## EXECUTIVE SUMMARY

## Source details

**Canonical URL:** [EXECUTIVE SUMMARY](https://www.imf.org/-/media/files/publications/sdn/2020/english/sdnea2020002.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/sdn/2020/english/sdnea2020002.pdf.md)
- [Structured JSON version](/-/media/files/publications/sdn/2020/english/sdnea2020002.pdf.json)

---

### Overview
- Research question: Which industries benefit (suffer) during credit booms (busts); what is the long-term impact of credit booms on industrial structure; can sectoral composition help distinguish good booms from bad ones?
- Sample and scope:
  - 55 countries between 1970 and 2014.
  - Industry-level data available for 33 industries, aggregated into 11 sectors.
- Boom definition:
  - Booms identified using the Dell’Ariccia, Igan, and Laeven (2016) methodology based on extraordinary positive deviations of the credit-to-GDP ratio from its backward-looking rolling trend.

### Sectoral findings: who wins, who loses
- Aggregate patterns:
  - Aggregate value-added and employment growth accelerate during credit booms, with substantial industry-level heterogeneity.
- Most procyclical sectors:
  - Construction leads; finance is a distant second.
- Least procyclical sectors:
  - Agriculture and services.
- Sector characteristics associated with greater procyclicality:
  - Less tradable.
  - More labor-intensive.
  - More reliant on external finance.
- Concentration:
  - Most extra value-added and employment growth during booms is concentrated in a few industries—specifically, construction and, as a distant second, finance.
- Boom–bust asymmetry:
  - Sectors that benefit most during booms tend to suffer most during busts, implying limited long-term footprints on industrial composition.

### Construction: uniqueness, predictive role, and economic costs
- Uniqueness:
  - Construction is the only sector that consistently overperforms in bad credit booms.
  - On average, output and employment growth in construction are roughly 3 percentage points higher in bad booms than in good ones.
  - Result holds in advanced economies and emerging markets, is robust to excluding the post-2003 period, and holds when including house (or stock) prices in specifications.
- Predictive power for bad booms:
  - An additional percentage point of value-added growth in construction during a boom raises the probability of the boom being bad by 2 percentage points.
  - An additional percentage point of employment growth in construction during a boom raises the probability of the boom being bad by 5 percentage points.
  - In the sample, long-lasting booms that featured rapid construction growth never ended well (conditional probability = 1.00).
- Predictive power for economic costs of bad booms:
  - One percentage point higher value-added or employment growth in construction during the boom corresponds to nearly a 0.1 percentage point drop in aggregate output growth during the bust.
  - Annual real GDP growth in the three-year period following bad credit booms with a construction boom is –3.3 percent, compared with –1.9 percent for bad credit booms without a construction boom—a difference of almost 1½ percentage points.
  - By contrast, the difference between costs of bad booms with high versus low household credit growth is smaller: annual real GDP growth is lower by ¾ percentage point in the high-household-credit growth case than in the low-household-credit growth case.
- Interpretation:
  - The extra procyclicality of finance may be driven by its link to construction.

### Good vs bad booms: persistence and trade-offs
- Definitions and prevalence:
  - A bad boom is one followed by subpar economic performance or a systemic financial crisis.
  - Of 59 credit booms identified, roughly two-thirds are classified as bad booms (overall probability that a boom is bad = 0.64).
- Persistence:
  - Negative growth effects of bad booms appear mostly transitory: 1 percentage point higher construction value-added growth during a bad boom predicts a 0.1 percentage point drop in annual output growth during the three-year period after the boom; this negative impact disappears over a six-year period.
  - Good booms may have positive, long-lasting impact: 1 percentage point higher construction value-added growth during a good boom predicts a 0.3 percentage point higher annual output growth over both three- and six-year horizons.
- Policy trade-off:
  - Policymakers face a mean-volatility trade-off: action to curb booms may spare the economy a recession or systemic crisis (with negative persistent effects) but may reduce average long-term growth.

