## _wp16174

## Source details

**Canonical URL:** [_wp16174](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16174.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16174.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16174.pdf.json)

---

### I. Introduction — purpose and headline findings
- Purpose: study the impact of uncertainty on productivity (TFP).
- Key empirical finding: the adverse impact of uncertainty on industry-level productivity growth is greater in industries that rely more heavily on external finance.
- Data scope for main empirical result: 25 industries for 18 advanced economies (Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Hungary, Korea, Ireland, Italy, Japan, the Netherlands, Spain, Sweden, the United Kingdom and the United States) over the period 1985-2010.
- Supporting evidence from related literature:
  - Bloom et al. (2014): a temporary increase in uncertainty lowers aggregate output, investment and productivity; impact on output is a drop of 3 percent within one quarter; TFP drops by about 0.5 percent within a year.
  - Mechanisms highlighted: firms adopt “wait and see” attitudes; increased use of temporary/short-term inputs and investments reduces aggregate productivity and growth.

### II. Theoretical framework — setup and formal predictions
- Model setup and key assumptions:
  - Two-period model with short-term (liquid) and long-term (more productive, delayed) investments.
  - Aggregate productivity: a_{t+1} = a_{t}(1+σ_{t} ε_{t+1}), where σ_{t} ∈ {σ_{L}, σ_{H}} and ε_{t} ~ N(0,1).
  - Human capital allocation: K_{t}=kH_{t}, Z_{t}=zH_{t} with w=k+z.
  - Firm types: fraction 1−μ unconstrained; fraction μ constrained; 0<μ<1.
  - Constrained firms face idiosyncratic liquidity shock C_{i,t+1}=c_{i,t+1}H_{t} with c_{i,t} ~ i.i.d. uniform(0,1); survival probability f_{t+1}=Pr(a_{t+1} k^{α} ≥ c) = min{a_{t+1} k^{α},1}.
  - Short-term profits π_{t+1}=a_{t+1} k^{α} H_{t}; long-term yields v_{t+1} H_{t} with probability λz.
- Formal results (selected):
  - Proposition 1: z_{nc}>0 and z_{nc} > z_{c}.
  - Lemma 1: A mean preserving spread of aggregate productivity distribution decreases z_{c}, but does not affect z_{nc}. (∂ z_{c} / ∂ σ_{t} < 0)
  - Growth decomposition: ln A_{i,t+1} − ln A_{i,t} = ln a_{t+1} − ln a_{t} + ln H_{i,t+1} − ln H_{i,t}.
  - Expected TFP growth: g_{t+1} = E[ln A_{i,t+1} − ln A_{i,t}] = λ((1−μ) z_{nc} + μ z_{c} min( a_{t+1} (w−z_{c})^{α}, 1 )).
  - Lemma 2: ∂ g_{t+1} / ∂ σ_{t} < 0 (an increase in aggregate uncertainty reduces expected TFP growth).
  - Lemma 3: ∂/∂μ (∂ g_{t+1} / ∂ σ_{t}) < 0 (reduction in expected TFP growth from increased uncertainty is larger the higher the fraction of constrained firms).
  - Lemma 4: ∂/∂ a_{t+1} ( ∂/∂μ (∂ g_{t+1} / ∂ σ_{t} )) < 0 (differential effect larger when realized productivity is lower).
- Theoretical implications tested empirically:
  - Main hypothesis: increase in aggregate uncertainty reduces productivity growth more in industries that are financially constrained.
  - Additional implication: interaction between uncertainty and financial constraints on TFP growth is larger in recessions than expansions.

### III. Data and measurement
- Sample: unbalanced panel of 25 industries for 18 advanced economies over 1985-2010.
- Country-specific uncertainty proxy:
  - Baseline measure: realized volatility of aggregate stock market returns for each country.
  - Annualized realized volatility: RV_{i,t} = 100 √( (1/T_{i}) ∑_{s=1}^{T_{i}} r_{i,s}^{2} ), where r_{i,s} are daily returns and T_{i} is the stock market i's number of trading days in a year.
  - Daily closing prices source: Global Financial Data.
  - Robustness: economic policy uncertainty indices (Baker et al., 2015) for a 10-country subsample.
- Sectoral data:
  - Sectoral TFP (labor productivity) growth from EU KLEMS and World KLEMS databases.
  - Industry dependence on external finance: Compustat US firm-level median of (total capital expenditures − current cash flow) / total capital expenditures, following Rajan and Zingales (1998).
- Stock market data coverage examples (Table A.1):
  - Australia: Australia ASX All-Ordinaries (w/GFD extension) 1985-2010
  - Hungary: Budapest Stock Exchange Index (BUX) 1992-2010
  - United States: S&P 500 Composite Price Index (w/GFD extension) 1985-2010
  - (Full table lists exchanges and coverage for all 18 countries.)

