BANCO DE PORTUGAL Economics Research Deparmen Why should Cenral Banks avoid he use of he underlying inflaion indicaor? Carlos Robalo Marques Pedro Duare Neves Afonso Gonçalves da Silva WP 5-00 Augus 2000 The analyses, opinions and findings of his paper represen he views of he auhors, hey are no necessarily hose of he Banco de Porugal. Please address correspondence o Carlos Robalo Marques, Economics Research Deparmen, Banco de Porugal, Av. Almirane Reis, nº 71, 1150-012 Lisboa, Porugal Tel.#351-1-3130000; Fax#351-1-3143841;e-mail:cmrmarques@bporugal.p.
Why should Cenral Banks avoid he use of he underlying inflaion indicaor? Carlos Robalo Marques (*) Pedro Duare Neves (*) Afonso Gonçalves da Silva (*) Augus 2000 Absrac This paper assesses he usefulness of he commonly used underlying inflaion indicaor, in ligh of he crieria proposed in Marques e al. (2000). Empirical evidence for a group of six counries srongly suggess ha he use of underlying inflaion as an indicaor of rend inflaion should be avoided. Keywords: core inflaion indicaor; underlying inflaion; evaluaion crieria JEL classificaion: C43, E31, E52 (*)We would like o hank Maximiano Pinheiro and José Ferreira Machado for helpful commens on a previous draf of his paper. The usual disclaimer applies. 1
1. Inroducion The CPI excluding unprocessed food and energy indicaor, also known as underlying inflaion, was one of he firs core inflaion indicaors ever proposed in he lieraure. Blinder (1982) uses he CPI excluding food, energy and morgage ineres coss o esimae he underlying inflaion for he USA in he 1970s, as hese componens were largely responsible for he inflaionary shocks of 74 and 78-80. During he 1980s and early 1990s, his ype of indicaors became exremely popular amongs Cenral Banks. 1 The procedure is moivaed by he high volailiy of he excluded caegories, which is supposed o be caused by emporary, non-moneary phenomena. 2 However, recen research on core inflaion has cas some doubs on he usefulness of his indicaor. Oher core inflaion indicaors have been developed such as rimmed means and mos of hem seem o ouperform he underlying inflaion measure. See, for insance, Bryan and Cecchei (1994), Freeman (1998) and Marques e al. (2000). On he oher hand, since here is no economic model explaining he consrucion of he indicaor, he selecion of he excluded iems is a purely subjecive decision. From a heoreical poin of view, alhough we can hink of core inflaion esimaion as a signal exracion problem, i seems highly unrealisic o assume ha some caegories conain no informaion a all. This paper presens furher evidence agains he use of he underlying inflaion indicaor, resoring o he concep of desirable properies of a measure of core inflaion, as defined in Marques e al. (2000). The empirical evidence presened below shows ha he underlying inflaion indicaor sysemaically fails hese condiions, for a reasonably large number of differen counries. 1 The monioring of such an indicaor is common pracice in he Unied Saes. Banco de Porugal [see Nascimeno (1990)], Banco de España, Banca d Ialia, he Deusche Bundesbank, De Nederlandsche Bank and Banque Naionale de Belgique have, among ohers, commened eiher on a regular basis or in a case-by-case basis he evoluion of such a ype of indicaor. See Álvarez and Maea (1999) for a complee lis of references. Finally, he European Cenral Bank regularly menions a similar measure of inflaion in is Monhly Bulleins. 