ESTUDIOS SOBRE LA ECONOMÍA ESPAÑOLA. Oil Shocks and the Business Cycle in Europe. Carlos De Miguel Baltasar Manzano José Mª Martín-Moreno

Transcription

1 ESTUDIOS SOBRE LA ECONOMÍA ESPAÑOLA Oil Shocks and he Business Cycle in Europe Carlos De Miguel Balasar Manzano José Mª Marín-Moreno EEE 215 Ocober 2005 ISSN Las opiniones conenidas en los Documenos de la Serie EEE, reflejan exclusivamene las de los auores y no necesariamene las de FEDEA. The opinions in he EEE Series are he responsibiliy of he auhors an herefore, do no necessarily coincide wih hose of he FEDEA.

2 Oil Shocks and he Business Cycle in Europe * Carlos De Miguel, Balasar Manzano and José Mª Marín-Moreno Universidad de Vigo and rede Absrac This paper analyzes he effecs of oil price shocks on he business cycle of he EU-15 counries using a sandard dynamic general equilibrium for a small open economy in which oil is included as an impored producive inpu. The resuls show ha oil shocks can accoun for a significan percenage of GDP flucuaions in many of hose counries. Furhermore, we show ha he increases in he relaive price of oil had a negaive effec on welfare, paricularly in souhern European counries, which are hisorically associaed wih a lax moneary policy during oil crisis. Key words: Oil crisis, aggregae flucuaions, welfare cos. JEL: Q43, E32. * The auhors would like o hank paricipans in he 1 s Alanic Workshop on Energy and Environmenal Economics for heir valuable suggesions. Financial suppor from he Spanish Minisry of Science and Technology and from FEDER hrough gran BEC and from Xuna de Galicia (PGIDIT03PXIC30001PN, PGIDIT03CSO30001PR) is graefully acknowledged. cmiguel@uvigo.es; bmanzano@uvigo.es; jmarin@uvigo.es Universidad de Vigo. Lagoas-Marcosende s/n Vigo (Spain).

3 1.- Inroducion The effecs of oil price changes on economic aciviy have been widely sudied. An increase of oil prices ends o reduce he level of economic aciviy, given is implicaions on he evoluion of imporan macroeconomic variables and due o he srong dependence of wesern economies on his inpu. Recenly, we have winessed a subsanial increase in he prices of oil, derived mainly from he high demand of emerging economies (China among ohers) and from he poliical and economic siuaion of Iraq. In his framework, he economic problems derived from rends in oil prices appear again, no only wih regard o he siuaion of he markes, bu also o heir possible macroeconomic effecs. Tensions in he oil markes have reurned, and we are perhaps enering a period ha some expers call he end of he cheap oil age. The ghos of old crises seems o be reappearing, sirring a fear of recessions in he main economies ha inegrae he European Union. As i is very well-known, for economies so dependen on oil, increases in he price of he barrel have differen adverse effecs. In he shor run, hey involve a reducion of GDP and an increase of he inflaion rae. In he medium and long run, indusrial producion is affeced, consumpion decreases due o he fall of purchasing power and invesmen also falls, affecing he cyclical posiion of he economy and ciizens welfare. Consequenly, wih oil prices beween 40 and 50 dollars per barrel, he economic implicaions should be sudied. However, o be able o carry ou a rigorous analysis of wha his new oil crisis in he European seing can represen, and o make an accurae analysis of economic effecs, we mus siuae ourselves in a cerain ime frame. Thus, in his paper, we analyze he effecs of he increase in oil prices since he firs crisis (70s), on economies so dependen on oil as he Europeans. The goal of he aricle is wofold. Firsly, we analyze he imporance of his source of shocks for boh he size and he shape of aggregae flucuaions in hose economies during he las hree decades. Secondly, we quanify he effecs of he changes in relaive oil prices on welfare in hese economies. The effecs of oil price shocks on indusrialized economies have been widely acknowledged in economic lieraure. Pindyck (1979), Hamilon (1983) and Olson (1988) sugges ha hese shocks affec growh as well as he business cycle, hereby becoming an addiional source of economic flucuaion. There is exensive empirical lieraure ha offers evidence of an asymmeric relaionship beween oil prices and 2