### IV. Empirical methodology
- Baseline specification (equation (5)):
  - TFP_{i,j,t} = α_{i,t} + γ_{i,j} + β fd_j U_{i,t} + ε_{i,j,t}, where fd_j = industry external finance dependence (US-based), α_{i,t} = country-time fixed effects, γ_{i,j} = country-industry fixed effects.
- Estimation: OLS with standard errors clustered at the country-industry level.
- Identification advantages of the j,i,t design: controls for aggregate and country-sector shocks via α_{i,t} and γ_{i,j}; mitigates reverse causality concerns; allows identification of non-linearities across industries.
- Controls and robustness checks: interactions of fd with credit-to-GDP, inflation, government consumption/GDP, budget balance/GDP; alternative dependent variable labor productivity growth; alternative uncertainty measure economic policy uncertainty.

### V. Empirical results — primary claims and magnitudes
- Primary empirical claim: the adverse impact of uncertainty on industry-level productivity growth is greater in industries that rely more heavily on external finance.
- Baseline coefficients (Uncertainty * financial dependence):
  - Table 2 column I: -4.102*** (t-statistic -3.24).
  - Table 2 column II: -2.889***.
  - Table 2 column III: -2.513***.
  - Table 2 column IV (lag): -2.945***.
  - Table 2 column V (ordinal fd): -0.214***.
- Differential TFP growth loss (25th to 75th percentile of uncertainty, comparing industry at 25th percentile of fd to one at 75th percentile):
  - Table 2 column I baseline: about 2.1 percentage points.
  - Other Table 2 columns: -1.5, -1.3, -1.5, -3.3 percentage points respectively.
- Aggregate-level correlations:
  - US quarterly sample 1970-2013: Coef = -2.16; t-statistics = -3.65 (Figure 3 Panel A).
  - 18 advanced economies annual sample 1985-2013: Coef = -0.76; t-statistics = -3.51 (Figure 3 Panel B).
- Industry-level raw-data pattern (Figure 4):
  - Industries below median financial dependence: Coef = 3.4; t-statistics = 3.6 (positive relationship).
  - Industries above median financial dependence: Coef = -2.8; t-statistics = -2.9 (negative relationship).
- Robustness checks and extensions (selected):
  - Specification robustness: results hold with (i) only country-time fixed effects + industry dummies, (ii) country, time, and industry dummies without interactions, (iii) lagged interaction, (iv) categorical ordinal fd.
  - Manufacturing vs non-manufacturing (Table 3): Manufacturing Uncertainty*financial dependence = -4.423*** (t = -3.23); differential TFP growth = -2.3 percentage points. Non-manufacturing coefficient = -1.107 (t = -0.43); differential = -0.6 percentage points.
  - Controls for omitted variables (Table 4): including Credit-to-GDP * fd, Inflation * fd, Government size * fd, Budget balance * fd separately or jointly leaves Uncertainty*fd negative and significant; differential TFP loss ≈ -2.0 percentage points (column I).
  - Alternative dependent variable — labor productivity growth (Table 5): Uncertainty*fd negative and significant (e.g., -4.251*** column I); differential productivity loss from 25th to 75th uncertainty percentile ≈ -2.2 percentage points in baseline.
  - Alternative uncertainty measure — economic policy uncertainty (Table 6): Economic policy uncertainty * fd coefficient: -0.081* (t = -1.62) for TFP and -0.093** (t = -2.03) for productivity. Differential effects from 25th to 75th percentiles: TFP growth ≈ -4.0 percentage points; productivity growth ≈ -4.6 percentage points. Re-estimating baseline for subsample yields TFP differential ≈ -2.6 percentage points.
  - Nonlinearities (Tables 7 and 8): Quadratic Uncertainty^2 * fd not statistically significant. Splitting by low vs high uncertainty (below/above historical median) shows negative and significant interactions in both regimes. Uncertainty * high fd (above 75th percentile) = -2.441* for TFP (t = -1.80) and -3.785*** for productivity (t = -2.62).
- Recessions vs expansions (Table 9):
  - Smooth transition specification (equation (9)) using contemporaneous GDP growth:
    - Uncertainty*fd*recessions = -8.002** for TFP (t = -2.24); productivity recessions = -8.960*** (t = -3.41).
    - Uncertainty*fd*expansions = -0.483 for TFP (not statistically different from zero).
  - Discrete output-gap split (median output gap = -0.3):
    - Negative output gaps: Uncertainty*fd = -4.859*** for TFP (t = -2.51), -5.636*** for productivity (t = -3.91).
    - Positive output gaps: Uncertainty*fd = -3.829** for TFP (t = -2.51), -3.324** for productivity (t = -2.55).
  - Conclusion: negative effects of uncertainty via financial constraints are stronger in downturns.

