## Appendix I.

## Source details

**Canonical URL:** [Appendix I.](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp1167.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp1167.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp1167.pdf.json)

---

### I. INTRODUCTION
- Since 2003, new member states (NMS) of the European Union received a particularly large level of capital inflows resulting in an unprecedented credit boom-bust cycle which created rapid growth and deep recessions.
- Question raised: What will be the impact of a new wave of capital inflows on emerging markets?
- Sluggish recoveries and monetary easing in advanced countries is unleashing a new wave of capital flows; the impact on emerging markets remains unclear.
- High output volatility stemming from capital inflows raises concerns because output volatility tends to be associated with lower long-term GDP growth (citing Ramey and Ramey (1995), Martin and Rogers (2000), and Cerra and Saxena (2008)).
- The effect of capital inflows on GDP is influenced by:
  - The maturity and currency of capital flows.
  - The size of capital flows and the channels through which they arrive.
  - The economic sector into which capital flows are ultimately invested (e.g., real estate vs. tradables).
- This paper distinguishes flows channeled to real estate and corporate investment via: mortgages, consumer credit, corporate real estate credit, real estate FDI, corporate non-real estate credit, and non-real estate FDI.

### II. STYLIZED FACTS: CAPITAL FLOWS AND GDP GROWTH IN NMS DURING THE RECENT BOOM-BUST CYCLE
- Sectors receiving capital flows determine nature and magnitude of GDP booms and busts: surge and sudden decline of capital inflows coincide with rapid increase and collapse of GDP growth.
- NMS with smaller swings in GDP growth received smaller capital inflows and tended to have export-led booms.
- NMS with the largest swings concentrated capital inflows into real estate; consumption-led booms and busts coincided with majority of capital inflows being channeled to real estate.
- NMS where majority of inflows were channeled towards non-real estate sectors had investment-led booms and busts with lower GDP fluctuations.

Findings from the Boom Years:
- Growth dispersion across NMS during boom:
  - Baltics: average boom GDP growth by 8 to 10 percent per year.
  - Czech Republic, Hungary, and Poland: around 3 to 5 percent per year.
  - Bulgaria, Romania, and the Slovak Republic: 6, 7, and close to 8 percent annual average growth, respectively.
- Net capital inflows in the five years leading to boom peak: ranged from 80 to 160 percent of 2003 GDP in the Baltics, Bulgaria, and Romania.
- Flow composition by country:
  - Baltics and Romania: bank inflows (direct and through FDI) dominated.
  - Bulgaria: FDI (non-bank) was the main inflow.
  - Hungary, Poland, Czech and Slovak Republics: inflows ranging from 30 to 60 percent of 2003 GDP, relatively evenly split between bank-related flows and FDI (non-bank). Exceptions: Hungary where bank flows were much larger; Czech Republic where non-bank FDI dominated.
- Sectoral allocation and outcomes:
  - Baltics and Romania: almost half of domestic credit went to households (mainly mortgages in the Baltics and consumer credit in Romania). Example: Latvia nominal real estate prices grew by 60 percent per annum during the boom period defined for Latvia.
  - Bulgaria: half of FDI financed corporate investment (manufacturing and trade infrastructure), one fifth to the financial sector, remainder to corporate real estate; average nominal real estate prices grew by 50 percent per annum during the boom.
- Contributions to GDP during boom (average annual boom values):
  - Consumption was the largest component of GDP growth in Latvia, Lithuania, and Romania (growing at over 10 percent per annum).
  - Investment drove the boom in Bulgaria and played an important role in Estonia; average annual boom investment growth in Hungary, Poland, Czech and Slovak Republics was around 5 percent, but in the Baltics it was almost three times as much, and over 20 percent in Bulgaria and Romania.
- Export-driven growth:
  - Hungary, Poland, Czech and Slovak Republics had export-led growth with more moderate booms.
  - Since the turn of the century, NMS annual exports grew by an average of 8½ percent per year.
- Import leakage:
  - The negative contribution of imports to GDP growth was about two thirds that of other components’ positive contribution across NMS, except Latvia and Lithuania where it was one half.