### Mechanisms and interpretation
- Misallocation channel:
  - Construction is typically not a high TFP growth sector; excessive expansion of construction during bad booms suggests resource misallocation (labor, capital, credit) as a channel linking booms to subsequent lower TFP and GDP growth.
  - In emerging markets, where household credit is a small share of total credit, misallocation toward corporate/real-estate developers may be especially important.
- Complementarity with earlier channels:
  - Findings add to collateral effects, strategic leverage, capacity constraints, and behavioral biases by emphasizing sectoral resource allocation and construction as an early indicator.
- Other mechanisms highlighted:
  - Tangibility and collateral encourage credit allocation to construction that may not yield high productivity.
  - Labor intensity and low skill requirements can mask structural labor-market problems.
  - Leverage and fire-sale dynamics amplify vulnerability in busts.

### Policy implications and early-warning uses
- Sectoral monitoring:
  - Monitoring disaggregated real-sector activity during booms can help identify bad credit booms and anticipate economic costs more accurately.
  - Rapid, lopsided expansion concentrated in construction is a red flag that could tilt policy toward tightening.
- Practical advantages:
  - Construction activity data are more broadly available across countries than some alternative indicators (e.g., household credit growth or house prices), making it a practical signal.
- Complementarity:
  - Construction monitoring complements other signals (size and duration of boom, household credit growth, house price growth).
- Policy recommendations when rapid construction expansion occurs during a credit boom:
  - Give consideration to tightening the macroeconomic policy stance and/or activating macroprudential tools.
  - Tighten limits on banks’ exposure to real estate developers and other construction firms.
  - Use real activity metrics and indicators such as rising exposure of banks to specific borrowers as additional triggers for action.
- Data and measurement recommendations:
  - Improve indicators of construction sector performance (for example, productivity).
  - Explore composing better measures of economic growth not contaminated by real estate bubbles or unusually rapid construction growth.

### Data, methodology, and robustness
- Industry aggregation:
  - 33 industries aggregated into 11 sectors: agriculture, construction, finance, information, manufacturing, mining, real estate, trade, utilities, other services, and public services.
- Boom identification:
  - Episodes identified using country-specific, path-dependent thresholds and absolute numerical thresholds applied to deviations of the credit-to-GDP ratio from its backward-looking rolling trend.
- Robustness:
  - Key results robust to excluding the global financial crisis (post-2003), to inclusion of house or stock prices, and hold across advanced and emerging market subsamples.

### Key statistics and empirical magnitudes
- Sectoral average cyclical gains:
  - On average, value added expands by roughly 1.2 percentage points more in all sectors during booms relative to tranquil times; employment expands by roughly 0.8 percentage point more.
  - Construction value added and employment grow by almost 4 percentage points more than in other sectors during booms (annual basis), and contract by almost the same amount during busts.
- Appendix Table 2 (growth dynamics in construction):
  - Construction value-added growth during booms exceeds growth in tranquil times by 5 percentage points on an annual average basis.
  - Construction employment growth during booms is 3.6 percentage points above its annual average growth rate in tranquil periods.
  - These differences correspond to roughly a 0.6 standard deviation of construction sector value-added and employment growth.
- Productivity (Table 1: Productivity Growth—adjusted residuals):
  - Tranquil times / Booms / Busts
    - Construction: -0.32 0.52 -0.89
    - All except construction: -0.24 0.62 0.15
  - TFP growth in construction typically lags that in other sectors in booms and tranquil times.
- Comovement regressions (sectoral correlations):
  - Value-added: y = -0.8376x -0.0571, R² = 0.7089
  - Employment: y = -1.0998x + 0.0055, R² = 0.7818
- Conditional probabilities and post-boom outcomes (Table 8 summaries):
  - Typical good boom: lasts about three years, credit-to-GDP ratio growing by 15 percent annually; Real GDP expands at an annual rate 1½ percentage points faster than trend.
  - Typical bad boom: lasts four years, annual credit-to-GDP growth rate reaching 17 percent; Real GDP grows 1¾ percentage points above trend (difference not statistically significant).
  - Summary three-year post–bad-boom average annual real GDP growth examples:
    - High Duration: -3.6; Low Duration: -2.0
    - High Construction Value-Added Growth: -2.9; Low Construction Value-Added Growth: -2.2
    - With Long Booms & High Construction Growth: -5.0; With Long Booms & Low Construction Growth: -2.1
    - With High Household Credit Growth: -3.3; With Low Household Credit Growth: -1.9
- Regression magnitudes on post-boom outcomes (Table 9 examples):
  - Construction Growth x Bad coefficients are negative and significant (for example, -0.466*** in column 1).
  - Quantitative interpretation: 1 additional percentage point of construction value-added growth during a bad boom implies more than a 0.1 percent decline in average sector growth over the three-year post-boom period (sum of coefficients yields approximately –0.1).
  - For six-year horizons, net effects of construction during bad booms become negligible or insignificant, suggesting adverse effects are predominantly medium-term.

