higher in developing countries An empirical investigation Alejandro Izquierdo Ruy Lama InterAmerican Development Bank International Monetary Fund Juan Pablo Medina Jorge Puig Universidad Adolfo Ibáñez ID: 1028793
Download Presentation The PPT/PDF document "Is the public investment multiplier" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
1. Is the public investment multiplier higher in developing countries? An empirical investigationAlejandro IzquierdoRuy LamaInter-American Development BankInternational Monetary FundJuan Pablo MedinaJorge PuigUniversidad Adolfo IbáñezUniversidad Nacional de La PlataDaniel Riera-CrichtonCarlos VeghWorld BankJohns Hopkins UniversityGuillermo VuletinWorld Bank
2. Road map Motivation and sneak peek of our findings2. Theoretical model 3. Empirics4. Final thoughts and policy implications
3. Road map Motivation and sneak peek of our findings2. Theoretical model 3. Empirics4. Final thoughts and policy implications
4. MotivationDuring the last decade, there has been a growing literature on spending multipliers (with a special focus on total aggregate spending or government consumption). Based on a well-known macroeconomic theory, the size of the spending multiplier is: - larger in recessions (Auerbach and Gorodnichenko, 2010 and 2012; Riera-Crichton, Vegh, and Vuletin, 2015) - larger when the exchange rate is fixed (Ilzetzki, Mendoza, and Vegh, 2013) - larger when debt is low (Ilzetzki, Mendoza, and Vegh, 2013; Huidrom, Kose, and Ohsorge, 2019) - larger when monetary policy is accommodative/zero lower bound (Lawrence, Eichenbaum, and Rebelo, 2011; Pennings and Giambattista, 2017) - smaller in open economies (Ilzetzki, Mendoza, and Vegh, 2013)
5. MotivationIn a recent survey, Ramey (2019) points out: “There is surprisingly little recent aggregate evidence on multipliers for public investment.”What do we know about public investment multipliers? - public investment multipliers > government consumption multipliers due to crowding-in effects on private investment. Theory: Aschauer (1989a, 1989b) and Baxter and King (1993) Empirics: Leduc and Wilson (2012), Eden and Kraay (2014), Calderon, Moral-Benito and Serven (2015), Furceri and Li (2017)
6. MotivationKnowing more about the public investment multiplier (and its determinants) is essential if we do not want policymakers “flying blind,” especially when taking into account the big push for public investment/infrastructure.For example, based on first principle arguments, there is a growing consensus about the importance of efficiency: - higher efficiency implies larger public investment multipliers Theory: Leeper, Walker, and Yang (2010) Empirics: Cavallo and Daude (2011), Leduc and Wilson (2012), Furceri and Li (2017), Izquierdo, Pessino, and Vuletin (2018)
7. What do we do?Motivated by a large policy discussion and agenda pointing to the dire need for public investment/infrastructure, especially in the developing world (due to large investment/infrastructure bottlenecks/gaps), in this paper, and also based on simple first principle arguments, we ask what we think is a natural and relevant question: Does the initial stock of public capital matter on the size of the public inv. multiplier? - In theory, and based on first principle arguments, it should! - In practice? The proof is in the pudding. We find it does!
8. What do we do? Theoretical front: Use mainstream Baxter and King (1993) calibrated neoclassical growth to analyze the macroeconomic effects of spending shocks (i.e., public consumption and investment shocks, like in Uhlig, 2010).Empirical front: Use alternative samples, levels of government, and identification strategies and find strong supporting arguments. - 31 European countries for 1987-2014 using Blanchard and Perotti (2002). - U.S. states for 1987-2016 using public spending forecast errors (like in Auerbach and Gorodnichenko, 2010 and 2012). - Argentinean provinces for 1964-2014 using IV approach. - Further international evidence comparing the results of 17 country-specific external studies with our estimates.
