DETERMINANTS OF FIRMS INVESTMENT IN SPAIN: THE ROLE OF POLICY UNCERTAINTY

DETERMINANTS OF FIRMS INVESTMENT IN SPAIN: THE ROLE OF POLICY UNCERTAINTY Daniel Dejuan and Corinna Ghirelli Bank of Spain European Network for Research on Investment EIB - Luxemburg 9 April 018 DG ECONOMICS, STATISTICS AND RESEARCH

Motivation Corporate investment: key for long-run growth and main driver of business cycle Renewed interest on investment s determinants due to low recovery after Great Recession Role of uncertainty on corporate investment debate Wait and see: postpone investment until additional information comes Financial frictions: for financially constrained firms, probability of default, cost of investment

Long run obstacles to investment (proportion of firms %)

This paper Study the effect of policy uncertainty on firm investment Firm panel 1998-014 Static investment model with firm FE Aggregated time-varying factors explicitly accounted for Contribution: e.g. Julio and Yook (01), Gulen and Ion (016), Baker et al. (016) SME coverage, as opposed to US literature Heterogeneous effects by financial position financial frictions story

The uncertainty measure 1. Current political situation (CIS survey). Expectations about future political situation (CIS survey) 3. Political risk, e.g. government stability, quality of institution (PRS group) 4. EPU index (Baker et al., 015) 5. Disagreement about public deficit among experts (FUNCAS) Correlation matrix policy U pol.sit. exp.pol.sit. pol.risk EPU dis.pub.def. policy U 1.00 pol.sit. -0.94 1.00 exp.pol.sit. -0.68 0.50 1.00 pol.risk -0.90 0.91 0.40 1.00 EPU 0.8-0.1-0.06-0.15 1.00 dis.pub.def. 0.8-0.13-0.1-0.18-0.1 1.00

The uncertainty measure (monthly variation) -3 - -1 0 1 3 4 EMU constitution 09/11 terrorist attack Spanish gen. elections 1997m1 1998m1 1999m1 000m1 001m1 Irak Invasion 11/03 terrorist attack & Spanish gen. elections Prestige sunk 00m1 003m1 004m1 Bear Sterns bailout & Spanish gen. elections Lehman Brothers bailout 005m1 006m1 007m1 008m1 009m1 Greece bailout Spain bailout Spanish general elections 010m1 011m1 01m1 013m1 014m1 015m1 016m1 017m1 Brexit Spanish gen. elections more

The empirical model (I/K ) it = α i + β 1 U t 1 + β X it 1 + γ 3 GVA it 1 + ɛ it 1998-014, 600, 000 firms more (Central Balance Sheet Data) I/K : gross fixed capital formation over total capital stock X: financial position, ROA, sales growth, size, export, quoted, corporate group Firm FE Business cycle: sector-specific GVA growth rates Tax changes indicator (Gil et al., 017) Narrative Index Standard errors clustered at year-sector level

Expected effects of uncertainty Leverage: if financially constrained, U, access to credit probability to default, cost of credit Profitability: No role for financial friction channel Optimal to delay when U and invest when U (?) SME: affected due to information & financial frictions Exporters: less sensitive to domestic uncertainty Quoted: less affected by credit supply shocks Corporate groups: More More affected if contagion effects Less affected if better access to credit or lower cost of debt

Heterogeneous effects (1) () (3) (4) (5) (6) (7) policy U -0.043-0.06-0.040-0.04-0.044-0.044-0.045 (0.005) (0.006) (0.006) (0.005) (0.005) (0.005) (0.006) U x SMEs -0.018 (0.003) U x debt rate p[50-100] -0.009 (0.001) U x ROA p[50-100] -0.00 (0.001) U x export U x quoted U x corp group tax changes post 008 0.010 (0.00) 0.03 (0.008) 0.06 (0.004) -0.030-0.09-0.09-0.09-0.030-0.030-0.030 (0.00) (0.00) (0.003) (0.00) (0.00) (0.00) (0.00) -0.046-0.045-0.046-0.046-0.045-0.046-0.045 (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) Observations 341365 341365 341365 341365 341365 341365 341365 R 0.050 0.050 0.051 0.050 0.050 0.050 0.050 X, sector-specific gross value added (GVA) growth rates included. Firm FE model. Standard errors clustered at the sector year level.