### VI. Interpretation and mechanisms
- Main mechanism: aggregate uncertainty reduces industry-level productivity growth more in industries that heavily depend on external finance; effect operates mainly through reduced technology adoption / TFP rather than through capital deepening.
- Consistency with literature: aligns with findings on financial development and industry growth (Rajan and Zingales, 1998; Aghion et al., 2014) and with uncertainty generating asymmetric industry and business-cycle effects.

### VII. Policy implications and recommendations
- Counter-cyclical policies may ameliorate the negative productivity effects of heightened uncertainty, especially for financially dependent sectors.
- Policies aimed at addressing weak corporate balance sheets and strengthening fiscal and monetary counter-cyclicality can support productivity growth in financially constrained industries during uncertain times.
- Targeted support for manufacturing industries and innovation-oriented sectors may be warranted given larger sensitivity of these sectors to uncertainty through external finance channels.

### VIII. Key quantitative magnitudes (selected exact figures preserved)
- Differential TFP growth loss from 25th to 75th percentile of uncertainty comparing low- vs high-fd industries:
  - Baseline: about 2.1 percentage points (Table 2 column I).
  - Robustness with credit-to-GDP control: about 2 percentage points (Table 4).
  - Labor productivity baseline: about 2.2 percentage points (Table 5).
  - Economic policy uncertainty measure: about 4 (TFP) and 4.6 (productivity) percentage points (Table 6).
- Selected regression coefficients:
  - Table 2 column I: -4.102*** (t-statistic -3.24).
  - Table 3 Manufacturing: -4.423*** (t-statistic -3.23).
  - Table 6 TFP (economic policy uncertainty): -0.081* (t-statistic -1.62).
  - Table 9 recessions interaction: -8.002** (Tfp, t-statistic -2.24); productivity recessions = -8.960*** (t-statistic -3.41).
- Sample dimensions and observation counts:
  - Unbalanced panel of 18 advanced economies and 25 industries over 1985-2010.
  - Observations in key regressions often reported as 10,654 (Tables 2–4, 7–9) and up to 11,083 for productivity regressions (Table 5).

*IMF working paper content unit _wp16174 (extracted text provided).*

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

### _wp16174 - References .............................................................................................................

### I. Introduction — purpose and headline findings
- Purpose: study the impact of uncertainty on productivity (TFP).
- Key empirical finding: the adverse impact of uncertainty on industry-level productivity growth is greater in industries that rely more heavily on external finance.
- Data scope for main empirical result: 25 industries for 18 advanced economies (Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Hungary, Korea, Ireland, Italy, Japan, the Netherlands, Spain, Sweden, the United Kingdom and the United States) over the period 1985-2010.
- Supporting evidence from related literature:
  - Bloom et al. (2014): a temporary increase in uncertainty lowers aggregate output, investment and productivity; impact on output is a drop of 3 percent within one quarter; TFP drops by about 0.5 percent within a year.
  - Mechanisms highlighted: firms adopt “wait and see” attitudes; increased use of temporary/short-term inputs and investments reduces aggregate productivity and growth.

### II. Literature review — strands and contributions
- Three strands of literature summarized:
  - Role of financial frictions in amplifying uncertainty shocks (Arellano et al., 2010; Bianchi et al., 2014; Christiano et al., 2014; Gilchrist et al., 2014; Cesa-Bianchi and Corugedo, 2014; Caldara et al., 2016; Popp and Zhang, 2016).
  - Use of disaggregated data to address endogeneity/reverse causality between uncertainty and aggregate variables (firm- and industry-level approaches).
  - Relationship between volatility and long-run growth (King and Levine, 1993; Obstfeld, 1994; Ramey and Ramey, 1995; Martin and Rogers, 2000; Acemoglu et al., 2003; Imbs, 2007).
- Contribution claimed: first attempt to use both cross country- and sector-level data to study macroeconomic effect of uncertainty shocks and the interaction with financial frictions using a three-dimensional data set (j industries, i countries, t time periods).

