EUROPEAN CENTRAL BANK...The WGF is composed of ECB and NCB macroeconomic and econometric experts. B An outline of Eurosystem macroeconomic projection exercises There are three main

A GUIDE TOEUROSYSTEM STAFFMACROECONOMIC

PROJECTION EXERCISES

June 2001

E U R O P E A N C E N T R A L B A N K

EC

B

EZ

B

EK

T

BC

E

EK

P

E U R O P E A N C E N T R A L B A N K

A GUIDE TOEUROSYSTEM STAFFMACROECONOMIC

PROJECTION EXERCISES

June 2001

ECB • A Gu i de t o Eu ro s y s t em s ta f f mac roeconom i c p ro j e c t i on exe r c i s e s • J une 20012

© European Central Bank, 2001

Address Kaiserstrasse 29

D-60311 Frankfurt am Main

Germany

Postal address Postfach 16 03 19

D-60066 Frankfurt am Main

Germany

Telephone +49 69 1344 0

Internet http://www.ecb.int

Fax +49 69 1344 6000

Telex 411 144 ecb d

All rights reserved.

Reproduction for educational and non-commercial purposes is permitted provided that the source is acknowledged.

ISBN 92-9181-177-7

3ECB • A Gu i de t o Eu ro s y s t em s ta f f mac roeconom i c p ro j e c t i on exe r c i s e s • J une 2001

Introduction 4

Chapter 1

An overview of Eurosystem staff macroeconomic projection exercises 5

A The framework for interaction between ECB and NCB staff 6

B An outline of Eurosystem macroeconomic projection exercises 6

1 Setting the underlying assumptions 7

2 Deriving the projections 7

3 Preparing the report 9

Chapter 2

The techniques and projection tools used in Eurosystem staff macroeconomicprojection exercises 11

A Techniques used for producing Eurosystem macroeconomic projections 12

B Eurosystem projection tools 14

1 Aggregating the projections 14

2 Ensuring trade consistency 14

3 Reflecting the uncertainty associated with economic projections 15

Conclusions and future developments 16

Bibliography 17

Contents


Introduction

1 See “Staff economic projections for the euro area”, in theDecember 2000 issue of the ECB Monthly Bulletin.

2 For a discussion of the role of projections within the monetarypolicy strategy of the ECB, see the ECB’s “Annual Report 2000”,“The two pillars of the ECB’s monetary policy strategy”, in theNovember 2000 issue of the ECB Monthly Bulletin and “Thestability-oriented monetary policy strategy of the Eurosystem”, inthe January 1999 issue of the ECB Monthly Bulletin.

The Governing Council of the ECB hasdecided to publish Eurosystem staffmacroeconomic projections for the euro areaon a biannual basis from December 2000.1

These projections will be published in theJune and December issues of the ECBMonthly Bulletin. The published figuresinclude projections for inflation in terms ofthe Harmonised Index of Consumer Prices(HICP), the growth of real GDP and its mainexpenditure components over a two-yearhorizon. The projections are accompanied bya description of their main features. In orderto reflect the degree of uncertainty attachedto such exercises, the Governing Councildecided to publish the projections in the formof ranges.

The present guide to Eurosystem staffmacroeconomic projection exercises (MPEs),prepared jointly by experts from the euroarea NCBs and from the ECB, describes theway in which the projections are produced.The projections are a convenient analyticaltool for helping bring together in a systematicmanner a range of information on currentand future economic developments.Conditioned on a set of assumptions, theycombine the use of conventional models with

economic experts’ knowledge. The outcomesof the MPEs are presented to the GoverningCouncil, to be used as input into its monetarypolicy deliberations.2 The Governing Councilitself has to make an overall assessment bothof the economic situation and of the risks toprice stability, using all the informationavailable, including, in particular, theinformation derived from the first pillar ofthe strategy, but also information other thanthe Eurosystem’s projections under thesecond pillar.

The report is organised in two chapters.Chapter 1 gives an overview of the maincharacteristics of the Eurosystem staff MPEs.Chapter 2 describes the tools used to derivethe projections, including the techniques usedto ensure that the euro area projections areobtained in a way fully consistent with theindividual country assessments.

Chapter I

An overview of the Eurosystemstaff macroeconomic projection

exercises


Within the second pillar of the ECB’smonetary policy strategy, the Eurosystemstaff macroeconomic projections play animportant role as a tool for aggregatingand organising existing information oncurrent and future economic developments.Conditioned on a set of assumptions,they provide projections for a range ofmacroeconomic variables, combining theresults of conventional models with economicexperts’ knowledge. This chapter deals withthe main characteristics of the Eurosystemprojection exercises, stressing how theprojections reflect the interaction betweenECB and NCB staff and the main stepsfollowed in their elaboration.

