How To Evaluate Recreation Benefits With Reference Dependent Preference

Journal of Econoics and International Finance Volue 8 Nuber January, 014 Volue ISSN 1993-833 7 Nuber 8 August 015 ISSN 006-981

ABOUT JEIF The Journal of Econoics and International Finance (JEIF) is published onthly (one volue per year) by Acadeic Journals. Journal of Econoics and International Finance (JEIF) is an open access journal that provides rapid publication (onthly) of articles in all areas of the subject such as econoetrics, trade balance, Mercantilis, Perfect copetition etc. The Journal welcoes the subission of anuscripts that eet the general criteria of significance and scientific ecellence. Papers will be published shortly after acceptance. All articles published in JEIF are peer reviewed. Contact Us Editorial Office: Help Desk: Website: Subit anuscript online jeif@acadeicjournals.org helpdesk@acadeicjournals.org http://www.acadeicjournals.org/journal/jeif http://s.acadeicjournals.e/

Editors Prof. Nathalie Jeanne Marie HILMI Professor of Econoics and Finance, International University of Monaco, Hedge Funds Research Institute, 98000 Monte Carlo, Principality of, Monaco. France Prof. Osaah M. Al Khazali Professor of Finance, School of Business and Manageent Aerican University of Sharjah, P.O. Bo 6666, United Arab Eirates, Dr. Guneratne B Wickreasinghe School of Accounting Faculty of Business and Law Victoria University Melbourne Victoria, 8001. Australia Dr. Melte Gürünlü Departent of Econoics and Finance University of Turin, G.Prato, Italy. Prof. Yongrok Choi Departent of International Trade, Inha university, Incheon, Korea. Prof. Mohaed Osan Ahed Bushara Departent of Agricultural Econoics; FAS; Gezira University; P. O. Bo 0; Wad Medani; Sudan. Prof. Anyanwu John Chukwudi Developent Research Departent African Developent Bank 15 Avenue du Ghana BP 33, 100 Tunis Belvedere Tunis Tunisia Prof. S. E. Neaie Departent of Econoics, Institute of Financial Econoics, Aerican University of Beirut, Beirut, Lebanon. Dr. Adrei Vernikov Banking Departent, Higher School of Econoics P.O. Bo 43, Moscow 15057, Russia. Prof. Keith Pilbea Departent of Econoics, City University, Northapton Square, London EC1V OHB. UK.

Editorial Board Dr. Gbadebo Olusegun ODULARU Regional Policies and Markets Analyst, Foru for Agricultural Research in, Africa (FARA), PMB CT 173 Cantonents, Gowa Close, Roan Ridge, Accra, Ghana. Dr ilhan Ozturk Çağ University, Faculty of Econoics and Adinistrative, Sciences, Adana Mersin karayolu, uzeri, 33800, Mersin,Turkey. Professor. Abdelkader BOUDRIGA Professor of finance, School of Econoics and Coerce, Tunis, Tunisia. Dr. Shivakuar Deene Dept. of Coerce and Manageent, Karnataka State Open University, Manasagangotri, Mysore 570 006, Karnataka India. Mohaed Oran The Egyptian Echange, 4 (A) El, Sherifein St, Down, Town, Postal Code 11513, P.O. Bo 358 Mohaed Farid, Cairo, Egypt. Dr. Kola Subair Adjunct Professor, Business and, Financial Econoics, Aerican Heritage University, California, USA. Dr. Bora Aktan Assistant Professor of Finance, Yasar University, Faculty of Econoics and, Adinistrative Sciences, Departent of Finance, Selcuk Yasar Capus, Universite Caddesi, No. 35 37, 35100 Bornova, Izir, Turkey. Dr. Davide Furceri Office of the Chief Econoist, Econoics Departent,, Rue André Pascal, 75775 Paris Cede 16, France. Dr. ABDUL JALIL Wuhan University, Econoics and Manageent School, Wuhan,Hubei, PR China. Prof. Silvia Ciotti Dept of Social and Environental Sciences, St. John International University, Della Rovere Castle Rey Square, 10048 Vinovo (Turin), Italy. Prof. Tessaleno Devezas Advanced Materials and Technological, Forecasting, University of Beira Interior, 600 001 Covilhã, Portugal. Dr. Nikolay Sukholin Autonoous University, Santo Doingo, Doinican Republic. Prof. Dev Tewari Deputy Dean, Faculty of Manageent Studies Professor, School of Econoics and Finance, Westville Capus, University of KwaZulu Natal Resource Econoics, Durban, 4001. South Africa. Dr. Tarek Chebbi Faculty of Law, Econoics and Political Science University of Sousse, Erriadh City 403 Sousse, Tunisia Professor Hichri Walid Gate & Uinversity of Lyon, LAREQUAD Gate, 93 Chein des ouilles, 69130 Ecully France. Dr.Sunderasan Srinivasan Navillu Road 7th Cross, Kuvepunagar, Mysore 57003, India. Dr. P. Malyadri Governent degree College,Osania University Tandur 501141,Rangareddy District India.

