## wp17129

## Source details

**Canonical URL:** [wp17129](https://www.imf.org/-/media/files/publications/wp/2017/wp17129.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2017/wp17129.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2017/wp17129.pdf.json)

---

### Introduction and research question
- Productivity growth has declined in advanced economies since the global financial crisis (GFC) (Adler et al., 2017; OECD 2015).
- Research question: whether the abruptness, magnitude and persistence of the post-GFC fall in total factor productivity (TFP) growth can be linked to the interaction between tighter credit conditions and weak corporate balance sheets following the collapse of Lehman Brothers in September 2008.
- Identification exploited: the sharp and unforeseen tightening of credit conditions immediately after Lehman as an exogenous credit supply shock.

### Data, sample, and key variable definitions
- Data source: cross-country firm-level dataset constructed by merging different waves of ORBIS.
- Sample: 15 advanced economies — Austria, Belgium, Denmark, Germany, Spain, France, Greece, Italy, Japan, Korea, the Netherlands, Portugal, Slovenia, the United Kingdom, and the United States.
- Sector scope: non-farm, non-financial business sector (NACE Rev.2 two-digit industry codes 5–82).
- Sample restrictions: exclude firms with less than 3 employees; require at least four consecutive periods.
- Key firm-level vulnerability measures:
  - Leverage Pre-Crisis: average pre-crisis debt-to-assets ratio (current liabilities + long-term debt) / total assets, averaged 2003–2007.
  - Debt Maturing 2008 (preferred rollover-risk measure): current liabilities at end-2007 divided by total sales in 2007.
- Country-level credit shock measure: change in the average CDS spread of domestic banks around Lehman (described both as change between first and second halves of 2008 and as difference between the two quarters before and the two quarters after the Lehman bankruptcy).
- Productivity measure: firm-level TFP growth computed using real value added, real capital stock, and number of employees applying the Wooldridge (2009) control function approach; robustness checks use labor productivity.
- Investment measures: capital stock and investment derived via perpetual inventory method; intangible investment measured as stock of intangible assets (from ORBIS) and investment in intangibles as ratio to value added.

### Empirical strategy
- Core design: differences-in-differences comparing change in firm-level average TFP growth between post-crisis (five years after 2008) and pre-crisis (five years until 2008) periods for firms with large versus low pre-existing balance sheet vulnerabilities.
- Baseline regression: ΔTFP Growth (post vs. pre) regressed on pre-crisis vulnerabilities, firm controls (age, log total assets, log EBITDA pre-crisis), country*sector fixed effects, clustered SE at country-sector level.
- Extended specification: includes interaction between firm-level vulnerability and country ΔCDS to exploit cross-country heterogeneity in credit shock severity.
- Identification assumption: timing of debt rollovers in 2008 and cross-country changes in bank CDS spreads are exogenous to unobserved determinants of within-firm post-crisis TFP growth.

### Main empirical findings (baseline and magnitudes)
- Baseline estimates (Table 2; N = 134,838; R-squared ≈ 0.151 when both vulnerabilities included):
  - Column (3) coefficients:
    - Leverage Pre-Crisis = -0.036*** (standard error 0.006).
    - Debt Maturing 2008 = -0.091*** (standard error 0.010).
  - Interpretation (from baseline column (3)):
    - A 10 percentage point higher leverage ratio associated with a 0.36 percentage point decline in average annual TFP growth after the crisis.
    - A 10 percentage points increase in the share of debt maturing in 2008 associated with a 0.9 percentage point drop in annual TFP growth.
- Extended specification interacting vulnerabilities with ΔCDS (Table 3; N = 104,275):
  - Representative reported coefficients:
    - Leverage Pre-Crisis (direct) = -0.046*** (0.010).
    - Leverage Pre-Crisis * ΔCDS = -0.051*** (0.021).
    - Debt Maturing 2008 (direct) = -0.108*** (0.008).
    - Debt Maturing 2008 * ΔCDS = -0.064*** (0.017).
  - Textual illustration (column (3) interpretation):
    - In a country with an average increase in bank CDS spreads:
      - A 10 percentage point higher leverage ratio associated with a 0.46 percentage point post-crisis decline in average annual TFP growth.
      - A 10 percentage points increase in the share of debt maturing in 2008 associated with a 1.08 percentage point drop in annual TFP growth.
    - In a country where the increase in CDS spreads was 100 basis points larger than the average:
      - The corresponding figures were 0.51 (rising to 0.97) and 0.64 (rising to 1.72) percentage points higher.
- Aggregate back-of-the-envelope (Table 2 column (3) debt-maturity coefficient = -0.091):
  - Aggregate TFP growth loss attributable to pre-existing financial vulnerabilities ≈ 2.39 percentage points.
  - Observed aggregate TFP growth drop in sample ≈ 6.37 percentage points.
  - Implied share explained ≈ 37% (~2.39/6.37) of the total within-firm TFP growth loss after the GFC.

