Research

THE INFLUENCE OF UNCERTAINTY ON INVESTMENT IN THE UK: A MACRO OR MICRO PHENOMENON? 1

Description
THE INFLUENCE OF UNCERTAINTY ON INVESTMENT IN THE UK: A MACRO OR MICRO PHENOMENON? 1 Absrac Paul Temple Deparmen of Economics, Universiy of Surrey Guildford, Surrey GU2 5XH (U.K.)
Categories
Published
of 26
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
THE INFLUENCE OF UNCERTAINTY ON INVESTMENT IN THE UK: A MACRO OR MICRO PHENOMENON? 1 Absrac Paul Temple Deparmen of Economics, Universiy of Surrey Guildford, Surrey GU2 5XH (U.K.) Giovanni Urga (*) Deparmen of Invesmen, Risk Managemen and Insurance Ciy Universiy Business School Frobisher Crescen, Barbican Cenre, London EC2Y 8HB Tel.: +/44/(0)20/ ; Fax.:+/44/(0)207/ Ciaran Driver Imperial College Managemen School London SW7 2PG While he heory examining he relaionship beween uncerainy and invesmen has suggesed new research avenues, i has no had srong predicive power. Neverheless, a he policy level he benefis for invesmen of a more sable economic climae are being emphasised. These consideraions poin o he need for empirical work. Accordingly, his paper draws on indusry level panel daa, obained by marrying he UK Census of Producion wih he CBI Indusrial Trends Survey, and applies dynamic panel daa mehods o disinguish beween macro and micro sources of uncerainy and o consider he role of financial facors. I is found ha boh sources of uncerainy exer a considerable negaive impac on invesmen, while financial facors may be imporan in some indusries. June 2000 Keywords: J.E.L. Invesmen, Uncerainy, Panel Daa Classificaion: C33, E22, D81 (*) Corresponding auhor 1 The auhors would like o hank Bob Chirinko, Sephen Hall, Brian Henry, Andrew Senance, and Ian Small for helpful commens and advice; also paricipans a research seminars a he London Business School, Hamburg Universiy, and London Guildhall Universiy. We wish o hank wo anonymous referees and he Edior for heir valuable commens and suggesions. Neverheless, he usual disclaimer applies. We hank Dr. M. Mamazakis for research assisance. ESRC funding under gran R is graefully acknowledged. 1. Inroducion 2 Weak physical invesmen has characerised many counries in he European Union in he 1990s. A he same ime, economiss of a variey of persuasions are increasingly emphasising he causal links running from invesmen o oupu. We may menion here no jus he endogenous growh heoriss, bu also hose recognising he role played by invesmen in shaping he inernaional compeiiveness of economies and hereby influencing exernal consrains on growh. Aemps however o influence he rae of invesmen and capure any favourable spillover effecs ha may accrue requires an undersanding of he deerminans of invesmen. Here heoreical undersanding has been enhanced by a new emphasis on he roles of uncerainy, irreversibiliy, and lumpiness in shaping he invesmen decision. Bu heoreical insighs have lacked predicive power, and much work remains for empirical analysis. This paper conribues o he empirical picure by esimaing invesmen funcions using survey-based informaion a an indusry level. These daa are used o explain he rae of invesmen in UK manufacuring, a secor hisorically characerised by raes of invesmen which are low by inernaional sandards (for a discussion see Kison and Michie 1996 or Temple 1997). While he use of survey daa is subjec o cerain well known limiaions, i does allow for he inroducion of a direc expecaional elemen ino he analysis. The paper addresses wo broad quesions relaing o he role of uncerainy. Firs, how imporan has uncerainy been in explaining he poor invesmen performance of UK manufacuring? Second, is he uncerainy repored by survey respondens primarily of a macro naure - i.e. a reflecion of he volailiy of he whole economy - or is i of a more micro characer, a resul of uncerainy a a secoral level? These quesions are of paricular ineres given he recen policy emphasis on creaing a sable macro environmen and he favourable impac ha such an environmen may have for he underlying growh of producive poenial. Beyond he quesion of uncerainy, he mehod adoped in his paper also permis an examinaion of he role ha financial facors play in he invesmen process, a possible influence which heoreical work on capial marke imperfecions has suggesed may be much more imporan han radiional approaches have suggesed. 3 The paper is arranged as follows. Secion 2 examines some relevan developmens in invesmen heory. Secion 3 oulines he use of he Confederaion of Briish Indusry s (CBI) Indusrial Trends Survey (ITS) in modelling uncerainy and financial effecs on invesmen. Secion 4 inroduces he chosen specificaion. Secion 5 discusses he resuls of he economeric analysis. Secion 6 concludes. 