A unique element of the Eurosystem staffmacroeconomic projection exercises is theircombination of both national and euro area-wide perspectives. The euro area consists ofa set of interacting and integrating economies.

However, monetary policy decisions arebased upon an assessment of euro area-wideconditions. The projection procedurestherefore need to take account of bothnational and area-wide perspectives.

Individual country-level assessments areprepared which make full use of the detailedknowledge and experience of countryexperts. These projections can take accountof the rich, but somewhat diverse, set ofnational data sources and incorporate thedetails of individual countries’ institutionalframeworks. The procedures of theprojection exercises are designed to integratesuch country assessments within an overalleuro area framework, ensuring that theindividual country trade flows are made fullyconsistent with each other and that the finalprojections represent the consensus ofEurosystem staff opinion through a technicalpeer-review process.

A The framework for interaction between ECB and NCB staff

Macroeconomic projection exercises (MPEs)involve close interaction between ECB andNCB staff to ensure that the euro areaprojections draw on all expertise availableand reflect a consensus among Eurosystemstaff.

The MPEs are carried out under theresponsibility and guidance of the MonetaryPolicy Committee (MPC). The MPC iscomposed of senior staff representatives ofthe ECB and the NCBs. The MPC providesbroad guidance for the production of the

projection figures and is responsible for thefinal draft of the report on the Eurosystemstaff macroeconomic projection exercise.

The Working Group on Forecasting (WGF),which is one of the three working groupsreporting to the MPC,3 is responsible forproducing the detailed figures for themacroeconomic projections and forproducing an initial version of the report.The WGF is composed of ECB and NCBmacroeconomic and econometric experts.

B An outline of Eurosystem macroeconomic projection exercises

There are three main steps in the productionof the MPEs. First is the setting of assumptionsunderlying the exercise; second, there is thederivation of and agreement on a set ofmacroeconomic projection figures, following aprocess of overview at area-wide level,projection consistency checks, and peerreview of the individual country projection

figures; finally, the third step is the preparationof the report for the Governing Council andof the publication of the projections in theMonthly Bulletin.

3 The other two are the Working Group on Econometric Modellingand the Working Group on Public Finance.


1 Setting the underlying assumptions

At the beginning of the exercise, a set ofprovisional assumptions is agreed, coveringinterest rates, exchange rates, the internationalenvironment and fiscal variables. Theseassumptions are reviewed and may bechanged in the course of the exercise.

The projections are based on the technicalassumption that three-month interest rates inthe euro area remain constant over thehorizon of the projection. The constant levelof the three-month interest rate correspondsto the prevailing level. This assumption ismade in order to illustrate the possibleconsequences of leaving monetary policyunchanged over the projection horizon.However, for this reason, the staff economicprojections will not necessarily be the bestunconditional predictor of future outcomes,particularly over longer horizons, sincemonetary policy will always act to containany threats to price stability.

For long-term interest rates, it is assumed thatten-year interest rates evolve in accordancewith the prevailing market expectations.These market projections for long-terminterest rates are obtained by pricing notionalten-year euro area coupon bonds. Long-terminterest rates may, however, be adjusted inthe context of the projection process if thisis needed to make them consistent with therest of the projection.

Exchange rate assumptions are purely technicaland based on an average of recent rates (thusthey do not in any way constitute a forecastfor the future evolution, or an assessment ofthe appropriate level, of the euro exchangerate). Oil and non-oil commodity priceassumptions are based on recently observedfutures market prices.

The assumptions for the internationalenvironment are agreed by the WGF on thebasis of a broad consensus. An initialassessment by ECB staff is discussed togetherwith contributions from the NCBs, which alsoreflect specialist knowledge concerning

external economies and regions. The NCBsof those European Union countries notparticipating in the euro area also contributeto this discussion as part of the co-operationwithin the European System of Central Banks,although they do not take part in thesubsequent macroeconomic projectionexercise.

For fiscal variables, assumptions for thecurrent year incorporate the most recentinformation available concerning budgetarydevelopments in the individual membercountries of the euro area, while theassumptions for the later years are based ona “most likely policy” scenario.

Whenever, during the course of the exercise,the initial assumptions become clearlyinconsistent with recently observedoutcomes, they are updated. A procedurefor quickly assessing the impact of theupdated assumptions on the projectionfigures is available, based on a set ofprojection update elasticities.

2 Deriving the projections

Following the agreement on the initialassumptions, NCB and ECB staff separatelyprepare initial projection figures. These aredocumented in accompanying reportsfollowing a standardised format. Thesereports focus on the key features of thecountry projections, describing theirunderlying economic rationale and thespecific assumptions made about theinstitutional setting at the national level (forinstance, labour market developments andwage prospects). Each NCB prepares aninitial projection for its own country, whereasECB staff prepare a set of projections for theindividual countries as well as for the euroarea as a whole, the latter being consistentwith the aggregation of the ECB staffindividual country projections. The NCBcountry projections are aggregated by ECBstaff.


