Diagnosing the cause is critical for devising policies to remedy the fall in investment.

Transcription

1 Privae Invesmen: Wha s he Holdup? Aqib Aslam, Samya Beidas-Srom, Daniel Leigh (lead), Seok Gil Park, and Hui Tong. Suppor from Gavin Asdorian, Joshua Bosshard, Angela Espiriu, Hao Jiang, Yun Liu, and Hong Yang. Consulan: Sebnem alemli-ozcan. 1 Moivaion Debae over why businesses are no invesing more. o Is low invesmen mainly sympom of weak economic environmen? (E.g., Chinn, 211; rugman, 211.) o Are special impedimens o blame, such as policy uncerainy or financial secor weaknesses? (E.g., European Invesmen Bank, 213; Bui and Mohl 214.) Diagnosing he cause is criical for devising policies o remedy he fall in invesmen.

2 Cenral quesions of he chaper 1. Is here a global slump in privae invesmen? 2. Is he slump in privae invesmen due o housing or is i broader? 3. How much of his slump reflecs he weakness of demand? 4. Which businesses have cu back more on invesmen and why? 3 1. Is here a global slump in privae invesmen? AEs. Real Privae Fixed Invesmen (Log index; 199 = ) World 2. Advanced Economies EMDEs 4. EMDEs excl. China

3 2. AEs: Jus housing or broader? Broader. Caegories of Real Fixed Invesmen (Log index; 199 = ) Residenial 2. Business Business Srucures 4. Business Equipmen Housing: A small share of oal invesmen. Shares and Relaive Prices of Invesmen Caegories (Percen of oal fixed invesmen, unless noed oherwise) 1. Residenial vs. Business 2. Privae vs. Public 1% 1% 9% 9% 8% 8% 7% 7% 6% 6% 5% 5% 4% 4% 3% 3% 2% 2% 1% 1% Business Residenial Privae Public % % % 9% 8% 7% 6% 5% 4% 3% 2% 3. Business Invesmen: Equipmen vs. Srucures 4. Relaive Prices: Equipmen vs. Srucures 1 1%.8 Equipmen Srucures %

4 Decomposing he slump: No mainly housing. 5 Decomposiion of he Invesmen Slump, (Average percen deviaion from spring 27 forecass) 1. Residenial vs. Business 2. Public vs. Privae Business Residenial Toal Advanced Unied Unied Japan Seleced Oher euro Oher Advanced Unied Unied Japan Seleced Oher Oher economies Saes ingdom euro area area economies Saes ingdom euro area euro area 1/ 1/ 1/ Euro area economies (Greece, Ireland, Ialy, Porugal, Spain) wih high borrowing spreads during he sovereign deb crisis. Public Privae Toal How much reflecs oupu? As menoned, imporan o diagnose correcly implicaons for policy. A. Has he comovemen of invesmen and oupu been unusual? o Is his ime differen from hisorical recessions? B. How much has weak economic aciviy driven he weakness in invesmen o Address reverse causaliy issues using insrumenal variables. 8

5 A. Unusual comovemen of invesmen and oupu? Responses of Business I and Y o Various Shocks Raio of Responses (Average I o Average Y) Hisorical recessions GFC AEs GFC crisis AEs GFC noncrisis AEs 1. Business Invesmen 2. Oupu Hisorical recessions GFC AEs GFC crisis AEs GFC noncrisis AEs 9 B. How much reflecs weak oupu? The bulk. Firs pass: Is his ime differen relaive o hisorical recessions? Advanced Economies: Real Business Invesmen (Percen deviaion from precrisis forecas; 9 percen C.I.) Acual Prediced Nex: Idenify effec of oupu on -1 business invesmen. (Challenge: Reverse causaliy.) -2 Approach: Focus on shocks no riggered by business invesmen. (Fiscal, housing.)