Findings from the Crisis Years:
- Timeline and triggers:
  - Mid-2007: Swedish banks began reducing capital flows to the Baltics.
  - Fall 2008: Following Lehman Brothers default, capital flows to NMS declined sharply.
- Declines in annual capital inflows during downturn (percent of GDP):
  - Latvia: 20 percent
  - Lithuania: 14 percent
  - Estonia: 12 percent
  - Bulgaria: 17 percent
  - Romania: 7 percent
  - Czech and Slovak Republics, and Poland: 1 to 3 percent
  - Hungary: small net increase (primarily due to increases in trade credits)
- Composition of the decline:
  - Reduced bank inflows accounted for almost 100 percent of the drop in capital inflows in Baltics and Romania; in Bulgaria the decline in nonbank FDI was slightly larger.
- Investment and credit contraction:
  - Investment fell by 65 percent in the Baltics relative to peak; slightly less in Bulgaria and Romania; around 15 percent in the Czech Republic and Poland.
  - Credit flows for all types of investment fell by 75 to 110 percent per annum in all NMS except Poland (25 percent).
- Consumption and prices:
  - Household consumption fell by 25 to 30 percent from its peak in the Baltics and Romania.
  - Housing prices: nominal values fell by around 25 percent per annum over two years in the Baltics.
  - Bulgaria: consumption fell by only half as much from its boom peak.

### Downturn GDP Growth and Sectoral Impacts
- GDP contractions in Hungary, and the Czech and Slovak Republics were even smaller (one quarter of the Baltics’)—driven by export contraction.
- In 2009, annual NMS export growth fell by 10 percent.
- Exports were the largest factor behind GDP declines in Hungary, the Czech and Slovak Republics.
- Poland:
  - Poland’s GDP grew by around 3 percent—supported by consumption growth.
  - Consumption grew by 8 percent.
  - Bank inflows and subsequent credit growth had been subdued; the crisis brought less reduction of bank inflows and lower credit contraction than in other NMS, with almost no change in credit to households.

Fiscal balances, reserves, and risk premia:
- Fiscal balances deteriorated sharply; Latvia’s fiscal balance deteriorated by 8 pps during the crisis.
- Other NMS (except Hungary) met with deteriorations of 3 to 6 pps of their fiscal balances.
- Lending rates and CDS spreads among fixed exchange rate NMS:
  - Latvia lending rates peaked at over 30 percent in 2009—rising 20 percentage points above its average pre-crisis value.
  - Bulgaria lending rates rose less than one percentage point from its pre-crisis value.
  - CDS spreads: Bulgaria within 660 basis points; Latvia rose to 1100 basis points.
- Flexible exchange rate NMS CDS spreads:
  - Romania: 740 basis points
  - Hungary: 530 basis points
  - Poland and the Czech and Slovak Republics saw smaller increases in CDS spreads.
- High public debt in Hungary: 65 percent of GDP in 2009.

### Channels: how capital inflows affect GDP
- Presumed channels:
  - Households via banks (mortgages and consumer credit) → boosts consumption and housing investment.
  - Corporate real estate via bank credit and FDI → supports investment (including vacation real estate targeted at foreigners).
  - Corporate non-real estate via bank credit and FDI → boosts investment and potentially exports.
- Some positive impact of capital inflows leaks through imports (inputs for investment and consumption of imported goods).

### Empirical approach
- Initial correlations: GDP growth is positively correlated with growth in flows of mortgages, consumer credit, corporate real estate credit, real estate FDI, non-real estate FDI, and corporate non-real estate credit during booms and busts.
- Econometric strategy:
  - Cross-section OLS on pre-boom (2003–4), boom (2005–7), and crisis periods for nine NMS (27 observations).
  - Panel OLS with fixed effects on annual data (2003–9) for nine NMS.
  - Instrumental variables and GMM estimation using lagged independent variables as instruments.
  - Fiscal and exchange rate policy influences were also tested.
- Data sources: central banks, ministries of finance, national statistical institutes, IMF’s International Financial Statistics, Haver, WEO.