2. Curren Issues in Invesmen Theory Recen heoreical lieraure has seen a renewed ineres in he role of irreversibiliy and uncerainy in shaping he invesmen decision of profi maximising firms (for a fuller discussion han is possible here see Driver and Temple [1999]). Much of he lieraure has been concerned wih undersanding he dynamics of invesmen. An early resul, esablishing he relevance of an arbirage condiion for he invesmen decision, bu in a siuaion of deerminisic demand, was ha coninually falling invesmen coss (or a rising value of invesmen opporuniies) would end o delay invesmen in he absence of rivalry (Barzel 1968; see also Dixi and Pindyck 1994, p.138). More recenly however, ineres has focussed on sochasic demand. I is now well known, for example, ha irreversible invesmen (underaking an invesmen projec enails sunk coss), combined wih ongoing uncerainy (he fuure benefis or coss of he invesmen projec are unknown, bu his is parially resolved by addiional informaion) and ime flexibiliy (i.e. if he invesmen projec is no underaken oday, he firm reains he opion of underaking he projec omorrow) may have a subsanial impac on invesmen behaviour (Abel e al 1996). The modern lieraure on he impac of uncerainy complemens he radiional view ha increased uncerainy will raise he invesmen of a risk-neural compeiive firm, unless subsiuabiliy beween facors is very high, because marginal profi is convex in he uncerain variable (Harman, 1972, 1976; Abel, 1983,1985). 2 However his finding is modified by irreversibiliy and imperfecly compeiive condiions (Berola, 1988; Pindyck, 1988; Dixi and Pindyck, 1994); if for example a risk neural, monopolisic firm canno dispose of insalled capial, increased uncerainy may reduce curren invesmen 2 I should also be noed ha even in perfec compeiion, he radiional convexiy resul only applies in he case of a risk neural firm and may be reversed under risk aversion (Aiginger 1987, Pleeer and Horowiz 1974, Nickell 1978, Salari and Travaglini 1999, and Nakamura 1999). Furhermore in models wih capaciy raioning or excess capaciy, and as is naural wih sochasic demand, he radiional resuls may again be reversed (Driver e al 1996a). 4 since he opion value of waiing is increased. 3 For he case of price-aking firms, he former (convexiy) effec ends o dominae and he parial equilibrium effec is sill for uncerainy o raise invesmen (Caballero 1991; see also Abel and Eberly 1993 and 1994). However, Pindyck (1993) has criicised his resul showing ha wih free enry he expeced price changes in response o marke demand shocks; a mean preserving spread in he sochasic erm in marke demand reduces he expeced value of price in laer periods because new enry will occur under favourable demand. 4 I can herefore be seen ha heoreical work does no lead o any clear-cu conclusions regarding he impac of invesmen, so ha he imporance of uncerainy is clearly an empirical maer. In he main, research ha has ried o evaluae he issue empirically has found a negaive effec for uncerainy (Caballero and Pindyck 1993, Pindyck and Solimano 1993, and Ferderer 1993a,b). For he U.K., despie he differen heoreical framework and measure of uncerainy, boh Driver and Moreon (1991) and Price (1995) conclude ha more uncerainy leads o less invesmen. Finally Alesina and Peroi (1993) esablished a negaive correlaion beween social and poliical insabiliy and invesmen. I seems clear ha he issue needs o be sudied a a more dis-aggregaed level. This is parly because uncerainy may acually arise a a firm specific - or indusry specific - level, bu also because exising heory suggess ha uncerainy has quie differen implicaions for differen indusries, depending, for example, on differen degrees of irreversibiliy or on marke srucure as discussed above. These of course are issues which he aggregae models of invesmen are incapable of addressing. Available papers, which do approach he issue a a micro level, include hose by Leahy and Whied (1996), Driver e al (1996b) and Guiso and Parigi (1999). In hese cases, uncerainy has a significan negaive effec on firm s invesmen. To complemen hese sudies, i is one of he key objecives of he curren paper o disinguish beween macro and micro sources of uncerainy, a disincion which we believe o be of considerable relevance from a policy perspecive given he emphasis currenly given o he advanages of a sable macro-economy. Our plan is o pursue a micro level approach which explicily considers 3 The combinaion of differen forms of adjusmen coss wih opion heory models makes i even more difficul o obain unambiguous predicion concerning he uncerainy-invesmen relaionship (Hammermesh and Pfann 1996; Pindyck 1991). 4 Leahy (1993) generaes he resul ha compeiion is irrelevan o he decision o delay; compeiion reduces he value of he opion by reducing he value of he invesmen. Thus he rigger value is he same for he firm ha ignores compeiors as for hose ha consider hem. 5 unobservable expecaions , following a suggesion made recenly by Chirinko (1996): Expecaions of he fuure are imporan, perhaps paramoun, in deermining invesmen expendiures. Complicaed models - regardless of heir descripive accuracy- will no make subsanial progress in improving undersanding of invesmen behaviour unless hey deal adequaely wih unobservable expecaions . Our approach o his problem is o use direc measures of expecaions from CBI survey daa, which we now consider in more deail. 