In addition to the preparation of theprojections themselves, ECB and NCB staffalso jointly produce a number of backgroundanalyses, considering in detail specificimportant technical or structural issues.

The initial NCB country projections and theiraggregation are discussed within the WGF,along with the initial country and area-wideprojections of ECB staff. The discussion ofand agreement on the final projection figuresare the result of the following three elements:first, a euro area overview is taken of thefigures and the implicit judgements theyembody; second, a number of technicalexercises are undertaken to ensure themutual consistency of the projection figures forindividual countries; and, third, a process ofpeer review is applied to come to a consensuson the final projection figures for individualcountries. The projections for the euro areaare then obtained by aggregating the countryfigures.

The euro area overview of the figures is basedupon the comparison of the initial ECB euroarea projection with the aggregation of theinitial NCB country projections, takingaccount of the background analyses that havebeen prepared. This comparison includes,inter alia, an analysis using the area-wideeconometric model4 of the differencesbetween the initial ECB staff and theaggregated NCB staff projections. Thisexercise identifies a set of area-wide issues,which helps to provide a framework for thesubsequent discussion of individual countryprojections.

Second, a number of consistency exercises arecarried out on the figures. A trade consistencyexercise is performed to ensure theconsistency between the projected tradevolumes and prices of the individual countryprojections. The assessment may also raiseissues concerning individual national tradevolume or price projections. The tradeconsistency exercise is repeated in the laterstages of the projection process. A financialconsistency exercise is also carried out toassess the implications of the MPE

assumptions and projections in the contextof the consolidated balance sheet of theMonetary Financial Institutions in the euroarea in the years ahead. This exercise willgain sophistication as financial statisticsbecome more fully available.

Finally, a detailed discussion of individualcountry projections is carried out within theWGF. It is based on a peer review of theinitial NCB country projection and the initialcountry projection prepared by ECB staff,and takes into account the outcome of theconsistency exercises. This discussion usuallyleads to further analyses aimed at clearingany remaining differences in the projectionfigures. In this context, detailed monthlyprojections of national HICPs are alsoconsidered along with the mainmacroeconomic projection figures. Thisensures that the model and expert-basedmacroeconomic discussions duly take intoaccount short-term sectoral and fiscal factorsrelevant for the price outlook.

On the basis of technical considerations,taking into account the outcome of theconsistency exercises and possibly uponconsultation with the MPC, a list ofadjustments needed to reconcile the viewson country projections and, implicitly, on theaggregated euro area projection is agreed.Usually one or two iterations are requiredto obtain a set of final country projections.The final euro area macroeconomic projectionsare the aggregation of the agreed revised countryprojections.

As mentioned earlier, assumptions may bereviewed in the course of the exercise. Whenthey are modified at an advanced stage of theprojection process, a set of projection updateelasticities, computed in a way that takes intoaccount trade volume and price consistencyrequirements, can be used for the update ofthe projection figures.

4 Fagan, G., Henry, J. and Mestre, R. (2001), “An Area-WideModel (AWM) for the euro area”, ECB Working Paper, No. 42,European Central Bank, Frankfurt am Main.


3 Preparing the report

The report for the Governing Council on theEurosystem staff macroeconomic projectionexercise is prepared by the MPC assisted bythe WGF. This report contains a statementof the assumptions made regarding interestrates, exchange rates, the internationalenvironment and fiscal policy. It then reviewsthe projection results for the euro area as awhole, looking at real GDP growth andemployment projections, price and costprospects; the prospects for individualeconomies are also briefly reviewed. Finally,the report is supplemented by a detailedassessment of short-term pricedevelopments.

The report is then submitted to theGoverning Council, via the ECB’s ExecutiveBoard.

A shorter description of the projectionsbased on the original report to the Councilis then published in the June and the Decemberissues of the ECB Monthly Bulletin. It is alsomade available on the ECB’s website. Thepublished figures include projection rangesfor HICP inflation, the growth of real GDPand its main expenditure components over atwo-year horizon. The projections areaccompanied by a description of their mainfeatures.


Chapter 2

The techniques and projection tools

used in Eurosystem staff

macroeconomic projection exercises


The scope of Eurosystem staff macroeconomicprojection exercises includes the HICP, realGDP, a detailed breakdown of expenditurecounterparts, as well as a number of othermacroeconomic variables, including the GDPdeflator. The projections are presented for ahorizon of up to two years ahead.

Conditioned on a set of assumptions, theprojections combine the use of econometricmodels with economic expert knowledge.