2 Empirical evidence ha unprocessed food and energy prices are more volaile, on average, han he remaining componens of he CPI is provided for insance in Álvarez and Maea (1999). Bryan e al. (1997) also provides empirical evidence ha fresh food and energy goods are more likely o be rimmed han he average componen of he CPI in he compuaion of heir seleced rimmed inflaion measure. 2
This paper is organised as follows. Secion 2 describes some condiions ha a core inflaion indicaor should verify. In secion 3, some argumens agains he use of he underlying inflaion indicaor are presened. The empirical resuls are shown in secion 4. Secion 5 concludes. 2. The properies of core inflaion indicaors Assume ha for any given period he inflaion rae, say π, is broken down ino he sum of wo componens: a permanen componen named core or rend inflaion, say π c, and a emporary componen represened by u. Therefore, we have: π c = π +. (1) u In equaion (1), we assume ha he emporary disurbances in he inflaion rae, u, are caused by developmens such as changes in weaher condiions, disurbances in he demand and supply of goods, ec. By definiion, u is expeced o have zero mean and finie variance, and herefore non-saionariy is excluded on heoreical grounds. ice for insance ha, if u were allowed o exhibi a nonzero mean, hen π c would no be capuring he whole sysemaic componen of π. According o Marques e al. (2000), when inflaion is an I(1) process, an appropriae measure of core inflaion, say π *, should verify he following: 3 * * i) π is I(1) and π and π are coinegraed wih uniary coefficien, i.e. * ( π π ) is saionary wih zero mean; 4 * ii) here is an error correcion mechanism for π given by z 1 = ( π 1 π 1), i.e. γ 0 in 3 Marques e al. (2000) also proposes a se of esable condiions when he inflaion rae is a saionary variable. 4 This condiion was firs proposed by Freeman (1998). 3
m * * π = α π + β π γ ( π π ) + ε ; (2) j j j= 1 m j= 1 j j 1 1 iii) π * is srongly exogenous for he parameers of equaion (2). Condiion i) is a direc consequence of he definiion of u in equaion (1). Condiion ii) may be inerpreed as he requiremen of π * being an aracor for π, as he error correcion mechanism forces inflaion o converge owards is rend. In oher words, π * should ac as a leading indicaor for π. Condiion iii) guaranees ha he pah of π * is no influenced by pas values of π, i.e. ha π does no Granger cause π *. We es condiion i) in wo seps. Firs, we use a uni roo es o esablish he * saionariy of z = ( π π ). We hen es he null α = 0 in he saic regression * π = α + π +u, (3) given ha z is saionary. Afer esablishing condiion i), he verificaion of ii) is simple, jus requiring he esimaion of model (2). The hypohesis γ = 0 can hen be esed wih he convenional -raio of $γ. Condiion iii) implies ha in he error correcion model for π *, r * * * π = δ π + θ π λ( π π ) + η j j j= 1 r j= 1 j j 1 1, (4) he null hypohesis λ = θ =... = θr = should no be rejeced. A necessary condiion for 1 0 iii) is weak exogeneiy of π *. This can be verified by esing λ = 0 in equaion (4). 4