6 B. How much reflecs weak economic aciviy? Invesmen-Oupu Relaion: Insrumenal Variables Esimaion Business Invesmen Growh ( ln I i )= α i + λ + β{insrumened d ln Y i } + ρ ln I i-1 + ε i (1) (2) (3) (4) β 2.445*** 2.633*** 1.719*** 2.243*** (.726) (.883) (.371) (.583) ρ.8*.179***.18*.138** (.66) (.62) (.64) (.64) R Number of Observaions Firs-Sage F-Saisic p-value <.1 <.1.9 <.1 Overidenificaion Resricionsp ions p-value Definiion of Y i GDP C + X GDP GDP Insrumens for ln Y i Fiscal shocks Fiscal shocks Housing shocks Fiscal and housing shocks Sources: Haver Analyics; naional auhoriies; and IMF saff calculaions. *p <.1, ** p<.5, *** p< Counry level: Acual invesmen close o predicion. Acceleraor Model: Real Business Invesmen (Log index) Unied Saes 1 2. Japan 3. Germany France : Q : Q : Q : Q3

7 Secondary in some: Financial consrains, policy uncerainy. I Counry-specific Acceleraor Model. 1 1 Y i i ix i 1 i 1 i 1 Excepions in some euro area economies afer Acual and Fied Real Business Invesmen (Percen deviaion from precrisis forecas; 9 percen C.I.) Seleced Euro Area Economies Acual 3.4 Spring 27 forecas 3.3 Acceleraor model predicion / Euro area economies (Greece, Ireland, Ialy, Porugal, Spain) wih high borrowing spreads during he sovereign deb crisis. 13 Secondary in some: Financial consrains, policy uncerainy. I Counry-specific Acceleraor Model. 1 1 Y i i ix i 1 i 1 i 1 Excepions in some euro area economies afer 211. Financial consrains. (Survey-based measure.) Acual and Fied Real Business Invesmen (Percen deviaion from precrisis forecas; 9 percen C.I.) Seleced Euro Area Economies 1 Acual 3.4 Spring 27 forecas 3.3 Acceleraor model predicion Adding financial consrains o he model / Euro area economies (Greece, Ireland, Ialy, Porugal, Spain) wih high borrowing spreads during he sovereign deb crisis. 14

8 Secondary in some: Financial consrains, policy uncerainy. I Counry-specific Acceleraor Model. 1 1 Y i i ix i 1 i 1 i 1 Excepions in some euro area economies afer 211. Financial consrains. (Survey-based measure.) Policy uncerainy. (News-based measure.) Acual and Fied Real Business Invesmen (Percen deviaion from precrisis forecas; 9 percen C.I.) Seleced Euro Area Economies 1 Acual 3.4 Spring 27 forecas 3.3 Acceleraor model predicion Adding financial consrains o he model Adding uncerainy o he model / Euro area economies (Greece, Ireland, Ialy, Porugal, Spain) wih high borrowing spreads during he sovereign deb crisis Which businesses have cu back more and why? From macro o micro (firm-level). Why? Focus on role of financial consrains and uncerainy. Use a difference-in-difference approach. 16

9 Firm-level surveys cie weak demand as dominan facor. Insufficien Demand 1 1. Europe 1 2. Sressed Euro Area 1 3. Unied Saes Financial Consrains 1 2. Sressed Euro Area Europe Unied Saes Esimaing he role of financial consrains. Esimae effec on I/ raio for firm i in secor j in counry k in year. Diff-in-diff approach of Dell Ariccia, Deragiache, and Rajan (28), applied o invesmen as in Claessens, Tong and Wei (2). I ijk, ijk, 1 Financial Dependence j Crisis k, l x l ijk, i k, k, d k, j, j, d j, ijk, Inuiion: If financial consrains play a significan role, hen firms in secors ha are more dependen on exernal finance should cu I more during a credi crunch. Daa: Thomson Reuers Worldscope; Sample: 28 AEs, 27,661 firms,