3. The CBI Indusrial Trends Survey The Indusrial Trends Survey (ITS) of he CBI (and is forerunner, he Federaion of Briish Indusry) has been producing daa on he curren sae of opinion in UK manufacuring since Of paricular ineres in he curren conex is quesion 16 of he ITS, for which daa exiss, on a quarerly basis since Par C of he quesion invies respondens o consider which facors, including uncerainy abou demand, are expeced o limi capial expendiure auhorisaions over he nex welve monhs. Available replies are: inadequae ne reurn on proposed invesmen (CBI16CA); a shorage of inernal finance (CBI16CB); an inabiliy o raise exernal finance (CBI16CC); he cos of finance (CBI16CD); uncerainy abou demand (CBI16CE); shorage of labour including managerial and echnical saff (CBI16CF); oher (CBI16CG). I should be noed ha respondens are able o give more han one of he above answers. The percenage replying posiively o each of hese possibiliies we designae CBI16CA, CBI16CB, CBI16CC, CBI16CD, CBI16CE, CBI16CF, CBI16CG respecively. The se of quesions can be regarded as consiuing a relaively complee se of expeced consrains which can be esed for explanaory power when enered in sandard specificaions. The informaion in hese survey replies reflecs a view of he invesmen process pained on a raher broader canvas han is common in economic heory. We may noe here ha he framework no only allows for he consideraion of uncerainy bu 6 also ha i is based on imperfec compeiion and allows for capial marke imperfecions. I is herefore possible o esimae effecs oher han uncerainy, he possible influence of financial facors. Here, conroversy daes from he famous Modigliani and Miller resul (ha he financing decisions of firms are essenially orhogonal o he invesmen decision) and which is based on arbirage possibiliies in perfec capial markes. While he lis of possible violaions o he assumpions of he Modigliani-Miller heorem is a subsanial one, he lieraure has focused on he idea ha capial srucures reflec informaion asymmeries beween managemen and owners and/or lenders. Typically, managemen may be expeced o possess more informaion abou he value of he firm s invesmens (and of heir own effors) han eiher deb or equiy holders. Hubbard (1998) provides a comprehensive overview of he role of such capial marke imperfecions. Figure 1 illusraes each of he possible survey answers for aggregae manufacuring for he period The wo mos imporan replies in erms of percenages responding posiively are Q16CA - an inadequae reurn, and Q16CE - uncerainy abou demand. The former rose srongly in he period o he middle 1980s bu has since sabilised. The demand uncerainy response shows somehing of a cyclical paern, ye a no sage during he recovery of he 1990s has repored uncerainy diminished o he exen winessed in he recovery of he 1980s. The remaining caegories of reply appear less imporan, a leas a his aggregae level. The percenage of firms being repored as being financially consrained remains relaively seady a around 20-25%, wih only a mild cyclical influence, while he cos of finance feaures srongly only for very limied periods of igh moneary policy; exernal finance does no appear o be imporan a all. [Figure 1 abou here] Of course here are ofen major problems in inerpreing survey daa and his is one explanaion of why more use has no been made of such evidence in empirical economics. In he presen case we may poin for example o he imprecision of answer CBI16CA an inadequae ne reurn on proposed invesmen; his would appear o cover aspecs of boh he expeced profiabiliy of an invesmen, he cos of finance, and possible risk premia. More imporan for his paper is he inerpreaion of answer CBI16CE uncerainy abou demand. I is quie possible ha respondens may be giving his as a reason on he basis of low expeced demand as well as on he basis of he variabiliy of demand. Since we are 7 ineresed of course in he laer s impac on invesmen, i is necessary ha we conrol as ighly as possible in our empirical analysis for he former facor. For his reason we make use of boh a measure of real demand as well as a measure of capaciy uilisaion also drawn from he CBI survey. Mos sudies uilise quesion 4 of he CBI Survey o capure uilisaion. However, previous work of ours has emphasised he ambiguiy of his quesion (Temple and Urga 1997) and so we use Quesion 14 of he Survey which explicily disinguishes beween skilled labour (CBI14B) and capial (CBI14D) as facors limiing oupu over he following four monhs. The CBI daa is available for fify indusries; his allows us o consider he ineracion beween uncerainy and he degree of monopoly wih he laer being measured by a sandard concenraion raio obained from ONS sources. 