A variety of structural macroeconometricmodels, as well as reduced-form equationsystems and time series models, areemployed to generate and provide analyticalsupport for the projections. The models usedinclude the area-wide model of the ECB andcountry models developed by the ECB andthe NCBs, including the multi-country modelof the European System of Central Banks(ESCB). Such structural models provide adetailed structure for the projections,ensuring that they are internally consistentand that the relationships between variablesare in line with economic theory andeconometric evidence.

Nevertheless, structural econometric modelsdo not always fully track the most recentdevelopments, reflect sector-specificbehaviour or capture exceptional factors orstructural changes. Therefore, the euro areaand country assessments of ECB and NCBstaff also make use of leading indicators andjudgemental methods. Knowledge of theinstitutional context and other specificinformation are thus also integrated into theprojections. In addition, a number ofconsistency tools are used to ensure themutual consistency of the projection figuresfor individual countries. As a result, theprojections incorporate both structuraleconometric model-based and judgementalmethods. In particular, consensus buildingamong experts – modellers and non-modellers – is facilitated by quantifying andanalysing the extent to which themacroeconomic projections may depart froma prediction implied by past behaviour, asderived from a structural macroeconometricmodel.

A Techniques used for producing Eurosystem macroeconomicprojections

The approaches followed to produce theEurosystem macroeconomic projections area combination of conjunctural analysis,econometric model projections andjudgmental assessments on the basis of expertknowledge. An overview of these techniquesis provided in this section. Detailedreferences are given in the bibliography.

Good conjunctural analysis of growth andinflation is one of the foundations ofmacroeconomic projection exercises. For theshort-term analysis of real GDP and its mainexpenditure counterparts, a number ofstandard coincident and leading indicators areused. Many NCBs have developed and useeconometric techniques to quantify theserelationships. Coincident indicators andnational accounts bridge models are widely

used (see Bravo and Sanchez (2001), Dias(1993), Irac and Sedillot (2001) and Parigiand Schlitzer (1995)). Leading indicatormodels have been developed by some NCBs(e.g. Rooij and Stokman (2000) and Altissimoet al. (1999)). Some NCBs have developedindicators for cyclical turning-points (e.g.Vanhaelen et al. (2000)).

A particular feature of Eurosystemmacroeconomic projection exercises is theconcentration on the detailed analysis ofshort-term price developments. For thispurpose, ECB and NCB staff use a number oftools ranging from relatively simple indicatorsto sophisticated econometric models. Suchanalysis is generally carried out at adisaggregated level, together with anaggregated projection to check consistency.


Vector auto-regressive models, including insome cases Bayesian priors, are used by anumber of NCBs for projecting the overallprice index or for particularly importantcomponents (see Ballabriga et al. (2000),Meyler et al. (1998), Cristadoro and Sabatinni(1999) and Wit (1998)).5 These are oftencombined with ARIMA time series, transferfunction models for particular pricecomponents, and provide an initial baselineprojection for judgemental input (e.g. Alvarez(1999)). Structural Phillips curve equations,relating inflation to a measure of excessdemand, are also used by some NCBs (e.g.Jondeau et al. (1999) and Zonzilos (2000)).Moreover, indicator models are widely usedby NCBs. Some NCBs have developedcomposite leading indicator models, where aselected set of leading indicators is weightedtogether to form a projection for a particularvariable, with the weights being determinedstatistically (for a published example seeGibson and Lazaretou (2001)).

In the context of the MPEs, the ECB staff anda majority of NCBs combine medium tolarge-scale structural econometric modelswith various sources of additionalinformation. With few exceptions, structuralmacroeconometric models are used torepresent the central aspects of theprojection in a coherent framework,incorporating additional information comingfrom different sources. In addition to themacroeconometric models used by NCBs,two models are also used to assess theoverall consistency of the projections: thearea-wide model and the multi-country modelof the ESCB. In addition to themacroeconomic models, staff also use off-model tools – such as reduced form equationsystems and time series models – forcomplementary assessments.

The variety of econometric tools used inEurosystem macroeconomic projectionexercises encompasses the area-wide model(AWM) (see Fagan et al. (2001)). The AWMis used in Eurosystem projection exercises,inter alia, to provide an assessment of averagejudgements embodied in the aggregation of

country projections and for runningscenarios. The AWM treats the euro area asa single economy and has the particularadvantage of helping to underpin an area-wide focus in general economic analysis andpolicy discussion within the Eurosystem.While the long-run properties of the modelare consistent with a basic neo-classicalsteady state, in which the long-run output isdetermined by technological progress and theavailable factors of production, the short-rundynamics are not explicitly derived from anoptimisation framework, but are insteadspecified in a more traditional “ad hoc” formand estimated on the basis of historical data.The relatively small scale of the model alsomakes it a tractable tool for the purpose ofscenarios.