3. Why should one no expec underlying inflaion o measure core inflaion The underlying inflaion indicaor is obained by excluding some componens of he CPI. Le us define 0 1 P = αp + ( 1 α) P (5) where P sands for he CPI, P 1 for he iems excluded from he CPI wih he argumen ha hey are more volaile, i.e. energeic producs and unprocessed food, and ( 1 α ) for he weigh of he remaining goods and services. Therefore, by definiion, P 0 represens he price index used o compue underlying inflaion. For monhly daa his equaion may be rewrien as π = α w + ( 1 α ) v, (6) where 0 1 0 P P P P 12 π = 1; w = 1; v = 1; α = α. (7) 0 1 P P P P 12 12 12 12 ice ha w is our convenional underlying inflaion indicaor. w we may ask under wha condiions does w mee he firs crierion for a core inflaion indicaor. Recall ha his will be saisfied if, in he saic regression π = βw + u, (8) we have β = 1 and u ~ I(0). Le us see under wha circumsances we can approximae (6) by (8). If we define α = µ + ε, (9) 5
expression (6) can be rewrien as π = µ w + ( 1 µ ) v + ( α µ )( w v ), (10) w, if π is I(1), one expecs w and v also o be I(1). 5 If v and w are no coinegraed, hen clearly π and w may no be coinegraed eiher. On he oher hand, if v and w are coinegraed, we may wrie v = θw + η, (11) wih η ~I(0). Insering (11) ino (10), one ges π = µ + ( 1 µ ) θ w + u, (12) where u = ( α µ )( w v ) + ( 1 µ ) η. (13) w, if θ = 1, we are back o our definiion of a core inflaion measure presened in equaion (1) of he previous secion. So, in order o have β = 1 and u ~I(0) in (8), i is necessary and sufficien ha boh α and ( w v ) are saionary variables. The fac ha his may no occur means ha he iems excluded from he CPI in order o compue w, i.e. P 1, may conain some informaion ha sysemaically differs from he one included in z. Therefore, in compuing he underlying inflaion indicaor, we may be excluding oo much informaion, i.e., we may be excluding from π no only noise bu also signal. If his is he case, w will no mee condiion i) in secion 2. I is also easy o undersand why w mus no be expeced o mee condiions ii) and iii), i.e. o be a leading indicaor for π and no o be Granger caused by π. In order o compue w, we exclude from he CPI he prices of goods ha ener as 5 If v ~ I(0), hen (6) saes ha π α w is saionary. Therefore, i is no possible o have β=1 in (8), as 0< α < 1. 6
inermediae inpus in he producion process (energeic goods and unprocessed food). Therefore, changes in v are expeced o direc and conemporaneously affec π while affecing w indirecly wih a lag. This being so, v is a leading indicaor for w and as long as i affecs π, his means ha π also appears as a leading indicaor for w. In pracice, his siuaion would cause condiions ii) and iii) of he previous secion no o be verified, since CPI inflaion would cause, raher han be caused by, underlying inflaion. As secion 4 shows, his is he ype of resuls obained for he six counries considered in his sudy. 4. Empirical resuls This paper assesses he empirical properies of underlying inflaion for six differen counries: USA, Germany, France, Ialy, Spain and Porugal. For he cases of he USA and France, however, here