4. Modelling Approach We now inroduce our invesmen specificaion which includes he direc expecaional series discussed in Secion 3. Given he broad specrum of answers possible in he ITS, we required a general specificaion. However, here is no single model in he lieraure which can be said o be ruly general. Mos models semming from sandard profi maximisaion under consrains include variables in demand, relaive prices and financial variables (Caina e al 1987). A common specificaion is he sandard Eulerype equaion or equaions based on solving he opimal conrol rajecory when convex coss of adjusmen of fixed capial are assumed. These specificaions can be modified o ake accoun of financial consrains. Alernaive specificaions may be based on delivery lags wihou convex adjusmen coss, hreshold models based on fixed or disconinuous adjusmen coss, or models based on real opions heory. A fairly general specificaion (for which a derivaion on he basis of an imperfecly compeiive firm is considered in Appendix 1), is however as follows: ln(i) = h [ln(w e ), ln(p e ), ln(y e ), A e, CF e, FV e, CU e, ln (PI e ), s 2 d ] (1) where: Superscrip e denoes expeced values I is he volume of invesmen W is he real produc wage per uni of labour inpu P is he price of oupu Y is real oupu (value added) A is an index of echnical progress CF is cash flow. FV are financial facors, such as real deb, ineres raes and he marke value of he firm. CU is he degree of capaciy uilisaion PI is he price of invesmen goods relaive o oupu prices s 2 d is he variance in demand 8 Daa on hese variables were all available o us in some form, excep for he echnology erm which is modelled by a ime rend. The relaive price and cash flow erms are aken from ONS sources (wih expecaions being capured by a lag srucure). Cash flow (CF) is no direcly available bu a proxy available from official sources which we employ is gross value added less labour cos expressed as a proporion of gross value added; his we designae as PROF. The CU erm and he remaining expeced consrain erms, including he uncerainy erm are represened by he various CBI series discussed in secion 3. The deails are given in Appendix 2. In order o make use of he CBI daa (which is only available for individual indusries), we canno esimae (1) using observaions on individual firms. In any case, we believe a srong case can be made for invesigaion a he level of he indusry. Firs, for manufacuring a leas, he annual Census of Producion (CoP) and oher official indusry level daa provide annual observaions a a qualiy level superior o ha conained in company accouns. Noe ha his daa source allows us o obain reasonable measures of relaive produc wages and he relaive price of invesmen goods a he level of he indusry. Generally hese can only be proxied a he level of he firm 5. Second, and more imporanly, i will sill enable us o pursue our objecive of disinguishing beween micro and macro sources of 5 Oher influences on he cos of capial namely ineres raes and depreciaion raes we model by ime dummies and indusry fixed effecs respecively. In he case of ineres raes, here is also quesion 16CD in he CBI indusrial rends survey asking respondens abou he cos of finance as a facor limiing capial expendiure auhorizaions. 9 uncerainy. Exising sudies of he role of uncerainy in UK manufacuring (Driver and Moreon 1991; Price 1995) are a an aggregae level and hence are incapable of drawing such a disincion. Our esimaion sraegy explicily allows us o exploi cross secional variaion o esimae how much of he impac is derived from macro uncerainy - exchange rae insabiliy, GDP or inflaion volailiy and so forh - and how much is aribuable o more local facors. As oulined in Secion 2, a furher objecive in he esimaion was o explore he links beween indusrial srucure and uncerainy. I seems clear ha while macro uncerainy is, o an exen, common across indusries, a a micro-level boh he exen of uncerainy, and is impac (in boh size and sign), may differ, iner-alia, according o he degree of monopoly. Our earlier discussion has poined o he limiaions as well as he richness of he ITS. Accordingly, we firs consider a core model which makes no use of he ITS daa on invesmen consrains, alhough i does include an ITS based measure of capaciy uilisaion (see above). We hen consider he impac of he inclusion of he CBI Survey quesion 16 variables in heir enirey. Invesigaion of micro versus macro influences on invesmen is faciliaed by dynamic panel daa mehods. This allows us o exploi differen levels of aggregaion across indusries and secors. The full panel consiss of 80 indusries a approximaely he hree digi SIC level; we also consider aggregaes of hese indusries formed ino 17 secors a approximaely he wo digi level. In order o disinguish macro from micro effecs we exploied common paerns of variaion across he secors and indusries using ime dummies. The simples dynamic form of equaion (1) may be wrien as follows: ln I 1 7 = α i + λ δ CBI16CA δ CBI16CG j + λ 2 + ε + β ln Y + δ CBI16CB 1 + β 3 2 ln( PI) + δ CBI16CC + β 3 PROF 4 + β + δ CBI16CD 4 ln W + δ CBI16CE 5 + γ CBI14B 1 + δ CBI16CF 6 + γ CBI14D wih i = 1,...,80; j = 1,...,17; = 1979,...,1992; (2) α i represens he individual indusry effec, λ j he secoral effec, and λ he common effec across indusries a ime. We leave he choice beween fixed and random effecs esimaors as an empirical issue which address below Discussion of Resuls A summary of resuls is presened in Tables 1 o 3. Table 1 presens boh he core model as discussed in he previous secion and he resuls when Quesion 16 respon
Search
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks