## wpiea2019101

## Source details

**Canonical URL:** [wpiea2019101](https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019101.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wpiea2019101.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wpiea2019101.pdf.json)

---

### I. Introduction — role, examples, and risks of public debt
- Public debt is used to:
  - deal with negative shocks,
  - undertake countercyclical fiscal policy,
  - finance exceptionally large expenditures (e.g., public infrastructure).
- Illustrative observations:
  - United States: American Recovery and Reinvestment Act in 2009 was USD 831 billion—about 5.5 percent of GDP.
  - US government debt rose from 64 percent of GDP in 2007 to above 100 percent in 2012.
  - Advanced economies: average debt-to-GDP rose from about 60 percent in 2007 to over 90 percent in 2016.
  - China: discretionary stimulus spending of more than 6 percent of GDP; public debt increased from 29 percent of GDP in 2007 to 44 percent of GDP by 2016.
- Risks:
  - reduced capacity to stabilize the business cycle (limited fiscal space),
  - potential impairment of growth via crowding out private investment or raising uncertainty about future tax and inflation rates.

### II. Good motives to borrow — tax smoothing, investment, safe assets
- Tax-smoothing logic and implications:
  - Smoothing taxes minimizes distortionary costs when taxes are distortionary with convex cost (Gallatin (1807); Barro (1979)).
  - Tax smoothing applies to temporary changes in spending; permanent changes imply no persistent deficit/surplus.
  - Governments may accumulate assets or pay down debt to hedge unexpected shocks (Aiyagari et al. 2002; Bhandari et al. 2016).
  - Domestic vs. foreign borrowing: domestic borrowing does not expand the resource envelope for a given GDP unless borrowing from foreigners.
  - Debt is a dead-weight loss equal to the present value of distortionary taxes required for repayment; higher inherited debt can reduce optimal public investment (Ostry, Ghosh, and Espinoza, 2015).
  - Cost of borrowing matters: higher interest rates reduce optimal borrowing for a given spending shock.
- Empirical correlations:
  - G7 average debt-to-GDP moved from about 40 percent in the 1970–1980 period to over 80 percent in 2007.
  - Simple regression (country- and year-fixed effects) in advanced economies: a one percentage point change in the debt-to-GDP ratio is associated with a 0.04 percentage point increase in public investment — implying typically only 4 percent of debt issuance is used to finance public investment.
  - In emerging and developing economies, correlation between public debt and public investment is negative but not statistically significant.
  - Bacchiocchi, Borghi, and Missale (2011): negative correlation between debt and public investment in high-debt countries and positive correlation in low-debt countries.
- Public debt as a safe asset and market development:
  - Moderate non-inflationary government debt can support financial market development and reduce output decline during crises (Abbas and Christensen, 2010; Gorton and Ordoñez, 2013).
  - Global safe-asset shortages occurred during the Global Financial Crisis (Caballero, Farhi, and Gourinchas, 2008; Brunnermeier et al., 2017).

### III. Keynesian demand stimulus and output stabilization
- Fiscal and monetary policy interactions:
  - Fiscal and monetary policy can be substitutes in stabilization models.
  - Monetary policy may be constrained (e.g., zero lower bound), shifting stabilization burden to fiscal policy (Eggertsson and Woodford, 2004).
  - Monetary policy may be assigned to external objectives while fiscal policy addresses internal objectives; with multiple distortions fiscal policy can be necessary to achieve first best (Blanchard and Gali, 2010).
  - Countercyclical fiscal policy can be problematic when countries face high default risk (Neumeyer and Perri, 2005; Hatchondo, Martinez and Roch, 2017).
- Empirical patterns:
  - Procyclical fiscal policy stronger among emerging markets and Latin America (Gavin and Perotti, 1997; Kaminsky, Reinhart, and Végh, 2004).
  - OECD/European evidence mixed; fiscal policy often countercyclical but can be procyclical in consolidations (Égert, 2012; Fatás, 2018).
  - Procyclical policy linked to higher output volatility and lower growth (Aghion et al. 2007).
- Measurement and debt linkage:
  - Fiscal stance measure: change in the inflation-adjusted budget balance as a ratio to GDP (Blanchard, 1993).
  - Countercyclical policy implies deficits when growth is below trend, leading to debt accumulation.
- Automatic vs. discretionary:
  - Automatic stabilizers arise from pre-existing rules; larger governments tend to have stronger automatic stabilizers.
  - Discretionary policy complements or substitutes for automatic stabilizers; discretionary measures used more where automatic stabilizers are weak (Fatás, 2009).

### IV. Dynamic inefficiency and debt as intergenerational vehicle
- Rationale:
  - Government debt can transfer wealth across generations when private markets fail to provide optimal vehicles (Blanchard, 1985, 2019).
  - Criterion: economy dynamically inefficient if rate of return on capital is below growth rate.
- Evidence:
  - Abel et al. (1989): criterion not met for 6 advanced economies historically.
  - Recent reductions in real returns on safe assets suggest revisiting dynamic inefficiency (Geerolf, 2017).
  - Whether evidence warrants policy change in some countries remains open (Blanchard and Summers, 2017; Blanchard, 2019).

### V. Bad reasons to issue debt — political economy and overborrowing
- Limits to borrowing rationale:
  - Deficits in bad times should be offset by surpluses in expansions; steady accumulation indicates failure.
  - Overborrowing occurs when social marginal cost of additional debt exceeds social return of expenditures (Yared, 2018).
- Four sources of excessive debt accumulation:
  - (i) political budget cycles and rent seeking,
  - (ii) intergenerational transfers,
  - (iii) strategic manipulation,
  - (iv) common pool problems.
- Political budget cycles and rent seeking:
  - Politicians may cut taxes and raise spending to boost reelection; fiscal illusion or imperfect information can enable cycles (Rogoff and Sibert, 1988).
  - Yared (2010): higher debt can in some models reduce rent extraction incentives.
- Intergenerational transfers:
  - Public debt can act as negative bequests when private negative bequests are unenforceable (Cukierman and Meltzer, 1989; Tabellini, 1991; Song et al., 2012; Yared, 2018).
- Strategic manipulation:
  - Parties may manipulate debt to bind successors (Persson and Svensson, 1989; Alesina and Tabellini, 1990).
- Common pool problems:
  - Concentrated beneficiary groups vs. diffuse payers create overspending bias (Olson, 1965; Mauro and Villafuerte, 2013).
  - Hierarchical budgeting (ministry of finance dominance) can reduce common pool overspending (Alesina and Perotti, 1996).
  - Political fragmentation and turnover amplify deficits (Woo, 2003; Aguiar and Amador, 2011).
  - Battaglini and Coate (2008): low debt -> political distortions dominate; high debt -> self-insurance motives increase, policy converges toward social-planner outcomes.

