hipping Marke Inegraion - he Case of icky Newbuilding Prices Roar Adland Deparmen of Economics, Norwegian chool of Economics, 535 Bergen, Norway Absrac In his paper we derive an equaion defining conemporaneous equilibrium prices across he four shipping markes: he newbuilding marke he second-hand marke he demoliion marke and he freigh marke. he model requires raher general condiions, namely ha (i) all second-hand ships are priced based on linear deprecaion down o scrap value, (ii) expecaions of he fuure scrap value and lifespan of he vessel are equal o curren values and (iii) he erm srucure of (imecharer) freigh raes is observable. In he empirical par of he paper we use his equilibrium relaionship o illusrae he presence of a erm srucure of newbuilding prices and show ha he lower volailiy of newbuilding prices compared o second-hand values is a resul of a ime-varying delivery lag which is posiively correlaed wih he alernaive cos of operaing in he freigh marke. hese observaions are imporan for saisical analysis of he dynamics of ship values. Keywords: Newbuilding Prices; Marke Inegraion; erm rucure Inroducion One of he remaining puzzles in mariime economics is he observaion ha newbuilding prices appear exogeneous or inelasic wih respec o he demand for new vessels. Pu differenly, newbuilding prices are oo sicky or insufficienly volaile given he ofen rapid changes in marke fundamenals. his long-held belief appears o have sared wih Zanneos (966) who saes (p.8) ha newbuilding prices, however, are neiher explosive nor are hey perpeually esablishing new lows hese shifs do no bring abou a balance in supply and demand bu are, insead, disequilibriaing. he explanaion has radiionally been one of exernaliies and organizaional rigidiies. Zanneos (966) argues ha in imes of shipbuilding overcapaciy, shipyards have an incenive o smooh ou producion across he cyclical low o avoid job cus and mainaining knowledge and compeency levels. Given ha such social funcions are lef ou of economic models, newbuilding prices will appear sub-opimal. randenes (22) suggess ha he srong presence of labour unions in he shipbuilding indusry has led o lower flexibiliy and ha he presence of subsidies disors newbuilding prices. Dikos (24) suggess ha an alernaive explanaion can be found wihin he framework of a funcioning compeiive equilibrium if he acions of shipyards are inerpreed as a resul of exercising real opions (o offer capaciy) under uncerainy. In paricular, wihin he radiional framework of invesmen under uncerainy (see, for insance, Dixi and Pindyck, 994), here will exis an upper hreshold level for he newbuilding price ha would immediaely rigger many yards o offer capaciy. Yards would raionally anicipae his, imposing an upper reflecing barrier o he price process ha can accoun for he lack of volailiy in newbuilding prices when compared wih charer raes. However, a newbuilding conrac a resale (he sale of a ship under consrucion) and a modern vessel are near perfec subsiues from he viewpoin of he shipowner. Ignoring echnological differences, hey will only differ in erms of price and ime o delivery (i.e. he ime he asse sars o generae revenue). Morever, mos shipowners are price akers compeing for yard slos and exising onnage in a compeiive marke. Consequenly, while subsidies may affec he level of he newbuilding price, his will insananeously ripple hrough o he second-hand marke such ha marke agens are indifferen beween ordering a newbuilding a he prevailing conracing price and buying a resale or an exising vessel. In his framework, none of he possible imperfecions menioned in he lieraure will acually affec he inegraion of he newbuilding, second-hand and freigh markes and, conversely, he corresponding volailiy of prices is compleely described by he dynamic equilibrium beween he markes. In he mariime economic lieraure, he assumed degree of marke inegraion and, consequenly, he approach o modeling he markes, has changed over ime.