is no available informaion for he unprocessed componen of food and, herefore, he whole class of food was excluded. Daa is described in Table 1. Table 1 Daa Descripion Series Excluded iems Sample period USA Food and energy 1987:1 2000:2 Germany Seasonal food and energy 1992:1 2000:4 France Food, energy and public uiliies 1987:1 1997:12 Ialy Fresh food and energy 1987:1 2000:5 Spain Unprocessed food and energy 1987:1 2000:2 Porugal Unprocessed food and energy 1987:1 1999:12 Preliminary esing showed ha he null of a uni roo in CPI inflaion could no be rejeced (by an ADF es) for any counry. Therefore, we are able o use he esing 7
procedure described in secion 2. Table 2 conains he main resuls of hese ess. 6 The saionariy ess on CPI inflaion are shown in column 1. Columns 2 and 3 repor he ess on condiion i). Column 4 refers o condiion ii), while columns 5 and 6 presen he resuls on boh versions of condiion iii). Finally, column 7 shows he conclusions drawn from he ess. Since he ess for condiions ii) and iii) are condiional on he verificaion of i), we did no es condiions ii) and iii) for he series ha failed he firs crierion (i.e. Ialy and Porugal). For Porugal, we conclude ha he series are no coinegraed wih uniary coefficien, so i may be possible for CPI inflaion o diverge from underlying inflaion for subsanial periods of ime. For Ialy, in spie of evidence showing a srong coinegraion relaionship, he es on column 3 repors a sysemaic bias of he underlying inflaion measure, which naurally lessens is ineres as a core inflaion indicaor. For all oher counries, condiion i) is verified; however, condiion ii) does no hold. Recall ha his condiion required he underlying inflaion indicaor o arac CPI inflaion. The fac ha his does no occur means ha knowing CPI inflaion is, say, below underlying inflaion in a given period, does no convey any informaion on he fuure pah of CPI inflaion. Given he resuls so far, one would of course expec condiion iii) no o hold also. * Since π and π are coinegraed and here is no ECM represenaion for π, he Granger Represenaion Theorem requires ha an ECM represenaion exiss for π *. This is in fac verified (in column 5) for all counries bu Spain. However, even in his case, he null would no be rejeced a he 10% level of significance. This means ha underlying inflaion is no a leading indicaor for inflaion and, moreover, ha i is he inflaion rae iself ha appears o lead he so-called indicaor of underlying inflaion, as he analysis in secion 3 suggess. 6 In he esing procedure, we used significance levels of 10% for he ADF ess, and 5% for he and F ess. We se he orders of he lag polynomials in he ADF and he ECM regressions such ha he residuals were no auocorrelaed. Alhough he ECM models presened in secion 2 (equaions (2) and (4)) do no include a consan erm, we also esed condiions ii) and iii) wih a nonzero consan. We show he resuls for his case only when he conclusions differ from he main es. 8