### Dynamics and persistence
- Temporal pattern for TFP (Tables 7–8; N ≈ 115,214):
  - Effects open in 2009 and grow or remain persistent through 2013; no evidence of closing by 2013.
  - Selected dynamics (Leverage Pre-Crisis coefficients, Table 7):
    - 2008: 0.00796 (insignificant).
    - 2009: -0.0325***.
    - 2010: -0.0523***.
    - 2011: -0.0446***.
    - 2012: -0.0433***.
    - 2013: -0.0439***.
  - Selected dynamics (Debt Maturing 2008 coefficients, Table 8):
    - 2008: -0.0525***.
    - 2009: -0.0760***.
    - 2010: -0.0683***.
    - 2011: -0.0730***.
    - 2012: -0.0737***.
    - 2013: -0.0779***.
- Dynamics of intangible investment (more front-loaded; Tables 9–10):
  - Investment effects largest in 2008 and decline over time but remain negative through 2013.
  - Example (Leverage Pre-Crisis on ΔInvestment in Intangible Assets, Table 9; N = 114,680):
    - 2008: -0.182***.
    - 2009: -0.112***.
    - 2010: -0.0915***.
    - 2013: -0.0696***.
  - Example (Debt Maturing 2008 on ΔInvestment in Intangible Assets, Table 10; N = 116,843):
    - 2008: -0.0868***.
    - 2009: -0.0592***.
    - 2013: -0.0465***.

### Role of intangible investment (channel evidence)
- Hypothesis: Firms facing greater rollover risk cut intangible investment because physical capital is more pledgeable and converts to sales faster.
- Empirical results:
  - Firms with greater pre-existing balance sheet vulnerabilities decreased intangible investment significantly more than less vulnerable firms (Table 5 summary; N = 97,487; R-squared ≈ 0.041–0.045).
  - Effects amplified in countries with larger ΔCDS.
  - Magnitudes (intangible-investment illustration, column (3) in text):
    - A 10 percentage point higher leverage ratio associated with a 0.23 percentage points larger drop in the intangible investment rate post-crisis in an average country.
    - That number rises to 0.66 percentage point in a country whose banking sector experienced a 100 basis point larger increase in CDS spreads than the average country.
  - Change in share of intangible assets (Table 6; N = 101,150; R-squared ≈ 0.38):
    - A firm with a 10 percentage point higher leverage ratio reduced the share of intangible assets by 1.29 percentage points more than the average firm in the average country.
    - The same firm reduced the share by 3.94 percentage points more than the average firm in a country where bank CDS spreads rose by 100 basis points more than in the average country.
    - For a 10 percentage point increase in the share of debt maturing in 2008, corresponding effects are 0.61 and 1.36 percentage points, respectively.

### Robustness and placebo evidence
- Robustness checks:
  - Results robust to using labor productivity instead of TFP (Table 11; N = 106,395; R-squared ≈ 0.037–0.038):
    - Leverage Pre-Crisis: -0.027*** (0.006) in (1); -0.022*** (0.006) in (2).
    - Debt Maturing 2008: -0.050*** (0.005) in (1); -0.048*** (0.005) in (2).
  - Results robust to alternative measures of financial vulnerabilities (e.g., interest coverage ratio), different pre-crisis windows, and alternative measures of country credit conditions (national surveys of senior loan officers) — additional results available upon request.
- Placebo test (2000 recession as placebo; N = 53,200; R-squared = 0.157):
  - Coefficients on pre-2000 leverage and share of debt maturing in 2000 are not statistically significant (examples: Leverage Pre-Crisis = -0.00383 (0.015); Debt Maturing 2000 = -0.0657 (0.046)).
  - Graphical evidence shows no divergence in post-recession TFP growth paths by pre-crisis vulnerability around the 2000 recession.
  - Interpretation: effects are specific to the GFC banking-crisis-driven shock rather than to regular recessions.

### Interpretation, economic significance, and implications
- Mechanism: pre-existing firm balance sheet weaknesses interacted with sharper post-Lehman credit supply tightening (as proxied by ΔCDS) to produce larger and persistent within-firm TFP slowdowns, with reduced intangible investment a plausible channel.
- Economic significance:
  - Median firm-level TFP growth fell from 1.77 percent pre-crisis to -4.35 percent post-crisis (ΔTFP Growth median -7.12).
  - Substantial cross-firm heterogeneity in vulnerabilities: median leverage 8.71% (Standard Deviation 17.83); median debt maturing in 2008 = 26.64% of 2007 sales (Standard Deviation 23.95).
  - Aggregate calculation suggests pre-existing vulnerabilities may explain up to 37% of the within-firm TFP growth slowdown in the sample.
- Temporal patterns:
  - TFP effects were gradual, built up beginning in 2009, and were highly persistent through 2013.
  - Reductions in intangible investment were more front-loaded and somewhat shorter-lived than TFP effects.
- Suggested direction: future research should further examine channels (including reduced intangible investment) through which credit conditions affect within-firm productivity.

*Source: wp17129 (IMF working paper excerpt).*

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

### wp17129 - References

### Introduction and research question
- Productivity growth has declined in advanced economies since the global financial crisis (GFC) (Adler et al., 2017; OECD 2015).
- The paper investigates whether the abruptness, magnitude and persistence of the post-GFC fall in total factor productivity (TFP) growth can be linked to the interaction between tighter credit conditions and weak corporate balance sheets following the collapse of Lehman Brothers in September 2008.
- The analysis exploits the sharp and unforeseen tightening of credit conditions immediately after Lehman as an exogenous credit supply shock.

### Data and empirical strategy
- Uses an extensive cross-country firm-level dataset constructed by merging different waves of ORBIS.
- Compares within-firm TFP growth between pre- and post-crisis periods, and examines heterogeneity by pre-crisis firm-level financial vulnerabilities within narrowly defined country-industry cells.
- Measures of pre-crisis financial vulnerability:
  - Leverage (debt overhang risk).
  - Share of debt maturing within a year at end-2007 (pre-crisis rollover-risk), the preferred firm-level vulnerability measure.