### Key regression results (impact on GDP growth per 10 pps increase in flow)
- Mortgage flows:
  - Panel OLS: 0.36 ***
  - GMM: 0.21 to 0.31 ***
  - Cross-section OLS: -0.25 (negative, not significant)
  - Interpretation: GMM range implies a 10 pps increase corresponds to a 0.21–0.31 pps increase in GDP growth.
- Consumer credit:
  - Cross-section OLS: 0.52 ***
  - Panel OLS: 0.14
  - GMM: 0.01 to 0.13
- Corporate real estate credit:
  - Cross-section OLS: -0.41 **
  - Panel OLS: -0.02
  - GMM: -0.05 to 0.18
- Real estate FDI:
  - Cross-section OLS: -0.06 ***
  - Panel OLS: 0.06
  - GMM: 0.17 to 0.28 ***
  - Interpretation: GMM implies a 10 pps increase corresponds to a 0.17–0.28 pps increase in GDP growth.
- Non-real estate FDI:
  - Cross-section OLS: 0.14
  - Panel OLS: 0.10 *
  - GMM: 0.32 to 0.47 **
  - Interpretation: GMM implies a 10 pps increase corresponds to a 0.32–0.47 pps increase in GDP growth.
- Corporate non-real estate credit:
  - Cross-section OLS: 0.92 ***
  - Panel OLS: -0.01
  - GMM: -0.03 to -0.02 *
  - Note: GMM suggests a small negative effect (significant at 10 percent).

Selected reported coefficient examples (preserved as reported):
- Mortgage coefficients in appendix Table 3 and Table 4: "-0.025", "0.036  ***", "0.021  ***", "0.043  **", "0.034  **", "0.033  **", "0.014  *".
- Consumption growth regressions (Table 6): Mortgage "0.084  ***", "0.096  **", "0.071  **"; Consumer credit "0.040  *", "0.036  *", "0.015  *".
- Investment growth regressions (Table 5): FDI real estate "0.084  ***" and "0.101  ***"; FDI non real estate "0.076  **" and "0.100  ***"; corporate real estate credit "0.025" and "0.058  ***".

Diagnostics and table-level figures:
- prob(F-statistic): "0.05" and "0.00".
- J-statistic values: "0.04", "0.03", "0.02", "0.04".
- R-square adjusted: "0.73" and "0.50".
- Number of observations: "27" and "63".

### Synthesis of empirical findings
- Capital inflows into real estate (mortgages and real estate FDI) show the strongest correlation with GDP growth.
  - Summing mid-points of GMM ranges: a 10 pps increase in mortgage flows (mid ~0.26 pps) plus a 10 pps increase in real estate FDI flows (mid ~0.225 pps) together correspond to about a 0.49 pps increase in GDP growth (as noted in the source).
- Non-real estate sectors have weaker but positive estimated impacts via non-real estate FDI (GMM 0.32–0.47 pps per 10 pps increase).
- The paper reports: "0.37 pps for a 10 pps increase in growth in each of FDI flows and corporate credit flows in non-real estate sectors," applying mid-points of GMM estimates.
- Results vary across estimation methods; endogeneity and small-sample concerns affect some estimates.

### Policy recommendations and implications
- To minimize sharp swings in GDP growth and support sustainable growth:
  - Focus policies on areas that improve the attractiveness of tradables for capital inflows (examples cited: infrastructure and education).
  - Strengthen financial sector supervision.
  - Strengthen corporate governance in emerging market economies.
- Role of fiscal and financial buffers:
  - Although fiscal policy showed little direct impact on GDP growth in regressions, "strong fiscal and financial sector buffers were key to softening the blow experienced during crisis and maintaining exchange rate stability."

### Suggested avenues for future research
- Further explore the role of sub-sectors within real estate and non-real estate sectors; differentiate impacts of tradables versus non-tradables within the non-real estate sector.
- Refine estimations with other instruments.
- Extend analysis to crisis episodes in other emerging markets to bolster results.