10 How do we measure financial dependence? Financial dependence a secor level (Rajan and Zingales, 1998). Fixed over ime. Financial Dependence Capial Expendiures Cash Flow Capial Expendiures Based on US firms. Apply o 3-digi secor level for all AEs. (Assumpion.) Inerac wih counry-level l credi crunch: Banking crisis (Laeven-Valencia); real credi growh. 19 Financial consrains Firm-Level Evidence: Financial Consrain Channel Raio of firm invesmen o lagged capial (1) (2) (3) Bank Crisis Financial Dependence.24***.23***.26*** (.7) (.7) (.8) Recession Financial Dependence.8 (.6) Sales o Lagged Capial Raio.8***.8*** (.) (.) Lagged Tobin'sQ.42***.42*** Fixed Effecs (.2) (.2) Firm Y Y Y Secor Year Y Y Y Counry Year Y Y Y Number of Observaions 161,73 16,239 16,239 R² Sources: Haver Analyics; naional auhoriies; Thomson Reuers Worldscope; and IMF saff calculaions. ***p <.1. 2

11 Financial consrains Firm-Level Evidence: Financial Consrain Channel Raio of firm invesmen o lagged capial (1) (2) (3) Bank Crisis Financial Dependence.24***.23***.26*** (.7) (.7) (.8) Recession Financial Dependence.8 (.6) Sales o Lagged Capial Raio.8***.8*** (.) (.) Lagged Tobin'sQ.42***.42*** Fixed Effecs (.2) (.2) Firm Y Y Y Secor Year Y Y Y Counry Year Y Y Y Number of Observaions 161,73 16,239 16,239 R² Sources: Haver Analyics; naional auhoriies; Thomson Reuers Worldscope; and IMF saff calculaions. ***p < Esimaing he role of uncerainy. Analogous diff-in-diff approach: I ijk, ijk, 1 Uncerain y Sensiiviy i i j Volailiyk, l xijk, i k, dk, j, d l k, j, j, ijk, Inuiion: If uncerainy has played a significan role, hen firms whose sock prices usually respond more wih aggregae measure of uncerainy ( sensiiviy ) should cu I more during periods of high aggregae uncerainy. Aggregae uncerainy: counry-specific (SD of counry sock index reurn). 22

12 Firm level: Measuring sensiiviy o uncerainy. Two measures of sensiiviy y( (fixed over ime): VIX-based. Regress sock reurn on marke reurn and VIX for US firms. Weekly daa (2-26). Collec coefficien for VIX, apply median of he 3-digi secor o all counries. (Assumpion.) News-based (Bloom e al.). Regress sock reurn on marke reurn and newsbased index for US firms. Weekly daa (2-26). Collec coefficien, apply median of 3-digi secor o all counries. (Assumpion.) 23 Policy Uncerainy Firm-Level Evidence: Policy Uncerainy Channel Raio of firm invesmen o lagged capial (1) (2) (3) Marke Volailiy Policy Uncerainy Sensiiviy.1*.28***.17** (.6) (.8) (.8) Bank Crisis Financial Dependence.24***.23** (.7) (.7) Sales o Lagged Capial Raio.8*** (.) Lagged Tobin'sQ 42***.42 Fixed Effecs Firm Y Y Y Secor Year Y Y Y Counry Year Y Y Y (.2) Number of Observaions 22,211 16, ,645 R² Sources: Haver Analyics; naional auhoriies; Thomson Reuers Worldscope; and IMF saff calculaions. *p <.1; ** p <.5; *** p <.1. 24

13 Firm level: Financial consrains: Some more inuiion. In banking crises, more financially dependen secors (op quarile) cu I// by 1.5pp more han less dependen secors (lowes quarile). Cauion: Diff-in-diff speaks o relaive I performance of differen firms. Illusraion: Relaive I performance of differen firms since he crisis. i 25 Firm level: Financial consrains and uncerainy play a role. Response of Firm Invesmen o GFC, By Firm Type (percen; based on local projecion model) ) 1 By Degree of Sensiiviy o Policy Uncerainy (percen; based on local projecion model) ) Less financially dependen More financially dependen Less sensiive o uncerainy 2 More sensiive o uncerainy