### Main empirical findings
- Firms with greater pre-existing balance sheet vulnerabilities experienced significantly larger declines in average within-firm TFP growth between the pre- and post-crisis periods, even within narrowly defined country-industry cells.
- Pre-crisis balance sheet weakness was associated with a larger TFP slowdown in countries where credit conditions tightened more, as measured by the increase in the average CDS spread of domestic banks immediately after Lehman.
- Temporal pattern:
  - The TFP gap between more and less vulnerable firms opens in 2009 and further increases in subsequent years, indicating buildup over time and ruling out that the result is purely cyclical.
- Magnitude:
  - A back-of-the-envelope calculation suggests these effects may account for up to a third of the within-firm TFP growth slowdown between the six years before and six years after the crisis.
- Robustness and causal interpretation:
  - Results are not driven by more vulnerable firms being less productive or having enjoyed slower productivity gains before the crisis—more and less vulnerable firms do not differ significantly along these dimensions.
  - A placebo test shows the change in within-firm TFP growth between the pre- and post-2000 recession periods was unrelated to pre-2000 balance sheet vulnerabilities, highlighting the peculiar role of the GFC and its associated massive credit supply shock.

### Channels: role of intangible investment
- Hypothesis: Firms facing greater rollover risk had stronger incentives to cut intangible investment (e.g., R&D, workforce training) because physical capital is more pledgeable as collateral and converts to sales faster.
- Evidence:
  - Firms with pre-existing balance sheet vulnerabilities cut back on intangible investment more than less vulnerable counterparts after the crisis.
  - The divergence in intangible investment was larger in countries where credit conditions tightened more during Lehman.
- The findings are consistent with theoretical predictions that reduced investments in intangible assets can slow within-firm productivity growth.

### Relation to literature
- Differs from the literature on misallocation (Hsieh and Klenow, 2009; Midrigan and Xu, 2014; Moll, 2014; Benigno et al., 2015; Borio et al., 2016; Gopinath et al., 2016) by focusing on within-firm productivity growth rather than cross-firm resource misallocation.
- Relates closely to work showing credit constraints lead firms to cut R&D and illiquid long-term investments during recessions (Aghion et al., 2010; Aghion et al., 2012) and to models linking reduced intangible investment to slower within-firm productivity growth (Garcia-Macia, 2016; Anzoategui et al., 2016).

### Paper structure
- Section 2: Empirical strategy and dataset description.
- Section 3: Main results.
- Section 4: Dynamics of the productivity effects identified.
- Section 5: Robustness checks.
- Section 6: Conclusion.

*Source: wp17129 - References*

### 2. Empirical Strategy

### 2. Empirical Strategy

### Identification Approach
- Strategy: differences-in-differences comparing the change in firm-level TFP growth between the post-crisis (five years after 2008) and pre-crisis (five years until 2008) periods for firms with large versus low pre-existing balance sheet vulnerabilities.
- Baseline regression: change in average TFP growth for firm i (in sector s and country c) regressed on pre-crisis balance sheet vulnerabilities, firm-level controls (age, log total assets, log earnings (EBITDA) before the financial crisis), country-sector fixed effects, and clustered standard errors at the country-sector level.
- Fixed effects and controls:
  - Country-sector fixed effects included to compare more and less vulnerable firms within narrowly defined country-sector cells and to control for sectoral differences in external finance reliance, trade intensity, sectoral cyclical impacts, and post-crisis policy changes.
  - Time-invariant firm characteristics are implicitly controlled for by focusing on within-firm change across two periods.
- Exploiting cross-country variation in credit shock severity:
  - Augmented specification includes interaction between firm-level vulnerability and a country-level measure of the credit shock.
  - Country-level shock measure: change in the average CDS spread of domestic banks in country c around the collapse of Lehman Brothers in September 2008.
    - The paper describes ܵܦܥ∆
௖ as the change in the average CDS spread of domestic banks between the first and second halves of 2008.
    - Elsewhere the change is computed as the difference in the average CDS spread between the two quarters before and the two quarters after the Lehman bankruptcy.
  - Interpretation: larger increases in domestic bank CDS spreads reflect greater exposure to the Lehman bankruptcy, larger perceived banking vulnerabilities, and a more adverse credit supply shock; this variation is treated as exogenous to changes in firm productivity during the crisis.
  - Implicit assumption: domestic firms rely heavily on banks in their home country for funding and cannot fully substitute with corporate bonds, syndicated lending, or cross-border bank lending (reasonable given the sample composition).
- Firm-level financial vulnerability measures:
  - Leverage: captures debt overhang risk; calculated as (current liabilities + long-term debt) / total assets, averaged 2003–2007.
  - Rollover risk (preferred measure): share of pre-crisis debt scheduled to mature during the crisis, measured as current liabilities (maturing within a year) at end-2007 divided by total sales in 2007. Rationale: maturity structure prior to the unforeseen September 2008 shock unlikely correlated with unobserved determinants of post-crisis TFP growth.