4 aggregae economic aciviy (see Mork, 1989 and 1994). Kim and Loungani (1992) and Finn (1995) analyze he role of energy price shocks using real business cycle s in closed economies, focusing on he US. These auhors find ha such shocks offer very lile help in explaining he aggregae flucuaions in he economy in quesion. However, De Miguel, Manzano and Marín-Moreno (2003) show ha, in a small open economy framework, oil price shocks are very imporan when explaining aggregae flucuaions. This paper presens a real business cycle for he differen European economies. The used is a sandard general equilibrium of a small open economy in which oil is included as an impored producive inpu. The relaive oil price as well as he real ineres rae is assumed o be se in inernaional markes, so we consider a small open economy in he sense of aking hose prices as given. Oil price shocks are he only source of flucuaion considered. Therefore, alhough he economy is hi by many shocks, our analysis is condiional on a single shock. Thus, his analysis would allow us o verify he exen o which oil price shocks can accoun for aggregae flucuaions in European economies. The resuls show ha oil shocks can accoun for a significan percenage of GDP flucuaions in many of hose counries, alhough he explanaory power is smaller for ohers. This wide range of variaion can be explained by differences in he srengh of moneary policies ha affec relaive oil prices in each counry. In addiion, he reproduces he cyclical pah of he European economies in periods of oil crisis. Finally, we show ha increases in he relaive price of oil had a negaive effec on welfare, paricularly in souhern European counries, which are hisorically associaed wih a lax moneary policy during oil crises. The remainder of his paper is organized as follows. Secion 2 describes he. Secion 3 discusses he choice of parameer values. We repor he main resuls in Secions 4 and 5. Finally, in he las secion, we presen he conclusions. 2.- The Model The described in his secion is a sochasic dynamic general equilibrium of a small open economy populaed by a large number of infinie-lived households, and firms ha need o impor oil o produce a consumpion good. The basic srucure of he, in erms of preferences and echnology, is similar o he De Miguel, Manzano and Marín-Moreno (2003) srucure. The general feaures of he are he following: 3

5 The producion of he final good, y, requires he use of labor, n, capial, k and energy, e. The producion echnology of firms is described by a nesed CES funcion wih consan reurns o scale: 1 θ υ υ υ [(1 a) k ] + ae, θ = F ( n, k, e ) n (1) where θ is he labor share and he parameer υ is equal o (1-s)/s, where s is he elasiciy of subsiuion beween capial and energy. The economy s resource consrain for period is given by: c + i + xn = y, (2) where c is privae consumpion, i is invesmen and xn are ne expors. Capial, k, accumulaes according o he law of moion: i k 1 δ ) k + ( k, k ), (3) = + 1 ( Φ + 1 where δ is he depreciaion rae and Φ is a capial adjusmen cos funcion which we assume o be quadraic: Φ( k, k + φ k 1 1 ) 2 = + k k 2. (4) The represenaive firm solves: Max F( n, k, e ) w n r k p e, (5) where w is he wage, r he capial rae of reurn and p he relaive oil price. In equilibrium, marginal produciviies are equal o inpu prices: w = F n, r = F k. The relaive oil price follows a saionary sochasic process: ln p p 1 p = F e = p + ρ ln + ε, ε ~ N(0, σ p ), ρ <1. (6) and We assume ha he individuals can buy or sell an inernaional asse, b, a an exogenous inernaional ineres rae, r*. The evoluion of ne expors is dicaed by: 4

6 * 1 (, xn = p e + b + 1+ r ) b, (7) where p e are oil purchases. Consumers maximize he expeced value of lifeime uiliy subjec o heir budge consrain: Max s.. c E 0 + k = β 1 σ (1 δ ) k ν ( c ψn ) + Φ( k 1 σ, k ) + b + 1 = w n + r k + (1 + r * ) b, (8) where 0< β <1 is he subjecive rae of ineremporal discoun, σ >0 is he parameer of relaive risk aversion, ν is one plus he inverse of he ineremporal elasiciy of subsiuion of labor supply and ψ is a posiive parameer. The condiions ha solve he consumer's problem are he following: U U n c + U c w = 0, (9) k+ k 1 1 1, 2 k+ k+ k+ + φ = βe + + U c δ r φ 2 (10) k k+ 1 k+ 1 c + 1 * { U (1 r )} Uc = β E +, (11) c + k + 1 (1 δ ) k + Φ( k, k + 1) + b + 1 = w n + r k + (1 + r ) b. (12) * 3.- Parameer values We now briefly describe our procedures for selecing parameer values lised in Table 1. We follow he sandard real business cycle lieraure in using seady-sae condiions o find parameer values maching average values observed in he daa, while oher parameers will be equal o sandard values used in he lieraure. The is calibraed o reproduce average values of he European Union in annual daa from , before he unificaion of 2004 (EU-15). The main sources for he daa used are he AMECO Daabase from Eurosa and Inernaional Energy Agency (IEA) Saisics. The depreciaion rae of capial is obained from equaion (3) in seady sae, δ = i / k, where i/k is he average value of EU-15 daa, while he discoun facor β was se by using equaion (11) in seady sae, 1 = β (1 + r*), so as o give a real annual ineres rae of 4%. 5