### VI. Controlling overborrowing — electoral systems, fiscal rules, institutions
- Three institutional avenues:
  - Electoral systems:
    - Proportional systems tend to show deficit bias relative to majoritarian systems (Battaglini, 2010; empirical work: Roubini and Sachs, 1989; Grilli et al., 1991).
    - Parliamentary vs. presidential: presidential democracies tend to have smaller governments; within parliamentary democracies, majoritarian systems have smaller governments (Persson and Tabellini, 2003, 2004).
  - Fiscal rules:
    - Aim to limit debt accumulation and address time inconsistency.
    - Trade-offs: constraints on current action vs. constraints on successors; balanced-budget rules can limit countercyclical policy.
    - Adoption: from fewer than 20 countries in mid-1990s to nearly 100 countries now.
    - Empirical evidence mixed: Debrun et al. (2008), Bergman et al. (2016) find rules effective in Europe; Von Hagen (2006) suggests limits for largest euro-area countries; identification challenges remain (Heinemann et al., 2018).
  - Budgetary institutions:
    - Hierarchical budget rules empower ministries of finance and are associated with smaller deficits.
    - Transparency reduces incentives and ability to overborrow (Rogoff and Sibert, 1988; Milesi-Ferretti, 1997).
    - Fiscal transparency correlated with lower public debt (Alt and Lassen, 2006; Alesina et al., 1999; Dabla-Norris et al., 2010).
- Unexplained stock-flow reconciliation residuals:
  - Large unexplained increases in debt relative to recorded deficits occur (examples: Uruguay 2001–2002; Argentina 2001–2002).
  - Drivers: foreign-currency debt valuation effects, banking crises, hidden deficits, contingent liabilities.
  - Mitigations: safer debt structures, contingent debt instruments (e.g., GDP-indexed bonds), but political and market failures limit uptake.

### VII. Debt, growth, and investment — mechanisms and empirical evidence
- Mechanisms for adverse effects of high debt:
  - limits countercyclical policy,
  - crowds out private investment via higher interest rates or tightened bank balance sheets,
  - raises uncertainty about future taxes and potential confiscation,
  - can precipitate financial crises or higher sovereign spreads.
- Empirical patterns:
  - Reinhart and Rogoff (2010a) and Global Debt Dataset (Mbaye et al., 2018):
    - average (median) growth declines from 3.7 percent for debt-to-GDP < 30 percent,
    - to 2.6 (2.7) percent when debt ratio between 30 and 60 percent,
    - to 1.2 (1.6) percent when debt surpasses 90 percent of GDP (20 advanced economies sample).
    - In 119 low- and middle-income countries, average growth declines from 4.4 percent (low-debt) to 2.6 percent (debt > 90 percent).
  - Authors find a strong negative correlation between debt-to-GDP in year t and real GDP growth between t and t+5 controlling for year- and country-fixed effects.
  - Causality concerns: slow growth can cause rising debt; omitted variables and reverse causality complicate inference.
- Micro and identification evidence:
  - Panizza and Presbitero (2014): valuation effects instrument finds no effect; instrument weak in advanced-country samples.
  - Firm-level studies (Huang et al. 2017, 2018; Croce et al. 2019): higher government debt tightens financing constraints and raises cost of capital for firms, especially R&D-intensive or credit-constrained firms.

### VIII. Non-linearities, thresholds, and heterogeneity in the debt–growth relationship
- Non-linearity and threshold issues:
  - Literature examines thresholds (commonly 90 percent) and trajectories (Pescatori et al., 2014; Chudik et al., 2017).
  - Statistical challenges: limited observations above thresholds, sensitivity to parametric assumptions and outliers.
  - Non-parametric evidence (20 advanced economies, 1960–2016) shows:
    - average negative correlation conceals large heterogeneity across countries,
    - no common threshold beyond which debt uniformly causes growth slowdown.
  - Reported turning point intervals and mixed findings:
    - Checherita-Westphal and Rother (2012): turning point confidence interval of 49 to 119 percent of GDP.
    - Woo and Kumar (2015): mixed evidence; cannot statistically establish a stronger correlation above 90 percent consistently.
- Country-specific determinants of tolerable debt levels:
  - historical credit and inflation experience,
  - historical fiscal adjustment track record (Ghosh et al., 2013),
  - debt composition (Eichengreen, Hausmann, and Panizza, 2005),
  - institutional quality (Kourtellos et al., 2013).
- Debt structure and measurement caveats:
  - “Not all debts are equal”: use, holder (residents vs. non-residents), currency, maturity matter.
  - Data limitations: most researchers observe only debt level; comprehensive structure data limited (Abbas et al., 2014 exception).
  - Examples:
    - public debt held by non-residents: Japan ~5-7 percent; Italy ~40 percent; Ireland higher (Abbas et al., 2014).
  - Gross vs. net and implicit liabilities: pensions, local government debt, and SOE debt can make public-sector obligations much larger than official measures indicate.
- Welfare implications and policy trade-offs:
  - Negative correlation between debt and growth in advanced economies is clear, but causality is not convincingly established.
  - Even if debt harms growth, repaying inherited debt is not always welfare-improving due to tax-smoothing considerations (Ostry, Ghosh, and Espinoza, 2015).
  - Policy must weigh flexibility to respond to shocks against disciplining policymakers from excessive borrowing.

### IX. Box 1 — Ethiopia: balancing investment needs with debt sustainability
- Investment scaling and financing:
  - Ethiopia public investment: above 7 percent of GDP in the 2000s; about 15 percent of GDP between 2014 and 2017.
  - Major projects: Grand Ethiopian Renaissance Dam (cost estimate almost USD 5 billion — about 5 percent of GDP); railway to Djibouti.
  - Financing: external concessional and non-concessional financing, including large Chinese investment flows; supported by restrained government consumption, financial repression and an overvalued exchange rate (World Bank, 2016).
- Outcomes:
  - Real GDP growth averaged 10 percent annually (last decade).
  - Electricity access: 14 percent in 2005; 43 percent in 2016.
- Debt dynamics and risks:
  - Debt-to-GDP: 107 percent in 2002; 38 percent in 2009; 62 percent in June 2018.
  - IMF and World Bank assessment: Ethiopia is at "high risk of debt distress (IMF 2018b)".
  - Concerns: debt sustainability limits, crowding out private credit, weak external competitiveness due to exchange rate appreciation, absorptive capacity constraints (Presbitero, 2018).
- Policy lessons and recommendations:
  - Scale-up must account for risks to debt sustainability and future growth.
  - Diversify financing sources and strengthen absorptive capacity to improve project success rates.
  - Consider trade-offs between rapid public investment expansion and potential crowding out, loss of competitiveness, and higher debt distress risk.
- Key statistics:
  - Public investment: above 7 percent of GDP in the 2000s; about 15 percent of GDP between 2014 and 2017.
  - Grand Ethiopian Renaissance Dam cost estimate: almost USD 5 billion — about 5 percent of GDP.
  - Electricity access: 14 percent in 2005; 43 percent in 2016.
  - Global Infrastructure Hub estimate of investment shortfall to achieve 2030 Agenda targets for Ethiopia: about USD 285 billion.
  - Debt-to-GDP: 107 percent in 2002; 38 percent in 2009; 62 percent in June 2018.
  - Real GDP growth: averaging 10 percent annually (last decade).

### X. Figures, quantitative highlights, and methodological notes (selected)
- Figure 2 (Correlation Between Change in Public Debt and Contemporaneous Public Investment):
  - Regression coefficient on the debt variable: 0.041 (p-value of 0.011).
  - Interpretation: "a 10 percent increase of the debt-to-GDP ratio is associated with 0.4 percent lower ratio of public investment over GDP."
  - Number of observations: 899.
- Figure 5 (Government Debt and Subsequent GDP Growth, Selected Advanced Economies; 1960–2016):
  - Regression coefficient on the debt variable: -0.016 (p-value of 0.001).
  - Interpretation: "10 percent higher debt-to-GDP ratios are associated with 0.2 percent lower future growth over 5 years."
  - Number of observations: 923.
- Methodological notes:
  - Regressions control for year and country fixed effects.
  - Binned scatterplots constructed by grouping x-residuals into 50 equal-sized bins and plotting mean y within each bin holding controls constant.