### Data and stylized facts
- Data sources and construction:
  - Firm-level variables from ORBIS (Bureau van Dijk): harmonized longitudinal data on listed and unlisted firms (value-added output, capital stock, employment, debt, assets, tangible and intangible fixed assets, long-term debt).
  - Real monetary variables: original USD data converted to local currency; nominal variables deflated using local currency deflators from OECD STAN (ISIC4) rebased to 2005 US dollars using country-industry level PPPs from Timmer and Inklaar (2014).
  - Capital stock and investment: derived via perpetual inventory method (PIM) using depreciation and tangible fixed assets (details per Gal, 2013; Gal and Hijzen, 2016).
  - Productivity measurement: firm-level TFP growth computed using real value added, real capital stock, and number of employees applying the Wooldridge (2009) control function approach; robustness check uses labor productivity (real value added per employed worker).
  - Aggregate country-level financial conditions: main measure is economy-wide average bank CDS spreads (simple average across a country’s banks); results robust to principal component alternative.
- Sample composition:
  - 15 advanced economies: Austria, Belgium, Denmark, Germany, Spain, France, Greece, Italy, Japan, Korea, the Netherlands, Portugal, Slovenia, the United Kingdom, and the United States.
  - Sector scope: non-farm, non-financial business sector (NACE Rev.2 two-digit industry codes 5–82), covering manufacturing and selected services including real estate and professional/scientific/technical activities.
  - Sample restrictions: exclude firms with less than 3 employees; require firms to report at least four consecutive periods.
- Key variable definitions:
  - Leverage Pre-Crisis: average pre-crisis debt-to-assets ratio (sum of current liabilities and long-term debt divided by total assets) averaged 2003–2007.
  - Debt Maturing 2008: amount of debt maturing in 2008 divided by average total sales pre-crisis (2007 balance sheet current liabilities divided by total sales).
  - Change in CDS: described both as change between first and second halves of 2008 and as difference between two quarters before and two quarters after Lehman bankruptcy.
- Summary statistics (Table 1, unweighted, within-firm focus):
  - Average TFP Growth Pre-Crisis: median 1.77 percent; 25th percentile -7.50 percent; 75th percentile 11.50 percent; Standard Deviation 17.38
  - Average TFP Growth Post-Crisis: median -4.35 percent; 25th percentile -17.0 percent; 75th percentile 4.73 percent; Standard Deviation 17.19
  - ΔTFP Growth (post vs. pre): median -7.12; 25th percentile -27.67; 75th percentile 9.22; Standard Deviation 32.80
  - Leverage Pre-Crisis: median 8.71%; 25th percentile 1.23%; 75th percentile 23.11%; Standard Deviation 17.83
  - Debt Maturing 2008: median 26.64%; 25th percentile 16.66%; 75th percentile 42.84%; Standard Deviation 23.95
  - Note: `Average TFP Growth Pre-Crisis’ is the average TFP growth rate pre-crisis. `Average TFP Growth Post-Crisis’ is the average TFP growth rate post-crisis. `ΔTFP Growth’ is the difference in the TFP growth rate post vs. pre-crisis. `Percentage Change in Leverage Ratio’ is the percentage change in the leverage ratio. `Leverage Pre-Crisis’ is the average pre-crisis debt-to-assets ratio. `Debt Maturing 2008’ is the amount of debt maturing in 2008 divided by average total sales pre-crisis. The post-crisis sample starts in 2008.
- Stylized patterns:
  - Median firm experienced a large drop in within-firm TFP growth after the 2008 financial crisis: from 1.77 percent pre-crisis to -4.35 percent post-crisis (ΔTFP Growth median -7.12).
  - Substantial cross-firm variation in financial vulnerabilities: median leverage 8.71% (Standard Deviation 17.83); median debt maturing in 2008 equals 26.64% of 2007 sales (Standard Deviation 23.95).
  - TFP level paths (Figure 1): before the crisis, “weak” and “resilient” firms had similar productivity growth; after 2008, trajectories diverged—especially by ex-ante rollover risk—with a large gap opening in 2009 that is not closed by 2013 and appears to widen.
- Measurement caveats and identification considerations:
  - Use of revenue-based productivity (Wooldridge control function) implies firm-specific price variations within sector affect productivity estimates; such effects could reflect quality changes or market power.
  - If more resilient firms increased prices since the crisis, this would mechanically raise measured productivity growth for those firms; however, Gilchrist et al. (2017) show financially constrained firms raised prices during the financial crisis, implying any bias would, if anything, be downward.
  - Rollover-risk measure preferred because future debt structure was pre-determined relative to the unforeseen September 2008 shock and thus less likely correlated with unobserved determinants of post-crisis TFP changes.

*Source: wp17129 - 2. Empirical Strategy (IMF working paper content provided).*

### 3. Empirical Results

### 3. Empirical Results

### Baseline and extended specifications

- Research design
  - Baseline regression: dependent variable `ΔTFP Growth` is the difference in the average TFP growth rate between post- and pre-crisis periods. Post-crisis period starts in 2008.
  - Key firm-level vulnerability measures:
    - `Leverage Pre-Crisis`: average pre-crisis debt-to-assets ratio.
    - `Debt Maturing 2008`: amount of debt maturing in 2008 divided by average total sales pre-crisis.
  - Controls: firm age, size of assets, earnings (EBITDA). Country*Sector fixed effects. Standard errors clustered at the country-sector level.