### III. Theoretical framework — setup and formal predictions
- Model features and assumptions:
  - Two-period model with two types of investment: short-term (immediate output, liquid) and long-term (more productive, delayed).
  - Aggregate productivity evolves as augmented geometric random walk: 
    - a_{t+1} = a_{t}(1+σ_{t} ε_{t+1}), where σ_{t} ∈ {σ_{L}, σ_{H}} and ε_{t} ~ N(0,1).
  - Entrepreneurs allocate human capital between short-term investment K_{t}=kH_{t} and long-term investment Z_{t}=zH_{t} with w=k+z.
  - Two types of firms: fraction 1−μ unconstrained (can borrow), fraction μ constrained (must rely on own cash flows), with 0<μ<1.
  - Constrained firms survive idiosyncratic liquidity shock C_{i,t+1}=c_{i,t+1}H_{t} where c_{i,t} ~ i.i.d. uniform(0,1); survival probability f_{t+1}=Pr(a_{t+1} k^{α} ≥ c) = min{a_{t+1} k^{α},1}.
  - Short-term profits π_{t+1}=a_{t+1} k^{α} H_{t}, long-term yields v_{t+1} H_{t} with probability λz.
- Formal results (Propositions / Lemmas):
  - Proposition 1: Unconstrained firms invest a positive amount in long-term investment, z_{nc}>0, and z_{nc} > z_{c}.
  - Lemma 1: A mean preserving spread of aggregate productivity distribution decreases z_{c}, but does not affect z_{nc}. (∂ z_{c} / ∂ σ_{t} < 0)
  - Growth decomposition: ln A_{i,t+1} − ln A_{i,t} = ln a_{t+1} − ln a_{t} + ln H_{i,t+1} − ln H_{i,t}.
  - Expected TFP growth: g_{t+1} = E[ln A_{i,t+1} − ln A_{i,t}] = λ((1−μ) z_{nc} + μ z_{c} min( a_{t+1} (w−z_{c})^{α}, 1 )).
  - Lemma 2: An increase in aggregate uncertainty reduces expected TFP growth: ∂ g_{t+1} / ∂ σ_{t} < 0.
  - Lemma 3: The reduction in expected TFP growth from increased uncertainty is larger the higher the fraction of constrained firms: ∂/∂μ (∂ g_{t+1} / ∂ σ_{t}) < 0.
  - Lemma 4: The differential effect of uncertainty is larger when realized productivity is lower: ∂/∂ a_{t+1} ( ∂/∂μ (∂ g_{t+1} / ∂ σ_{t} )) < 0.
- Theoretical implications tested empirically:
  - Main hypothesis: increase in aggregate uncertainty reduces productivity growth more in industries that are financially constrained.
  - Additional implication: interaction between uncertainty and financial constraints on TFP growth is larger in recessions than expansions.

### IV. Data and methodology — construction and empirical design
- Sample: unbalanced panel of 25 industries for 18 advanced economies over 1985-2010.
- Country-specific uncertainty proxy:
  - Baseline measure: realized volatility of aggregate stock market returns for each country.
  - Annualized realized volatility calculated using daily returns:
    - RV_{i,t} = 100 √( (1/T_{i}) ∑_{s=1}^{T_{i}} r_{i,s}^{2} )
    - where r_{i,s} are daily returns and T_{i} is the stock market i's number of trading days in a year.
  - Data source for daily closing prices: Global Financial Data (longest international time-series on stock prices).
  - Robustness check: results also tested using economic policy uncertainty indices constructed by Baker et al. (2015).
- Empirical strategy advantages of three-dimensional data (j,i,t):
  - Controls for aggregate and country-sector shocks via country-time (i,t) and industry-country (j,i) fixed effects.
  - Mitigates reverse causality concerns; main independent variable is interaction between uncertainty and industry dependence on external finance.
  - Identifies non-linearities in volatility–TFP relationship across industries, reconciling mixed evidence in the literature.

### V. Empirical results and robustness (summary of primary claims)
- Primary empirical claim: the adverse impact of uncertainty on industry-level productivity growth is greater in industries that rely more heavily on external finance.
- Approach: cross-country/cross-industry difference-in-difference analysis with controls and robustness checks (details of tables and figures listed in document: Figures 1–4; Tables 1–9; Appendix Table A.1).
- Non-linearities documented: relationship between volatility and TFP growth can be both positive and negative at the industry level depending on degree of credit constraints; larger negative interaction effects in recessions.