9. Road map Motivation and sneak peek of our findings2. Theoretical model 3. Empirics4. Final thoughts and policy implications
10. Theoretical model - Based on Baxter and King (1993) Model setup:
11. Theoretical model - Based on Baxter and King (1993) Model setup:
12. Theoretical model - Based on Baxter and King (1993) We consider a Cobb-Douglas production function:We follow Baxter and King (1993), Glomm and Ravikumar (1997), and Leeper et al. (2010) by assuming a production function with increasing returns to scale with respect to private and public inputs. Alternatively, we could assume the possibility of congestion of public capital as in Glomm and Ravikumar (1997). Under this alternative assumption, the simulated public investment multiplier is marginally smaller for an empirically plausible parametrization of G, in a range around 0.05.
13. Theoretical model - Based on Baxter and King (1993) Model equilibrium conditions:
14. Theoretical model - Based on Baxter and King (1993) Calibration - baseline:αG=0.07 is close to benchmark value in Baxter and King (1993) and consistent with cross-section empirical evidence from Calderon, Moral-Benito, and Serven (2015)
15. Theoretical model - Based on Baxter and King (1993) Calibration: cross-country statistics for stock of public capital over GDP in Europe
16. Theoretical model - Based on Baxter and King (1993) Public cons. and public investment multipliers (median stock of public capital/GDP) Effect on GDPUnlike public consumption, public investment directly improves the productive capacity of the economy by increasing the marginal product of private capital and labor, generating positive effects both on private investment and private consumption.
17. Theoretical model - Based on Baxter and King (1993) Public cons. and public investment multipliers (median stock of public capital/GDP) Additional effect on private consumption and private investment of a public investment shock (vis-à-vis a public consumption shock)
18. Theoretical model - Based on Baxter and King (1993) Public cons. and public investment multipliers (high/low stock of public capital/GDP) Effect on GDP
19. Theoretical model - Based on Baxter and King (1993) Public cons. and public investment multipliers (high/low stock of public capital/GDP) Additional effect on private consumption and private investment of a public investment shock (vis-à-vis a public consumption shock) Low stock of public capital/GDP High stock of public capital/GDP
20. Road map Motivation and sneak peek of our findings2. Theoretical model3. Empirics - Europe4. Final thoughts and policy implications
21. Empirics – EuropeSample: 31 European countries (quarterly) for 1987:Q1-2014:Q4.Methodology: Blanchard and Perotti (2002) based on Local Projections (Jorda, 2005). Specifications: #1: #2:- Stock of public capital from IMF-FAD.
22. Empirics – Europe Specifications: #3:
23. Empirics – EuropeInitial stock of public capital over GDP ratio in European countries.
24. Empirics – Europe (specification #1)We find that public investment multipliers are larger than consumption multipliers. Effect of public Effect of public consumption on GDP investment on GDP
25. Empirics – Europe (specification #1)Mechanism: In line with the model. Effect of public consumption Effect of public investment on private consumption on private consumption Effect of public consumption Effect of public investment on private investment on private investment
26. Empirics – Europe (specification #2)Public investment multiplier after two years of the spending shock:Mechanism: In line with the model.
27. Empirics – Europe (specification #2)Policy implications: Evaluated in 1990 Evaluated in 2014
28. Empirics – Europe (specification #3)Does efficiency matter on the size of the public investment multiplier? Yes it does, in the expected way.
29. Empirics – Europe (specification #3)Does efficiency affect our findings? No, it does not.Other macro variables are not behind the role of the initial stock of public capital.
30. Road map Motivation and sneak peek of our findings2. Theoretical model3. Empirics – U.S. states4. Final thoughts and policy implications
31. Empirics – U.S. statesSample: All U.S. states, but Hawaii and Alaska, (annual) for 1987-2016.Methodology: public spending forecast errors based on Local Projec. (Jorda, 2005). Specifications: #1: where #2:- Stock of public capital from Aklilu A. Zegeye (2000) for 1980-1985.