- Baseline results (Table 2)
  - Column (1): Leverage Pre-Crisis coefficient = -0.047*** (standard error 0.008).
  - Column (2): Debt Maturing 2008 coefficient = -0.094*** (standard error 0.010).
  - Column (3): Both variables included:
    - Leverage Pre-Crisis coefficient = -0.036*** (standard error 0.006).
    - Debt Maturing 2008 coefficient = -0.091*** (standard error 0.010).
  - Fit and sample:
    - R-squared: 0.149 (col 1), 0.151 (col 2), 0.151 (col 3).
    - N = 134,838 (all columns).
  - Magnitudes (from baseline column (3)):
    - A 10 percentage point higher leverage ratio associated with a 0.36 percentage point decline in average annual TFP growth after the crisis.
    - A 10 percentage points increase in the share of debt maturing in 2008 associated with a 0.9 percentage point drop in annual TFP growth.
  - Robustness: results robust to controlling for firms’ average rollover risk over 2003-2007; share of debt maturing in 2008 is uncorrelated with leverage ratio.

- Extended specification exploiting cross-country heterogeneity (interaction with ΔCDS; Table 3)
  - `∆CDS`: difference in the average CDS spread of banks in each country between the two quarters after and the two quarters before the Lehman bankruptcy. Demeaned by country sample mean so a one unit change in ΔCDS = 1 percentage point = 100 basis points larger increase relative to sample mean.
  - Table 3 coefficients (column numbering corresponds to table):
    - Leverage Pre-Crisis (direct) = -0.046*** (standard error 0.010) in a specification reported.
    - Leverage Pre-Crisis * ΔCDS = -0.051*** (standard error 0.021) in a specification reported.
    - Debt Maturing 2008 (direct) = -0.108*** (standard error 0.008) in a specification reported.
    - Debt Maturing 2008 * ΔCDS = -0.064*** (standard error 0.017) in a specification reported.
  - Fit and sample:
    - R-squared: 0.163 (col 1), 0.167 (col 2), 0.167 (col 3).
    - N = 104,275 (all columns).
  - Interpreting magnitudes (text illustration based on column (3)):
    - In a country with an average increase in bank CDS spreads:
      - A 10 percentage point higher leverage ratio associated with a 0.46 percentage point post-crisis decline in average annual TFP growth.
      - A 10 percentage points increase in the share of debt maturing in 2008 associated with a 1.08 percentage point drop in annual TFP growth.
    - In a country where the increase in CDS spreads was 100 basis points larger than the average:
      - The corresponding figures were 0.51 (rising to 0.97) and 0.64 (rising to 1.72) percentage points higher. 
        - (Textual presentation: “the corresponding figures were 0.51 (rising to 0.97) and 0.64 (rising to 1.72) percentage points higher.”)
  - Graphical illustration: Figure 2 and Figure 3 show that firms at the 75th percentile of pre-crisis vulnerability indicators experienced substantially larger post-crisis TFP declines than firms at the 25th percentile, with the gap larger in countries where bank CDS spreads rose more.

- Aggregate illustrative calculation (back-of-the-envelope)
  - Method: assume firms with zero debt maturing in 2008 experienced no finance-related TFP slowdown; use coefficient of debt maturity 2008 from Table 2 column (3) = -0.091 multiplied by each firm’s share of debt maturing in 2008, aggregate using value added weights.
  - Results:
    - Aggregate TFP growth loss attributable to pre-existing financial vulnerabilities ≈ 2.39 percentage points.
    - Observed aggregate TFP growth drop in sample ≈ 6.37 percentage points.
    - Implied share explained ≈ 37% (~2.39/6.37) of the total within-firm TFP growth loss after the GFC.

### Placebo Test

- Purpose: test whether effects are specific to the GFC (a banking-crisis-driven recession) by using the 2000 recession (dot-com bust) as placebo (assume crisis year = 2000; placebo post-crisis period runs 2000–2005).
- Results (Table 4):
  - None of the coefficients on pre-2000 leverage or share of debt maturing in 2000 are statistically significant.
  - Table 4 coefficients (columns reported):
    - Leverage Pre-Crisis = -0.00383 (standard error 0.015) in column (1); = 0.00620 (standard error 0.017) in column (2).
    - Debt Maturing 2000 = -0.0657 (standard error 0.046) in column (2); = -0.0690 (standard error 0.050) in column (3) as presented.
  - Fit and sample:
    - R-squared = 0.157 (all columns).
    - N = 53,200 (all columns).
  - Graphical evidence: Figure 4 shows no divergence in post-recession TFP growth paths by pre-crisis vulnerability around the 2000 recession.
- Interpretation: absence of statistically significant effects in the placebo test supports that the documented TFP impacts are related to the financial-crisis credit supply shock of the GFC rather than to regular recessions.

### The Role of Intangible Investment

- Hypothesis: credit-constrained firms cut investment in intangible assets (which cannot be pledged as collateral, have longer gestation, and are riskier), contributing to the post-crisis productivity slowdown.
- Measurement:
  - Dependent variable: change in the investment rate in intangible assets = change in the stock of intangible assets divided by value added (from ORBIS).
  - Estimation uses same difference-in-differences strategy as for TFP.
- Results (Table 5 summary as described in text):
  - Firms with more vulnerable balance sheets decreased intangible investment significantly more than less vulnerable firms.
  - Effect amplified in countries whose banking sectors were more exposed to the collapse of Lehman Brothers (i.e., larger ΔCDS).
  - Magnitudes (column (3) reported in text):
    - A 10 percentage point higher leverage ratio in the pre-crisis period is associated with a 0.23 percentage points larger drop in the intangible investment rate post-crisis in a country with average exposure to the Lehman bankruptcy.
    - That number rises to 0.66 percentage point in a country whose banking sector experienced a 100 basis point larger increase in CDS spreads than the average country.
    - A 10 percentage point higher share of debt maturing in 2008 is associated with a 0.19 larger decline in TFP growth in the average country (text presents this coefficient for the intangible-investment analysis before truncation).
  - Economic significance: because investment rates are typically much lower for intangible assets than for tangible ones, these estimated declines are economically meaningful.
- Interpretation: weaker intangible investment is a plausible channel through which tighter credit conditions and pre-existing firm vulnerabilities contributed to the post-GFC TFP slowdown.

