition explains around 16 percent of the estimated TFP growth.

### D. What Drove Firm Level Differences?
- Methodology:
  - A difference-in-means approach ranks firms by TFP growth between 2010 and 2013, divides them into three buckets, and compares averages of top and bottom buckets (results presented as percent of the average firm value in 2010). Statistical significance assessed at the 5 percent level.
- Catch-up by less productive firms:
  - Firms that increased productivity the most between 2010–2013 had on average 30 percent lower TFP in 2010 relative to firms that saw a relatively smaller increase in productivity; difference statistically significant.
  - The gap in initial labor productivity was over 60 percent (less productive firms being much lower initially); difference statistically significant.
  - Interpretation: a “catch-up” pattern—relatively low productivity firms increased productivity faster and moved closer to the frontier.
- Performance metrics:
  - Firms that increased productivity the most had markedly weaker performance metrics in 2010 (differences statistically significant):
    - Return on equity: less than one-twentieth (compared to firms that saw the least increase in TFP).
    - Return on assets: just above one-tenth.
    - Profit margin: less than a quarter.
  - Note: analysis subject to survivorship bias as it is based on firms that survived and had data in both 2010 and 2013.
- Role of firm characteristics:
  - Larger firms (value added and employees) improved TFP relatively more, but differences were not statistically significant.
  - Firms with greater capital intensity (higher share of capital stock to value added) improved productivity relatively more.
  - Firms with lower average cost of employees improved productivity relatively more; average labor cost used as a proxy for worker skill level, suggesting firms employing relatively lower skilled workers saw greater TFP improvements.

*Source: IMF staff analysis in “1. Macroeconomic Adjustment” (May 17, 2016).*

### 16.      Financial constraints do not appear to have hampered productivity improvements.

### 16.      Financial constraints do not appear to have hampered productivity improvements.

### Financial constraints and TFP
- Firms with greater TFP growth during the post-crisis period on average had significantly worse current ratio, liquidity ratio and solvency ratio and greater gearing.
- Credit conditions were extremely tight, with overall credit to the corporate sector declining over the entire period; firms may have been cut off from the credit market irrespective of their financial ratios.
- It is surprising financially constrained firms were able to make investments to improve TFP, unless there was considerable slack and sub-optimal use of resources in such firms.

### Gains in x-efficiency and labor use
- Overall, there were significant gains in “x-efficiency”, whereby firms were able to maintain output with a smaller workforce.
- Difference-in-means results:
  - Firms that increased productivity relatively more increased employment less (holds for 2010–2013 and 2010–2014).
  - Ranking firms by new employment generated: firms that increased employment relatively more saw lower gains in TFP compared with firms that did not aggressively increase the size of their labor force.
- Interpreted outcome: firms that improved TFP the most better utilized available resources and did not expand hiring aggressively.

### Export orientation: did export-oriented firms benefit?
- Trade theory predicts internal devaluation (wage restraint + productivity gains lowering unit labor costs) should benefit export-oriented firms and reallocate resources toward them.
- Empirical findings (manufacturing, 2-digit NACE export-share proxy):
  - Value added share of manufacturing firms with least exposure to the non-domestic market increased, while those of medium and high export orientation decreased.
  - For employment: the share of the most export oriented firms decreased marginally; the share of firms with medium export orientation increased.
  - Labor productivity increased for low export oriented manufacturing firms and decreased for medium and high export orientation firms.
  - TFP increased for all categories of manufacturing firms, irrespective of export orientation.
- Caveats: analysis uses two-digit industry export shares as a proxy for firm export orientation and does not capture within-sector heterogeneity or exports via intermediaries.
- Difference-in-means using export orientation: firms that increased productivity the most were relatively more export oriented in 2010–2013 and 2010–2014, but the difference was not statistically significant. No evidence of a shift in labor resources toward more export oriented firms.