*IMF working paper: Appendix I.*

### Appendix I. ............................................................................................................

### Appendix I.

### I. INTRODUCTION
- Since 2003, new member states (NMS) of the European Union received a particularly large level of capital inflows resulting in an unprecedented credit boom-bust cycle which created rapid growth and deep recessions.
- Question raised: What will be the impact of a new wave of capital inflows on emerging markets?
- Sluggish recoveries and monetary easing in advanced countries is unleashing a new wave of capital flows; the impact on emerging markets remains unclear.
- High output volatility stemming from capital inflows raises concerns because output volatility tends to be associated with lower long-term GDP growth (citing Ramey and Ramey (1995), Martin and Rogers (2000), and Cerra and Saxena (2008)).
- In the case of NMS, sensitivity to capital inflows implied greater GDP growth when capital inflows were increasing year to year, but also deeper recessions when they suddenly fell—leaving these NMS with lower average GDP growth than before the surge of capital inflows (Table 1). Hungary was an exception, where GDP growth slowed drastically near the end of the boom years, largely reflecting the impact of fiscal consolidation on domestic demand.
- The effect of capital inflows on GDP is influenced by:
  - The maturity and currency of capital flows (citing Rodrik and Velasco (1999), Allen et al. (2002)).
  - The size of capital flows and the channels through which they arrive.
  - The economic sector into which capital flows are ultimately invested (e.g., real estate vs. tradables).
- This paper studies the influence of economic sectors on the impact of capital flows on GDP, distinguishing flows channeled to real estate and corporate investment via:
  - mortgages, consumer credit, corporate real estate credit, real estate FDI, corporate non-real estate credit, and non-real estate FDI.

### II. STYLIZED FACTS: CAPITAL FLOWS AND GDP GROWTH IN NMS DURING THE RECENT BOOM-BUST CYCLE
- General observation: The sectors to which capital flows are directed matter for the nature and magnitude of GDP booms and busts. Surge and sudden decline of capital inflows coincide with rapid increase and collapse of GDP growth.
- NMS with smaller swings in GDP growth received smaller capital inflows and tended to have export-led booms influenced by global growth.
- NMS with the largest swings in GDP growth had the largest concentration of capital inflows into real estate; consumption-led booms and busts coincided with majority of capital inflows being channeled to real estate (directly via FDI or via bank-fueled household mortgage and consumer credit).
- NMS where majority of capital inflows were channeled towards non-real estate sectors had investment-led booms and busts with lower GDP fluctuations.

Findings from the Boom Years:
- Growth dispersion across NMS during boom:
  - Baltics grew fastest—on average by 8 to 10 percent per year.
  - Czech Republic, Hungary, and Poland: around 3 to 5 percent per year.
  - Bulgaria, Romania, and the Slovak Republic: annual average growth of 6, 7, and close to 8 percent, respectively.
- In the five years leading to peak of the boom, net capital inflows ranged from 80 to 160 percent of 2003 GDP in the Baltics, Bulgaria, and Romania.
- Flow composition by country:
  - Baltics and Romania: bank inflows (direct and through FDI) dominated.
  - Bulgaria: FDI (non-bank) was the main inflow.
  - Hungary, Poland, Czech and Slovak Republics: inflows ranging from 30 to 60 percent of 2003 GDP, relatively evenly split between bank-related flows and FDI (non-bank). Exceptions: Hungary where bank flows were much larger; Czech Republic where non-bank FDI dominated.
- Sectoral allocation and outcomes:
  - Baltics and Romania:
    - Bank credit and FDI flowed to household real estate and consumption supporting consumption boom, and to non-real estate corporate investments fueling investment boom.
    - Almost half of domestic credit went to households (mainly mortgages in the Baltics and consumer credit in Romania).
    - Example: Latvia nominal real estate prices grew by 60 percent per annum during the boom period defined for Latvia.
  - Bulgaria:
    - Bank credit and FDI fueled corporate investments; investment grew by 120 percent since boom began at its peak.
    - Half of FDI financed corporate investment (manufacturing and trade infrastructure), one fifth to the financial sector (fueling domestic credit), remainder to corporate real estate (vacation real estate aimed at foreigners).
    - Average nominal real estate prices grew by 50 percent per annum during the boom.
    - Consumption was only 30 percent higher than when the boom began.
- Contributions to GDP during boom (average annual boom values):
  - Consumption was the largest component of GDP growth in Latvia, Lithuania, and Romania (growing at over 10 percent per annum).
  - Consumption growth was about half as much in Bulgaria and one third as high in Hungary, Poland, and the Czech and Slovak Republics.
  - Investment drove the boom in Bulgaria and played an important role in Estonia; average annual boom investment growth in Hungary, Poland, Czech and Slovak Republics was around 5 percent, but in the Baltics it was almost three times as much, and over 20 percent in Bulgaria and Romania.
- Export-driven growth:
  - Hungary, Poland, Czech and Slovak Republics had export-led growth with more moderate booms.
  - Since the turn of the century, NMS annual exports grew by an average of 8½ percent per year.
- Import leakage:
  - Investment and export-led booms had less impact on GDP growth than consumption-led booms largely due to higher import growth draining positive impacts.
  - The negative contribution of imports to GDP growth was about two thirds that of other components’ positive contribution across NMS, except Latvia and Lithuania where it was one half.