*Source: wp17129 - 3. Empirical Results (IMF working paper content).*

### 0.33  larger  decline  in  a  country  where  CDS  spreads  rose  100  basis  points  more  than  in  the

### wp17129 - 0.33  larger  decline  in  a  country  where  CDS  spreads  rose  100  basis  points  more  than  in  the

### Intangible Investment Regression Results
- Dependent variable: `ΔInvestment in Intangible Assets` (difference in investment in intangible assets as a ratio of value added post vs. pre-crisis).
- Key regressors and estimated coefficients (Table 3):
  - `Leverage Pre-Crisis`: -0.025*** (0.004) in (1); -0.023*** (0.004) in (2).
  - `Leverage Pre-Crisis * ΔCDS`: -0.047*** (0.009) in (1); -0.043*** (0.009) in (2).
  - `Debt Maturing 2008`: -0.02*** (0.002) in (1); -0.019*** (0.002) in (2).
  - `Debt Maturing 2008 * ΔCDS`: -0.018*** (0.004) in (1); -0.014*** (0.004) in (2).
- Controls: Yes. Country*Sector FE: Yes.
- Goodness of fit and sample size:
  - R-squared: 0.041 (cols 1–2), 0.045 (col 3).
  - N: 97,487.

- Interpretation and graphical evidence (Figure 5 summary):
  - Firms at the 25th percentile of financial vulnerability reduced intangible investment substantially less than firms at the 75th percentile.
  - The relationship is stronger in countries at the 75th percentile of change in CDS spreads around the Lehman Bankruptcy relative to the average country.

### Change in Share of Intangible Assets
- Dependent variable: `ΔShare of Intangible Assets` (difference in share of intangible assets in total capital post vs. pre-crisis).
- Key magnitudes (Table 6, column (3) referenced in text):
  - A firm with a 10 percentage point higher leverage ratio reduced the share of intangible assets by 1.29 percentage points more than the average firm in the average country.
  - The same firm reduced the share by 3.94 percentage points more than the average firm in a country where bank CDS spreads rose by 100 basis points more than in the average country.
  - For a 10 percentage point increase in the share of debt maturing in 2008, the corresponding effects are 0.61 and 1.36 percentage points, respectively.
- Estimated coefficients (Table 6):
  - `Leverage Pre-Crisis`: -0.136*** (0.018) in (1); -0.129*** (0.018) in (2).
  - `Leverage Pre-Crisis * ΔCDS`: -0.277*** (0.043) in (1); -0.265*** (0.043) in (2).
  - `Debt Maturing 2008`: -0.069*** (0.009) in (1); -0.061*** (0.009) in (2).
  - `Debt Maturing 2008 * ΔCDS`: -0.103*** (0.017) in (1); -0.075*** (0.016) in (2).
- Controls: Yes. Country*Sector FE: Yes.
- Goodness of fit and sample size:
  - R-squared: 0.382 (col 1), 0.3795 (col 2), 0.3835 (col 3).
  - N: 101,150.

### Dynamic Responses of Productivity Growth
- Objective: Assess persistence of effects from 2008 to 2013; show effects are not solely transitory.
- Summary of patterns:
  - TFP growth response to pre-crisis leverage and its interaction with `ΔCDS` is insignificant in 2008 (column (1)), negative and highly significant for 2008–2009 (column (2)), grows as the horizon lengthens, then stabilizes without evidence of decline by 2013.
  - Conclusion: cumulative TFP level gap between more and less vulnerable firms that opened up in 2009 does not start shrinking even by 2013.
  - Similar pattern when using `Debt Maturing 2008` as vulnerability measure: effects begin in 2008 and grow or remain persistent over time.
- Selected dynamic coefficient estimates (Table 7: leverage; Table 8: debt maturity):
  - Table 7 (`ΔTFP Growth`):
    - `Leverage Pre-Crisis`: 0.00796 (2008); -0.0325*** (2009); -0.0523*** (2010); -0.0446*** (2011); -0.0433*** (2012); -0.0439*** (2013). Standard errors shown in table.
    - `Leverage Pre-Crisis * ΔCDS`: -0.0226 (2008); -0.0621** (2009); -0.0899*** (2010); -0.0703*** (2011); -0.0623*** (2012); -0.0615*** (2013).
    - R-squared range: 0.0379 to 0.225 across columns. N: 115,214.
  - Table 8 (`ΔTFP Growth` to `Debt Maturit y 2008`):
    - `Debt Maturity 2008`: -0.0525*** (2008); -0.0760*** (2009); -0.0683*** (2010); -0.0730*** (2011); -0.0737*** (2012); -0.0779*** (2013).
    - `Debt Maturity 2008 * ΔCDS`: -0.0308** (2008); -0.0681*** (2009); -0.0562*** (2010); -0.0653*** (2011); -0.0634*** (2012); -0.0727*** (2013).
    - R-squared range: 0.0388 to 0.226. N: 115,214.