### Key quantitative findings (selected)
- Economic growth since 1995 has averaged 4.1 percent per year.
- Simple growth accounting: bulk of growth driven by TFP growth of about 4 percent (higher than other Baltics and CEE average).
- Post-crisis period: TFP grew at an average of 3.2 percent while growth averaged around 3.6 percent.
- Real total factor productivity (in logs) increased by around 10 percent for an average firm between 2010 and 2014 (controlling for sectoral composition).
- TFP dynamics by initial productivity group:
  - Low productivity group TFP increased by over 35 percent between 2010 and 2014.
  - High productivity group TFP declined marginally by around 3 percent.
  - In 2010 the average firm in the low productivity group was only half as productive as the high productivity group; by 2014 the difference was less than one-fourth.
- Latvia’s labor productivity is only a third of the EU15 average; labor in manufacturing only a forth as productive.
- Potential reform benefits (selected):
  - Estimates suggest better property rights and upgrading the legal system has the potential to improve Latvia’s TFP by close to 20 percent.
  - Authorities aim to double R&D spending from current 0.6 percent of GDP to 1.5 percent by 2020.

### Policy implications and recommended reform areas
- Strengthen business environment and enhance state-owned enterprise (SoE) governance:
  - Address dominance of SoEs in network sectors (electricity, postal services, airlines, railways, telecom) and foster competition (e.g., unbundling in energy markets).
  - Address OECD-identified shortcomings: non-transparent accountability and shareholder power; perception that SoEs operate less effectively than private enterprises.
  - Simplify license and permit systems; remove compulsory chamber membership of professional services; reduce complexity of regulatory procedure.
- Improve insolvency regime effectiveness and efficiency:
  - Recent steps: increase in number of courts; better specialization and allocation of judges; greater income disclosure for insolvency administrators; stricter supervision.
  - Outcomes: length of proceedings and backlog have fallen; clearance rates have risen.
  - Remaining issues: banks report procedures can be lengthy and administrators obstructive; further implementation and refinement needed; strengthen out-of-court settlement and arbitration; combat grey economy to strengthen property rights.
- Upgrade public infrastructure to attract FDI and technology diffusion:
  - Rail connectivity limited, infrastructure outdated and incompatible with European neighbours.
  - Majority of regional and local roads (over 90 percent of the network) are in poor condition; only a quarter of roads (in km) are categorized in good condition.
  - Energy transmission and distribution costs relatively high; links to wider European gas and energy markets still being developed.
- Promote structural transformation, innovation, and R&D:
  - Latvia’s share of high-technology exports: 9.2% in 2014.
  - Current R&D intensity: 0.6 percent of GDP; target 1.5 percent of GDP by 2020 (doubling).
  - Address low commercialization potential of research and weak cooperation between research, education and industry.
- Improve labor resource use and skill matching:
  - Latvia experiences significant skill shortages; demographic pressures (low birth rate, emigration) may worsen shortages.
  - Authorities emphasize reforms focused on science, technology, engineering, and mathematics; policies needed to stem emigration and encourage skilled return migration.

### Selected difference-in-means statistics (Top third vs Bottom third)
- Total factor productivity: Top third 0.87, Bottom third 1.20, Difference -0.33, t-stat -4.51
- Labor productivity: Top third 0.45, Bottom third 1.22, Difference -0.77, t-stat -6.47
- Return on assets: Top third -0.62, Bottom third 2.42, Difference -3.04, t-stat -11.68
- Return on equity: Top third 0.08, Bottom third 1.66, Difference -1.58, t-stat -5.44
- Profit margin: Top third -0.30, Bottom third 2.02, Difference -2.32, t-stat -11.44
- Capital intensity: Top third 1.09, Bottom third 0.89, Difference 0.20, t-stat 5.72
- Long term debt to total assets: Top third 2.95, Bottom third 0.01, Difference 2.94, t-stat 1.01
- Current ratio: Top third 0.78, Bottom third 1.17, Difference -0.39, t-stat -2.12
- Liquidity ratio: Top third 0.69, Bottom third 1.15, Difference -0.46, t-stat -2.52
- Solvency ratio: Top third 0.70, Bottom third 1.18, Difference -0.48, t-stat -5.81
- Gearing: Top third 1.14, Bottom third 0.86, Difference 0.28, t-stat 1.82
- Change in employment (2010–2013): Top third 0.33, Bottom third 1.52, Difference -1.19, t-stat -2.50
- Share of exports: Top third 1.07, Bottom third 1.03, Difference 0.05, t-stat 0.52