A second econometric tool is the multi-countrymodel (MCM), which is used by ECB staff andsome NCBs. The MCM is an ESCB project,i.e. it is the result of collaboration betweenECB and NCB staff. The country blocs of theMCM share many features of the AWM,regarding their theoretical foundations, theirrelatively small size and the approach takento distinguish clearly between long-runspecifications and dynamic adjustment in theshort-run. Whereas the AWM is for the euroarea as a whole, the country blocs of theMCM provide a tool for forecasting themacroeconomic developments in thecountries of this area.

A range of structural national macroeconometricmodels are used by NCB staff. These modelsdiffer in size, scope, degree of underlyingmicrofoundations, and in their treatment ofthe financial sector and forward-lookingexpectations. Most of the models are of astandard type based on the neo-classical

5 In VAR (Vector Auto-Regressive) models, every endogenousvariable is modelled as a function of its own lagged values andthe lagged values of all the other endogenous variables in thesystem. In Bayesian VAR models, prior beliefs (based e.g. oneconomic theory) are used – in addition to sample information –to estimate values of the parameters in a VAR model and toindicate the degree of confidence with which these beliefs areheld. ARIMA (Auto-Regressive Integrated Moving Average)models or univariate time-series models are models where avariable is only expressed in terms of its own past values alongwith current and past errors.


synthesis, combining short-run demand-drivenbusiness cycle dynamics with steady-stateproperties as derived from standard growththeory (e.g. Banque de France et al. (1996-97), De Nederlandsche Bank (1985)/Fase etal. (1992), Banca d’Italia (1986)). Some NCBsuse as their main tool a version of therespective country bloc of the MCM, oftencomplemented by other econometric tools(McGuire and Ryan (2000)). In addition, NCBshave themselves developed multi-countrymodels which can be used in the peer-reviewprocess (e.g. Deutsche Bundesbank (2000)and de Bondt et al. (1997)/De NederlandscheBank (2000)). NCBs also use, in parallel, anumber of different macroeconometricmodels that focus on different aspects of the

economy. For instance, a number of NCBshave developed models based on optimisingbehaviour and forward-looking expectationswhich are often used for scenario analysis(e.g. Jeanfils (2000), Kortelainen (2001)) andWillman et al. (1998; 2000)).

Whilst ECB and NCB staff use a variety ofapproaches to project economicdevelopments, they review their techniqueson an ongoing basis. In particular, theexchange of experience and ideas taking placein the WGF and the Working Group onEconometric Modelling, as well as joint effortsto further develop macroeconomic projectiontools, will lead to further development in themedium term.

B Eurosystem projection tools

6 Winder, C. (1997), “On the Construction of European Area-Wide Aggregates: A Review of the Issues and EmpiricalEvidence”, DNB Staff Report No. 499.

1 Aggregating the projections

In the context of the discussions betweenthe ECB and the NCB staff an overall euroarea projection is obtained. This projectionis consistent with the aggregation of countryfigures, as described below.

The euro area aggregation of the countryprojections uses methods analogous withthose of Eurostat, which involves taking thesum of national concepts in levels expressedas a common unit of measure. The raw dataused for the aggregation are the projectionsin levels in national currencies, the euroconversion rates at the 1995 ECU exchangerate, as well as some specific weights.

The aggregation of GDP and its expenditurecomponents at constant prices is performed in acommon currency, in this case the 1995 ECU.The GDP deflator and its demand componentsare derived as the ratio of variables in nominalterms (also obtained by simple sum) divided bythe corresponding variables at constant prices.The nominal variables are derived by convertingthe original nominal variables (in nationalcurrency) into a single currency using the 1995ECU exchange rate. In practice, this amounts toa fixed exchange rate-weighted aggregation, but

with variable-dependent and time-varyingweights.6

The HICP aggregation is computed using theexact Eurostat methodology, i.e. as an annualchain index with changing country weights.The weight of a country is its share of theprivate final domestic consumptionexpenditure expressed in euro. For theprojection period, the latest available set ofweights is used.

2 Ensuring trade consistency

In the context of the MPEs, projections areprepared for the individual countries and thenaggregated at the euro area level. However,in order for these individual countryprojections to serve as a reliable guide forarea-wide conclusions, it is necessary forthem to be fully consistent with each other.The trade consistency exercise ensures thatindividual country projections of tradevolume and price variables are consistent witheach other as well as being consistent with


the assumptions made about the internationalenvironment.

Trade consistency can be divided into whatcan be termed “cross-trade consistency” and“ex ante/ex post trade consistency”.

“Cross-trade consistency” ensures that eachcountry’s export projection is consistent withthe import projections for its trading partnersin both volume and price terms. This isassessed using trade shares and various tradeequations. Convergence with regard to“cross-trade consistency” is achieved when(i) projected real exports are expected togrow in line with world demand and/ordeviations can be explained by changes incompetitiveness and/or other factors and(ii) projected export and import prices imply“reasonable” profiles for export and importcompetitiveness.