*Source: wpiea2019101 - References*

### References .............................................................................................................

### References

### I. Introduction
- Public debt is an important tool of economic policy, used to:
  - deal with negative shocks,
  - undertake countercyclical fiscal policy,
  - finance exceptionally large expenditures (e.g., public infrastructure).
- Examples and observed movements:
  - United States: American Recovery and Reinvestment Act in 2009 was USD 831 billion—about 5.5 percent of GDP.
  - US government debt rose from 64 percent of GDP in 2007 to above 100 percent in 2012.
  - Advanced economies: average debt-to-GDP rose from about 60 percent in 2007 to over 90 percent in 2016.
  - China: discretionary stimulus spending of more than 6 percent of GDP; public debt increased from 29 percent of GDP in 2007 to 44 percent of GDP by 2016.
- Risks noted:
  - Large debts may limit future capacity to stabilize the business cycle (limited fiscal space).
  - Large debts may impair economic growth via crowding out private investment or increasing uncertainty about future tax and inflation rates.
- Paper structure:
  - Section 2: why governments borrow (good reasons).
  - Section 3: political failures and bad motives to borrow.
  - Section 4: link between public debt and economic growth.
  - Section 5: conclusion.

### II. Good Motives to Borrow
- Broad rationale:
  - Budget deficits act as a buffer to delink spending and revenues over time.
  - Tax smoothing: run deficits during periods of exceptionally high spending to finance expenditures with future tax revenues.
  - Applicable cases include wars, natural disasters, recessions (to smooth the business cycle), large public investment projects, and temporary accommodation for structural reforms.

#### A. The Logic of Tax-Smoothing
- Fundamental points and citations:
  - Principle: if taxes are distortionary with convex cost, smoothing taxes over time minimizes total distortionary cost (Gallatin (1807); Barro (1979)).
  - Barro (1979) formalized the concept: equalizing marginal cost of levying taxes over time implies a constant tax rate; deficits in bad times, surpluses in good times.
- Implications and nuances:
  - Tax smoothing applies to temporary changes in spending; for permanent changes, expected change is zero and government should run neither deficit nor surplus.
  - Governments may "save for a rainy day" by accumulating assets or paying down debt to hedge unexpected shocks (Aiyagari et al. 2002; Bhandari et al. 2016 suggest small positive long-run net asset position).
  - Optimal portfolio choice: hold debt and financial assets to minimize risk of altering tax rates across time or states of nature (Bohn, 1990; Barro, 1995; other literature cited).
- Three further points emphasized:
  1. Domestic vs. foreign borrowing:
     - Domestic borrowing does not increase the economy's resource envelope for given GDP unless borrowing from foreigners.
     - Borrowing from foreigners expands the real resource constraint and is associated with current account deficits.
  2. Debt as dead-weight loss:
     - Issued debt represents an economic loss equal to present value of distortionary taxes required to repay it.
     - Higher inherited debt raises taxes required to service it and, if taxes fall on factors complementary to public capital productivity, governments optimally undertake less public investment (Ostry, Ghosh, and Espinoza, 2015).
  3. Cost of borrowing matters:
     - For a given spending shock, governments facing higher interest rates will borrow less.
     - Tax-smoothing still holds but higher interest rates raise permanent tax levels needed to service accumulated debt, reducing optimal borrowing.
- Empirical observations and correlations:
  - G7 average debt-to-GDP moved from about 40 percent in the 1970-1980 period to over 80 percent in 2007 (Figure 1).
  - Simple regression (country- and year-fixed effects) in advanced economies: a one percentage point change in the debt-to-GDP ratio is associated with a 0.04 percentage point increase in public investment (Figure 2) — implying that typically only 4 percent of debt issuance is used to finance public investment projects.
  - In emerging and developing economies, the correlation between public debt and public investment is negative but not statistically significant.
  - Heterogeneity: Bacchiocchi, Borghi, and Missale (2011) find negative correlation between debt and public investment in high-debt countries and positive correlation in low-debt countries.

#### B. Keynesian Demand Stimulus
- Countercyclical fiscal policy context:
  - Countercyclical fiscal policy during recessions fits tax-smoothing logic and adds objective of influencing output.
  - Monetary and fiscal policy are standard tools to stabilize the business cycle; in open economies with fixed exchange rates and open capital account, monetary policy may be ineffective, leaving fiscal policy as the available tool (Mundell-Fleming).
  - The IS-LM model provides intuition: fiscal changes stabilize aggregate demand by offsetting private spending changes; New Keynesian models validate IS-LM intuition in dynamic frameworks (Beetsma and Jensen, 2005).
- Policy implication:
  - This intuition underlies most policy discussions on need for countercyclical policy (International Monetary Fund, 2008).

*Italic source attribution: wpiea2019101 - References*

### 17.      In discussing output stabilization, there is a sense in which fiscal policy and monetary

### 17. In discussing output stabilization, there is a sense in which fiscal policy and monetary

### Output stabilization: monetary versus fiscal policy
- Fiscal policy and monetary policy can be seen as substitutes in stabilization models.
- Monetary policy traditionally viewed as quicker and less subject to political interference than fiscal policy.
- Automatic stabilizers (endogenous fiscal policy changes) differ from discretionary fiscal policy and are typically seen as superior to discretionary policy (Taylor, 2000).
- Instances exist where monetary policy cannot achieve first-best outcomes even with a flexible exchange rate:
  - Monetary policy constrained by the zero lower bound on interest rates shifts the stabilization burden to fiscal policy (Eggertsson and Woodford, 2004).
  - When the government is not indifferent to the exchange rate (or its external balance), monetary policy has traditionally been “assigned” to the external objective and fiscal policy to the internal objective (i.e., minimizing the output gap and maintaining full employment).
  - In the presence of more than one distortion (not just price rigidity), monetary policy alone may not suffice and fiscal policy can help achieve first best (Blanchard and Gali, 2010).
- Caveat: countercyclical fiscal policy can be problematic when countries face high default risk (Neumeyer and Perri, 2005; Hatchondo, Martinez and Roch, 2017).

### Empirical patterns in fiscal cyclicality
- Evidence of procyclical fiscal policy is stronger among emerging markets and Latin American economies (Gavin and Perotti, 1997; Kaminsky, Reinhart, and Végh, 2004).
  - Possible causes: limited scope for deficit financing during downturns; lack of fiscal discipline during upswings due to political economy considerations.
  - Recent evidence post-Global Financial Crisis is more encouraging (Frankel, Vegh, and Vuletin, 2013).
- OECD/European economies: mixed evidence.
  - Most times fiscal policy is countercyclical but can turn procyclical (e.g., recent fiscal consolidations in European countries) (Égert, 2012; Fatás, 2018).
  - Real-time judgment of the output gap is difficult; after financial crises, potential output trajectory may change.
  - Procyclical policy leads to higher output volatility and lower growth (Aghion et al. 2007).