5. Conclusions This paper shows ha he so-called measure of underlying inflaion, which is used by several Cenral Banks as a measure of rend inflaion, does no mee he necessary condiions se ou in Marques e al. (2000). These condiions posi ha any core inflaion measure should be coinegraed wih inflaion (wih a uni coefficien) and ac as an aracor for inflaion, i.e., o Granger cause inflaion bu no o be Granger caused by i. Therefore, i appears o be inappropriae o use his indicaor o analyse he curren saus of inflaion or o make inference abou is likely fuure pah. References Álvarez, L.J., Maea, M.L. (1999) Underlying Inflaion Measures in Spain, Banco de España, Working Paper 9911. Blinder, A.S. (1982) The Anaomy of Double-Digi Inflaion in he 1970s, in Inflaion: Causes and Effecs, R.E. Hall (ed.), Universiy of Chicago Press for NBER, pp. 261-282. Bryan, M.F., Cecchei, S.G. (1994) Measuring Core Inflaion, in Moneary Policy, N.G. Mankiw (ed.), Universiy of Chicago Press for NBER, pp. 195-215. Bryan, M.F., Cecchei, S.G., Wiggins II, R.L. (1997) Efficien Inflaion Esimaion, NBER, Working Paper 6183, Sepember. Freeman, D.G. (1998) Do core inflaion measures help forecas inflaion?, Economics Leers 58, pp. 143-147. Marques, C.R., Neves, P.D., Sarmeno, L.M. (2000) Evaluaing Core Inflaion Indicaors, Banco de Porugal, Working Paper 3/00, April. Nascimeno, T. (1990), Indicadores de Inflação, Boleim Trimesral do Banco de Porugal, vol. 12, nº 4, December. 9
Saionariy of π Table 2 Evaluaing he underlying inflaion indicaor (a) Saionariy of (π - π * ) α = 0 given β = 1 γ = 0 λ = 0 Srong exogeneiy λ = θ 1 = = θ r = 0 Conclusion Column (1) (2) (3) (4) (5) (6) (7) USA ADF(1) = -2.00 ADF(1) = -2.81 * P = 0.163 P = 0.250 (b) P = 0.024 (c) P = 0.045 Fails condiions ii) and iii) Germany ADF(1) = -2.46 ADF(1) = -3.12 ** P = 0.277 P = 0.378 P = 0.001 P = 0.018 Fails condiions ii) and iii) France ADF(1) = 0.52 ADF(1) = -2.73 * P = 0.079 P = 0.824 (d) P = 0.038 P = 0.520 Fails condiions ii) and iii) Ialy ADF(1) = -0.28 ADF(1) = -3.62 *** P = 0.033 --- --- --- Fails condiion i) Spain ADF(1) = -0.53 ADF(1) = -2.93 ** P = 0.099 P = 0.287 P = 0.093 P = 0.258 Fails condiion ii) Porugal ADF(1) = -0.39 ADF(1) = -1.91 --- --- --- --- Fails condiion i) (a) The significance level of he ADF ess is marked: * for 10%, ** for 5%, *** for 1%. In all oher ess, P sands for he corresponding p-value. (b) We have λ=0 in he model wih consan erm, wih P=0.064 (c) We have λ=θ 1 =θ 2 = =θ r =0 in he model wih consan erm, wih P=0.071 (d) We have λ=0 in he model wih consan erm, wih P=0.131
WORKING PAPERS 1/90 PRODUTO POTENCIAL, DESEMPREGO E INFLAÇÃO EM PORTUGAL Um esudo para o período 1974-1989 Carlos Robalo Marques 2/90 INFLAÇÃO EM PORTUGAL Um esudo economérico para o período 1965-1989, com projecções para 1990 e 1991 Carlos Robalo Marques 3/92 THE EFFECTS OF LIQUIDITY CONSTRAINTS ON CONSUMPTION BEHAVIOUR The Poruguese Experience Sílvia Luz 4/92 LOW FREQUENCY FILTERING AND REAL BUSINESS CYCLES Rober G. King, Sérgio T. Rebelo 5/92 GROWTH IN OPEN ECONOMIES Sérgio Rebelo 6/92 DYNAMIC OPTIMAL TAXATION IN SMALL OPEN ECONOMIES Isabel H. Correia 7/92 EXTERNAL DEBT AND ECONOMIC GROWTH Isabel H. Correia 8/92 BUSINESS CYCLES FROM 1850 TO 1950: NEW FACTS ABOUT OLD DATA Isabel H. Correia, João L. Neves, Sérgio Rebelo 9/92 LABOUR HOARDING AND THE BUSINESS CYCLE Craig Burnside, Marin Eichenbaum, Sérgio Rebelo 10/92 ANALYSIS OF FOREIGN DIRECT INVESTMENT FLOWS IN PORTUGAL USING PANEL DATA Luísa Farinha 11/92 INFLATION IN FIXED EXCHANGE RATE REGIMES: THE RECENT PORTUGUESE EXPERIENCE Sérgio Rebelo 12/92 TERM STRUCTURE OF INTEREST RATES IN PORTUGAL Armindo Escalda 13/92 AUCTIONING INCENTIVE CONTRACTS: THE COMMON COST CASE Fernando Branco 14/92 INDEXED DEBT AND PRODUCTION EFFICIENCY Anónio S. Mello, John Parsons