Findings from the Crisis Years:
- Timeline and triggers:
  - Mid-2007: Swedish banks began reducing capital flows to the Baltics due to concerns about over-exposure.
  - Fall 2008: Following Lehman Brothers default, capital flows to NMS declined sharply, accelerating downturns.
- Downturn changes in net capital inflows (comparison with year preceding crisis):
  - Latvia had the largest decline of annual capital inflows: 20 percent of GDP.
  - Lithuania: reduction of 14 percent of GDP.
  - Estonia: reduction of 12 percent of GDP.
  - Bulgaria and Romania: reductions of 17 and 7 percent of GDP, respectively.
  - Czech and Slovak Republics, and Poland: declines limited (1 to 3 percent of GDP).
  - Hungary saw a small net increase primarily due to increases in trade credits.
- Composition of the decline:
  - Reduced bank inflows accounted for almost 100 percent of the drop in capital inflows in Baltics and Romania; in Bulgaria the decline in nonbank FDI was slightly larger.
- Investment and credit contraction:
  - Relative to peak boom values, investment fell by 65 percent in the Baltics; slightly less in Bulgaria and Romania; around 15 percent in the Czech Republic and Poland.
  - Sharp drops in FDI and bank flows resulted in strong credit declines for all types of investment—with pronounced decline of mortgage credit in the Baltics and Bulgaria and consumer credit declines in Romania.
  - Credit flows for all types of investment fell by 75 to 110 percent per annum in all NMS except Poland (25 percent).
- Consumption declines:
  - Baltics and Romania: household consumption fell by 25 to 30 percent from its peak.
  - Housing prices: nominal values fell by around 25 percent per annum over two years in the Baltics.
  - Bulgaria: consumption fell by only half as much from its boom peak because household consumption had not grown as much during the boom; some fall in mortgage credit was offset by increase in consumer credit.

### Key statistics and exact figures (as reported)
- Baltics: average boom GDP growth by 8 to 10 percent per year.
- Czech Republic, Hungary, Poland: around 3 to 5 percent per year during boom.
- Bulgaria, Romania, Slovak Republic: 6, 7, and close to 8 percent annual average growth, respectively.
- Net capital inflows in five years leading to boom peak: ranged from 80 to 160 percent of 2003 GDP in the Baltics, Bulgaria, and Romania.
- Latvia nominal real estate prices growth during boom: 60 percent per annum.
- Bulgaria average nominal real estate price growth during boom: 50 percent per annum.
- NMS annual exports growth since turn of century: average of 8½ percent per year.
- Declines in annual capital inflows during downturn (percent of GDP): Latvia 20 percent, Lithuania 14 percent, Estonia 12 percent, Bulgaria 17 percent, Romania 7 percent; Czech, Slovak, Poland: 1 to 3 percent; Hungary: small net increase (trade credits).
- Investment declines relative to peak: Baltics 65 percent; Czech Republic and Poland around 15 percent.
- Household consumption declines from peak in Baltics and Romania: 25 to 30 percent.
- Housing price nominal declines in Baltics during downturn: around 25 percent per annum over two years.
- Credit flow declines per annum during downturn: 75 to 110 percent in all NMS except Poland (25 percent).