“Ex ante/ex post trade consistency” ensures thatthe original assumptions made for each countryfor external demand variables are updated toremain consistent with the projections obtainedfor the imports of other euro area countries,with the same kind of consistency being ensuredwith regard to external competitors’ prices.Hence, in successive iterations of the projection,the euro area trade prices will be updated andnew measures of competitors’ prices will becalculated. Convergence with regard to “exante/ex post trade consistency” is achievedwhen the differences between successive valuesof world demand and competitors’ prices aresmall enough to be of no practical importance.

Depending on the results of the tradeconsistency analysis, additional projectionrounds or iterations could be required.

3 Reflecting the uncertaintyassociated with economicprojections

As with any similar exercise, the Eurosystemmacroeconomic projections are surroundedby considerable uncertainty. To take intoaccount this uncertainty, the GoverningCouncil has decided to publish theprojections in the form of projection ranges.

The technique used to produce projectionranges is simple. For each variable, a range isapplied with a size equal to twice the averageabsolute value of the differences between actualoutcomes and previous macroeconomicprojections carried out over a number of yearsby euro area central banks. In general, the rangesdiffer depending on the variables and the timehorizons involved. They reflect both thedifferent degrees of difficulty in projectingindividual variables at different horizons and theeffects of discrepancies between theassumptions made for conditioning variables andtheir subsequent actual values. It should benoted that the tendency of most ranges to widenover the projection horizon reflects theincreased uncertainty surrounding projectionsfor the later years. In addition, the ranges tendto be greater for variables with greater intrinsicvolatility, such as the growth of gross fixedcapital formation.

In addition, to further quantify the impact ofvarious sources of uncertainty, the Eurosystemstaff may also present scenario exercises to theGoverning Council. Such scenario exercises areperformed at euro area aggregate level bothusing the area-wide model and by examiningthe aggregate of the country effects. The latterexercise is performed by the NCBs, who firstcompute the direct impacts of the scenario fortheir respective countries, before ECB staffderive the aggregate euro area impact, takinginto account the second-round spillover effectsthrough intra-euro area trade volumes andprices. The results are then discussed by theWGF, as a result of which modifications to theimpacts may be made. Both the area-wide modeland aggregated country results are presentedto the Governing Council.


This guide to Eurosystem staffmacroeconomic projection exercises followsthe publication for the first time in December2000 of the outcome of the staffmacroeconomic projections. The projectionsrepresent Eurosystem staff views regardingfuture prospects. As an input twice a yearinto the monetary policy decisions of theGoverning Council, the projections serve asa useful tool for summarising the existinginformation on current and future economicdevelopments, in addition to the analysis,under the first pillar and other regularanalysis, of information under the secondpillar. The projections derive from a processthat involves close interaction between stafffrom the NCBs and the ECB, ensuring thatthe projections are fully consistent bothacross countries and for the euro area andthat they draw on all expertise available.

This guide has described the currentprocedures and techniques used to derive

Conclusions and future developments

Eurosystem staff macroeconomic projectionexercises. However, there are a number ofareas where further developments will takeplace. In particular, tools and procedures willbe updated in the light of newly availableinformation and data, particularly asadditional statistical series are released byEurostat, allowing for more detailedprojections of financial accounts forhouseholds and enterprises. Theimprovement of balance of paymentsprojections, including the disaggregation ofintra-euro area and extra-euro area tradeflows, is another example. Ongoing economicresearch leads constantly to the developmentof new and better projection techniques,whilst new economic issues also keep arisingwhich call for developing new approachesand projection tools. An important task ofthe Eurosystem is to continue to be at theforefront of economic research, and therebyto develop and improve techniques to assistmonetary policy.


Bibliography

This bibliography includes references to documents on tools and techniques used by the ECBand the NCBs in the context of their macroeconomic projection exercises. The list onlyincludes published documentation and has been restricted to specific references to tools andtechniques used in the context of forecasts and projections within the Eurosystem.

Altissimo, F., Marchetti, D.J. and Oneto, G. (2000), “The Italian Business Cycle:Coincident and Leading Indicators and Some Stylized Facts”, Temi di discussione, No. 377,Banca d’Italia.

Ballabriga, F., Álvarez, L.J. and Jareño, J. (2000), “A BVAR macroeconometric modelfor the Spanish economy: methodology and results”, Economic Studies, No. 64, Banco deEspaña.

Banca d’Italia (1986), “Modello trimestrale dell’economia italiana”, Temi di discussione, No.80, Banca d’Italia.

Banque de France, CEPREMAP, Direction de la Prévision, Erasme, Insee, OFCE (1996/1997), “Structures et propriétés de cinq modèles macroéconométriques français”, Notesd’Etudes et de Recherche, Vol. 38, June 1996, and Economie et Prévision.