Beensock (985) assumes ha new and second-hand ship prices are perfecly correlaed, hough he observes ha his condiion is unlikely o hold because newbuilding prices are sicky compared o second-hand prices. In subsequen work (Beensock and Vergois, 989), his sringen assumpion is relaxed by separae dynamic modeling of he newbuilding marke. randenes (984, 986) defines he long-run expeced earnings of a vessel based on is newbuilding price and assumes ha he second-hand value is a weighed average of shor and long-erm freigh earnings. solakis e al (23) invesigae second-hand prices in an Error Correcion Model wih a heoreical basis and find ha newbuilding prices and imecharer raes are he main drivers. Adland e al. (26) es he insananeous equilibrium relaionship beween second-hand values, conracing prices and freigh raes in a Vecor Error Correcion model framework and accouns for he ime-varying delivery lag. heir empirical ess sugges ha he second-hand marke for bulk carriers is closely coinegraed wih he fundamenal freigh and newbuilding marke wih no evidence of a shor-erm asse bubble even in he boom years in he early par of he cenury. Greenwood and Hanson (23) argue ha ship prices are way oo volaile given he observed mean reversion in freigh earnings and propose a behavioral model where firms over-exrapolae exogeneous demand shocks and parially neglec he endogeneous invesmen responses of heir compeiors. heir formal esimaion of he model confirms ha boh ypes of expecaional errors are needed o accoun for he empirical evidence. Recen research has followed up on he imporance of ime o build in erms of marke volailiy. Kaloupsidi (24) examine he impac of consrucion lags and heir lenghening in imes of high invesmen aciviy by consrucing a dynamic model of enry and exi from he shipping markes. Kaloupsidi finds ha moving from ime-varying o consan o no ime o build reduces prices while increasing boh he level and volailiy of invesmen. In his paper we propose an enirely new line of reasoning ha does no conradic wih marke efficiency namely ha newbuilding prices are no acually comparable across ime because he parameers of he newbuilding conracs (primarily he delivery lag and paymen schedules) will vary. In oher words, he observaion ha newbuilding prices are sicky is largely a resul of sysemaic changes in he specificaion of he newbuilding conrac underlying he price index.. he conribuion of his paper o he lieraure is hreefold. Firsly, we derive a new and simple single-equaion model ha relaes conemporaneous prices in all four shipping markes (newbuilding, secondhand, freigh and demoliion) in equilibrium. econdly, we use his model o illusrae he presence of a erm srucure of newbuilding prices and why a ime-varying delivery lag means newbuilding and second-hand values are no direcly comparable. hirdly, we show how a imevarying delivery lag ha is posiively correlaed wih he freigh marke will resul in sicky newbuilding prices and ha his is no a sign of marke inefficiency. he remainder of his paper is organized as follows. ecion 2 derives he equilibrium relaionship beween conemporaneous prices in he four markes, ecion 3 illusraes empirical applicaions relaed o he erm srucure and volailiy of newbuilding prices and ecion 4 concludes. 2 Defining marke inegraion In his secion we define mahemaically he equilibrium condiion where invesors are indifferen beween invesing in he second-hand and newbuilding markes. Le F (τ) be he ne presen value a ime (curren ime) of he paymens o he shipyard according o a newbuilding conrac for a ship delivered a he fuure ime τ. We noe ha in general, F (τ) need no correspond o he quoed newbuilding price in he marke as paymen schedules (boh iming and percenages) end o vary wih marke condiions. Furhermore, le and τ be he second-hand value a ime of ships of age zero and τ, respecively. We define he ne presen value of he profi earned by a modern ship beween ime and ime τ as I i C r OPEX () where C τ is he prevailing period imecharer rae a ime for he duraion (τ ) years, r τ is he risk-free ineres rae for he corresponding mauriy and OPEX is he daily operaing cos for a modern vessel. Finally, he wo would only be idenical in he case of full upfron paymen o he shipyard; a rare occurrence.