### Conclusions and directions for future research
- Construction sector displays unique patterns across booms and busts:
  - Strongest acceleration (deceleration) in value-added and employment growth during booms (busts).
  - The only sector that consistently overperforms during bad booms.
  - Pace of construction activity during the boom phase is a better predictor of economic costs associated with bad booms than other variables identified in previous studies.
- Monitoring construction activity during booms can provide a litmus test for policy action; high-frequency indicators (for example, construction permit applications) could serve as valuable signals.
- Suggested future research directions:
  - Use more granular data to improve understanding of macro-financial dynamics.
  - Introduce construction sector indicators in the growth-at-risk framework.
  - Explore effectiveness and side effects of policy options to curb excessive construction-sector developments.
  - Assess whether the trade-off faced during booms changes over longer horizons as longer time series become available.

*Source: EXECUTIVE SUMMARY and appendices of sdnea2020002 (IMF).*

### EXECUTIVE SUMMARY __________________________________________________________________________ 4

### EXECUTIVE SUMMARY

### Overview
- Research question: Which industries benefit (suffer) during credit booms (busts); what is the long-term impact of credit booms on industrial structure; can sectoral composition help distinguish good booms from bad ones?
- Sample and scope:
  - 55 countries between 1970 and 2014.
  - Industry-level data available for 33 industries, aggregated into 11 sectors.
- Booms defined using the Dell’Ariccia, Igan, and Laeven (2016) methodology based on extraordinary positive deviations of the credit-to-GDP ratio from its backward-looking rolling trend.

### Sectoral Findings: Who Wins, Who Loses
- Aggregate outcomes:
  - Aggregate value-added and employment growth accelerate during credit booms, but industry-level heterogeneity is substantial.
- Most procyclical sectors:
  - Construction and finance grow the most during booms (construction leads; finance is a distant second).
- Least procyclical sectors:
  - Agriculture and services are at the opposite end of the spectrum.
- Sector characteristics associated with greater sensitivity to the credit cycle:
  - Less tradable.
  - More labor-intensive.
  - More reliant on external finance.
- Concentration:
  - Most of the extra value-added and employment growth during booms is concentrated in a few industries—specifically, construction and, as a distant second, finance.
- Boom–bust asymmetry:
  - The sectors that benefit most during booms tend to suffer the most during busts, implying limited long-term footprints on industrial composition.

### Construction: Predictive Role and Economic Costs
- Construction is unique:
  - Construction is the only sector that consistently overperforms in bad credit booms.
  - On average, output and employment growth in construction are roughly 3 percentage points higher in bad booms than in good ones.
  - This result holds in advanced economies and emerging markets, is robust to the exclusion of the post-2003 period, and holds when including house (or stock) prices in specifications.
- Predictive power for bad booms:
  - An additional percentage point of value-added growth in construction during a boom raises the probability of the boom being bad by 2 percentage points.
  - An additional percentage point of employment growth in construction during a boom raises the probability of the boom being bad by 5 percentage points.
  - In the sample, long-lasting booms that featured rapid construction growth never ended well.
- Predictive power for economic costs of bad booms:
  - One percentage point higher value-added or employment growth in construction during the boom corresponds to nearly a 0.1 percentage point drop in aggregate output growth during the bust.
  - Annual real GDP growth in the three-year period following bad credit booms with a construction boom is –3.3 percent, compared with –1.9 percent for bad credit booms without a construction boom—a difference of almost 1½ percentage points.
  - By contrast, the difference between costs of bad booms with high versus low household credit growth is smaller: annual real GDP growth is lower by ¾ percentage point in the high-household-credit growth case than in the low-household-credit growth case.
- Interpretive note:
  - The extra procyclicality of finance may be driven by its link to construction.