15/92 TESTING FOR MEAN AND VARIANCE BREAKS WITH DEPENDENT DATA José A. F. Machado 16/92 COINTEGRATION AND DYNAMIC SPECIFICATION Carlos Robalo Marques 17/92 FIRM GROWTH DURING INFANCY José Maa 18/92 THE DISTRIBUTION OF HOUSEHOLD INCOME AND EXPENDITURE IN PORTUGAL: 1980 and 1990 Miguel Gouveia, José Tavares 19/92 THE DESIGN OF MULTIDIMENSIONAL AUCTIONS Fernando Branco 20/92 MARGINAL INCOME TAX RATES AND ECONOMIC GROWTH IN DEVELOPING COUNTRIES Sérgio Rebelo, William Easerly 21/92 THE EFFECT OF DEMAND AND TECHNOLOGICAL CONDITIONS ON THE LIFE EXPECTANCY OF NEW FIRMS José Maa, Pedro Porugal 22/92 TRANSITIONAL DYNAMICS AND ECONOMIC GROWTH IN THE NEOCLASSICAL MODEL Rober G. King, Sérgio Rebelo 23/92 AN INTEGRATED MODEL OF MULTINATIONAL FLEXIBILITY AND FINANCIAL HEDGING Anónio S. Mello, Alexander J. Trianis 24/92 CHOOSING AN AGGREGATE FOR MONETARY POLICY: A COINTEGRATION APPROACH Carlos Robalo Marques, Margarida Caalão Lopes 25/92 INVESTMENT: CREDIT CONSTRAINTS, REGULATED INTEREST RATES AND EXPECTATIONS OF FINANCIAL LIBERALIZATION THE PORTUGUESE EXPERIENCE Koleman Srumpf 1/93 SUNK COSTS AND THE DYNAMICS OF ENTRY José Maa 2/93 POLICY, TECHNOLOGY ADOPTION AND GROWTH William Easerly, Rober King, Ross Levine, Sérgio Rebelo 3/93 OPTIMAL AUCTIONS OF A DIVISIBLE GOOD Fernando Branco 4/93 EXCHANGE RATE EXPECTATIONS IN INTERNATIONAL OLIGOLOPY Luís Cabral, Anónio S. Mello 5/93 A MODEL OF BRANCHING WITH AN APPLICATION TO PORTUGUESE BANKING Luís Cabral, W. Rober Majure
6/93 HOW DOES NEW FIRM SURVIVAL VARY ACROSS INDUSTRIES AND TIME? José Maa, Pedro Porugal 7/93 DO NOISE TRADERS CREATE THEIR OWN SPACE? Ravi Bhushan, David P. Brown, Anónio S. Mello 8/93 MARKET POWER MEASUREMENT AN APPLICATION TO THE PORTUGUESE CREDIT MARKET Margarida Caalão Lopes 9/93 CURRENCY SUBSTITUTABILITY AS A SOURCE OF INFLATIONARY DISCIPLINE Pedro Teles 10/93 BUDGET IMPLICATIONS OF MONETARY COORDINATION IN THE EUROPEAN COMMUNITY Pedro Teles 11/93 THE DETERMINANTS OF FIRM START-UP SIZE José Maa 12/93 FIRM START-UP SIZE: A CONDITIONAL QUANTILE APPROACH José Maa, José A. F. Machado 13/93 FISCAL POLICY AND ECONOMIC GROWTH: AN EMPIRICAL INVESTIGATION William Easerly, Sérgio Rebelo 14/93 BETA ESTIMATION IN THE PORTUGUESE THIN STOCK MARKET Armindo Escalda 15/93 SHOULD CAPITAL INCOME BE TAXED IN THE STEADY STATE? Isabel H. Correia 16/93 BUSINESS CYCLES IN A SMALL OPEN ECONOMY Isabel H. Correia, João C. Neves, Sérgio Rebelo 17/93 OPTIMAL TAXATION AND CAPITAL MOBILITY Isabel H. Correia 18/93 A COMPOSITE COINCIDENT INDICATOR FOR THE PORTUGUESE ECONOMY Francisco Craveiro Dias 19/93 PORTUGUESE PRICES BEFORE 1947: INCONSISTENCY BETWEEN THE OBSERVED COST OF LIVING INDEX AND THE GDP PRICE ESTIMATION OF NUNES, MATA AND VALÉRIO (1989) Paulo Soares Eseves 20/93 EVOLUTION OF PORTUGUESE EXPORT MARKET SHARES (1981-91) Crisina Maneu, Ildebera Abreu 1/94 PROCUREMENT FAVORITISM AND TECHNOLOGY ADOPTION Fernando Branco