Conclusions and policy implications Effect of policy uncertainty on firm investment, 1998-014 On average: policy U by standard deviation, I/K by 4 pp More negative effect for financially constrained firms, small firms (access to credit) Belonging to corporate group seems to improve access to credit for SMEs Financial friction story Policy implications: decision, coherence, clarity, transparency in public policy

THANK YOU FOR YOUR ATTENTION! DG ECONOMICS, STATISTICS AND RESEARCH

Short run obstacles to investment (net effect %)

Counterfactual exercise Counterfactual - policy uncertainty in 006.05.1.15..5.3 avg. observed investment rate avg. predicted investment rate with policy U fixed in 006 as from 006 1997 1998 1999 000 001 00 003 004 005 006 007 008 009 010 011 01 013 014 015 016

Firm FE model. Standard errors clustered at the sector year level. (1) () (3) (4) (5) (6) (7) ROA 0.037 0.040 0.039 0.041 0.040 0.04 0.041 (0.003) (0.003) (0.003) (0.004) (0.004) (0.004) (0.004) debt burden -0.009-0.009-0.009-0.009-0.009-0.009-0.009 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) debt rate -0.033-0.03-0.03-0.031-0.031-0.030-0.030 (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) sales growth rate 0.011 0.01 0.011 0.011 0.011 0.01 0.01 (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) empl. growth rate 0.010 0.010 0.010 0.010 0.010 0.010 0.011 (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) policy U -0.041-0.043-0.03-0.01-0.05-0.06 (0.005) (0.005) (0.00) (0.00) (0.006) (0.006) GVAsec1-0.000-0.000 (0.000) (0.000) GVAsec 0.003 0.004 (0.000) (0.000) GVAsec3 0.001-0.001 (0.001) (0.001) GVAsec4-0.007-0.006 (0.001) (0.001) GVAsec5 0.006 0.006 (0.001) (0.001) GVAsec6-0.004-0.004 (0.001) (0.001) ESI 0.016 0.01 (0.00) (0.00) GDPgrowth 0.00 0.001 (0.001) (0.000) tax changes -0.030-0.019-0.09 (0.00) (0.004) (0.00) Time FE yes no no no no no no Observations 341365 341365 341365 341365 341365 341365 341365 R 0.050 0.049 0.050 0.048 0.049 0.047 0.049

Robustness exercise Add time FE with U X (as is Baker et al., 016) β 1 absorbed by time FE Omitted variable bias ruled out Can still identify the interaction Shift interest to channels through which effect of U materializes

Heterogeneous effect including time FE (1) () (3) ROA 0.037 0.039 0.041 (0.003) (0.003) (0.004) debt rate -0.033-0.030-0.035 (0.00) (0.003) (0.00) corp group 0.030 0.030 0.031 (0.003) (0.003) (0.003) U x ROA U x debt rate -0.005 (0.00) -0.01 (0.003) Observations 341365 341365 341365 R 0.050 0.050 0.050 X included. Firm FE model. Time FE included. Standard errors clustered at the sector year level.

Balanced panel: Heterogeneous effects (1) () (3) (4) (5) (6) (7) policy U -0.01-0.015-0.00-0.05-0.04-0.01-0.01 (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) (0.007) U x SMEs -0.009 (0.003) U x debt rate p[50-100] -0.006 (0.00) U x ROA p[50-100] 0.007 (0.004) U x export 0.011 (0.003) U x quoted -0.01 (0.01) U x corp group -0.001 (0.005) tax changes -0.03-0.033-0.03-0.033-0.033-0.03-0.03 (0.006) (0.006) (0.006) (0.006) (0.006) (0.006) (0.006) Observations 30991 30991 30991 30991 30991 30991 30991 R 0.078 0.078 0.078 0.078 0.078 0.078 0.078 Based on sub-sample of firms observed throughout the entire period 1998-014 (balanced panel). X included, sector-specific gross value added (GVA) growth rates included. Firm FE model. Standard errors clustered at the sector year level. more

Balanced panel: Heterogeneous effects including time FE (1) () (3) ROA 0.175 0.181 0.175 (0.0) (0.03) (0.0) debt rate -0.033-0.08-0.034 (0.013) (0.014) (0.013) corp group 0.034 0.033 0.035 (0.006) (0.006) (0.006) U x ROA -0.00 (0.014) U x debt rate -0.014 (0.005) Observations 30991 30991 30991 R 0.079 0.079 0.079 Based on sub-sample of firms observed throughout the entire period 1998-014 (balanced panel). X included. Firm FE model. Time FE included. Standard errors clustered at the sector year level.

Robustness: placebos (1) () (3) policy U (t-1) -0.0435*** -0.0435*** -0.0454*** (0.0055) (0.005) (0.0053) policy U (t+1) -0.0001 (0.0036) policy U (t+) 0.0047 (0.009) ROA 0.0386*** 0.0386*** 0.0386*** (0.0035) (0.0035) (0.0035) debt rate -0.030*** -0.030*** -0.030*** (0.00) (0.00) (0.00) tax changes -0.096*** -0.095*** -0.087*** (0.005) (0.008) (0.005) debt burden -0.0088*** -0.0088*** -0.0088*** (0.0004) (0.0004) (0.0004) sales growth rate 0.0113*** 0.0113*** 0.0113*** (0.0008) (0.0008) (0.0008) empl. growth rate 0.0100*** 0.0100*** 0.0100*** (0.0008) (0.0008) (0.0008) Observations 3,41,365 3,41,365 3,41,365 R-squared 0.0496 0.0496 0.0497 Number of id 618,811 618,811 618,811 Based on entire sample. X included, sector-specific gross value added (GVA) growth rates included. Firm FE model. Standard errors clustered at the sector year level.