### Good vs Bad Booms: Persistence and Trade-offs
- Persistence of effects:
  - Negative growth effects of bad booms appear mostly transitory: a 1 percentage point higher value-added growth in construction during a bad boom predicts a 0.1 percentage point drop in annual output growth during the three-year period after the boom; this negative impact disappears over a six-year period.
  - Good booms may have positive, long-lasting impact: 1 percentage point higher construction value-added growth during a good boom predicts 0.3 percentage point higher annual output growth over both three- and six-year horizons.
- Trade-off facing policymakers:
  - Policy action to curb a boom may spare the economy a recession or a systemic crisis (with negative persistent effects) but may reduce average long-term growth.
  - Policymakers therefore face a mean-volatility trade-off, with different net growth effects depending on country circumstances.

### Mechanisms and Interpretation
- Misallocation channel highlighted:
  - Construction is typically not a high total factor productivity (TFP) growth sector; excessive expansion of construction during bad booms suggests resource misallocation (labor, capital, credit) as a channel linking booms to subsequent lower TFP and GDP growth.
  - In emerging markets, where household credit is a small share of total credit, misallocation toward corporate/real-estate developers may be especially important.
- Complementarity with earlier literature:
  - Findings add to known channels (collateral effects, strategic leverage, capacity constraints, behavioral biases) by emphasizing sectoral resource allocation and the construction sector as an early indicator.

### Policy Implications and Early Warning Uses
- Sectoral monitoring:
  - Monitoring disaggregated real-sector activity during booms can help identify bad credit booms and anticipate their economic costs more accurately.
  - Rapid, lopsided expansion concentrated in construction is a red flag that could tilt policy toward tightening.
- Practical advantages:
  - Construction activity data are more broadly available across countries than some alternative indicators (for example, household credit growth or house prices), making it a practical signal for policymakers.
- Complementarity:
  - Construction monitoring complements other signals (size and duration of boom, household credit growth, house price growth) already identified as predictive of problematic booms.

### Data and Methodology (brief)
- Industry aggregation and data:
  - 33 industries aggregated into 11 sectors: agriculture, construction, finance, information, manufacturing, mining, real estate, trade, utilities, other services, and public services.
- Boom identification:
  - Episodes identified using country-specific, path-dependent thresholds and absolute numerical thresholds applied to deviations of the credit-to-GDP ratio from its backward-looking rolling trend, following Dell’Ariccia, Igan, and Laeven (2016).
- Robustness:
  - Key results robust to excluding the global financial crisis (post-2003), to inclusion of house or stock prices, and hold across advanced and emerging market subsamples.

*Source: EXECUTIVE SUMMARY of sdnea2020002 (IMF).*

### Appendix Table 2.

### Appendix Table 2

### Growth dynamics in construction
- The growth rates are adjusted to purge out country and sector fixed effects to allow comparability across countries and sectors.
- Value-added growth in construction during booms is strong, exceeding growth in tranquil times by 5 percentage points on an annual average basis.
- Employment growth in construction during booms is 3.6 percentage points above its annual average growth rate in tranquil periods.
- These differences correspond to roughly a 0.6 standard deviation of construction sector value-added and employment growth.
- Taken from another perspective, construction value added and employment grow almost twice as fast as finance—which ranks second.

### Productivity and employment
- An interesting question is whether relatively stronger employment growth during booms is associated with faster productivity growth. The answer seems to be no, as TFP growth in construction typically lags that in other sectors not only during booms but in tranquil times as well (Table 1).
- Yet the dramatic growth in employment during booms more than compensates for the slow TFP growth the construction sector typically displays and places construction at the top in terms of value-added growth.
- Note on data limitations and interpretation:
  - Systematic data on labor productivity growth are not available. The coverage of TFP growth series also has significant gaps (especially for emerging market economies), so these findings should be taken with a grain of salt.
  - It is also worth noting that the numbers reported in Table 1 are obtained after purging out country-industry averages and should not be interpreted as productivity growth in tranquil times being negative.
  - The unadjusted TFP growth series has an average (median) of 0.34 (0.25) for all industries at all times and of 0.09 (0.13) in tranquil times.

### Sectoral procyclicality and bust dynamics
- On the flip side, no sector is immune to the bust. All of them contract during this phase, but there is again significant variation across sectors.
- Construction and finance again top the chart during busts (with finance again a distant second, especially when it comes to employment growth), making them the most procyclical sectors.
- The real estate sector shows a strong decline in employment but a much smaller decline in terms of value added.
- Trade, information, and manufacturing also show procyclicality, but primarily in terms of value added rather than employment.
- Overall finding: the sectors that benefit the most during booms also experience the most severe downturn during busts. This negative correlation between performance during a boom and performance during a bust is evident in the results.