7 Table 1.- Parameers of he economy. Preferences Subjecive Discoun Rae β 0.96 Parameer of he Uiliy Funcion ψ 1.53 Risk Aversion σ 11 Parameer of he Uiliy Funcion v 1.7 Technology Labor Share θ 0.64 Rae of Depreciaion δ 0.06 Parameer of he Producion Funcion υ 0.7 Inernaional Ineres Rae r * 0 The value of a, represening he imporance of oil wih respec o capial in he producion funcion, is obained from he firs order condiions of he firm s problem in he seady sae, ( ) ( υ+ 1 a = 1 [ e k ) (( r * + δ ) p) + 1], where e/k and p represen average values over he sample, and υ is borrowed from Kim and Loungani (1992). Table 2.- Process. Consan p Persisence Coefficien Ρ Sandard Deviaion σ Porugal Spain Greece Ialy France Unied Kingdom Ireland Germany Belgium Luxembourg Neherlands Ausria Sweden Finland Denmark p ψ = θ The value of he parameer ψ is chosen from (9) in he seady sae: ( υ 1) ( y n) vn, assuming ha he producive ime is 5476 hours per year. The oupu per worker, y/n, and he labor share, θ, represen averages for he EU-15 economies. 6

8 The parameer of he adjusmen coss φ is calibraed so he variabiliy of invesmen relaive o oupu in daa would be reproduced by he. The remaining parameers are chosen in conformiy wih earlier sudies. The parameer of he uiliy funcion v is aken from Greenwood e al. (1988), while he risk aversion parameer σ is obained from Mendoza (1991). Finally, he sochasic process parameers of relaive oil prices for each counry of he EU-15 are esimaed, in domesic currency, from equaion (6). Table 2 repors he resuls of he esimaion. 4.- Oil shocks and EU-15 aggregae flucuaions In his secion, we es how accuraely he driven by oil price shocks can fi he business cycle of he EU-15 beween 1970 and We sar by running simulaions, including for each counry he corresponding sochasic process for he relaive oil price presened in able 2. Such an experimen allows us o compare he acual GDP daa wih he corresponding flucuaions of he oupu in he, obaining he percenage of GDP volailiy ha can be explained by he wih oil shocks. Thus, we can explore o wha exen oupu flucuaions a each European counry could be generaed by oil shocks. The simulaion resuls are summarized in able 3. The relaive price of oil in each counry is obained as he raio beween he price of he Bren barrel expressed in domesic currency and he corresponding GDP deflaor. Therefore, he differences of relaive oil prices beween counries arise eiher from inflaion or from he exchange rae. In his sense, boh he exchange rae and he moneary policies would have been he main ool o accommodae he effecs of oil crises in each counry. In ligh of he simulaion resuls, we can consider several groups of counries. The firs group includes Porugal, Spain, Greece and Ialy, ha is, counries wih large GDP flucuaions, where he role of oil shocks is remarkable, responsible from 30% o 42% of oal oupu flucuaions in Porugal and Ialy, respecively. Those resuls are consisen wih he common view abou souhern European counries excessively lax moneary policy, which led o more difficulies in accommodaing oil shocks. There is anoher group of counries (Ausria, Denmark and Sweden) in which oil disurbances also play an imporan role, explaining around 30% of oupu flucuaions, alhough heir oupu is quie less volaile. 7