2/94 WAGE RIGIDITY AND JOB MISMATCH IN EUROPE: SOME EVIDENCE Sílvia Luz, Maximiano Pinheiro 3/94 A CORRECTION OF THE CURRENT CONSUMPTION INDICATOR AN APPLICATION OF THE INTERVENTION ANALYSIS APPROACH Renaa Mesquia 4/94 PORTUGUESE GDP AND ITS DEFLATOR BEFORE 1947: A REVISION OF THE DATA PRODUCED BY NUNES, MATA AND VALÉRIO (1989) Carlos Robalo Marques, Paulo Soares Eseves 5/94 EXCHANGE RATE RISK IN THE EMS AFTER THE WIDENING OF THE BANDS IN AUGUST 1993 Joaquim Pires Pina 6/94 FINANCIAL CONSTRAINTS AND FIRM POST-ENTRY PERFORMANCE Paulo Brio, Anónio S. Mello 7/94 STRUCTURAL VAR ESTIMATION WITH EXOGENEITY RESTRICTIONS Francisco C. Dias, José A. F. Machado, Maximiano R. Pinheiro 8/94 TREASURY BILL AUCTIONS WITH UNINFORMED BIDDERS Fernando Branco 9/94 AUCTIONS OF SHARES WITH A SECONDARY MARKET AND TENDER OFFERS Anónio S. Mello, John E. Parsons 10/94 MONEY AS AN INTERMEDIATE GOOD AND THE WELFARE COST OF THE INFLATION TAX Isabel Correia, Pedro Teles 11/94 THE STABILITY OF PORTUGUESE RISK MEASURES Armindo Escalda 1/95 THE SURVIVAL OF NEW PLANTS: START-UP CONDITIONS AND POST-ENTRY EVOLUTION José Maa, Pedro Porugal, Paulo Guimarães 2/95 MULTI-OBJECT AUCTIONS: ON THE USE OF COMBINATIONAL BIDS Fernando Branco 3/95 AN INDEX OF LEADING INDICATORS FOR THE PORTUGUESE ECONOMY Francisco Ferreira Gomes 4/95 IS THE FRIEDMAN RULE OPTIMAL WHEN MONEY IS AN INTERMEDIATE GOOD? Isabel Correia, Pedro Teles 5/95 HOW DO NEW FIRM STARTS VARY ACROSS INDUSTRIES AND OVER TIME? José Maa 6/95 PROCUREMENT FAVORITISM IN HIGH TECHNOLOGY Fernando Branco
7/95 MARKETS, ENTREPRENEURS AND THE SIZE OF NEW FIRMS José Maa 1/96 CONVERGENCE ACROSS EU COUNTRIES: INFLATION AND SAVINGS RATES ON PHYSICAL AND HUMAN CAPITAL Paulo Soares Eseves 2/96 THE OPTIMAL INFLATION TAX Isabel Correia, Pedro Teles 3/96 FISCAL RULES OF INCOME TRANSFORMATION Isabel H. Correia 4/96 ON THE EFFICIENCY AND EQUITY TRADE-OFF Isabel H. Correia 5/96 DISTRIBUTIONAL EFFECTS OF THE ELIMINATION OF CAPITAL TAXATION Isabel H. Correia 6/96 LOCAL DYNAMICS FOR SPHERICAL OPTIMAL CONTROL PROBLEMS Paulo Brio 7/96 A MONEY DEMAND FUNCTION FOR PORTUGAL João Sousa 8/96 COMPARATIVE EXPORT BEHAVIOUR OF FOREIGN AND DOMESTIC FIRMS IN PORTUGAL Sónia Cabral 9/96 PUBLIC CAPITAL ACCUMULATION AND PRIVATE SECTOR PERFORMANCE IN THE US Alfredo Marvão Pereira, Rafael Flores de Fruos 10/96 IMPORTED CAPITAL AND DOMESTIC GROWTH: A COMPARISON BETWEEN EAST ASIA AND LATIN AMERICA Ling-ling Huang, Alfredo Marvão Pereira 11/96 ON THE EFFECTS OF PUBLIC AND PRIVATE R&D Rober B. Archibald, Alfredo Marvão Pereira 12/96 EXPORT GROWTH AND DOMESTIC PERFORMANCE Alfredo Marvão Pereira, Zhenhui Xu 13/96 INFRASTRUCTURES AND PRIVATE SECTOR PERFORMANCE IN SPAIN Alfredo Marvão Pereira, Oriol Roca Sagales 14/96 PUBLIC INVESTMENT AND PRIVATE SECTOR PERFORMANCE: INTERNATIONAL EVIDENCE Alfredo Marvão Pereira, rman Morin 15/96 COMPETITION POLICY IN PORTUGAL Pedro P. Barros, José Maa