### Dynamic Responses of Investment in Intangible Assets
- Investment in intangibles: effects are more front-loaded and shorter-lived than TFP effects (Tables 9 and 10).
- Table 9 (`ΔInvestment in Intangible Assets` to `Leverage Pre-Crisis`):
  - `Leverage Pre-Crisis`: -0.182*** (2008); -0.112*** (2009); -0.0915*** (2010); -0.0777*** (2011); -0.0727*** (2012); -0.0696*** (2013).
  - `Leverage Pre-Crisis * ΔCDS`: -0.321*** (2008); -0.187*** (2009); -0.154*** (2010); -0.132*** (2011); -0.121*** (2012); -0.115*** (2013).
  - R-squared range reported around 0.0793–0.0926. N: 114,680.
- Table 10 (`ΔInvestment in Intangible Assets` to `Debt Maturing 2008`):
  - `Debt Maturity 2008`: -0.0868*** (2008); -0.0592*** (2009); -0.0517*** (2010); -0.0534*** (2011); -0.0490*** (2012); -0.0465*** (2013).
  - `Debt Maturity 2008 * ΔCDS`: -0.122*** (2008); -0.0804*** (2009); -0.0666*** (2010); -0.0661*** (2011); -0.0645*** (2012); -0.0607*** (2013).
  - R-squared range reported around 0.0731–0.106. N: 116,843.

### Robustness: Labor Productivity versus TFP
- Re-estimated baseline regressions using change in labor productivity (real value-added output per employee).
- Table 11 results confirm baseline findings:
  - `Leverage Pre-Crisis`: -0.027*** (0.006) in (1); -0.022*** (0.006) in (2).
  - `Debt Maturing 2008`: -0.050*** (0.005) in (1); -0.048*** (0.005) in (2).
  - Controls: Yes. Country*Sector FE: Yes.
  - R-squared: 0.037–0.038. N: 106,395.
- Note in text: results are robust to other measures of financial vulnerabilities (e.g., interest coverage ratio), different pre-crisis windows, and alternative measures of country credit conditions (national surveys of senior loan officers); these additional results are available upon request.

### Conclusion — Main Findings and Implications
- Pre-existing financial fragilities interacted with tighter credit conditions after the 2008 shock to:
  - Weaken within-firm productivity growth, especially in countries where credit conditions tightened more.
  - Lead more vulnerable firms to cut back on investment in intangible assets more than less vulnerable firms.
- Temporal patterns:
  - TFP effects were gradual, highly persistent, and did not show evidence of closing by 2013.
  - Reductions in intangible investment were more front-loaded and shorter-lived.
- Suggested direction: Future research should examine the channels (including reduced intangible investment) through which credit conditions affect within-firm productivity.

*Source: Excerpt from wp17129 (IMF working paper content provided).*

### References

### wp17129 - References

### Productivity and Technology
- Adler, Gustavo, Romain Duval, Davide Furceri, Ksenia Koloskova and Marcos Poplawski-Ribeiro, 2017. “Gone with the Headwinds: The Global Productivity Slowdown,” IMF Staff Discussion Note, forthcoming.
- Alcalá, F. and A. Ciccone, 2014. “Trade and Productivity,” The Quarterly Journal of Economics, 119 (2), 613-646.
- Andrews, Dan, Chiara Criscuolo and Peter Gal. 2015. Frontier Firms, Technology Diffusion and Public Policy: Micro Evidence from OECD Countries,” OECD Productivity Working Papers No. 2, OECD Publishing.
- Anzoategui, D., D. Comin, M. Gertler and J. Martinez, 2016. “Endogenous Technology Adoption and R&D as Sources of Business Cycle Persistence,” NBER Working Paper Series 22005.
- Cette, G., J. Fernald and B. Mojon. 2016. “The pre-Great Recession slowdown in productivity,” European Economic Review (88) 1-120.
- Corrado, C., C. Hulten and D. Sichel, 2005. Measuring Capital and Technology: An Expanded Framework. In Measuring Capital in New Economy, ed. John Haltiwanger, Carol Corrado, and Dan Sichel. Chicago, IL: University of Chicago Press pp. 114-146.
- Corrado, C., C. Hulten and D. Sichel, 2009. “Intangible Capital and U.S. Economic Wealth,” Review of Income and Wealth, 55, 661-685.
- Decker, R.A., J. Haltiwanger, R.S Jarmin and J. Miranda, 2016. “Where has all the skewness gone? The decline in high-growth (young) firms in the US". European Economic Review, 86, 4-23.
- Fernald, John. 2015. Chapter in NBER book NBER Macroeconomics Annual 2014, Volume 29 (2015), Jonathan A. Parker and Michael Woodford, editors (p. 1 - 51).
- Gal, Peter. 2013. “Measuring Total Factor Productivity at the Firm Level using OECD-Orbis,” OECD Economics Department Working Paper 1049, OECD Publishing.
- Gal, P. and A. Hijzen, 2016. “The Short-Term Impact of Product Market Reforms: A Cross-Country Firm-Level Analysis,” IMF Working Paper Series, WP/16/116. International Monetary Fund.
- Garcia-Macia, Daniel. 2016. “The Financing of Ideas and the Great Deviation”, Mimeo, Stanford University.
- Gopinath, G., S. Kalemli-Ozcan, L.Karabarbounis and C. Villegas-Sanchez. 2015. “Capital Allocation and Productivity in South Europe,” NBER Working Paper Series 21453.
- Gordon, Robert, 2016. “The Rise and Fall of American Growth: The U.S. Standard of Living since the Civil War,” Princeton: Princeton University Press.
- OECD. 2015. “The Future of Productivity,” Joint Economics Department and the Directorate for Science, Technology and Innovation Policy Note, July, OECD publishing.
- Syverson, C. 2011. “What determines productivity,” Journal of Economic Literature, 49 (2), 326-65.