Descriptive Statistics Des. stat. firm data: full sample mean sd min max N ROA 0.051 0.171-1.675 0.76 3,41,365 debt burden 0.575 0.950 0.000.774 3,41,365 debt rate 0.690 0.449 0.000 4.783 3,41,365 sales growth rate 0.053 0.537-1.000 10.03 3,41,365 empl. growth rate 0.09 0.380-1.000 4.500 3,41,365 SMEs 0.984 0.15 0.000 1.000 3,41,365 export 0.034 0.180 0.000 1.000 3,41,365 corp group 0.015 0.13 0.000 1.000 3,41,365 quoted 0.001 0.04 0.000 1.000 3,41,365 1(Gross Inv.> 0) 0.71 0.448 0.000 1.000 3,41,365 I/K 0.139 0.63-1.05 1.506 3,41,365 back

Average investment rate 0.04.08.1.16..4.8 1998 1999 000 001 00 003 004 005 006 007 008 009 010 011 01 013 014

The uncertainty measure (yearly) back more -.00-1.00 0.00 1.00.00 1997m6 1998m6 1999m6 weighted avg Pol.Eco.U Pol.Eco.U 000m6 001m6 00m6 003m6 004m6 005m6 006m6 007m6 008m6 009m6 010m6 011m6 01m6 013m6 014m6 015m6 1997 1998 1999 000 001 00 003 004 005 006 007 008 009 010 011 01 013 014 015 016 016m6

Heterogeneous effects of group by size All No SME SME (1) () (3) policy U -0.095-0.035-0.099 (0.015) (0.05) (0.015) U x corp group 0.068 0.006 0.088 (0.013) (0.01) (0.03) corp group -0.049 0.010-0.064 (0.017) (0.016) (0.05) debt rate -0.03-0.030-0.03 (0.00) (0.016) (0.00) ROA 0.040 0.010 0.040 (0.004) (0.05) (0.004) debt burden -0.009-0.008-0.009 (0.000) (0.00) (0.000) sales growth rate 0.01 0.013 0.01 (0.001) (0.004) (0.001) empl. growth rate 0.010 0.03 0.010 Time FE no no no Observations 341365 5816 318339 R 0.049 0.0 0.049 Firm FE model. Standard errors clustered at the sector year level.

The uncertainty measure (monthly) Correlation matrix policy U ESI financial U Vol.index 10 years gov.bond Vol.index IBEX35 policy U 1.00 ESI -0.61 1.00 financial U 0.03-0.35 1.00 Vol.index 10 years gov.bond 0.14-0.57 0.54 1.00 Vol.index IBEX35 0.11-0.8 0.90 0.39 1.00 Vol.index 10 years gov.bond: calculated by IESI, based on daily growth rate 10 years gov.bond returns Vol.index IBEX35: monthly average of daily volatility of IBEX35 index reported by Bloomberg back

Index of permanent and exogenous tax changes Constructed by Gil et al. (017) for Spain Narrative approach (Romer and Romer, 010) Based on record of legislated tax changes Each change is attributed to year of adoption t Quantified as change in tax revenues collected in t wrt counterfactual Classified as permanent/temporary, exogenous/endogenous All changes in t are expressed as % of GDP of t 1 Back

Indicator of tax changes Tax Revenues[t] as % of GDP[t-1] -1 -.5 0.5 1 1991 199 1993 1994 1995 1996 1997 1998 1999 000 001 00 003 004 005 006 007 008 009 010 011 01 013 014 015 016 Back

Financial vs political uncertainty Correlation matrix (monthly) financial U vol.ibex35 ind.vol.ibex35 ind.vol.exchange ind.vol.brent ind.vol.10y.bond financial U 1.00 vol.ibex35 0.90 1.00 ind.vol.ibex35 0.91 0.89 1.00 ind.vol.exchange 0.7 0.45 0.47 1.00 ind.vol.brent 0.37 0.7 0.1 0.4 1.00 ind.vol.10y.bond 0.54 0.39 0.41 0.30-0.3 1.00 Financial U: High frequency shocks, not persistent