### Key statistics (Table 1: Productivity Growth)
- Note: Productivity growth is computed as the residuals from regressing total factor productivity growth on country-sector dummies. Annual averages are calculated separately for tranquil times, booms, and busts.
- Tranquil times / Booms / Busts
  - Construction: -0.32 0.52 -0.89
  - All except construction: -0.24 0.62 0.15

*Source: Appendix Table 2.*

### Appendix I for definition of booms and busts). The "all

### sdnea2020002 - Appendix I for definition of booms and busts

### Sectoral activity during booms and busts
- On average, value added expands by roughly 1.2 percentage points more in all sectors during booms relative to tranquil times; employment expands by roughly 0.8 percentage point more.  
- These gains are largely reversed during busts: value added contracts by almost the same amount (Table 2).  
- Construction value added and employment grow by almost 4 percentage points more than in other sectors during booms (annual basis), and contract by almost the same amount during busts.  
- Finance expands in booms and contracts in busts but with a significantly smaller cycle magnitude than construction.  
- Mining is an outlier: it grows at the average rate during booms but does not shrink as much during busts; however, neither the boom nor the bust effect on mining is significant once controlling for country and time fixed effects (see Appendix Table 3).

### Comovement: booms versus busts (correlation)
- Sectoral adjusted value-added and employment growth rates in booms are strongly negatively correlated with growth in busts (Figure 2): regressions yield
  - Value-added: y = -0.8376x -0.0571, R² = 0.7089
  - Employment: y = -1.0998x + 0.0055, R² = 0.7818

### Which sector characteristics increase procyclicality?
- Industries that are (i) less tradable, (ii) more labor-intensive, and (iii) more reliant on external finance tend to be more sensitive to the credit cycle (Table 3).  
- Key quantitative findings from Table 3 (coefficients and significance preserved):
  - Tradability: negative baseline coefficients (e.g., -1.550***, -2.151***, -3.168***, -2.783*** across specifications); Boom x Tradability shows negative coefficients (for example, -0.927**, -1.400***).
  - Labor Intensity: positive baseline and Boom x Labor Intensity positive and significant in several specifications (for example, Boom x Labor Intensity 2.703***).
  - External Finance Dependence: negative baseline coefficients (for example, -1.534***, -3.162***), but Boom x External Finance Dependence positive and significant in some specifications (for example, 1.518**, 1.875***).
  - GDP Growth and its interactions enter significantly in several specifications (e.g., GDP Growth 0.789***; GDP Growth x Tradability 0.209***; GDP Growth x Labor Intensity 0.696***; GDP Growth x External Finance Dependence 0.556***).
- Interpretation of mechanisms:
  - External-finance-dependent industries relax constraints during credit booms and expand faster.
  - Labor-intensive sectors generate local feedback effects (hiring increases local consumption).
  - Nontradables are more exposed to domestic credit cycles and may suffer less from exchange-rate-driven dampeners, amplifying cycles.

### Long-term relation between credit booms and industrial structure
- No significant, robust relationship is detected between a country’s credit-boom experience and long-term changes in industrial structure (Figure 4).  
- Countries that go through more or longer booms tend to reach higher credit-to-GDP ratios (financial deepening) (Figure 5).  
- There is a nonlinear relationship between credit boom experience and long-term economic growth: countries that have booms grow faster, but growth benefits taper off when a country spends more than five years in a boom (Table 4).

### Defining good and bad booms; prevalence
- A bad boom is one followed by subpar economic performance or a systemic financial crisis.  
- Of 59 credit booms identified, roughly two-thirds are classified as bad booms.  
- Typical durations and growth:
  - A typical good boom lasts about three years, with credit-to-GDP ratio growing by 15 percent annually; Real GDP expands at an annual rate 1½ percentage points faster than trend.  
  - A typical bad boom lasts four years, with annual credit-to-GDP growth rate reaching 17 percent; Real GDP grows 1¾ percentage points above trend (difference not statistically significant).