Botas S., Marques, C.R. and Neves, P.D. (1998), “Estimation of Potential Output for thePortuguese Economy”, Quarterly Bulletin, Banco de Portugal, December 1998.

Bravo, M. and Sánchez, P. (2001), “Forecasting and Quarterly disaggregation of NationalAccounts (ESA 95) aggregates”, forthcoming Working Paper, Banco de España.

Bruneau, C. and De Bandt, O. (1999), “La modélisation VAR structurel: application à lapolitique monétaire en France”, Economie et Prévision, January 1999.

Buisán, A. and Gordo, E. (1997), “El sector exterior en España”, Economic Studies, No. 60,Banco de España.

Craveiro Dias, F. (1993), “A Composite Coincident Indicator for the Portuguese Economy”,Working Paper, No. 18/93, Banco de Portugal.

Cristadoro, R. and Sabbatini, R. (1999), “Seasonality in the HICP and its Implications forthe Short-Term Monitoring and Forecasting of Inflation”, in Silver, M., and Fenwick, D. (eds.),Proceedings of the Measurement of Inflation Conference, Cardiff University, Cardiff 31 August-1September 1999.

de Bondt, G.J., van Els, P.J.A. and Stokman, A.C.J. (1997), “EUROMON: amacroeconometric multi-country model for the EU”, DNB Staff Reports, No. 17, DeNederlandsche Bank.

De Nederlandsche Bank (1985), “MORKMON: A Quarterly Model of the NetherlandsEconomy for Macro-Economic Policy Analysis”, Monetary Monographs, No. 2, DeNederlandsche Bank, Amsterdam.


De Nederlandsche Bank (2000), “EUROMON: The Nederlandsche Bank’s multi-countrymodel for policy analysis in Europe”, Monetary Monographs, No. 19, De Nederlandsche Bank,Amsterdam.

Deutsche Bundesbank (2000), Macro-Econometric Multi-Country Model: MEMMOD, Frankfurtam Main.

Estrada, A., de Castro, F., Hernando, I. and Vallés, J. (1997), “La inversión en España”,Economic Studies, No. 61, Banco de España.

Fagan, G., Henry, J. and Mestre, R. (2001), “An Area-Wide Model (AWM) for the euroarea”, ECB Working Paper, No. 42, European Central Bank, Frankfurt am Main.

Fase, M.M.G., Kramer, P. and Boeschoten, W.C. (1992), “De Nederlandsche Bank’squarterly model of the Netherlands’ economy”, Economic Modelling, Vol. 9, pp. 146-204.

Garganas N.C. (1992), The Bank of Greece Econometric Model of the Greek Economy, Bank ofGreece.

Gibson, H. and Lazaretou, S. (2001), “Leading Inflation Indicators for Greece”, forthcomingin Journal of Economic Modelling.

Hall, S.G. and Zonzilos, N.G. (1997), “The output gap and inflation in Greece”, EconomicBulletin, Vol. 9, Bank of Greece.

Hall, S.G. and Zonzilos, N.G. (2001), “The determination of wage and price inflation inGreece: An application of modern cointegration techniques”, paper presented at theConference on Greece’s Economic Performance and Prospects (Dec. 2000). (To be publishedin a book by the Bank of Greece and the Brookings Institution in spring/summer 2001.)

Hella, H. and Takala, K. (1998), “Estimating Current Account by using System Methodsand Barometer Variables”, paper presented at Ciret conference in New Zealand in 1999, IFOStudien Journal, No. 4.

Irac, D. (2000), “Estimation of a time varying Nairu in France”, Notes d’Etudes et deRecherche, Vol. 75, Banque de France, July 2000.

Irac, D. (2001), “Using output gap measures based on statistical methods to predict inflation”,mimeo, Banque de France, expected to be published by November 2001.

Irac, D. and Jacquinot, P. (1999), “L’investissement en France depuis le début des années80”, Notes d’Etudes et de Recherche, Vol. 63, Banque de France, April 1999.

Irac, D. and Sédillot, F. (2001), “Short run assessment of economic activity: the use ofsurvey data”, expected to be published in Notes d’Etudes et de Recherche, Banque de France.

Jacquinot, P. and Mihoubi, F. (2000), “Modèles à anticipation rationnelles de la conjoncturesimulée: Marcos”, Notes d’Etudes et de Recherche, Vol. 78, Banque de France, November 2000.

Jeanfils, P. (2000), “A Model with Explicit Expectations for Belgium”, NBB Working Paper,No. 4, National Bank of Belgium, March 2000.


Jondeau, E., Le Bihan, H. and Sédillot, F. (1999), “Modélisation et prévision des indicesde prix sectoriels”, Notes d’Etudes et de Recherche, Vol. 68, Banque de France, September1999.