Financial vs political U (yearly) Here: yearly data, RHS lagged yearly avg. U in t 1 explains investment in t Policy uncertainty better suited since more persistent -.00-1.00 0.00 1.00.00 1997 1998 1997m6 1998m6 1999m6 000m6 001m6 00m6 003m6 004m6 005m6 weighted avg Pol.Eco.U Pol.Eco.U 006m6 007m6 1999 000 001 00 003 004 005 006 007 008 008m6 009m6 009 010 010m6 011m6 011 01 01m6 013m6 013 014 014m6 015m6 016m6 015 016 -.00 0.00.00 4.00 1997 1998 1997m6 1998m6 1999m6 000m6 001m6 00m6 003m6 004m6 005m6 weighted avg Fin.U Fin.U 1999 000 001 00 003 004 005 006 006m6 007m6 007 008 008m6 009m6 009 010 010m6 011m6 011 01 01m6 013m6 013 014 014m6 015m6 016m6 015 016 back

Cost of debt and % of firms belonging to corporate group median cost of debt.0.04.06.08.1 000 005 010 015 year Group=1 Proportion Group=1 Group=0 0..4.6.8 1 prortion of firms (%) cost of debt (1) corp-group -0.04 (0.001) Large company -0.03 (0.001) debt burden 0.017 (0.000) sales growth -0.001 (0.000) ROA -0.011 (0.001) Observations 31553 R 0.035 OLS with time FE and sector FE. Standard errors clustered at the sector year level. Back

Debate EC & OECD: high uncertainty is all the more damaging for growth as it magnifies the effect of credit constraints and weak balance sheets, forcing banks to rein in credit further and companies to hold back investment (Buti and Padoan, 013) ILO: indecision of policy makers in several countries has led to uncertainty about future conditions and reinforced corporate tendencies to increase cash holdings or pay dividends rather than expand capacity and hire new workers (ILO, 013) back

Q&A, Malta, October 015 - ECB Question: Could you comment on the possible political uncertainty in Spain and Portugal and if that could derail the process of reform that you ve alluded to so many times today? And is it a worry for the ECB, the fact that some of the parties that may be called to form a government, especially in Portugal, reject the euro? Draghi: Well, the answer is no, I can t comment. But a more gentle way of saying no would be the following: uncertainty, for economics, is bad. It s bad for consumption; it s bad for investment. Political uncertainty, however, is part of democracy. back

Table: Correlation matrix pol U fin U ESI GVAsec1 GVAsec GVAsec3 GVAsec4 GVAsec5 GVAsec6 pol U 1.00 fin U -0.0 1.00 ESI -0.71-0.31 1.00 GVAsec1-0.7-0.01-0.11 1.00 GVAsec -0.65-0.1 0.84 0.14 1.00 GVAsec3-0.87-0.16 0.69 0.4 0.70 1.00 GVAsec4-0.87-0.0 0.7 0.40 0.80 0.9 1.00 GVAsec5-0.9 0.08 0.69 0.18 0.63 0.87 0.8 1.00 GVAsec6-0.79-0.14 0.56 0.4 0.63 0.86 0.86 0.79 1.00 The table reports correlations of the time-varying factors used in the estimations. All variables are defined yearly. pol U fin U are the policy and financial uncertainty measures. They are expressed as the log of the yearly weighted average of monthly original indexes. ESI is the yearly weighted average of monthly ESI, as provided by the European Commission. GVA refers to gross value added growth rate at the sectoral level. Sectors from 1 to 6are as follows: Energy, Industry, Construction and real estate, trade and hotel, informatics and communications, Others. more

Table: Descriptive Statistics: sub-sample of firms observed between 1998-014 mean sd min max N ROA 0.094 0.111-1.460 0.71 30,991.000 debt burden 0.306 0.66 0.000.774 30,991.000 debt rate 0.504 0.65 0.000 4.359 30,991.000 sales growth rate 0.045 0.75-1.000 8.760 30,991.000 empl. growth rate 0.00 0. -1.000 4.490 30,991.000 SMEs 0.874 0.33 0.000 1.000 30,991.000 export 0.19 0.394 0.000 1.000 30,991.000 corp group 0.078 0.68 0.000 1.000 30,991.000 quoted 0.015 0.10 0.000 1.000 30,991.000 1(Net Inv.> 0) 0.396 0.489 0.000 1.000 30,991.000 1(Gross Inv.> 0) 0.867 0.339 0.000 1.000 30,991.000 Gross Inv.Rate 0.175 0.8-1.16 1.500 30,991.000 Gross Inv.Rate-Intang.Assets 0.014 0.087-0.443 0.950 30,991.000 Gross Inv.Rate-Tang.Assets 0.161 0.1-1.137 1.47 30,991.000 back

THANK YOU FOR YOUR ATTENTION! DG ECONOMICS, STATISTICS AND RESEARCH