### Sectoral activity: distinguishing good vs bad booms
- Overall sectoral rankings are similar between good and bad booms: top sectors include construction, finance, and information.  
- Construction displays the starkest asymmetry between good and bad booms:
  - Construction value added expands by 2.8 percentage points more in bad booms than in good ones (relative to other sectors) (Table 5).  
  - Construction employment expands by 3.3 percentage points more in bad booms than in good ones (Table 5).  
- The construction asymmetry is robust to:
  - Excluding the global financial crisis period (Table 5, column 3).  
  - Subsamples of advanced economies and emerging markets (Table 5, columns 6 and 7).  
- The asymmetry largely survives controls for house price growth and asset price growth, though sample size reductions weaken statistical significance in some specifications (Table 6).

### Why construction is uniquely informative
- Potential channels and distortions:
  - Tangibility and collateral: construction produces pledgeable assets, encouraging heavy credit allocation to construction that may not yield high productivity, leading to resource misallocation and lower TFP (Reis 2013; Ebrahimy, forthcoming).  
  - Labor intensity and low skill requirements: rapid construction employment growth can mask structural labor-market problems and reduce incentives for skill accumulation (Charles, Hurst, and Notowidigdo 2016).  
  - Measurement and governance: construction’s growth can obscure fundamentals and delay prudent policy responses (Fernandez-Villaverde, Garicano, and Santos 2013).  
  - Leverage and fire-sale dynamics: construction projects have large up-front financing; booms that shift leverage to construction increase vulnerability and can amplify busts via fire sales and balance-sheet feedbacks (Caballero, Hoshi, and Kashyap 2008; Kiyotaki and Moore 1997; Brunnermeier and Sannikov 2014).

### Is construction activity a predictive signal for bad booms?
- Conditional probabilities (Table 7):
  - Overall probability that a boom is bad: 0.64.
  - Conditional on long booms: 0.74.
  - Conditional on long booms AND high construction growth: 1.00 (in-sample: long booms with rapid construction growth never ended well in the sample).  
  - Conditional on long booms AND low construction growth: 0.50.
- Regression evidence:
  - Construction growth (value-added and employment) raises the odds of a boom being bad even after controlling for duration, initial credit-to-GDP, and credit growth during the boom.  
  - Approximate marginal effects: a 1 percentage point increase in value-added or employment growth in construction raises the probability of a bad boom by (approximately) 2 and 5 percentage points, respectively, after controlling for duration or size.  
  - Construction employment growth remains a significant predictor of bad booms even after controlling for duration, household credit growth, house price growth, and asset price growth (Appendix Tables 5a–5d).  
  - In this sample, household credit growth, house price growth, and asset price growth are not individually statistically significant predictors of bad booms.

### Costs of bad booms and construction growth
- Summary statistics (Table 8): average annual real GDP growth over the three years after bad booms (examples):
  - High Duration: -3.6; Low Duration: -2.0
  - High Construction Value-Added Growth: -2.9; Low Construction Value-Added Growth: -2.2
  - With Long Booms & High Construction Growth: -5.0; With Long Booms & Low Construction Growth: -2.1
  - With High Household Credit Growth: -3.3; With Low Household Credit Growth: -1.9
- Regression evidence on post-boom outcomes (Table 9):
  - Higher construction growth during booms predicts worse three-year outcomes after bad booms: e.g., Construction Growth x Bad coefficients are negative and significant (for example, -0.466*** in column 1).  
  - Quantitative interpretation: 1 additional percentage point of construction value-added growth during a bad boom implies more than a 0.1 percent decline in average sector growth over the three-year post-boom period (sum of coefficients yields approximately –0.1; see Table 9 discussion).  
  - For six-year horizons, the net effects of construction during bad booms become negligible or insignificant (column 6), suggesting the adverse effects are predominantly medium-term rather than permanent on average.  
  - Construction growth predicts worse post-boom performance even after controlling for household credit growth; when both are included, household credit often loses significance.

### Policy implications and the trade-off
- Trade-off highlighted:
  - Credit booms can be good (continued growth) but have a high chance of turning bad: bad booms are twice as likely as good booms in the sample.  
  - Bad booms do not show statistically significantly better performance during the boom phase than good booms, but their costs (tail losses) can be large: the worst quartile of bad booms has annual average GDP growth of –3.8 percent versus +2.1 percent for the best quartile of good booms.  
- Practical implication:
  - Monitoring construction-sector activity during booms provides additional early-warning information beyond traditional indicators (boom duration, boom size, household credit growth, house-price and asset-price increases).  
  - Given the asymmetric downside risks associated with high construction growth during booms—especially when booms are long or accompanied by rising leverage—policy makers face a trade-off between tolerating short-term gains and preventing severe medium-term downturns.  