Kortelainen, M. (2001), “Actual and perceived monetary policy rules in a dynamic generalequilibrium model of the euro area”, Bank of Finland Discussion Papers, No. 3.

Luís, J.B. (2000), “The Estimation of Risk Premium Implicit in Oil Prices”, Working Paper,No. 2/00, Banco de Portugal.

Malo de Molina, J.L., Viñals, J. and Gutierrez, F. (eds.) (1998), Monetary Policy andinflation in Spain, Macmillan Press, various chapters.

Marques, C.R. and Botas S. (1997), “Estimation of the NAIRU for the PortugueseEconomy”, Working Paper, No. 6/97, Banco de Portugal.

McGuire, M. and Ryan, M. (2000), “Macro-modelling Developments in the Central Bank”,Quarterly Bulletin, Spring 2000, Central Bank of Ireland.

Meyler, A. (1999), “A Statistical Measure of Core Inflation”, Technical Paper, No. 2/RT/99,Central Bank of Ireland.

Meyler, A., Quinn, T. and Kenny, G. (1998a), “Forecasting Irish Inflation using ArimaModels”, Technical Paper, No. 3/RT/98, Central Bank of Ireland.

Meyler, A., Quinn, T. and Kenny, G. (1998b), “Bayesian VAR Models for forecasting IrishInflation”, Technical Paper, No. 4/RT/98, Central Bank of Ireland.

Parigi, G. and Schlitzer, G. (1995), “Quarterly Forecasts of the Italian Business Cycle byMeans of Monthly Economic Indicators”, Journal of Forecasting, Vol. 14, pp. 117-141.

Pinheiro, M. (1998), “Estimation of the output gap: a bivariate approach”, Quarterly Bulletin,Banco de Portugal, December 1998.

Quinn, T., Kenny, G. and Meyler, A. (1999), “Inflation Analysis: An Overview”, TechnicalPaper, No. 1/RT/99, Central Bank of Ireland

Quinn, T. and Mawdsley, A. (1996), “Forecasting Irish Inflation: A composite leadingindicator”, Technical Paper, No. 4/RT/96, Central Bank of Ireland.

Rooij, M.C.J. van and Stokman, A.C.J. (2000), “Voorspellers voor de bbp-groei in de VS,Japan en de EU op basis van indicatoren”, Onderzoeksrapport WO&E, No. 636, DeNederlandsche Bank.

Sicsic, P. and Villetelle, J.P. (1995), “Du nouveau sur le taux d’épargne des ménages”,Economie et Prévision.

Siviero, S., and Terlizzese, D. (1997), “Crisi di cambio e discontinuità strutturali: un’analisicon il modello econometrico della Banca d’Italia”, in Ricerche quantitative per la politicaeconomica 1995, Banca d’Italia, Rome.


Takala, K. and Tsupari, P. (1999), “The Predictive Power of Finnish Business Surveyamong Industrial Sectors”, in Oppenländer, K.H., Poser, G. and Schips, B. (eds.), The Use ofSurvey Data for Industry, Research, and Economic Policy, pp. 583-613.

Terlizzese, D. (1994), “Il modello econometrico della Banca d’Italia: una versione in scala1:15”, in Ricerche quantitative per la politica economica 1993, Banca d’Italia, Rome.

Vanhaelen, J.-J., Dresse, L. and De Mulder, J. (2000), “The Belgian Industrial ConfidenceIndicator: Leading Indicator of Economic Activity in the Euro Area?”, NBB Working Paper,No.12, National Bank of Belgium, November 2000.

Willman, A., Kortelainen, M., Männistö, H.-L., Tujula, M. (1998), “The BOF5Macroeconomic Model of Finland, Structure and Equations”, Bank of Finland Discussion Papers,No. 10.

Willman, A., Kortelainen, M., Männistö, H.-L., Tujula, M. (2000), “The BOF5macroeconomic model of Finland, structure and dynamic microfoundations”, EconomicModelling, No. 17, pp. 275-303.

Wit, J.N.M. (1998), “Voorspelling geharmoniseerde index van consumentprijzen: de HICPvoor Nederland”, Onderzoeksrapport WO&E, No. 562, De Nederlandsche Bank.

Wouters, R. and Dombrecht, M. (2000), “Model-based inflation forecasts and monetarypolicy rules”, NBB Working Paper, No. 1, National Bank of Belgium, March 2000.

Zonzilos, N.G. (2000), “The Phillips curve of the Greek economy and the time varyingNAIRU”, Economic Bulletin, Vol. 15, Bank of Greece.

EUROPEAN CENTRAL BANK...The WGF is composed of ECB and NCB macroeconomic and econometric experts. B An outline of Eurosystem macroeconomic projection exercises There are three main

Documents