9 Table 3.- Comparison beween he Predicions of he Model and he Daa. Daa GDP (sandard Model Explanaion of Oupu volailiy deviaion) Porugal 5.51% 30% Spain 3.47% 37% Greece 4.07% 42.6% Ialy 3.17% 41.8% France 2.06% 22% Unied Kingdom 1.51% 10.6% Ireland 2.36% 8.9% Germany 2.69% 10% Belgium 2% 19.8% Luxembourg 2.82% 15.6% Neherlands 2.16% 16.7% Ausria 2.14% 28.5% Sweden 1.30% 33.1% Finland 4.08% 13.2% Denmark 1.38% 30.4% The main group includes he res of he EU-15 counries, in which he conribuion of oil shocks o explain GDP flucuaions scores from 10% o 20%. Tha group includes counries like France, UK, Neherlands, Belgium, Luxembourg and Germany, whose cenral banks have implemened sronger moneary policies o face he oil crises of he 70s. Therefore, as i was poined ou above, he role of moneary policies has been crucial in accommodaing he oil shocks ha hi wesern economies in he las hree decades. Oil shocks are an imporan source for explaining aggregae flucuaions across Europe, alhough wih a wide range of variaion. Bu i is also imporan o es wheher he inroducion of oil shocks can mimic he business cycle shape. In order o do ha, we simulae he wih he acual pah of he relaive price of oil as he only source of flucuaion. The analysis allows us o verify he exen o which oil prices acually give rise o he business cycle pah of he EU-15 counries. Figures 1-15 represen he evoluion of he relaive price of impored oil, and he comparison beween he evoluion of he acual GDP and he oupu pah simulaed by he for each counry. Given ha we focus on he business cycle, boh series have been derended using he Hodrick-Presco filer. 8

10 90 Oupu.2 Poruguese escudos daa Figure 1. Porugal. Spanish peseas Oupu daa Figure 2. Spain. 140 Oupu.15 Greek drachmas daa Figure 3. Greece. 800 Oupu Ialian lire daa Figure 4. Ialy. 9

11 2.4 Oupu French francs daa Figure 5. France..45 Oupu UK serling daa Figure 6.Unied Kingdom..28 Oupu.24 Irish pounds daa Figure 7. Ireland. Oupu Deusche Mark daa Figure 8. Germany. 10

12 Oupu Belgian francs daa Figure 9. Belgium. 16 Oupu Luxembourg francs daa Figure 10. Luxembourg..8 Oupu.7 Duch guilders daa Figure 11. Neherlands. Oupu Ausrian schillings daa Figure 12. Ausria. 11

13 .5 Oupu.4 Swedish krona daa Figure 13. Sweden. 2.4 Oupu Finnish markkas daa Figure 14. Finland..40 Oupu.35 Dannish krone daa Figure 15. Denmark. The char of he relaive oil prices roughly shows he same pah for all counries analyzed. The firs grea oil crisis was riggered by he Arab-Israeli Yom Kippur War when he OPEC imposed an oil embargo on wesern counries, hereby dramaically increasing oil prices a he beginning of The increase had immediae effecs on he evoluion of he oupu in he as well as on he GDP daa, hough wih some delay in several counries. Therefore, he saisfacorily reproduces he negaive effec on oupu during he firs oil crisis. The nex period which calls for our aenion is he second oil crisis, which akes us back o 1979 and was riggered by he Iranian 12

14 Revoluion, causing he relaive price of oil o double in less han a year. Again, he increase in oil prices reduces he oupu in boh he and he economy in mos counries, which is accuraely depiced by he. Some counries responded o he oil crisis wih some delay (Germany, Belgium, Neherlands, Ausria and Ialy). A he beginning of 1986, he oil marke collapsed, bringing abou a significan drop in oil prices. This siuaion affeced oupu in a posiive way, which quickly recovered. The reacs insanly when confroned wih he fall in oil prices, while he recovery in he daa begins some periods laer in mos counries. Summing up, he guided by changes in oil prices is able o mimic he shape of he oupu flucuaions in episodes of dramaic changes in he oil marke, especially for he wo crises in he 70s and he siuaion in he mid-80s. However, when oil marke condiions were sable, as hey became in 1986, here are oher disurbances ha explain aggregae flucuaions, so he driven by oil shocks fails o reproduce he cyclical pah of he counries analyzed. The resuls show he vulnerabiliy of small open economies heavily dependen on impored oil when hey are confroned wih large changes in he condiions ha conrol he inernaional oil marke. This highlighs he imporance of considering he behavior of relaive oil prices when analyzing he business cycle. 5.- Shocks and Welfare The previous secion sresses he influence of oil shocks on boh he size and he shape of aggregae flucuaions. Neverheless, oil shocks affec no only GDP flucuaions bu also welfare. As De Miguel, Manzano and Marín-Moreno (2003) emphasized, hose effecs are paricularly imporan in he framework of a small open economy dependen on oil impors because of no domesic oil producion, and he possibiliies of subsiuing his inpu wih capial are limied due o he complemenariy beween capial and oil. In his secion, we evaluae he loss of welfare occasioned by oil price increases regisered from he firs half of he 70s unil he second half of he 80s. The relaive price in he las quarer of 1973 was very similar o he value regisered in he hird quarer of 1986 in each counry; herefore, we can consider he inermediae period as a emporary price increase whose welfare cos we calculae. The welfare loss of oil crises is esimaed by he welfare cos derived from he increase in oil prices wih respec o he iniial siuaion. This cos is defined a each momen as he percenage of he increase in consumpion ha an individual would 13