*Source: IMF staff analysis as presented in the chapter.*

### References

### _cr16172 - References

### References (bibliography)
- Blanchard, Olivier, Griffiths, Mark, Gruss, Bertrand (2013) – Boom, Bust, Recovery: Forensics of the Latvia Crisis. Brookings Papers on Economic Activity, Economic Studies Program, The Brookings Institution, vol. 47, issue 2 (Fall), pp. 325–388
- Gal, Peter N. (2013), “Measuring Total Factor Productivity at the Firm Level using OECD-ORBIS.”, No. 1049 OECD Publishing.
- Gilhooly, Bob, (2009), Firm-level estimates of capital stock and productivity. Economic and Labour Market Review, 3(5), 36–41.
- IMF (2016), CESEE Convergence: How to Get Back on the Fast Track, European Regional Economic Issues Report, April 2015
- Javorcik, B., 2010, “Foreign Direct Investment and International Technology Transfer,” Encyclopedia of Financial Globalization.
- OECD (2015a), “Review of Latvia’s position relative to the OECD guidelines on corporate governance of state-owned enterprises,” OECD Publishing.
- OECD (2015b), “Policy areas for increasing productivity in Latvia economics,” OECD Economics Department Working Papers, No 1255.
- Wooldridge, Jeffrey M. (2009) "On estimating firm-level production functions using proxy variables to control for unobservables." Economics Letters 104.3: 112–114.

### Executive summary and key findings
- Latvia has experienced a prolonged "creditless recovery"; credit continues to decline 5 years after economic growth turned positive.
- Recovery has been driven mainly by consumption rather than investment; bank credit growth will be needed to fund investment and productivity-enhancing projects.
- Demand factors explain subdued lending to larger, cash-rich firms; supply factors are likely more important for SMEs and households.
- Legacy of the crisis: banks are cautious due to past losses and difficulties recovering collateral; excessive risk-aversion risks becoming entrenched.
- In a foreign-dominated banking sector, the business cycle in parent (home) countries should not become an impediment to lending in the host country.

### Financial sector landscape and structure
- Commercial banks’ assets accounted for almost 130 percent of GDP at end-2015.
- Alternative corporate funding is limited: total stock market capitalization is close to 4 percent of GDP; outstanding non-financial corporate debt is less than 1 percent of GDP; government securities are close to 5 percent of GDP.
- Out of 27 commercial banks, around half are foreign: 3 are subsidiaries from the Nordic region, and 10 are branches of foreign banks.
- Foreign banks account for around 81 percent of domestic lending.
- Domestic banks rely mostly on non-resident deposits (NRDs) and hold highly liquid foreign assets.
- NRD banks have around 13 percent share of domestic loan portfolio (as of February 2016) and hold close to 60 percent liquid assets (liquidity ratio).
- Non-resident deposits form more than half of the deposit base in Latvia.
- NRD deposit outflow example: suffered a 30 percent outflow during 2008.
- Regulatory responses: NRD banks face larger capital (up to an additional 9.5 percent) and liquidity hurdles (up to 60 percent).

### AML and supervisory reforms
- Authorities widened definition of Politically Exposed Persons (PEP) and FCMC adopted a risk-based supervision strategy.
- FCMC employed external resources for in-depth audits to impose minimum standards on banks’ risk monitoring systems.
- Staff encouraged continued efforts and allocation of sufficient resources to AML supervision.

### Financial sector repair: progress and indicators
- Pre-crisis peak: 2006 Q3 credit growth reached almost 70 percent.
- Crisis impact: NPLs peaked at 20 percent.
- Repair outcomes: NPLs decreased to 6 percent as of end 2015; provisions to NPLs have increased substantially.
- Banking system capital adequacy ratio at 22.7 percent (CET1 of 19.7 percent) as of end-2015.
- Banking sector liquidity ratio stood above 65 percent.
- For banks focused on residents, CET1 was 26 percent.
- Debt deleveraging: loans to households and companies have significantly decreased as a percent of GDP; debt-to-income ratios and debt repayment capacity have improved.
- Credit gap remains significantly negative (credit as percentage deviation from long-run trend).

### Credit cycle positioning and financial conditions
- Methodology (GFSR October 2015) places Latvia ahead of the Euro Area in the credit cycle, but behind Japan and the United States.
- Financial Conditions Index (FCI) constructed from NPLs, credit growth, profitability, house prices, and funding ratios; first principal component explains over 75 percent of covariance (sample 2007–2015). FCI has mean zero and variance 1 over sample.
- FCI shows easiest conditions in 2007, deterioration thereafter, and a slow return toward neutral; credit growth remains far behind neutral levels.
- Latvijas Banka bank lending survey indicates lending standards have not eased since 2014; banks expect loosening but observed standards remain tight.

### Why has credit been restrained? (supply-side analysis)
- Supply constraints that can tighten credit include: (i) shortages in capital or liquidity (or increased requirements); (ii) increased risk aversion; (iii) absence of good collateral; (iv) uncertainty about borrowers’ creditworthiness (asymmetric information); (v) encumbered balance sheets preoccupied with workout of bad loans.
- Capital and liquidity do not appear binding: capital adequacy 22.7 percent, CET1 19.7 percent, liquidity ratio above 65 percent; resident-focused banks CET1 26 percent.
- Higher post-crisis regulatory requirements increased marginal cost of lending; example: Sweden total capital requirements for groups range between 19–25 percent of RWA.
- Crisis legacy effects tightening supply:
  - Re-evaluation of risk: Nordic banks withdrew from all but top-tier clients; some banks withdrew from non-top clients and are more cautious with new clients due to recovery problems with collateral.
  - Calibration of internal risk-based (IRB) models using limited local data implies higher projected probability of losses and loss given default for Latvian portfolios, mechanically producing tighter conditions compared to parent banks’ home countries.
- Real credit contraction context:
  - Real credit has shrunk for over 5 consecutive years despite positive GDP growth.
  - Comparative durations: Estonia ceased real credit contraction after 3 years; Lithuania after 4 years.
  - Bijsterbosch and Dahlhaus (2011): out of 211 recoveries studied, only 12 experienced 3 consecutive years of negative credit growth.

### Supply-side conclusions
- Capital and liquidity availability at aggregate level are adequate, but higher regulatory costs and risk model calibration raise lending costs and tighten standards.
- Legacy issues (collateral recovery, heightened risk aversion, centralized risk management) are significant contributors to persistent tight lending conditions, particularly for SMEs and households lacking collateral or documentable income.
- Continued reform implementation (courts, insolvency procedures) and balanced supervision are vital to restore sustained credit growth without embedding excessive risk-aversion.

*Source: _cr16172 - References (IMF staff text provided in content unit).*

### 14.      Structural factors further contribute to the tightness of credit standards. The significant

### _cr16172 - 14.      Structural factors further contribute to the tightness of credit standards. The significant

### Structural factors constraining credit access
- Large size of the gray economy means many firms and individuals are effectively excluded from seeking bank credit because their “official” income is inadequate.
- Slow process of reclaiming collateral is seen by banks as an impediment to lending to riskier borrowers.
- Insolvency framework reform has made commendable progress, but implementation lags; administrators and judges could be better trained on insolvency-specific issues.

### Banking sector structure and foreign bank behavior
- Countries with more international banking linkages tend to see larger contractions in lending as foreign banks deleverage disproportionately from non-strategic stressed regions or markets (April 2015 GFSR, Chapter 2).
- Latvia constitutes 2 percent of the Swedish banks’ assets.
- Bank-level data (Bankscope) for Swedbank and SEB show loan growth is much more sensitive to GDP growth in the Baltics than in Sweden (Figure 5), generating highly cyclical behavior in subsidiaries.
- Resulting patterns:
  - Reduction in net loans since 2008 has been significantly faster in Baltic subsidiaries compared to parent banks.
  - This procyclical behavior could reinforce the business cycle and volatility of output and credit growth.
  - The behavior is similar for subsidiaries in Estonia and Lithuania and is independent of level of capitalization; subsidiaries on average enjoy higher levels of capital compared to the overall group.
- Shared supervisory responsibility (home supervisor, host supervisor FCMC, and SSM) may contribute to constrained lending:
  - Small size of subsidiaries relative to parents could imply group and home-supervisor priorities override local considerations.
  - Activation of macro-prudential levers by Swedish FSA (appropriate for strong credit growth in Sweden) may be less appropriate for some host countries; while some tools (e.g., Counter Cyclical Buffer) only apply to Swedish exposures, a risk-off attitude could percolate through group business models.

### Domestic banks and market structure implications
- Domestic banks have not broadly picked up the slack left by foreign banks:
  - Except for one small domestic bank (less than 7 percent of total loans) gaining market share, domestic banks are mostly concentrated on non-resident lending, which is growing more strongly than domestic lending (Figure 6).
  - NRD sector’s lending to the domestic economy remains limited at 13 percent of total lending and typically targets special big-ticket projects such as luxury real estate.
- Implications:
  - Large parts of the economy (first-time borrowers, SMEs) have limited access to credit.
  - The banking structure constrains policy levers available to encourage credit growth.

### Demand-side factors affecting credit
- Monetary policy and general macroeconomic conditions are unlikely to be a drag on demand for credit given current monetary policy stance, improving property prices, and strengthening household and company balance sheets.
- Crisis legacy and uncertainties have affected borrowing attitudes:
  - Large number of insolvencies during the crisis likely reduced willingness to take on debt.
  - Anecdotal evidence: actual duration of mortgage portfolios is significantly shorter than time-to-maturity at origination as borrowers amortize faster than required.
  - Estonia and Lithuania, which did not see NPL increases to the same extent as Latvia, saw credit returning earlier (Figure 5).
  - Ongoing uncertainties: economic prospects in key trading partners, geo-political risks, delays in arrival of EU structural funds may drag on credit demand.

### Sectoral reallocation of credit and recent developments
- Overall outstanding credit reduced by 26 percent since end-2010, masking sectoral differences.
- Construction and real estate deleveraged strongly; these two sectors constituted 36 percent of domestic corporate credit at end-2015, compared to 44 percent at end-2010.
- Excluding real estate and construction, year-on-year credit growth turned positive in the final quarter of 2015.
- NRD (non-resident) lending has grown strongly while loans to the domestic economy continue to shrink.

### Lessons and policy implications
- Resuscitating credit growth is key to maintaining recovery:
  - Creditless recoveries tend to be weak and short-lived; investment contributes disproportionately less to growth in such episodes.
  - Latvia has experienced low investment growth and output growth driven by consumption.
  - Future investment—and productivity gains beyond labor shedding—will require credit access; absorption of EU funds requires private co-financing.
- Policy focus: facilitate access to credit for SMEs and first-time borrowers where market failures are largest.
  - Government programs such as loan guarantee schemes and subordinated loans by the Single Development Institute (SDI) are welcome and well received.
  - Effective credit guarantee schemes are typically:
    - Targeted to financially constrained sectors.
    - Operated on a commercial basis via on-lending through competent, financially sound banks with adequate staff to screen and monitor SME loans.
  - Credit guarantees should not become a vehicle for misallocation of funds to SMEs with little future or to firms that do not need support.
  - These programs do not help firms and individuals in the gray economy who under-report income and are therefore excluded from credit access.
- Recommendations on bank risk models and supervisory practices:
  - Centralized risk management by parent banks may contribute to overly-constrained lending in Latvia.
  - Banks’ internal risk models may place too much weight on recent crisis experience; bank lending standards have not changed substantially after sharp tightening since the crisis.
  - Subsidiary-level evidence suggests Nordic banks’ lending is more volatile in Baltic business cycles than in home country.
  - Staff urged prudential authorities to review benchmarking results of banks’ risk models (in cooperation with other competent authorities) to ensure models appropriately capture risk and that crisis legacy does not unduly constrain lending.
- Institutional improvements to restore trust and facilitate lending:
  - Ensure uniformity in quality of insolvency administration and improve implementation of insolvency reforms.
  - Establishment of a credit bureau (this year) is an important step to facilitate access to credit history, particularly for first-time borrowers.
  - Authorities should ensure bureaus become a “one-stop shop” with access to positive and negative data, including land register and information on court proceedings.
- International experience on policies to reignite credit may be helpful; examples include:
  - Measures to prevent discrimination against SME lending (e.g., lender codes of conduct, public disclosure of application procedures, written reasons for rejection and appeal processes).
  - Awareness programs informing SMEs of financing options.

*IMF staff analysis in "REPUBLIC OF LATVIA" (excerpts)._*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr16172.pdf_