16/96 THE IMPACT OF FOREIGN DIRECT INVESTMENT IN THE PORTUGUESE ECONOMY Luísa Farinha, José Maa 17/96 THE TERM STRUCTURE OF INTEREST RATES: A COMPARISON OF ALTERNATIVE ESTIMATION METHODS WITH AN APPLICATION TO PORTUGAL Nuno Cassola, Jorge Barros Luís 18/96 SHORT-AND LONG-TERM JOBLESSNESS: A SEMI-PARAMETRIC MODEL WITH TIME -VARYING EFFECTS Pedro Porugal, John T. Addison 19/96 SOME SPECIFICATION ISSUES IN UNEMPLOYMENT DURATION ANALYSIS Pedro Porugal, John T. Addison 20/96 SEQUENTIAL AUCTIONS WITH SYNERGIES: AN EXAMPLE Fernando Branco 21/96 HEDGING WINNER'S CURSE WITH MULTIPLE BIDS: EVIDENCE FROM THE PORTUGUESE TREASURY BILL AUCTION Michael B. Gordy 22/96 THE BRICKS OF AN EMPIRE 1415-1999: 585 YEARS OF PORTUGUESE EMIGRATION Sanley L. Engerman, João César das Neves 1/97 LOCAL DYNAMICS FOR PLANAR OPTIMAL CONTROL PROBLEMS: A COMPLETE CHARACTERIZATION Paulo Brio 2/97 INTERNATIONAL PORTFOLIO CHOICE Bernardino Adão, Nuno Ribeiro 3/97 UNEMPLOYMENT INSURANCE AND JOBLESSNESS: A DISCRETE DURATION MODEL WITH MULTIPLE DESTINATIONS Pedro Porugal, John T. Addison 4/97 THE TREASURY BILL MARKET IN PORTUGAL: INSTITUTIONAL ISSUES AND PROFIT MARGINS OF FINANCIAL INSTITUTIONS Bernardino Adão, Jorge Barros Luís 5/97 ECONOMETRIC MODELLING OF THE SHORT-TERM INTEREST RATE: AN APPLICATION TO PORTUGAL Nuno Cassola, João Nicolau, João Sousa 6/97 ESTIMATION OF THE NAIRU FOR THE PORTUGUESE ECONOMY Carlos Robalo Marques, Susana Boas 7/97 EXTRACTION OF INTEREST RATE DIFFERENTIALS IMPLICIT IN OPTIONS: THE CASE OF SPAIN AND ITALY IN THE EUROPEAN MONETARY UNION Bernardino Adão, Jorge Barros Luís
1/98 A COMPARATIVE STUDY OF THE PORTUGUESE AND SPANISH LABOUR MARKETS Olympia Bover, Pilar García-Perea, Pedro Porugal 2/98 EARNING FUNCTIONS IN PORTUGAL 1982-1994: EVIDENCE FROM QUANTILE REGRESSIONS José A. F. Machado, José Maa 3/98 WHAT HIDES BEHIND AN UNEMPLOYMENT RATE: COMPARING PORTUGUESE AND US UNEMPLOYMENT Olivier Blanchard, Pedro Porugal 4/98 UNEMPLOYMENT INSURANCE AND JOBLESSNESS IN PORTUGAL Pedro Porugal, John T. Addison 5/98 EMU, EXCHANGE RATE VOLATILITY AND BID-ASK SPREADS Nuno Cassola, Carlos Sanos 6/98 CONSUMER EXPENDITURE AND COINTEGRATION Carlos Robalo Marques, Pedro Duare Neves 7/98 ON THE TIME-VARYING EFFECTS OF UNEMPLOYMENT INSURANCE ON JOBLESSNESS John T. Addison, Pedro Porugal 8/98 JOB SEARCH METHODS AND OUTCOMES John T. Addison, Pedro Porugal 1/99 PRICE STABILITY AND INTERMEDIATE TARGETS FOR MONETARY POLICY Víor Gaspar, Ildebera Abreu 2/99 THE OPTIMAL MIX OF TAXES ON MONEY, CONSUMPTION AND INCOME Fiorella De Fiore, Pedro Teles 3/99 OPTIMAL EXECUTIVE COMPENSATION: BONUS, GOLDEN PARACHUTES, STOCK OWNERSHIP AND STOCK OPTIONS Chongwoo Choe 4/99 SIMULATED LIKELIHOOD ESTIMATION OF NON-LINEAR DIFFUSION PROCESSES THROUGH NON-PARAMETRIC PROCEDURE WITH AN APPLICATION TO THE PORTUGUESE INTEREST RATE João Nicolau 5/99 IBERIAN FINANCIAL INTEGRATION Bernardino Adão 6/99 CLOSURE AND DIVESTITURE BY FOREIGN ENTRANTS: THE IMPACT OF ENTRY AND POST-ENTRY STRATEGIES José Maa, Pedro Porugal 1/00 UNEMPLOYMENT DURATION: COMPETING AND DEFECTIVE RISKS John T. Addison, Pedro Porugal
2/00 THE ESTIMATION OF RISK PREMIUM IMPLICIT IN OIL PRICES Jorge Barros Luís 3/00 EVALUATING CORE INFLATION INDICATORS Carlos Robalo Marques, Pedro Duare Neves, Luís Morais Sarmeno 4/00 LABOR MARKETS AND KALEIDOSCOPIC COMPARATIVE ADVANTAGE Daniel A. Traça 5/00 WHY SHOULD CENTRAL BANKS AVOID THE USE OF THE UNDERLYING INFLATION INDICATOR? Carlos Robalo Marques, Pedro Duare Neves, Afonso Gonçalves da Silva