### Financial Frictions, Credit, and Crisis Effects
- Aghion, Philippe. George-Marios Angeletos, Abhijit Banerjee and Kalina Manova, 2010. “Volatility and Growth: Credit Constraints and the Composition of Investment,” Journal of Monetary Economics, 57 (3), 246-65.
- Aghion, P, P. Askenazy, N. Berman, G. Cette, and L. Eymard, 2012. “Credit Constraints and the Cyclicality of R&D Investment: Evidence from France,” Journal of the European Economic Association, 10 (5), 1001-1024.
- Almeida, H. M. Campello, B. Laranjeira and S. Weisbenner, 2012. “Corporate Debt Maturity and the Real Effects of the 2007 Credit Crisis,” Critical Finance Review, 1(1), 3-58.
- Benigno, G., N. Converse, and L. Fornaro, 2015. “Large Capital Flows, Sectoral Allocation, and Economic Performace,” Journal of International Money and Finance, 55, 60-87.
- Borio, C. E., E. Kharroubi, C. Upper, and F. Zampolli, 2016. “Labour Reallocation and Productivity Dynamics: Financial Causes, Real Consequences.” BIS Working Paper
- Cerra, V. and S. Saxena, 2008. “The Myth of Economic Recovery,” American Economic Review, 98 (1).
- Gilchrist, S., R. Schoenle, J. Sim and E. Zakrajske, 2017. “Inflation Dynamics during the Financial Crisis,” American Economic Review, 107 (3): 785-823.
- Giroud, X. and H. Mueller, 2017. “Firm Leverage, Consumer Demand, and Employment Losses during the Great Recession,” Quarterly Journal of Economics, 132 (1): 271-316..
- Khwaja, A. and A. Mian, 2008, “Tracing the Impact of Bank Liquidity Shocks: Evidence from an Emerging Market,” American Economic Review, 98 (4), 1413-1442.
- Matsuyama, Kiminori, 2007. “Credit Traps and Credit Cycles,” American Economic Review, 97 (1), 503-516.
- Mian, A. and A. Sufi, 2014. “What explains the 207-2009 Drop in Employment,” Econometrica, 82 (6), 2197-2223.
- Midrigan, V. and D. Xu, 2014. “Finance and Misallocation: Evidence from Plant-Level Data,” American Economic Review, 104 (2), 422-458.
- Moll, B., 2014. “Productivity Losses from Financial Frictions: can self-financing undo Capital Misallocation?”, American Economic Review, 104 (10), 3186-3221.
- Rajan, R. and L. Zingales, 1998. “Financial Dependence and Growth,” American Economic Review, 88 (3), 559-586.
- Rioja, F, F. Rios and N. Valev, forthcoming. “The Persistent Effect of Banking Crises on Investment and the Role of Financial Markets,” Journal of Financial Economic Policy.

### Firm-level Measurement, Estimation Methods, and Micro Evidence
- Bertrand, M., E. Duflo and S. Mullainathan, 2004. “How much should we trust differences-in-differences estimates,” The Quarterly Journal of Economics, 119 (1), 249-275.
- Decker, R.A., J. Haltiwanger, R.S Jarmin and J. Miranda, 2016. “Where has all the skewness gone? The decline in high-growth (young) firms in the US". European Economic Review, 86, 4-23.
- Gal, Peter. 2013. “Measuring Total Factor Productivity at the Firm Level using OECD-Orbis,” OECD Economics Department Working Paper 1049, OECD Publishing.
- Haltiwanger, J., Hathaway, I., and J. Miranda, 2014, “Declining Business Dynamism in the U.S. High-Technology Sector”
- Hsieh, C.T. and P.J.Klenow, P.J., 2009. “Misallocation and manufacturing TFP in China and India.” The Quarterly Journal of Economics, 124(4), 1403-1448.
- Kalemli-Ozcan, S., B. Sorensen, C. Villegas-Sanchez, V.Volosovych and S. Yesiltas, 2015. “How to Construct Nationally Representative Firm Level Data from the Orbis Global Database,” NBER Working Paper Series 21558, National Bureau of Economic Research.
- Levinsohn, J. and A. Petrin. 2003. “Estimating Production Functions using Inputs to Control for Unobservables,” Review of Economic Studies, 70 (2), 317-341.
- Olley, S. and A. Pakes, 1996. “The Dynamics of Productivity in the Telecommunications Equipment Industry,” Econometrica, 64 (6), 1263-1297.
- Wooldridge, Jeffrey, 2009. “On Estimating Firm-Level Production Functions using Proxy Variables to Control for Unobservables,” Economic Letters, 104(3), 112-114.

*References list from wp17129 - References*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2017/wp17129.pdf_