*Source: IMF staff calculations and figures from the IMF paper "Discern ing Good from Bad Credit Booms" (appendix material provided).*

### Appendix Figure 3).

### Appendix Figure 3)

### Key findings on credit and construction booms
- Credit booms generally carry high risk but offer relatively lower rewards.
- A construction boom along with a credit boom makes a favorable outcome even less likely.
- The robust and persistent association of the construction sector with the severity of bad booms is observed above and beyond household credit growth.
- Even when household credit grows fast in emerging markets, most of the credit allocated during booms goes through the corporate sector (see Figure 9; Appendix Figures 5 and 6). 35

### The role of construction (mechanisms and implications)
- High household leverage can exacerbate the bust through its negative impact on aggregate demand (Mian, Sufi, and Verner 2017).
- A fast-growing construction sector may be associated with significant misallocation of capital and labor, which can:
  - Negatively affect TFP and output growth during the bust.
  - Be consistent with a corporate sector overexposed to real estate risk.
- In emerging market economies, where household credit markets are small, construction activity may be a more robust signal than household credit for policymaking.
- Distinct channels:
  - Misallocation channel: construction is not a high-TFP-growth sector; booms may redirect resources away from higher-TFP uses.
  - Collateral amplification channel: real estate used as collateral can amplify credit booms and increase leverage, strengthening amplification mechanisms.

### Policy discussion and recommendations
- Two caveats:
  - The definition of good and bad booms is intrinsically after the fact; an ideal evaluation would require a counterfactual via a calibrated structural model, which the cross-country setup makes difficult.
  - Association between construction activity during booms and subsequent outcomes does not necessarily indicate causality; construction may capture omitted factors (for example, lack of other investment opportunities or shifts in bank portfolio allocation).
- From a predictive and policy standpoint, empirical regularities remain useful for discerning good from bad credit booms.
- Policy considerations when rapid expansion in construction occurs during a credit boom:
  - Give consideration to tightening the macroeconomic policy stance and/or activating macroprudential tools.
  - Activation of tools may be triggered by other indicators (for example, house price surges or rapid growth in mortgage loan markets), but sometimes these may not sound the alarm (for example, when construction booms are financed by the corporate sector or by foreigners).
  - Real activity metrics and indicators such as rising exposure of banks to specific borrowers could signal the need for action beyond credit speed limits or tighter lending standards.
  - Examples of possible measures: limits on banks’ exposure to real estate developers and other construction firms could be tightened.
- Data and measurement recommendations:
  - Improve indicators of construction sector performance (for example, productivity).
  - Explore composing better measures of economic growth that are not contaminated by real estate bubbles, unsustainable credit market dynamics, or unusually rapid construction growth. 36

### External/internal balance considerations
- Nontradable-sector prominence (such as construction) in credit cycles implies different dynamics in restoring internal and external balances during credit booms:
  - If a country has an excessive current account deficit, credit booms favoring nontradables could worsen the trade balance; curbing the credit boom to restore internal balances could also restore external balance.
  - If a country has an excessive current account surplus, curbing the credit boom may raise the surplus by reallocating resources from nontradable to tradable sectors.

### Conclusion and directions for future research
- Construction sector displays unique patterns across booms and busts:
  - Strongest acceleration (deceleration) in value-added and employment growth during booms (busts).
  - The only sector that consistently overperforms during bad booms.
  - Pace of construction activity during the boom phase is a better predictor of economic costs associated with bad booms than other variables identified in previous studies.
- Monitoring construction activity during booms can provide a litmus test for policy action; high-frequency indicators (for example, construction permit applications) could serve as valuable signals.
- Suggested future research directions:
  - Use more granular data to improve understanding of macro-financial dynamics.
  - Introduce construction sector indicators in the growth-at-risk framework (Adrian and others, forthcoming).
  - Explore effectiveness and side effects of policy options to curb excessive developments in the construction sector.
  - Assess whether the trade-off faced during booms changes over longer horizons as longer time series become available.

*International Monetary Fund — sdnea2020002 (Appendix Figure 3))*

---


_Source: https://www.imf.org/-/media/files/publications/sdn/2020/english/sdnea2020002.pdf_