### Measuring countercyclical fiscal policy and link to debt
- A measure of the fiscal policy stance: change in the inflation-adjusted budget balance as a ratio to GDP (Blanchard, 1993).
  - Rationale: spending affects aggregate demand; taxes stabilize disposable income and private spending; combined effect captured by budget balance change.
- Direct connection: countercyclical fiscal policy implies running deficits when growth is below trend, leading to debt accumulation.

### Automatic vs. discretionary changes in the budget balance
- Distinction important though aggregate demand cares about the overall balance.
- Automatic stabilizers:
  - Changes in the budget balance resulting from pre-existing tax or spending laws not decided in response to current conditions.
  - Stronger automatic stabilizers come from rules that generate larger swings in the budget balance.
  - Major source in advanced economies: acyclicality of public spending (i.e., maintaining spending constant when GDP falls increases deficits).
  - Magnitude of automatic stabilizers proportional to size of government: larger governments = stronger stabilizers and larger deficits during downturns.
  - Empirical finding: majority of automatic stabilizers among advanced economies comes from this acyclicality effect (Fatás and Mihov, 2012).
  - Additional contributions: spending that automatically increases during downturns; taxes with elasticity larger than one.
- Discretionary fiscal policy:
  - Contributes to deficits and debt accumulation during downturns.
  - Evidence that discretionary policy is used more aggressively where automatic stabilizers are weakest (smaller government size), highlighting substitutability between automatic stabilizers and discretionary policy (Fatás, 2009).

### From cyclical deficits to accumulation of debt
- If governments plan for the right balance over the business cycle, cyclical fiscal behavior should not affect long-run debt.
  - Debt may rise during below-average growth and fall when growth is above-average.
- Observation: public debt levels were increasing in many advanced economies even before the Global Financial Crisis and Great Recession.
- Two hypotheses for asymmetries causing debt drift:
  1. Political asymmetry: governments apply countercyclical policies during recessions but fail to follow symmetric consolidation during expansions (evidence in some countries; see Fatás and Mihov (2010); Alesina, Campante, and Tabellini (2008)). Political distortions discussed in next section.
  2. Forecast bias asymmetry: excessive optimism during strong growth leads to overly optimistic potential output forecasts and planning, producing procyclical fiscal policy in good times.
     - Empirical support: estimates of potential output and growth rates tend to be highly procyclical, leading to excessive expansionary fiscal policy in good times (Mc Morrow, Roeger, and Vandermeulen, 2017).
     - Example: December 28, 2000, President Clinton announced the US was on course to eliminate government debt within the following 10 years; scenario did not foresee the 2001 and 2008 recessions (nor Afghanistan and Iraq Wars).
- Additional asymmetries that produce net debt increase:
  - Interaction between procyclical GDP forecasts and political economy: procyclical optimism during expansions may be acted upon while procyclical pessimism during recessions is ignored, producing bias toward higher debt.
  - Asymmetric macro response: fiscal policy multipliers tend to be larger during recessions than booms (Freedman et al., 2010; Auerbach and Gorodnichenko, 2013; Jordà and Taylor, 2016).
    - Procyclical fiscal policy thus causes more damage to GDP during recessions.
    - Worst-case: hysteresis (permanent effects of cyclical shocks) can make negative GDP effects permanent, validating pessimistic forecasts and potentially increasing debt-to-GDP ratios—a phenomenon termed self-defeating fiscal consolidations (Fatás, 2018; DeLong and Summers, 2012).
    - Policy implication in these scenarios: a more aggressive policy (larger deficits during crises) is the solution, assuming no borrowing constraints.
- Crisis-related debt accumulation:
  - Debt accumulation in crises also stems from government support to repair weak financial and banking systems; these supports can change government debt levels as much or more than demand-supporting fiscal measures (International Monetary Fund, 2015; Campos, Jaimovich, and Panizza, 2006).
  - If fiscal projections omit occasional financial system support, they will be overly optimistic; once such events occur, debt levels rise and tax-smoothing logic implies higher debt will remain for a long time as adjustment is spread over many years (Ostry, Ghosh, and Espinoza, 2015).
  - Empirical statistic: median cost of direct government intervention in the banking sector over 1970–2011 amounted to about 7 percent of GDP; factoring in indirect fiscal costs raises the impact of banking crises to 12 percent of GDP (International Monetary Fund, 2015).

### Long-term investment, deficits, and welfare implications
- Lumpy government investment projects align with the tax-smoothing argument favoring deficit finance.
  - Distinction from government consumption: governments acquire assets that deliver future services, strengthening the case for debt financing.
- Welfare effects depend on social returns of projects; risk of "white elephant" projects if social benefits are nebulous.
- Limited absorptive capacity can constrain growth dividends from public investment, especially during rapid public investment acceleration, increasing likelihood of later debt-servicing difficulties (Presbitero, 2018).
- Government acquisition of financial assets via banking sector recapitalization:
  - When recapitalizing banks, governments acquire equity stakes financed by issuing debt (recapitalization often takes the form of a government bond).
  - Justifications: stabilizing financial sector reduces GDP decline and raises tax revenues; assets may be undervalued during panic and could deliver returns that offset debt costs.
- Deficits can also be justified to support structural reforms:
  - Political economy of reforms makes them hard to implement in democratic settings due to uncertain distribution of winners and losers (Fernandez and Rodrik, 1991).
  - Spending or tax measures that make benefits immediate can secure support and credibility for reforms that yield social/financial returns in the long run (Banerji et al., 2017).

### Asset management and government debt as a safe asset
- Consideration of the asset side of government balance sheets is key to understanding debt.
- Governments may hold foreign exchange reserves while issuing external debt to provide FX liquidity in normal times or during sudden stops, export shortfalls, or terms-of-trade shocks.
- Public debt can be issued not solely to meet borrowing needs but to provide financial markets with risk-free instruments; government debt markets historically aided financial market development (e.g., extending yield curves, providing benchmarks).
  - Moderate levels of non-inflationary government debt have a positive overall impact on economic growth (Abbas and Christensen, 2010).
  - In models, greater availability of government bonds reduces output decline during crises (Gorton and Ordoñez, 2013).
- Internationally, safe assets are associated with major reserve currencies (US dollar, Euro, Yen).
  - The Global Financial Crisis produced a shortage of global safe assets due to flight-to-safety and several sovereign downgrades, with notable consequences (Caballero, Farhi, and Gourinchas, 2008; Brunnermeier et al., 2017; Gourinchas and Jeanne, 2012).

*Source: wpiea2019101*

### 39.      A final potential argument for the issuance of government debt is the possibility that

### 39.      A final potential argument for the issuance of government debt is the possibility that 

### Dynamic inefficiency and the rationale for issuing debt
- Government debt can serve as a vehicle to transfer wealth across generations when the private sector cannot optimally provide such vehicles.
- In this environment, issuing additional government debt can be sustainable and optimal (Blanchard, 1985, 2019).
- Criterion for dynamic inefficiency: the rate of return of an economy must be below its growth rate. What matters is the rate of return on capital, not just interest rates on government debt.
- Empirical evidence and interpretation:
  - Abel et al. (1989) provided strong evidence for 6 advanced economies that the criterion for dynamic inefficiency was not met.
  - Recent decades have seen substantial reductions in real interest rates on safe assets, suggesting the question should be revisited.
  - Geerolf (2017) concludes that dynamic inefficiency cannot be ruled out for several advanced economies.
  - Whether the evidence is sufficiently compelling to warrant clear policy recommendations in some countries remains an open question (Blanchard and Summers, 2017; Blanchard, 2019).