*IMF working paper: Appendix I.*

### 2010. Nonetheless, the decline in consumption in Bulgaria has still been much smaller than that in the Baltics.

### Downturn GDP Growth and Sectoral Impacts in New Member States (NMS)

### Downturn patterns and contributions to GDP growth
- GDP contractions in Hungary, and the Czech and Slovak Republics were even smaller (one quarter of the Baltics’)—driven by export contraction, a consequence of global recession.
- In 2009, annual NMS export growth fell by 10 percent.
- Exports were the largest factor behind GDP declines in Hungary, the Czech and Slovak Republics.
- While investment declines further worsened GDP contraction in these countries, consumption had little impact.
- In other NMS, falling exports exacerbated the impact of sudden domestic demand declines on GDP.
- Figure labels and components (as presented): C = consumption; X = exports; I = investment; M = imports.

### Poland's experience
- Poland’s GDP grew by around 3 percent—supported by consumption growth.
- After the global crisis, Poland’s GDP declined slightly for one quarter before continuing to grow from the first quarter of 2009.
- Consumption grew by 8 percent (largely reflecting pre-crisis fiscal stimulus measures that came into effect during the crisis).
- Bank inflows and subsequent credit growth (especially to households) had been subdued relative to other NMS during the boom years—so consumption also grew less.
- Consequently, the crisis brought less reduction of bank inflows and lower credit contraction than in other NMS, with almost no change in credit to households—supporting consumption stability.

### Fiscal balances, reserves, and risk premia
- Large declines in consumption, business activity and profitability, and compliance during the downturn meant substantially lower government revenues in all NMS.
- Fiscal balances deteriorated sharply; Latvia’s fiscal balance deteriorated by 8 pps during the crisis—resulting in the highest headline deficit of all NMS.
- All other NMS (except Hungary) met with deteriorations of 3 to 6 pps of their fiscal balances.
- Among fixed exchange rate NMS:
  - Lending rates in Latvia peaked at over 30 percent in 2009—rising 20 percentage points above its average pre-crisis value.
  - In Bulgaria lending rates rose less than one percentage point from its pre-crisis value.
  - CDS spreads in Bulgaria remained within 660 basis points, while in Latvia they rose to 1100 basis points.
- For flexible exchange rate NMS:
  - CDS spreads rose to 740 basis points in Romania and 530 in Hungary.
  - With lower and improving pre-crisis fiscal deficits, CDS spreads rose by less than half as much in Poland and the Czech and Slovak Republics.
  - Poland’s precautionary Flexible Credit Line with the IMF and the Slovak Republic’s 2009 Euro area entry further buttressed their market positions.
- High public debt (65 percent of GDP in 2009) compounded market fears in Hungary.

### Channels: how capital inflows affect GDP
- Capital inflows impact GDP by flowing to economic sectors that affect consumption, investment, exports, and imports.
- Presumed channels described:
  - Households through banks via mortgages and consumer credit, boosting consumption and potentially investment in housing.
  - Corporate real estate, via bank credit and FDI, supporting investment (e.g., vacation homes targeted at foreigners increase investment but not consumption).
  - Corporate sectors outside real estate, via bank credit and FDI, boosting investment and potentially exports.
- Some positive impact of capital inflows is leaked through imports (inputs for investment and consumption of imported goods).