15 require o enjoy he same level of welfare wih respec o he saring poin. The welfare cos a each momen is calculaed hrough he x variable ha solves he following equaion: [ c ~ ( 1 + x), n] U = U ~, (13) where U is he level of uiliy reached a he iniial siuaion, in his case 1973, while c ~ and n ~ represen consumpion and hours worked a each momen, which provide a level of welfare U. Thus, he produc c ~ x indicaes he oal increase in consumpion required o resore he iniial level of welfare. This welfare cos measure is usually expressed as a percenage of oupu. Table 4.- Welfare Cos: Average cos per period (percen of he Oupu) Porugal 8.14% Spain 5.98% Greece 7.93% Ialy 6.97% France 1.74% Unied Kingdom 0.69% Ireland 0.74% Germany 1.12% Belgium 3.59% Luxembourg 3.67% Neherlands 1% Ausria 2.33% Sweden 0.87% Finland 1.51% Denmark 0.79% The resuls, summarized in able 4, show ha he welfare cos from oil crises was very differen beween counries. Again, souhern European counries, wih a lax moneary policy, had a larger welfare cos, ranging from 6% of oupu in Spain o more han 8% in Porugal, so consumers should have been compensaed in each period wih a significan fracion of GDP in erms of consumpion in order o make up for he loss in welfare derived from he differen oil crises. The welfare cos in he res of he counries was moderae (Belgium and Luxembourg) or even small. This gives us an idea of he significan amoun of he loss in welfare brough abou by he differen oil crises for a 14

16 small open economy, and he imporance of managing moneary policy o accommodae such kinds of shocks. 6.- Conclusions In his paper, we have analyzed he effecs of oil price shocks on he business cycle of he EU-15 counries. The used for his analysis is based on he sandard dynamic general equilibrium for a small open economy in which oil is included as an impored producive inpu. The price of oil and he ineres rae are assumed o be se by inernaional markes. The calibraion of he parameers of he is carried ou by aking daa from he EU-15 economies during he period The resuls show ha oil shocks can accoun for a significan percenage of GDP flucuaions in many of hose counries, bu he explanaory power is quie smaller for ohers. Tha wide range of variaion can be explained by differences in he srengh of moneary policies. In addiion, he reproduces he cyclical pah of he European economies in periods of oil crisis. Finally, we have shown ha he increases in he relaive price of oil had a negaive effec on welfare, paricularly in souhern European counries, which are hisorically associaed wih a lax moneary policy during oil crisis. References De Miguel, C., B. Manzano and J. M. Marín-Moreno (2003), Oil price shocks and aggregae flucuaions, The Energy Journal, 24(2), Finn, M. (1995), Variance properies of Solow s produciviy residual and heir cyclical implicaions, Journal of Economic Dynamics and Conrol, 19, Greenwood, J., Z. Hercowiz and G. Huffman (1988), Invesmen, capaciy uilizaion and he real business cycle, American Economic Review, 78, Hamilon, J. (1983), Oil and he macroeconomy since World War II, Journal of Poliical Economy, 91, Kim, I. and P. Loungani (1992), The role of energy in real business cycle s, Journal of Moneary Economics, 29, Mendoza, E. G. (1991), Real business cycles in a small open economy, American Economic Review, 81,

17 Mork, K. A. (1989), Oil and he macroeconomy when prices Go Up and Down: An exension of Hamilon s resuls, Journal of Poliical Economy, 97, Mork, K. A. (1994), Business cycles and he oil marke, The Energy Journal, Special issue, The Changing World Oil Marke, Olson, M. (1988), The produciviy slowdown, he oil shocks, and he real cycle, The Journal of Economic Perspecives, 2, Pindyck, R. S. (1979), The Srucure of World Energy Demand, Cambridge MA: The MIT Press. 16