14 Conclusions 1. Slump in privae invesmen: Mosly in AEs, broad-based. (No jus housing.) 2. Firms acing normally given weak economic environmen. Lile unexplained weakness. 3. Some excepions: financial consrains, uncerainy. 4. Comprehensive se of policies required. Suppor overall demand. Faser recovery would lif invesmen. 27 Japan

15 Japan Caegories of Real Fixed Invesmen (Log index; 199 = ) Residenial 2. Business Business Srucures 4. Business Equipmen Sources: Haver Analyics; naional auhoriies; and IMF saff calculaions. Japan. Privae Invesmen and Componens-o-GDP Raio Source: IMF, World Economic Oulook. Noes: Relaive Price of Invesmen = Privae Invesmen Value/ Privae Invesmen Volume; Base year is

16 Housing even smaller in Japan. Price of equipmen fla. Shares and Relaive Prices of Invesmen Caegories (Percen of oal fixed invesmen, unless noed oherwise) 1. Residenial vs. Business 2. Privae vs. Public 1% 1% 9% 9% 8% 8% 7% 7% 6% 6% 5% 5% 4% 4% 3% 3% 2% 2% 1% 1% Business Residenial Privae Public % % % 9% 8% 7% 6% 5% 4% 3% 2% 3. Business Invesmen: Equipmen vs. Srucures 4. Relaive Prices: Equipmen vs. Srucures 1 1%.8 Equipmen Srucures % OTHER

17 Invesmen raios o GDP: Lile sign of global slump Privae Invesmen-o-GDP Raio 1. World 2. Advanced Economies 3. EMDEs EMDEs excl. China Source: IMF, World Economic Oulook. Noes: Relaive Price of Invesmen = Privae Invesmen Value/ Privae Invesmen Volume; Base year is 25; AEs included are: Ausralia, Ausria, Belgium, Canada, Cyprus, Czech Republic, Denmark, Esonia, Finland, France, Germany, Greece, Hong ong SAR, Iceland, Ireland, Israel, Ialy, Japan, orea, Lavia, Lihuania, Luxembourg, Mala, Neherlands, New Zealand, Norway, Porugal, San Marino, Singapore, Slovak Republic, Slovenia, Spain, Sweden, Swizerland, Taiwan 33 Province of China, Unied ingdom and Unied Saes; Brisk invesmen growh in he 2s across EMDEs, bu slowdown in recen years even relaive o pre-boom forecass Real Privae Fixed Invesmen (Log index; 199 = ) EMDEs excl. China Lain America and he Caribbean 3. Commonwealh of Independen Saes Emerging and Developing Europe

18 A. Has he comovemen of invesmen and oupu been unusual? Emerging and Developing Economies Responses of Business I and Y o Various Shocks Raio of Responses (Average I o Average Y) Emerging and Developing Economies (EMDEs) Emerging and Developing g( (ED) Asia Emerging and Developing (ED) Asia excl. China Lain America and Caribbean (LAC) 5 1. Privae Invesmen 5 2. Oupu Hisorical EMDEs EMDEs ED Asia ED Asia excl. CHN LAC 35 Real Business Invesmen: Acceleraor Model Residuals and Invesmen Losses Relaive o Precrisis Forecass, percen confidence inerval Residual Toal Invesmen Loss

19 Financial markes unusual given firms invesmen decisions? Tobin s Q and Real Business Invesmen-o-Capial Raios 4 1. Unied Saes 4 2. Japan 3. Germany 4 4. Unied ingdom Sources: Haver Analyics; naional auhoriies; and IMF saff calculaions Tobin s Q no in lock sep, bu can predic fuure invesmen Wha we can say: I/ and Tobin s Q -.5 no in lock sep Bu is his unusual? No. Blanchard, Rhee, Summers (1993). Some evidence: Q has lagged relaion wih I. 1. Unied Saes Germany Japan Unied ingdom