### Empirical approach
- Initial correlations: GDP growth is positively correlated with growth in flows of mortgages, consumer credit, corporate real estate credit, real estate FDI, non-real estate FDI, and corporate non-real estate credit during booms and busts.
- Econometric strategy:
  - Cross-section OLS on pre-boom (2003–4), boom (2005–7), and crisis periods for nine NMS (27 observations).
  - Panel OLS with fixed effects on annual data (2003–9) for nine NMS.
  - Instrumental variables and GMM estimation to address endogeneity, using lagged independent variables as instruments.
  - Fiscal and exchange rate policy influences were also tested.
- Data sources: central banks, ministries of finance, national statistical institutes, IMF’s International Financial Statistics, Haver, WEO.

### Key regression results (impact on GDP growth per 10 pps increase in flow)
- Mortgage flows:
  - Panel OLS: 0.36 ***
  - GMM: 0.21 to 0.31 ***
  - Cross-section OLS: -0.25 (negative, not significant)
  - Interpretation: mortgage credit flows boost household consumption and thus GDP; GMM range indicates a 10 pps increase corresponds to a 0.21–0.31 pps increase in GDP growth.
- Consumer credit:
  - Cross-section OLS: 0.52 ***
  - Panel OLS: 0.14
  - GMM: 0.01 to 0.13
- Corporate real estate credit:
  - Cross-section OLS: -0.41 **
  - Panel OLS: -0.02
  - GMM: -0.05 to 0.18
- Real estate FDI:
  - Cross-section OLS: -0.06 ***
  - Panel OLS: 0.06
  - GMM: 0.17 to 0.28 ***
  - Interpretation: GMM implies a 10 pps increase in real estate FDI flows corresponds to a 0.17–0.28 pps increase in GDP growth.
- Non-real estate FDI:
  - Cross-section OLS: 0.14
  - Panel OLS: 0.10 *
  - GMM: 0.32 to 0.47 **
  - Interpretation: GMM implies a 10 pps increase in non-real estate FDI flows corresponds to a 0.32–0.47 pps increase in GDP growth.
- Corporate non-real estate credit:
  - Cross-section OLS: 0.92 ***
  - Panel OLS: -0.01
  - GMM: -0.03 to -0.02 *
  - Note: GMM suggests a small negative effect (significant at 10 percent) where a 10 pps increase corresponds to a 0.03 pps decline in GDP growth.
- Statistical notes: ***, **, and * indicate 1, 5, and 10 percent significance levels, respectively. All panel regressions include a constant and control for cross-sectional fixed effects. The GMM coefficient ranges summarize nine sets of GMM regressions.

### Synthesis of empirical findings
- Capital inflows into real estate (mortgages and real estate FDI) show the strongest correlation with GDP growth.
  - Summing mid-points of GMM ranges: a 10 pps increase in mortgage flows (mid ~0.26 pps) plus a 10 pps increase in real estate FDI flows (mid ~0.225 pps) together correspond to about a 0.49 pps increase in GDP growth (as noted in the source).
- Non-real estate sectors have weaker but positive estimated impacts via non-real estate FDI (GMM 0.32–0.47 pps per 10 pps increase).
- Consumer credit, corporate credit, and some cross-section OLS estimates show mixed significance and signs; results vary by estimation method, highlighting endogeneity and small-sample concerns.
- Regressions of consumption growth indicate strong correlation to mortgage flows (a 10 pps increase yields 0.71–0.96 pps rise in consumption growth) and to a lesser extent consumer credit flows (less than half the impact of mortgage flows).
- Investment growth is most highly correlated with growth in both real estate and non-real estate FDI flows (a 10 pps increase of each corresponds to about a 1 ppt rise in investment growth), followed by corporate real estate credit and mortgage flows (see Appendix references in the source).