32. Empirics – U.S. statesInitial stock of public capital over GDP ratio in U.S. states.
33. Empirics – U.S. statesWe find that public investment multipliers are larger than consumption multipliers.Public consumption mult. ≈ 0 is consistent with Clemens and Miran (2012).Public inv. mult. ≈ 1 is consistent with Coenen et al. (2012) based on DSGE models.When using European data, we obtain public inv. mult. = 0.85
34. Empirics – U.S. statesPublic investment multiplier:
35. Empirics – U.S. statesPolicy implications: Evaluated in 1980-1985 Evaluated in 2012
36. Road map Motivation and sneak peek of our findings2. Theoretical model3. Empirics – Argentinean provinces4. Final thoughts and policy implications
37. Empirics – Argentinean provincesSample: All, but City of Buenos Aires, (annual) for 1964-2014.Methodology: Instrumental variables. Specifications: #1: We instrument ga with provincial over-representation in the Chamber of Deputies (defined as actual number of prov. legislators pc less prov. legislators based on proportional representation pc) at the times of constitutional changes (1949, 1972, and 1983). Motivated by Porto and Sanguinetti (2001) and Vegh and Vuletin (2015). Why? 1) Importance of federal transfers in financing prov. spending. 2) Political economy support ↑ over-representation ↑ federal transfers pc. 3) Constitutional changes driving changes in over-representation are driven by political economy considerations and not by macro circumstances (Vegh and Vuletin, 2015).
38. Empirics – Argentinean provincesSpecification #1 works in practice.
39. Empirics – Argentinean provinces Specifications: #2: where and We instrument θ with prov. population dependency ratio. Why? 1) Large literature points out that current spending increases with old (health and social security) and young (education and health) 2) Plausible exogeneity: (i) based on demographics (not labor) which is less sensitive to short-macro fluctuations. (ii) there is no statistical relationship between population dependency ratio and real GDP in Argentinean provinces and at a global level.
40. Empirics – Argentinean provincesSpecification #2 works in practice.
41. Empirics – Argentinean provinces Specifications: #3:
42. Empirics – Argentinean provincesInitial stock of public capital over GDP ratio in Argentinean provinces:
43. Empirics – Argentinean provincesWe find that public investment multipliers are larger than consumption multipliers.Public cons. mult. ≈ 0 is consistent with Avramovich et al. (2006) and Puig (2014).Public inv. mult. > 1 is consistent with Puig (2014).When using European data, we obtain public investment mult. = 1.71
44. Empirics – Argentinean provincesPublic investment multiplier:
45. Empirics – Argentinean provincesPolicy implications: Evaluated in 1964 Evaluated in 1990 Evaluated in 2014
46. Road map Motivation and sneak peek of our findings2. Theoretical model3. Empirics – Further evidence4. Final thoughts and policy implications
47. Empirics – Further empirical support based on external studiesCompile public investment mult. calculated in 17 country-specific external studies: Algeria, Argentina, Australia, Cameroon, Colombia, Czech Republic, Germany, Italy, Japan, Korea, Netherlands, Paraguay, Peru, Saudi Arabia, Spain, Tunisia, Unites States. We compare these findings with those obtained with our non-linear estimates of Section 3 (i.e., based on a sample of 31 European countries) and relevant stock of public capital over GDP from IMF-FAD.
48. Empirics – Further empirical support based on external studiesVis-à-vis comparison for each country:
49. Empirics – Further empirical support based on external studiesRegression between our public investment multiplier based on a non-linear strategy and public investment multiplier from external sources:We cannot reject that β=1
50. Road map Motivation and sneak peek of our findings2. Theoretical model3. Empirics – Further evidence4. Final thoughts and policy implications
51. Final thoughts and policy implicationsWe contribute to the literature studying the aggregate macroeconomic effects of public investment. In particular, we analyze whether the size of the public investment multiplier depends on the initial stock of public capital. We show both theoretically and empirically (using alternative samples, levels of government, and identification strategies) that:1) Public inv. multipliers are, on average, larger than those of public consumption.2) The smaller the initial stock of public capital, the larger the public inv. multiplier.3) We also show that 1) and 2) occur due to basic crowding-in effects.4) Large public investment multipliers are not a universal regularity. They crucially depend on the initial stock of public capital.
52. Therefore, doing a little bit of a stretch…Policy implications (using 2014 ratios):