*Source: _wp16174 - References.*

### 2010. Although our measure of uncertainty shows some degree of co-movements across

### _wp16174 - 2010. Although our measure of uncertainty shows some degree of co-movements across

### Data and measurement
- Country- and time-varying measure of uncertainty constructed for an unbalanced panel of 18 advanced economies over 1985-2010; descriptive statistics (Table 1) report country Means, SDs, and Obs (e.g., Hungary Mean 26.37 SD 9.39 Obs 19; Australia Mean 14.66 SD 6.74 Obs 26; United States Mean 17.08 SD 7.99 Obs 26).
- Sectoral TFP (labor productivity) growth from EU KLEMS and World KLEMS databases.
- Industry dependence on external finance measured using Compustat US firm-level data following Rajan and Zingales (1998): median across US firms in an industry of (total capital expenditures − current cash flow) / total capital expenditures.
- Alternative uncertainty measure: economic policy uncertainty index of Baker et al. (2015), available for 10 countries in sample (Canada, France, Germany, Italy, Japan, Korea, the Netherlands, Spain, the UK, the US), mainly since the mid-90s.

### Empirical methodology
- Baseline specification (equation (5)):
  - TFP_{i,j,t} = α_{i,t} + γ_{i,j} + β fd_j U_{i,t} + ε_{i,j,t}, where i=countries, j=industries, t=years.
  - fd_j = industry external finance dependence (US-based).
  - α_{i,t} = country-time fixed effects; γ_{i,j} = country-industry fixed effects.
- Estimation: OLS with standard errors clustered at the country-industry level.
- Controls and checks: inclusion of country-time and country-industry fixed effects to address common shocks and industry-specific factors; additional robustness checks interact fd with credit-to-GDP, inflation, government size (government consumption/GDP), and budget balance/GDP.

### Baseline empirical findings
- Main baseline result (Table 2): interaction Uncertainty * financial dependence is negatively correlated with industry TFP growth.
  - Estimated coefficients (selected): -4.102*** (column I), -2.889*** (column II), -2.513*** (column III), -2.945*** (lag, column IV), -0.214*** (ordinal fd, column V).
  - Reported differential TFP growth loss when uncertainty increases from the 25th to the 75th percentile (≈ one standard deviation) comparing an industry at the 25th percentile of fd to one at the 75th percentile: about 2.1 percentage points (Table 2 column I). Other columns: -1.5, -1.3, -1.5, -3.3 percentage points respectively.
- Aggregate-level correlations:
  - US quarterly sample 1970-2013: Coef = -2.16; t-statistics = -3.65 (Figure 3 Panel A).
  - 18 advanced economies annual sample 1985-2013: Coef = -0.76; t-statistics = -3.51 (Figure 3 Panel B).
- Industry-level raw-data pattern (Figure 4):
  - Industries below median financial dependence: Coef = 3.4; t-statistics = 3.6 (positive relationship between uncertainty and sector-level TFP growth).
  - Industries above median financial dependence: Coef = -2.8; t-statistics = -2.9 (negative relationship).