*Canonical source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp1167.pdf*

### 0.37 pps for a 10 pps increase in growth in each of FDI flows and corporate credit flows in

### _wp1167 - 0.37 pps for a 10 pps increase in growth in each of FDI flows and corporate credit flows in

### Sectoral impact of capital inflows on GDP growth
- Estimated impact: "0.37 pps for a 10 pps increase in growth in each of FDI flows and corporate credit flows in non-real estate sectors," applying mid-points of GMM estimates.
- Real estate vs non-real estate:
  - Capital inflows targeted at the real estate sector (in particular via mortgage flows and real estate related FDI flows) have the most sizeable impact on GDP growth, outweighing inflows destined to non-real estate activities.
  - Non-real estate FDI flows are reported as having "the largest impact of any single variable."
- Mechanisms:
  - Many inflows went to FDI or banks fueling credit growth; the paper emphasizes that it is the destination, not the form of capital inflow, that most influences GDP growth.
  - Non-real estate FDI impact may reflect FDI in nontradables (restaurants, hotels, retail trade) that follow consumption booms and busts or FDI in tradables.

### Policy control parameters: fiscal policy and exchange rate regime
- Fiscal policy:
  - The change in fiscal balance over GDP was included as a RHS variable in GMM estimation.
  - Regression results indicate fiscal policy "did not have a significant direct impact on GDP growth" during the recent boom-bust episodes.
  - Stylized facts: strong fiscal policy during boom years helped provide a buffer for the crisis, but did not appear to add significantly to overheating pressures via a direct growth effect.
- Exchange rate:
  - The exchange rate "neither has a significant impact on GDP growth nor does it enhance the effect of credit flows or FDI flows in a particular sector on GDP."
  - A dummy for fixed exchange rate regimes "lacks significance and is slightly negative."
  - Table result: Exchange rate dummy value shown as "-3.591" (table footnote 2/ 1=Fixed exchange rate).
  - Interactive terms testing whether fixed regimes enhance impact of sectoral inflows: coefficients "are not significant and mostly near zero."

### Empirical regression highlights and diagnostics (selected exact figures from tables)
- Table-level diagnostics:
  - prob(F-statistic): "0.05" and "0.00" (in listed results).
  - J-statistic entries: "0.04", "0.03", "0.02", "0.04" (multiple GMM specifications show J-statistic values around these figures).
  - R-square adjusted: "0.73" and "0.50" for reported models.
  - Number of observations: "27" and "63" (cross-section and panel regressions).
- Selected coefficient examples (preserved as reported):
  - Mortgage coefficients in appendix Table 3 and Table 4: values include "-0.025", "0.036  ***", "0.021  ***", "0.043  **", "0.034  **", "0.033  **", "0.014  *".
  - Consumer credit, corporate credit, and FDI coefficients reported across specifications (examples): "0.052  ***", "0.014", "0.092  ***", "-0.041  **", "-0.006  ***", "0.014", "0.023", "0.049  *".
  - Investment growth regressions (Table 5): FDI real estate "0.084  ***" and "0.101  ***"; FDI non real estate "0.076  **" and "0.100  ***"; corporate real estate credit "0.025" and "0.058  ***".
  - Consumption growth regressions (Table 6): Mortgage "0.084  ***", "0.096  **", "0.071  **"; Consumer credit "0.040  *", "0.036  *", "0.015  *".
- Significance notation preserved: ***, **, and * indicate 1, 5, and 10 percent significance levels, respectively.

### Policy recommendations and implications
- To minimize sharp swings in GDP growth and support sustainable growth:
  - Focus policies on areas that improve the attractiveness of tradables for capital inflows (examples cited: infrastructure and education).
  - Strengthen financial sector supervision.
  - Strengthen corporate governance in emerging market economies.
- Role of fiscal and financial buffers:
  - Although fiscal policy showed little direct impact on GDP growth in regressions, "strong fiscal and financial sector buffers were key to softening the blow experienced during crisis and maintaining exchange rate stability."

### Suggested avenues for future research (as stated)
- Further explore the role of sub-sectors within real estate and non-real estate sectors.
  - Differentiate impacts of tradables versus non-tradables within the non-real estate sector.
- Refine estimations with other instruments.
- Extend analysis to crisis episodes in other emerging markets to bolster results.

*Source: _wp1167 - 0.37 pps for a 10 pps increase in growth in each of FDI flows and corporate credit flows in (IMF PDF content provided).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2011/_wp1167.pdf_