le Z be he scrap value of he vessel and be is life expecancy. We follow he resuls in Adland and Koekebakker (27) and assume ha all second-hand ships are priced based on linear deprecaion down o scrap value. he linear depreciaion funcion a any fuure ime τ [, ) can be wrien as:, Z, Z (2) In he remainder of he paper we will assume ha invesors have myopic expecaions wih regards o he vessel life expecancy and scrap value such ha E (Z τ ) = Z and E ( τ ) =. Pu differenly, invesors do no expec he scrap value or he expeced useful economic life of he vessel ype o change beween ime and τ. Given ha fuure vessel values are uncerain, we can inerpre he invesmen in a newbuilding conrac or, alernaively, a brand new vessel, as having uncerain payoffs. If you inves in a newbuilding conrac he known paymen sream wih ne presen value of F τ will resul in he ownership of a brand new (age zero) vessel upon delivery a ime τ, In oher words, he payoff in presen value erms is given by: F (3), r Alernaively, an invesmen in an exising brand new vessel provides a known freigh income from he period imecharer marke plus an uncerain residual value of a vessel wih age τ a he fuure ime τ. Formally, we can wrie he laer as: I (4) r, In equilibrium, prices have adjused such ha shipowners are indifferen beween hese wo invesmen alernaives. herefore, subsiuing Equaion 2 ino Equaion 4 we ge I r Z, Z (5) which, using Equaion 3, can be rewrien as, I Z r, F r Z (6) which, assuming Z τ = Z and τ = for all τ, can be furher shorened o I F Z r (7) Under he above assumpions, Equaion 7 defines he equilibrium relaionship beween he prices in all he four shipping markes a ime. I follows ha if we know, for insance, he curren value of a new vessel, he applicable imecharer rae and he scrap value Z, we can derive he heoreically consisen newbuilding conrac value F τ. We noe ha in he heoreical case of immediae delivery (τ = ), Equaion 7 is reduced o a simple equaliy beween he resale price for an age-zero vessel and he newbuilding price, as expeced. Imporanly, Equaion 7 allows for he possibiliy of a ime-varying delivery lag (τ ). Our equilibrium equaion requires raher general condiions, namely ha (i) all second-hand ships are
priced based on linear deprecaion down o scrap value, (ii) expecaions of he fuure scrap value and lifespan of he vessel are equal o curren values and (iii) he erm srucure of (imecharer) freigh raes is observable. 3 he erm srucure of newbuilding prices When discussing shipbuilding capaciy i is useful o disinguish beween capaciy o build (say, global compensaed gross onnes produced per year) and he ordering capaciy. he former is clearly consrained in he shor run by he number of drydocks and oher areas of ship assembly, while he laer has no heoreical capaciy since a yard will book is building slos also for fuure use. here exiss, in fac a erm srucure of newbuilding prices, describing he combinaions of cos and ime o delivery beween which shipowners would be indifferen. If he opporuniy cos of ime for he operaion of modern vessels is posiive, his erm srucure will be downward sloping such ha early delivery slos (and resales) command a premium over deliveries furher ino he fuure. Given ha he quoed newbuilding price in he marke refers o he prevailing ypical ime o delivery, which will necessarily vary wih he size of he orderbook and developmens in shipyard produciviy, he assumpions underlying he observaions will herefore resul in a ime series ha is inconsisen across ime. Drawing empirical conclusions on his basis, in relaion o price volailiy or oherwise, may herefore be severely misleading, hough he lieraure has hihero no acknowledged his problem. he delivery lag is ofen inerpreed as he ime required building a ship (cf. ime-o-build as in Kaloupsidi, 24), hough in pracice his is usually no he main componen. Rapid developmen of shipyard echnology and produciviy over he las 2 years, wih he aid of compuerised seel cuing and modular ship secions, means a large anker can be buil in a maer of several monhs. he acual delivery ime beween he conracing dae and handover is insead largely decided by he availabiliy of ime slos in he drydock faciliies of he individual yard, which again depends on he size and composiion of is orderbook