### Robustness checks and extensions
- Specification robustness:
  - Results robust to (i) only country-time fixed effects + industry dummies (Table 2 column II), (ii) country, time, and industry dummies without interactions (Table 2 column III), (iii) lagged interaction (Table 2 column IV), (iv) categorical ordinal fd 1–25 (Table 2 column V).
- Sample of industries:
  - Manufacturing vs non-manufacturing (Table 3): Manufacturing Uncertainty*financial dependence = -4.423*** (t = -3.23), differential TFP growth = -2.3 percentage points; Non-manufacturing coefficient = -1.107 (t = -0.43), differential = -0.6 percentage points.
- Controls for omitted variable bias (Table 4):
  - Including Credit-to-GDP * financial dependence: Uncertainty*financial dependence remains negative and significant; differential TFP loss ≈ -2.0 percentage points (column I).
  - Inflation*financial dependence: Uncertainty effect unchanged; inflation interaction not statistically significant (Table 4 column II).
  - Government size * financial dependence and Budget balance * financial dependence included separately: Uncertainty effect remains statistically significant; Government size interaction positive, Budget balance interaction negative (Table 4 columns III and IV). All four controls jointly included: Uncertainty effect remains (Table 4 column V).
- Alternative dependent variable — labor productivity growth (Table 5):
  - Uncertainty*financial dependence negative and significant across specifications (e.g., -4.251*** column I; differential productivity loss from 25th to 75th uncertainty percentile ≈ -2.2 percentage points in baseline).
  - Magnitude slightly larger than TFP effect, suggesting limited impact on capital deepening and highlighting technology adoption channel.
- Alternative uncertainty measure — economic policy uncertainty (Table 6):
  - Economic policy uncertainty * financial dependence coefficient: -0.081* for TFP growth (t = -1.62) and -0.093** for productivity growth (t = -2.03).
  - Differential effect from 25th to 75th percentiles: TFP growth ≈ -4.0 percentage points; productivity growth ≈ -4.6 percentage points.
  - Re-estimating baseline for the subsample with this measure yields a TFP differential ≈ -2.6 percentage points (note: larger effects only partly driven by sample composition).
- Nonlinearities (Tables 7 and 8):
  - Quadratic term Uncertainty^2 * fd not statistically significant (Table 7 columns I-II).
  - Split by low vs high uncertainty (below/above historical median) shows negative and significant interactions in both regimes, but overall results suggest effect does not significantly depend on uncertainty level itself (Table 7 columns III-IV).
  - Testing degree of external finance constraint (equation (8), Table 8): Uncertainty * high financial dependence (above 75th percentile) = -2.441* for TFP (t = -1.80) and -3.785*** for productivity (t = -2.62); effect larger and statistically significant in more financially constrained industries.
- Recessions vs expansions (Table 9):
  - Smooth transition specification (equation (9)) using contemporaneous GDP growth as state indicator:
    - Uncertainty*financial dependence * recessions = -8.002** for TFP (t = -2.24); for productivity = -8.960*** (t = -3.41).
    - Uncertainty*financial dependence * expansions not statistically different from zero (Uncertainty*financial dependence*expansions = -0.483 for TFP).
  - Discrete output-gap split (negative vs positive output gaps; median output gap = -0.3):
    - Uncertainty*financial dependence * negative output gaps = -4.859*** for TFP (t = -2.51), -5.636*** for productivity (t = -3.91).
    - Uncertainty*financial dependence * positive output gaps = -3.829** for TFP (t = -2.51), -3.324** for productivity (t = -2.55).
  - Conclusion: negative effects of uncertainty via financial constraints are stronger in downturns.

### Interpretation and mechanisms
- Main mechanism: aggregate uncertainty reduces industry-level productivity growth more in industries that heavily depend on external finance; effect operates mainly through reduced technology adoption / TFP rather than through capital deepening.
- Consistent with literature: financial development interaction raises value-added growth more for financially dependent industries (Rajan and Zingales, 1998; Aghion et al., 2014), while uncertainty generates asymmetric effects across industries and across business cycle states.

### Policy implications and recommendations
- Counter-cyclical policies may ameliorate the negative productivity effects of heightened uncertainty, especially for financially dependent sectors.
- Policies aimed at addressing weak corporate balance sheets and strengthening fiscal and monetary counter-cyclicality can support productivity growth in financially constrained industries during uncertain times.
- Targeted support for manufacturing industries and innovation-oriented sectors may be warranted given larger sensitivity of these sectors to uncertainty through external finance channels.

### Key quantitative magnitudes (selected exact figures preserved)
- Differential TFP growth loss from 25th to 75th percentile of uncertainty comparing low- vs high-fd industries:
  - Baseline: about 2.1 percentage points (Table 2 column I).
  - Robustness with credit-to-GDP control: about 2 percentage points (Table 4).
  - Labor productivity baseline: about 2.2 percentage points (Table 5).
  - Economic policy uncertainty measure: about 4 (TFP) and 4.6 (productivity) percentage points (Table 6).
- Selected regression coefficients (Uncertainty * financial dependence):
  - Table 2 column I: -4.102*** (t-statistic -3.24).
  - Table 3 Manufacturing: -4.423*** (t-statistic -3.23).
  - Table 6 TFP (economic policy uncertainty): -0.081* (t-statistic -1.62).
  - Table 9 recessions interaction: -8.002** (Tfp, t-statistic -2.24); productivity recessions = -8.960*** (t-statistic -3.41).
- Sample dimensions: unbalanced panel of 18 advanced economies and 25 industries over 1985-2010; observations in key regressions often reported as 10,654 (Tables 2–4, 7–9) and up to 11,083 for productivity regressions (Table 5).