### III. BAD REASONS TO ISSUE DEBT — overview of limits to borrowing rationale
- Good reasons to borrow: cyclical stabilization, exceptional events (war, natural disasters, financial crises), and financing large investment projects.
- Limits and cautions:
  - Cyclical stabilization should not lead to a steady accumulation of debt; deficits in bad times should be offset by surpluses in expansions.
  - There is only limited evidence of a link between public debt accumulation and surges of public investment (Figure 2 referenced).
  - Overborrowing: when government borrows more than socially optimal. The social marginal cost of additional debt (principal, interest repayment, and any externality) should equal the social return of the additional debt-financed expenditure.
  - Yared (2018) suggests recent public debt accumulation is due to overborrowing driven by political distortions that bias toward present consumption.

### A. Why Do Countries Overborrow?
- Shift from normative to positive theories of public debt: policymakers may not maximize social welfare; model politicians as self-interested agents (Alesina and Tabellini, 1992).
- Four potential sources of excessive debt accumulation:
  - (i) political budget cycles and rent seeking;
  - (ii) intergenerational transfers;
  - (iii) strategic manipulation;
  - (iv) common pool problems.
- Note: Some politicians and policymakers are well-intentioned; literature models deviations as self-interested behavior.

#### Political Budget Cycles and Rent Seeking
- Politicians may cut taxes and increase spending to boost reelection chances; requires some form of "fiscal illusion" among voters.
- Asymmetric application of Keynesian stabilization policies amplifies fiscal illusion: deficits in recessions, reluctance to run surpluses in expansions (Buchanan and Wagner, 1977).
- Models do not require irrational voters: e.g., Rogoff and Sibert (1988) show political business cycles can arise with rational but imperfectly informed voters as signaling mechanisms.
- Politicians may remain in power for ego-rents, policy implementation, or resource extraction:
  - Yared (2010): higher debt can reduce incentives for rent extraction by a politician, making behavior closer to a social planner.

#### Intergenerational Transfers
- Individuals can leave positive bequests; negative private bequests are unenforceable, but public debt can be used to effect negative bequests (redistribute from future to current generations).
- Cukierman and Meltzer (1989): overlapping generations model with two types—those wanting positive bequests and bequest-constrained individuals who prefer more debt. Debt level depends on the bequest constraint faced by the median voter.
- Other models:
  - Tabellini (1991): defaultable debt with wealthy and poor voters—higher debt can create incentives to repay via inter- and intragenerational transfers.
  - Song et al. (2012): young and old have different preferences; debt level determined by preferences and political power of groups.
  - Jackson and Yariv (2015): if the old care less about the future, aggregation can cause present bias.
  - Yared (2018): theory consistent with positive cross-country correlation between growth rate of public debt and aging of the population; standard social planner model would predict the opposite.

#### Strategic Manipulation
- Historical example: President Reagan (February 18, 1981) noted public debt approaching $1 trillion (total US Federal debt; Federal debt held by the public was about $770 million or 25 percent of GDP). Eight years later federal debt held by the public had surpassed $2.1 trillion (a 100 percent increase in real terms) and reached 39 percent of GDP.
- Persson and Svensson (1989): left-of-center parties may run surpluses so right-wing successors inherit low debt and are constrained; right-wing parties may increase debt so left-wing successors are constrained.
- Alesina and Tabellini (1990): parties accumulate debt to constrain future governments; governments sure to stay in power behave like social planners and issue no debt; governments with low probability of reappointment will overborrow.
- Strategic debt use: debt is a state variable used to influence successor governments. Weakness: these models often rule out output shocks and tax-smoothing motives, complicating empirical predictions.

#### Common Pool Problems
- Common pool problems arise from externalities: private benefit of public expenditure differs from social marginal cost (Olson, 1965; Ostrom, 1990).
- Concentrated interests amplify the problem: small benefiting groups lobby strongly; larger dispersed payers have weaker incentives to oppose.
- Empirical manifestation:
  - Mauro and Villafuerte (2013): almost 90 percent of fiscal adjustment plans by EU countries envisaged large spending cuts partly compensated by lower taxes; expenditure cuts often did not materialize, producing smaller surpluses or temporary revenue measures.
  - Expenditure-based fiscal adjustments are preferable and more likely to be long-lasting than revenue-based ones, yet hard to implement due to common pool dynamics.
- Institutional context matters:
  - Models of pork-barrel and legislative common pool behavior (Weingast, Shepsle, and Johnse, 1981; Baron and Ferejohn, 1989).
  - Interaction between Ministry of Finance and line ministries can matter; hierarchical budgeting (Finance sets overall envelope, then line ministries allocate) may reduce excessive spending (Alesina and Perotti, 1996).
- Common pool problems can lead to overspending but not necessarily deficits if matched by higher taxes; however, weak property rights and fears of appropriation can lead groups to demand large transfers and push government to borrowing limits (Tornell and Lane, 1999; Velasco, 2000).
- Political turnover amplifies common pool problems: parties uncertain of future power are more likely to overspend (Aguiar and Amador, 2011).
- Empirical support: deficits tend to be larger in countries with deeper political cleavages and party fractionalization (Woo, 2003).
- Battaglini and Coate (2008): adding uncertainty to dynamic common pool models generates two forces:
  - Self-insurance motive: policymakers accumulate assets to insure against future shocks (Aiyagari et al., 2002).
  - Political distortion: policymakers accumulate debt to constrain future policymakers (strategic models).
  - When debt is low, political distortions dominate and governments overborrow; as debt increases, self-insurance becomes more important and policy converges toward social planner choices (with higher equilibrium debt).

### B. Controlling Overborrowing
- (Section begins in the source but content beyond the header is not included in this excerpt.)

*Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019101.pdf*

### 62.      The economics literature identifies three possible avenues to limit overborrowing.

### 62.      The economics literature identifies three possible avenues to limit overborrowing.

### Electoral Systems
- Three avenues to limit overborrowing: the electoral system, fiscal rules, and budgetary institutions.
- Battaglini (2010) (and Battaglini (2014) dynamics) predicts that proportional electoral systems suffer from a deficit bias relative to majoritarian systems.
- Empirical literature: democracies with a proportional electoral system accumulate more debt than democracies with a majoritarian system (examples: Roubini and Sachs, 1989; Grilli et al., 1991).
- Literature comparing presidential and parliamentary systems:
  - Presidential democracies tend to have smaller governments than parliamentary democracies.
  - Within parliamentary democracies, majoritarian systems have smaller governments than proportional systems.
  - In parliamentary democracies, increases in government spending during recessions are less likely to be reversed during expansions (Persson and Tabellini, 2003, 2004), potentially creating a ratchet effect and a deficit bias.

### Fiscal Rules
- Fiscal rules aim to address time inconsistency and limit debt accumulation by imposing an upper limit on budget deficits.
- Trade-off for governments: constraint on current action versus constraint on successor governments.
- Adoption timeline: mid-1990s fewer than 20 countries had national or international fiscal rules; now nearly 100 countries adhere to some type of fiscal rule.
- Types and trade-offs:
  - Balanced-budget rule (zero deficits in every period) is the most extreme and may reduce welfare by limiting countercyclical policy or tax smoothing.
  - Cyclically balanced rules address countercyclicality at the cost of transparency.
  - Yared (2018) surveys tradeoffs: role of public information, degree of enforcement (including escape clauses), costs and benefits of rules based on specific targets (total or primary deficits) versus rules concentrating on policy instruments (such as spending).
- Empirical evidence (mixed):
  - Debrun et al. (2008) and Bergman et al. (2016) find fiscal rules significantly limit budget deficits in European countries.
  - Von Hagen (2006) suggests Maastricht Treaty fiscal rules did not constrain the largest euro-area countries.
  - Identification challenge: establishing causal effects beyond correlations (Heinemann et al., 2018).
  - Caselli and Wingender (2018): Growth and Stability Pact led to a bunching of fiscal deficits around the 3 percent Maastricht deficit ceiling.

### Budgetary Institutions
- Budget preparation involves multiple government players and executive-legislative interaction; institutions regulating preparation and transparency impact fiscal outcomes.
- Hierarchical vs. collegial budget rules:
  - Hierarchical rules give more power to the ministry of finance, mitigate common pool problems, and are associated with smaller deficits and debt accumulation.
  - Collegial rules are more inclusive, give more power to spending ministries, and allow legislative amendment — trade-off with democratic accountability.
- Transparency matters:
  - Imperfect information can produce political business cycles (Rogoff and Sibert, 1988).
  - Politicians seeking to overborrow have incentives to window-dress budgets, especially when corrupt (Milesi-Ferretti, 1997).
  - Budget manipulation strategies include off-budget items and overoptimistic projections on the economy or policy effects on revenues/expenditures.
- Models and empirical findings:
  - Beetsma et al. (2017): transparent budgets mitigate incentives to overborrow.
  - Empirical literature: fiscal transparency is associated with lower public debt in advanced, emerging market, and low-income countries (Alt and Lassen, 2006; Alesina et al., 1999; Dabla-Norris et al., 2010).

### C. The Unexplained Part of Public Debt
- Debt accumulation equals the sum of past deficits plus an unexplained residual (stock-flow reconciliation or adjustment); this residual can be very large.
- Case examples:
  - Uruguay: end-2001 net debt-to-GDP ratio 35 percent (gross debt 55 percent of GDP); end-2002 net debt-to-GDP 76 percent (gross debt 106 percent). Over 2002 total budget deficit was 3.7 percent of GDP; the growth in debt was 17 percentage points of GDP higher than the deficit.
  - Argentina: end-2001 gross public debt 49 percent of GDP; end-2002 gross public debt 152 percent of GDP. 2002 public deficit recorded just above 2 percent. Debt grew 101 percent percentage points of GDP more than the deficit.
- Broader evidence:
  - Campos et al. (2006): sudden debt explosions unexplained by recorded deficits are not limited to a few cases.
  - Cafiso (2012): phenomenon not limited to emerging markets; stock-flow reconciliation accounted for nearly one third of public debt growth in EU countries over 2008–10.
- Main drivers of the unexplained part of debt:
  - Balance sheet effects linked to foreign currency debt (currency depreciations with dollar debt explain Argentina and Uruguay cases; Eichengreen et al., 2005).
  - Banking crises (Amaglobeli et al., 2017).
  - Hidden deficits and misreporting of public expenditure (Weber, 2012: more transparent budgets correlated with lower stock-flow adjustments).
  - Contingent liabilities linked to implicit subsidies and public guarantees.
- Mitigation policies:
  - Safer debt structures can reduce balance-sheet linked risks.
  - Contingent debt instruments (e.g., GDP indexed bonds) could reduce fiscal costs in crises (Borensztein and Mauro, 2004).
- Reasons for underuse of safer instruments:
  - Local currency or indexed debt tends to be more expensive than fair insurance would predict (market failures and incentives linked to local currency debt—Calvo, 1988; Tirole, 2003).
  - Political failures: domestic currency debt and contingent instruments act as insurance whose upfront cost is paid now while payoff may accrue to successors, making them unattractive to incumbent policymakers (Borensztein et al., 2006).

### IV. Debt, Growth and Investment
- High government debt can have adverse effects by:
  - Limiting capacity for counter-cyclical fiscal policy.
  - Reducing private investment through crowding-out, tightening credit constraints, raising expectations of future distortionary taxation, or increasing uncertainty.
- Public borrowing can be growth-supporting if directed to productive investment (e.g., infrastructure) or to stimulate aggregate demand, but must balance future debt sustainability—especially pressing for many developing countries needing large investment to meet 2030 Sustainable Development Agenda targets.
- Mechanisms and evidence:
  - If Ricardian Equivalence fails, decreased public saving from debt accumulation is not fully offset by higher private saving, lowering capital stock, raising interest rates, and reducing growth (Diamond, 1965; Blanchard, 1985).
  - Panizza and Presbitero (2013) using Elmendorf and Mankiw (1999) calculations: effect not quantitatively large in their estimates.
  - Crowding-out could be large if government debt tightens credit constraints via financial frictions (Broner, Erce, Martin, and Ventura, 2014).
  - European sovereign debt crisis evidence: expansion of government debt held by banking sector during crisis crowds out private lending (Altavilla, Pagano and Simonelli 2017; Becker and Ivashina 2018).
  - High public debt may increase uncertainty about future taxes, lead to expectations of confiscation (via inflation or financial repression; Cochrane 2011), or precipitate financial crises.
  - High public debt could signal sustainability concerns and translate into higher sovereign yield spreads (Codogno et al., 2003; Laubach, 2009; Baum et al., 2013), transmitted to the private sector.
  - Debt overhang literature (Krugman, 1988; Sachs, 1989; Aguiar, Amador and Gopinath, 2009) suggests a debt level at which growth effects make debt relief beneficial to debtors and creditors.
- Policy reactions and implications:
  - Governments typically respond to rising public debt with austerity—smaller deficits or larger surpluses (Bohn, 1998; Mendoza and Ostry 2008; Ghosh et al., 2013; Mauro et al., 2015).
  - High public debt can hinder countercyclical policy and may cause self-defeating austerity, increasing output volatility and reducing growth.
  - The relationship between debt levels and ability to conduct countercyclical policy depends on debt composition (Hausmann and Panizza, 2011; De Grauwe, 2011); different debt structures imply problems may arise at very different debt levels.

### A. What Do the Data Say?
- Post-Global Financial Crisis rise in public debt spurred literature on growth effects of public debt.
- Reinhart and Rogoff (2010a) (20 advanced economies, 1946-2009) finding: average and median growth substantially lower when public debt surpasses 90 percent of GDP.
  - Using Global Debt Dataset (Mbaye et al., 2018), average (median) growth declines:
    - From 3.7 percent in country-year pairs with debt-to-GDP < 30 percent,
    - To 2.6 (2.7) percent when debt ratio is between 30 and 60 percent,
    - To 1.2 (1.6) percent when debt surpasses 90 percent of GDP.
  - In a larger sample of 119 low- and middle-income countries, average growth declines from 4.4 percent for low-debt countries to 2.6 percent in high-debt (above 90 percent) countries.
- Subsequent research investigates robustness to controls, causal inference, and non-linearities.
- Overall evidence: strong negative correlation between public debt and future economic growth.
  - Authors corroborate negative correlation: debt-to-GDP in year t strongly negatively correlated with real GDP growth between t and t+5 controlling for year- and country-fixed effects.
  - Correlation does not prove causation: could reflect omitted variables or reverse causality (slow growth drives debt increases).
  - Slow growth contributes to rising debt via denominator effects and lower primary surplus; absent policy measures, low growth constrains revenues while expenditures rise with inflation, leading to larger deficits and rising debt (Mauro and Zilinsky, 2016).
  - Permanent growth slowdowns may be misinterpreted as temporary, prompting expansionary fiscal policy that increases debt (Mauro et al., 2015).
- Identification strategies and micro evidence:
  - Instrumental variable/natural experiment required to isolate exogenous changes in public debt; persistent debt ratios complicate standard instruments.
  - Panizza and Presbitero (2014): use valuation effects of foreign currency debt as instrument and find no effect of public debt on future growth; instrument weak in advanced-country samples due to small share of foreign-currency debt.
  - Micro-level studies:
    - Huang et al. (2017, 2018): matching firm-level balance sheets with public debt data across 69 countries or local government debt across 270 Chinese cities shows government debt tightens financing constraints for private manufacturing firms.
    - Croce et al. (2019): in the US, higher government debt increases cost of capital and negatively affects investment by R&D-intensive firms.
  - Trade-off: micro studies improve identification for specific channels but may obscure broader macro links; e.g., debt could increase investment for all firms but less so for credit-constrained firms as per Huang et al. (2017, 2018).

*Source: wpiea2019101 - 62.      The economics literature identifies three possible avenues to limit overborrowing.*

### 87.      Besides studying the average correlation between debt and growth, the economics

### wpiea2019101 - 87.      Besides studying the average correlation between debt and growth, the economics

### Non-linearities, thresholds, and statistical challenges
- Literature explores non-linearities and threshold effects in the debt–growth relationship; some analyses emphasize the debt trajectory over the level (Pescatori et al., 2014; Chudik et al. 2017) and link public debt to debt sustainability and market access (Bassanetti et al., 2018).
- The notion of a common threshold (often taken to be 90 percent) has become popular.
- Assessing non-linearities is complicated by:
  - lack of statistical power due to the limited number of observations above the relevant threshold;
  - potential influence of parametric assumptions and a few outliers.
- Cross-country studies often impose common coefficients and thresholds, but data show substantial heterogeneity, especially in large samples pooling developing and emerging economies (Eberhardt and Presbitero, 2015; Chudik et al. 2017).
- Example evidence:
  - A non-parametric regression on a sample of 20 advanced economies over the period 1960-2016 shows:
    - (i) the average negative correlation between debt and future (5 years ahead) growth hides a large degree of heterogeneity across countries, and
    - (ii) while the relationship is nonlinear there is no common threshold beyond which an increase in debt is associated with a growth slowdown.
- Confidence intervals and robustness:
  - Checherita-Westphal and Rother (2012) report a turning point confidence interval of 49 to 119 percent of GDP.
  - Woo and Kumar (2015) find mixed evidence: in 2 (out of 4) specifications the debt coefficient is negative and significant when larger than 90 percent, but that coefficient is lower than (equal to) that for debt between 30 and 90 percent in OLS (GMM) estimates—hence they cannot statistically establish a stronger correlation above 90 percent.

### Country-specific thresholds and explanatory factors
- The level at which public debt becomes “too high” depends on country characteristics:
  - Historical credit and inflation experience and classifications into clubs and “debt intolerance” regions (Reinhart, Rogoff and Savastano, 2003).
  - Historical track record of fiscal adjustment affects debt-sustainability thresholds (Ghosh et al., 2013).
  - Debt composition matters (Eichengreen, Hausmann, and Panizza, 2005).
  - Institutional quality: Kourtellos et al. (2013) find that only when institutions are below a certain level does higher debt translate into lower GDP growth.
- Mechanisms linking institutions and debt:
  - Countries with low-quality institutions may be more prone to political budget cycles and overborrowing.
  - Such countries may finance consumption rather than productive investment, raising debt and reducing growth.
- Historical studies:
  - Esteve and Tamarit (1851–2013, Spain) find some support for a negative relationship but no clear threshold.
  - Balassone et al. (Italy since 1861) find stronger negative effects when debt exceeds 100 percent of GDP.
  - Eberhardt (using data over more than two centuries for Great Britain, Japan, Sweden, and the United States) finds no evidence for any long-run non-linear relationship between debt and growth, criticizing time-series approaches for implying causation where none may exist.

### Debt structure, measurement, and heterogeneity
- “Not all debts are equal”: the impact of debt depends on:
  - what the debt was used for;
  - who holds government debt;
  - currency composition;
  - maturity profile.
- Data limitations:
  - Applied researchers typically observe only the level of government debt, not its structure.
  - Lack of comprehensive data impedes understanding of heterogeneity in results (Eberhardt and Presbitero, 2015; Chudik et al., 2017).
  - Abbas et al. (2014) is a notable exception, collecting historical structure data for 13 advanced economies.
- Examples of structural differences and vulnerabilities:
  - Share of public debt held by non-residents: Japan around 5-7 percent; Italy close to 40 percent; even higher in Ireland (Abbas et al., 2014).
  - Greater foreign creditor exposure or shorter maturities increase vulnerabilities and can hamper growth.
- Gross versus net and implicit debt:
  - Gross debt measures ignore financial assets held by the government and cross-holdings; net debt can be significantly smaller in some countries.
  - Implicit liabilities (pensions), local government debt, and state-owned enterprise debt can make public-sector debt much larger than official numbers suggest.
  - Capacity constraints and data gaps hinder comparable measures of net and implicit debts across countries (Panizza and Presbitero, 2013).

### Causality, welfare implications, and policy trade-offs
- Empirical summary:
  - In advanced economies there is a negative correlation between public debt and subsequent economic growth, but no convincing evidence of causality; high debt and low growth may both reflect a weak macroeconomic framework.
  - Cross-country averages mask complex, country-specific drivers; no clear evidence of a common tipping point beyond which additional debt uniformly harms growth.
  - The “endogeneity conundrum has not been fully resolved.”
- Welfare and optimal policy on existing debt:
  - Even if “debt is bad for growth” causally, repaying inherited debt is not necessarily welfare-improving (Ostry, Ghosh, and Espinoza, 2015).
  - Tax-smoothing logic: paying down debt today imposes distortionary tax costs now; the government discounts future costs at the market interest rate (1+r) and is indifferent between paying down debt today or tomorrow. The steady-state result that it may be optimal to “just live with the inherited debt” can hold even out of steady-state for iso-elastic utility functions.
  - The government faces a trade-off: increase taxes to service growing debt, or decrease taxes after repayment—both can violate tax-smoothing and increase distortionary costs.
- Political economy drivers of overborrowing:
  - Counter-cyclical fiscal policy implementation delays and forecast biases.
  - Political business cycles aimed at re-election incentives.
  - Strategic manipulation to constrain successors by running up debt.
  - Common pool problems where private benefit from spending exceeds the social marginal cost of funding it.
  - When inflation is unavailable as a “safety-valve” (e.g., under a currency board or monetary union), competing demands are often resolved through rising public debt.
- Policy avenues to limit overborrowing in democracies:
  - Electoral systems;
  - Fiscal rules;
  - Budgetary institutions.
  - Effectiveness depends on country circumstances and involves trade-offs between flexibility to respond to shocks and disciplining policymakers from excessive borrowing.