and he bargaining power of he cusomer. However, subjec o differences in qualiy and cusomisaion, he compeiive naure of he shipbuilding marke and he disseminaion of informaion by newbuilding shipbrokers means ha conracing prices and chronological slo availabiliy do no differ much among he main shipbuilding groups. I is worh emphasising ha he differen ypes of ships will compee for he same slos, and so he delivery lag for bulk carriers, for insance, will be influenced by he demand for oher ship ypes such as ankers and gas carriers. he empirical resuls in Adland e al (26) suggess ha here is no clear relaionship beween he size of he ship and he delivery lag, presumably as smaller vessels will end o be buil by smaller less efficien yards. I follows from he above ha esimaes of he ime-varying delivery lag should be based on he enire orderbook across vessel ypes and shipyards, bu wih a focus on a paricular size range for he sake of consisency. For he sake of illusraing he shape of he erm srucure of newbuilding prices we rewrie Equaion 7 in erms of F τ : F I Z r (8) Keep in mind ha F(τ) is he ne presen value a ime he curren ime of he paymens o he shipyard according o a newbuilding conrac for a ship delivered a he fuure ime τ, which is generally no equal o he quoed newbuilding price in he marke as paymen erms are no sandardised. In he numerical examples below we have assumed 2% paymen upfron and he remaining 8% upon delivery of he vessel. Our empirical esimaes concerns he Capesize bulk carrier segmen (, DW+). Clarkson Research Ld. provided daa for period imecharer raes, resale prices, average scrapping age by year and average delivery lags by order year, he laer wo of which are illusraed in Figure below. he risk-free ineres raes on 6-monh, -year, 3-year and 5-year consan-mauriy U.. reasury securiies were obained from he H.5 saisical release of he Federal Reserve. We assume daily operaing coss (OPEX) of $8,/day.
Years 3 25 2 5 Years 6 5 4 3 2 5 993 994 995 996 997 998 999 crapping age 2 2 22 23 24 25 Delivery lag 26 27 28 29 Figure : Capesize scrapping age and delivery lag ource: Clarkson Research Ld. 2 2 22 23 24 We noe ha boh delivery lag and scrapping age increased sharply during he drybulk freigh marke boom beween 23 and 28, wih owners having o wai up o four years for delivery. Nex using Equaion 8 we can illusrae he shape of he erm srucure of newbuilding prices as shown in Figure 2 below. A hese poins in ime, he acual delivery lag was around 2.4 years as indicaed by he arrow. We noe ha he erm srucure is downward sloping, reflecing mainly he alernaive cos of operaing an exising ship a period imecharer raes ha a leas for periods longer han a year, will always cover our assumed operaing cos level. Figure 2 already poins o why he volailiy of newbuilding prices should be lower han ha of exising vessels (resale ships wih immediae delivery, in his case): Good markes, which would be associaed wih boh long delivery lags (high τ values) and a seep erm srucures of freigh raes, and herefore a seep erm srucure of newbuilding prices (ref. he impac of I τ in Equaion 8) will naurally resul in a newbuilding price ha is much lower compared o second-hand onnage. A ime series of newbuilding prices will herefore appear much sickier and have lower volailiy han he price of exising modern vessels. Figure 2 can be inerpreed as he combinaions of delivery imes and coss where an owner would be indifferen beween a resale and a conrac for fuure delivery. $m 8 7 6 5 4 3 2 F(τ) 7-Jan- F(τ) 4-May- 2 3 4 5 Delivery lag (years) Figure 2: he erm srucure of newbuilding prices
he above argumen becomes clear when looking a he difference beween he acual price of exising ships ( ), he observed newbuilding price (adjused for paymen erms, F τ obs ) and he esimaed newbuilding price given observed delivery lags and Equaion 8 (F τ es ) shown in Figure 3 below. We see ha proper adjusmen for ime-varying delivery lags and alernaive cos smoohs ou he newbuilding prices, cerainly in he sronger markes in he firs par of he sample. Indeed, he sample sandard deviaion of hese hree ime series are: : $.77m, F τ es : $6.3m and F τ obs : $4.27m. In oher words, applying basic mariime economic heory has seemingly reduced much of he sicky newbuilding prices paradox as discussed in he lieraure. 