*Italicized source attribution: IMF working paper content unit _wp16174 (extracted text provided).*

### References

### _wp16174 - References

### Major themes covered by the cited literature
- Uncertainty and investment dynamics
  - Abel, Andrew B. “Optimal investment under uncertainty.” American Economic Review 73.1 (1983): 228-233.
  - Bernanke, Ben. “Irreversibility, Uncertainty, and Cyclical Investment.” Quarterly Journal of Economics, Vol. 97, No. 1, (1983), pp. 85-106.
  - Bloom, Nicholas. “The impact of uncertainty shocks.” Econometrica 77.3 (2009): 623-685.
  - Bloom, Nicholas, Stephen Bond, and John Van Reenen. “Uncertainty and investment dynamics.” The Review of Economic Studies 74.2 (2007): 391-415.
  - Pindyck, Robert S. “Irreversible Investment, Capacity Choice, and the Value of the Firm,” American Economic Review, vol. 78(5), (1988), 969-85, December.
  - Leahy, John, and Toni M. Whited. “The Effects of Uncertainty on Investment: Some Stylized Facts.” Journal of Money, Credit, and Banking 28.1 (1996): 64-83.
  - Stein, Luke CD, and Elizabeth Stone. “The effect of uncertainty on investment, hiring, and R&D: Causal evidence from equity options.” Working paper. (2013).

- Uncertainty, macroeconomic fluctuations, and growth
  - Acemoglu, Daron, Simon Johnson, James Robinson, and Yunyong Thaicharoen. “Institutional causes, macroeconomic symptoms: volatility, crises and growth.” Journal of Monetary Economics 50 (2003): 49-123.
  - Aghion, Philippe, George-Marios Angeletos, Abhijit Banerjee, and Kalina Manova. “Volatility and growth: Credit constraints and the composition of investment.” Journal of Monetary Economics 57, no. 3 (2010): 246-265.
  - Martin, Philippe, and Carol Ann Rogers. “Long-term growth and short-term economic instability.” European Economic Review 44.2 (2000): 359-381.
  - Ramey, Garey, and Valerie A. Ramey. “Cross-Country Evidence on the Link Between Volatility and Growth.” The American Economic Review 85.5 (1995): 1138-1151.
  - IMF World Economic Outlook, April 2015, Chapter 3. “Where Are We Headed? Perspectives on Potential Output,” International Monetary Fund, (2015)

- Financial frictions, asset markets, and risk
  - Christiano, Lawrence J., Roberto Motto, and Massimo Rostagno. “Risk Shocks.” American Economic Review 104.1 (2014): 27-65.
  - Gilchrist, Simon, Jae W. Sim, and Egon Zakrajšek. “Uncertainty, financial frictions, and investment dynamics.” No. w20038. National Bureau of Economic Research, (2014).
  - Bianchi, Francesco, Cosmin L. Ilut, and Martin Schneider. “Uncertainty shocks, asset supply and pricing over the business cycle.” No. w20081. National Bureau of Economic Research, (2014).
  - Gourio, Francois, Michael Siemer, and Adrien Verdelhan. “International risk cycles.” Journal of International Economics 89.2 (2013): 471-484.
  - Tong, Hui, and Shang-Jin Wei. “The composition matters: capital inflows and liquidity crunch during a global economic crisis.” Review of Financial Studies 24.6 (2011): 2023-2052.

- Policy uncertainty, political risk, and firm behavior
  - Baker, Scott R., and Nicholas Bloom. “Does uncertainty reduce growth? Using disasters as natural experiments.” No. w19475. National Bureau of Economic Research (2013).
  - Baker, Scott, Nicholas Bloom and Steven Davis, “Measuring Economic Policy Uncertainty.” Mimeo (2015).
  - Julio, Brandon, and Youngsuk Yook. “Political uncertainty and corporate investment cycles.” The Journal of Finance 67.1 (2012): 45-83.
  - Pástor, Ľuboš, and Pietro Veronesi. “Political uncertainty and risk premia.” Journal of Financial Economics 110.3 (2013): 520-545.
  - Handley, Kyle, and Nuno Limao. “Trade and investment under policy uncertainty: theory and firm evidence.” American Economic Journal: Economic Policy 7.4 (2015): 189-222.
  - Gulen, Huseyin, and Mihai Ion. "Policy uncertainty and corporate investment." Review of Financial Studies 29.3 (2016): 523-564.