*Source: wpiea2019101 - 87.*

### REFERENCES

### REFERENCES

### Major themes covered in the references
- Domestic debt markets, sovereign debt composition, and sovereign debt restructuring
- Fiscal policy design: fiscal rules, budget institutions, fiscal transparency, independent fiscal councils
- Public debt dynamics: debt sustainability, debt overhang, debt thresholds, fiscal adjustments, fiscal fatigue, fiscal space
- Macroeconomic interactions: public investment, fiscal multipliers, monetary–fiscal interactions, liquidity traps
- Political economy of fiscal policy: electoral competition, redistributive politics, fragmented policymaking, common-pool problems
- Financial stability and safe assets: banking crises, sovereign–bank linkages, global safe assets
- Methodological and empirical issues: panel evidence, non-linearities, heterogeneity, lagged explanatory variables, dataset construction

### Data sources and datasets cited
- Global Debt Dataset (Mbaye et al., 2018)
- World Economic Outlook
- World Development Indicators
- Systemic Banking Crises Database (Laeven and Valencia, 2013)
- Global Financial Stability Report (IMF, 2012)
- Fiscal Monitor (IMF, 2018a)
- IMF Working Papers and Staff Papers (multiple entries)

### Notable quantitative findings and figure notes (preserve original wording and numerics)
- Figure 2 (Correlation Between Change in Public Debt and Contemporaneous Public Investment)
  - Regression coefficient on the debt variable: 0.041 (p-value of 0.011). Interpretation given: "a 10 percent increase of the debt-to-GDP ratio is associated with 0.4 percent lower ratio of public investment over GDP."
  - Binned scatterplot construction: starting from the sample of 19 OECD economies (data on general government for New Zealand are not available); x-residuals grouped into 50 equal-sized bins.
  - Number of observations: 899.
- Figure 3 (Government Debt and Growth, Selected Advanced Economies; 1960–2016)
  - Sample includes 20 advanced economies as in Reinhart and Rogoff (2010, Figure 2): Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Japan, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, the United Kingdom, and the United States.
  - Data refer to central government debt, apart from the Netherlands, for which general government data have been used, because of data availability.
- Figure 4 (Government Debt and Growth, Low and Middle Income Countries; 1960–2016)
  - Sample includes 131 low- and middle-income countries.
  - Data refer to central government debt.
- Figure 5 (Government Debt and Subsequent GDP Growth, Selected Advanced Economies; 1960–2016)
  - Regression coefficient on the debt variable: -0.016 (p-value of 0.001). Interpretation given: "10 percent higher debt-to-GDP ratios are associated with 0.2 percent lower future growth over 5 years."
  - Binned scatterplot construction: starting from the sample of 19 OECD economies (data on general government for New Zealand are not available); x-residuals grouped into 50 equal-sized bins.
  - Number of observations: 923.
- Figure 6 (Non-linearities and Heterogeneity in the Debt-Growth Relationship)
  - Solid black line: smoothed values of the locally weighted regression of the annual real GDP growth between t+5 and t against the ratio of general government debt over GDP at time t for the whole sample.
  - Thin lines: smoothed values for single countries.
  - Histogram shows the density distribution of the ratio of general government debt over GDP (x-axis).
  - Sample includes the same 20 advanced economies listed for Figure 3; data refer to central government debt, apart from the Netherlands, for which general government data have been used, because of data availability.

### Representative methodological notes appearing in figure captions
- Regressions control for year and country fixed effects.
- Binned scatterplots are generated by grouping x-residuals into 50 equal-sized bins and plotting mean y within each bin, holding controls constant.
- Outcomes examined include contemporaneous public investment (as a percent of GDP) and subsequent real GDP growth over 5 years (annual real GDP growth between t+5 and t).

*Source: REFERENCES (wpiea2019101 - REFERENCES).*

### Box 1. Balancing Investment Needs with Debt Sustainability, the Case of Ethiopia

### Box 1. Balancing Investment Needs with Debt Sustainability, the Case of Ethiopia

### Context and investment scaling
- The 2030 Sustainable Development Agenda requires a large scale up of investment over a long period; private sector should play a key role but public investment is expected to increase significantly in several countries.
- Ethiopia’s public investment was above 7 percent of GDP in the 2000s and accelerated to about 15 percent of GDP between 2014 and 2017.
- This scale up was funded by external concessional and non-concessional financing (including large Chinese investment flows), and partly facilitated by restrained government consumption, financial repression and an overvalued exchange rate (World Bank, 2016).
- Major projects financed include the Grand Ethiopian Renaissance Dam and the railway connecting Addis Ababa with the port of Doraleh in Djibouti.

### Outcomes and benefits
- Infrastructure expansion contributed to sustained rapid growth: real GDP growth averaged 10 percent annually over the last decade.
- Poverty declined substantially and key public services improved.
- Access to electricity increased from 14 percent in 2005 to 43 percent in 2016.
- The Grand Ethiopian Renaissance Dam is estimated to cost almost USD 5 billion—about 5 percent of GDP—and once completed will be the largest hydroelectric power plant in Africa, supplying energy also to Sudan and Egypt.
- The railway project is expected to significantly reduce trade costs and improve access to global markets for Ethiopian firms.

### Debt dynamics and risks
- The model used to finance infrastructure expansion is showing limits in terms of debt sustainability, crowding out of private credit, and weak external competitiveness due to exchange rate appreciation (World Bank 2016).
- Debt-to-GDP dynamics: declined from 107 percent in 2002 to 38 percent in 2009 (thanks to debt relief), then increased sharply and reached 62 percent in June 2018.
- Despite sustained economic growth, adverse debt dynamics led the IMF and World Bank to assess Ethiopia as at "high risk of debt distress (IMF 2018b)".
- Absorptive capacity constraints could undermine project success rates and reduce the dividend of public investment (Presbitero 2018).
- As a result of these constraints and risks, investment has recently started declining and large external imbalances and the public debt burden are constraining future growth.

### Policy lessons and recommendations
- Even with strong growth and large investment needs, scaling up investment must account for the risks that debt-financed public investment and higher public debt could pose to debt sustainability and future economic growth.
- Diversify financing sources and strengthen absorptive capacity to improve project success rates and maximize returns from public investment.
- Consider the trade-off between rapid public investment expansion and potential crowding out of private credit, loss of external competitiveness, and higher debt distress risk.

### Key statistics and factual points
- Public investment: above 7 percent of GDP in the 2000s; about 15 percent of GDP between 2014 and 2017.
- Grand Ethiopian Renaissance Dam cost estimate: almost USD 5 billion — about 5 percent of GDP.
- Electricity access: 14 percent in 2005; 43 percent in 2016.
- Global Infrastructure Hub estimate of investment shortfall to achieve 2030 Agenda targets for Ethiopia: about USD 285 billion.
- Debt-to-GDP: 107 percent in 2002; 38 percent in 2009; 62 percent in June 2018.
- Real GDP growth: averaging 10 percent annually (last decade).
- IMF and World Bank assessment: Ethiopia is at high risk of debt distress (IMF 2018b).

*Source: Box 1. Balancing Investment Needs with Debt Sustainability, the Case of Ethiopia (wpiea2019101).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019101.pdf_