8 $m 7 6 5 4 3 2 _ F_au_obs F_au_es //2 //2 //22 //23 //24 4 Concluding remarks Figure 3: Observed vs esimaed equilibrium asse values ource: Clarkson Research and auhor s calculaions he lieraure conains several suggesions ha he volailiy of conracing prices is oo low compared o second-hand values ( sicky prices), and his is ofen inerpreed as a sign of marke inefficiency or a resul of exernaliies. However, as long as a newbuilding and an exising ship are largely fungible, differing only in he ime unil hey generae revenue, equilibrium prices will adap such ha invesors are indifferen beween alernaives. In his paper we have derived a single-equaion mahemaical relaionship defining conemporaneous equilibrium prices across he four shipping markes: he newbuilding marke he secondhand marke he demoliion marke and he freigh marke. Our equilibrium equaion requires raher general condiions, namely ha (i) all second-hand ships are priced based on linear deprecaion down o scrap value, (ii) expecaions of he fuure scrap value and lifespan of he vessel are equal o curren values and (iii) he erm srucure of (imecharer) freigh raes is observable. We use his equilibrium relaionship o illusrae he presence of a erm srucure of newbuilding prices which means ha newbuilding prices are no comparable across ime because he parameers of he newbuilding conracs (primarily he delivery lag and paymen schedules) will vary. When properly accouning for he alernaive cos of operaion and he ime-varying delivery lags we show ha he price for he newbuilding conrac which is merely a kind of fuures conrac - mus necessarily have lower volailiy han ha of exising new ships. While his approach seemingly solves par of he puzzle of sicky newbuilding prices i also leaves a leas one remaining open quesion for fuure research. During very poor freigh markes, such as hose experienced in 22, newbuilding prices appear oo elevaed compared o fundamenals. One way o resolve his is o assume ha asse prices raionally reflec he defaul risk inheren in period imecharers ha hey are no risk free as implicily assumed herein. his opens up he possibiliy of measuring he defaul risk premium in
period imecharers as he residual ha aligns he equilibrium prices in he freigh and asse markes, bu we leave his for fuure research. Anoher possibiliy is measuremen error arising from low liquidiy in boh period imecharer markes and newbuilding markes during imes of freigh marke sress. Given he lack of ransacions, brokers esimaes underlying he imeseries provided may be subjec o greaer noise han usual. References Adland, R. and Koekebakker,. (27), hip valuaion using cross-secional sales daa: A mulivariae nonparameric approach, Mariime Economics and Logisics, 9: 5 8. Adland, R., Jia, H. and randenes,.p. (26), Asse bubbles in shipping? An analysis of recen hisory in he drybulk marke Mariime Economics and Logisics, 8: 223 233. Beensock, M. (985), A heory of ship prices, Mariime Policy and Managemen 2: 25 225. Beensock, M. and Vergois, A. (989), An economeric model of he world shipping marke for dry cargo, freigh and shipping, Applied Economics, 2: 339 356. Dikos, G. (24), Newbuilding prices: Demand inelasic or perfecly compeiive? Mariime Economics and Logisics, 6: 32 32. Dixi A. and Pindyck, R. (994), Invesmen under uncerainy, Princeon Universiy Press, Princeon. Greenwood, R. and Hanson,. (23), Waves in ship prices and invesmen NBER working paper no. 9246 Kaloupsidi, M. (24), ime o build and flucuaions in shipping, American Economics Review, 4(2): 564 68. randenes,.p. (984), Price deerminaion in he ime charer and second hand markes, discussion paper 5/84, Norwegian chool of Economics and Business Adminisraion: Bergen, Norway randenes,.p. (986), Norship: A simulaion model of markes in bulk shipping, discussion paper /86, Norwegian chool of Economics and Business Adminisraion: Bergen, Norway randenes,. P. (22), Economics of he Markes for hips. he Handbook of Mariime Economics and Business, 86-22. solakis,., Cridland, C., and Haralambides, H. (23), Economeric modelling of second-hand ship prices, Mariime Economics and Logisics, 5: 347 377. Zanneos, Z. (966), he heory of oil ankership raes, MI Press: Cambridge, MA.