- Labour market, temporary employment, and productivity
  - Cappellari, Lorenzo, Carlo Dell’Aringa, and Marco Leonardi. “Temporary Employment, Job Flows and Productivity: A Tale of Two Reforms.” The Economic Journal 122.562 (2012): F188-F215.
  - Foote, David A., and Timothy Folta. “Temporary Workers as Real Options.” Human Resource Management Review 12(4), 579-97 (2002).
  - Lotti, Francesca, and Eliana Viviano. “Temporary workers, uncertainty and productivity.” The Society of Labor Economists, Mimeo (2012).
  - Dolado, Juan J. Salvador Ortiguira, and Rodolfo Stucchi. “Does Dual Employment Protection Affect TFP? Evidence from Spanish Manufacturing Firms.” CEPR Discussion Papers no. 8763. (2012)
  - Kleinknecht, Alfred, Remco M. Oostendorp, Menno P. Pradhan, and C. W. M. Naastepad. “Flexible labour, firm performance and the Dutch job creation miracle.” International Review of Applied Economics 20, no. 02 (2006): 171-187.

- Methods, measurement, and data-based studies
  - Bloom, Nicholas. “Fluctuations in Uncertainty.” The Journal of Economic Perspectives (2014): 153-175.
  - Bachmann, Rüdiger, Steffen Elstner, and Eric R. Sims. “Uncertainty and Economic Activity: Evidence from Business Survey Data.” American Economic Journal: Macroeconomics 5.2 (2013): 217-49.
  - Baker, Scott, Nicholas Bloom and Steven Davis, “Measuring Economic Policy Uncertainty.” Mimeo (2015).
  - O'Mahony, Mary, and Marcel P. Timmer. “Output, input and productivity measures at the industry level: The EU KLEMS database.” The Economic Journal 119.538 (2009): F374-F403.
  - Bernal, Oscar, Jean-Yves Gnabo, and Grégory Guilmin. “Economic policy uncertainty and risk spillovers in the Eurozone.” Journal of International Money and Finance 65 (2016): 24-45.

### Appendix content and analytical results
- Proof 1
  - The appendix presents Proof 1 with objective functions and first-order conditions labeled (A.1) through (A.4).
  - Key elements reproduced from the appendix:
    - Firms maximize end of t+1 consumption by choosing k and z after observing 푎푡 = 푎, subject to 푘+푧=푤. (Equations referenced as A.1 and A.2.)
    - For constrained firms, an expression (A.4) is written where 푓푡+1 < 1 and  푓푡+1 /  푡 < 0 under reasonable assumptions on the support of a, k, and c.
    - Maximization of A.4 leads to the conclusion that 푧푛푐 > 푧푐.

- Table A.1 — Description of stock market data used to construct uncertainty indices
  - Country / Stock exchanges / Coverage
    - Australia / Australia ASX All-Ordinaries (w/GFD extension) / 1985-2010
    - Austria / Austria Trading Index (ATX) / 1987-2010
    - Belgium / Brussels All-Share Price Index (w/GFD extension) / 1985-2010
    - Canada / Canada S&P/TSX 300 Composite (w/GFD extension) / 1985-2010
    - Denmark / OMX Copenhagen All-Share Price Index / 1985-2010
    - Finland / OMX Helsinki All-Share Price Index / 1987-2010
    - France / Paris CAC-40 Index / 1988-2010
    - Germany / Germany DAX Price Index / 1985-2010
    - Hungary / Budapest Stock Exchange Index (BUX) / 1992-2010
    - Ireland / Ireland ISEQ Overall Price Index (w/GFD extension) / 1985-2010
    - Italy / Banca Commerciale Italiana Index (w/GFD extension) / 1985-2010
    - Japan / Tokyo SE Price Index (TOPIX) (w/GFD extension) / 1985-2010
    - Korea / Korea SE Stock Price Index (KOSPI) / 1985-2010
    - Netherland / Amsterdam AEX Stock Index / 1985-2010
    - Spain / Madrid SE General Index (w/GFD extension) / 1985-2010
    - Sweden / OMX Stockholm All-Share Price Index / 1987-2010
    - United Kingdom / UK FTSE All-Share Index (w/GFD extension) / 1985-2010
    - United States / S&P 500 Composite Price Index (w/GFD extension) / 1985-2010

*Content derived from the "_wp16174 - References" PDF chapter/section.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16174.pdf_