Journal of Econoics and International Finance Table of Contents: Volue 7 Nuber 8 August 015 ARTICLES Research A odel for evaluating recreation benefits with reference 167 dependent preference Tadahiro Okuyaa* and Yasuhisa Hayashiyaa Banking sector refors and output growth of anufacturing 183 sector in Nigeria (1970 011) Olanrewaju, Oluwagbenga Gideon*, Areo, Adeleke Gabriel and Aiyegbusi Oluwole Oladipo

Vol. 7(8), pp. 167-18, August, 015 DOI: 10.5897/JEIF015.0681 Article Nuber: 33EDAF754740 ISSN 141-667 Copyright 015 Author(s) retain the copyright of this article http://www.acadeicjournals.org/jeif Journal of Econoics and International Finance Full Length Research Paper A odel for evaluating recreation benefits with reference dependent preference Tadahiro Okuyaa 1* and Yasuhisa Hayashiyaa 1 Departent of Regional Policy, Faculty of Econoics, University of Nagasaki, 13 Kawashio-cho, Sasebo-Shi, Nagasaki, 858-8580; Japan. Faculty of Econoics and Manageent, Tohoku University, 980-8576, Aoba-ku, Sendai, Miyagi, Japan. Received 4 May, 015; Accepted 7 July, 015 In environental valuation studies, it is coonly assued that a utility arises fro an absolute aount of environental quality. This criterion, called absolute evaluation, is used in ethods including the travel cost ethod and the contingent valuation ethod. Studies in eperiental econoics, however, have indicated that an individual s criterion depends on reference dependent preference (RDP) a relative evaluation rather than absolute evaluation. This criterion is used ainly in analysis of biases such as fraing effects or brand choice. The purpose of this paper is to construct a odel for evaluating recreational benefit with RDP. The odel focuses ainly on RDP for an environental quality so as not to conflict with the aio of choices, and the travel cost ethod is used as the odel s basis. First, a structure of utility function is discussed and the benefit with RDP is defined and analyzed based on the relation between the level of RDP and the agnitude of the benefit. Second, the calculating forula of the benefit is derived by the integrating-back ethod and tests for consistency between the results of static analysis and the nuerical eaple are perfored. Key words: Benefit analysis, environental quality, reference dependent preference, travel cost ethod. INTRODUCTION In environental econoics, it is a coon assuption that a utility arises fro the aounts of consuing a good and of an environental quality (Freean III et al., 014). For eaple of the environental quality, nitrogen or sulfur dioide is used as inde of air quality; biocheical oygen deand or phosphorus as water quality; area or a nuber of spices as forest or wetland qualities. Here, the weak copleentarity assuption that the increent of an environental quality leads to the increent of the aount of deand enables researchers to easure the positive or negative benefit of the environental quality change (Mäler, 1974). Thus, ost environental valuation studies have eployed this assuption for valuation ethods (Bockstael and McConnell, 007; Freean III et al., 014). The travel cost ethod (hereafter TCM) is representative of this *Corresponding author. E-ail: okuyaa@sun.ac.jp. Tel./Fa: +08-956-47-611. JEL classification: Q6, Q51. Authors agree that this article reain peranently open access under the ters of the Creative Coons Attribution License 4.0 International License

168 J. Econ. Int. Finance tendency because it relies on the assuption that there is a closed relationship between the aount of the recreational deand and the level of recreation site s environental quality (Shrestha et al., 00; Herriges et al. 004; Phaneuf and Siderelis, 003; Whitehead et al., 009). However, soe eperiental studies in econoics and psychology pointed out that an individual s decision aking is influenced by reference dependent preference (RDP), which has three characteristics. First, a visitor s behavior is influenced by facts including the reference point constructed by that visitor s previous purchase, knowledge, or initial endowent of a good. Second, a visitor s utility (value) function ay enter the negative diension (this is called a loss). Therefore, a visitor s preference is non-transitive. Third, a visitor s utility function is conve in the negative diension. This is called loss aversion. The purposes of this paper are 1) to forulate a recreation behavior odel with RDP for environental quality and ) to forulate an application odel for benefit calculation. As for the base odel, the travel cost ethod is eployed. This for of preference was presented in the prospect theory by Kahnean and Tversky (1979) and Tversky and Kahnean (1991). In environental valuation studies, RDP is a plausible cause of biases such as a fraing effect i, a status quo bias ii, and an endowent effect iii. These biases occur ost often when a stated preference ethod (SPM) is used for valuing an environental quality iv. As for the theoretical studies, Munro and Sugden (003) iproved the (prospect) theory of Tversky and Kahnean (1991) and added restrictions to the preference condition to epress an eogenous reference point. Bowan et al. (1999) used a gain-loss function to epress RDP and analyzed the effect when the reference point is endogenously deterined. Kőszegi and Rabin (006) analyzed the effect of the reference level of consuption under uncertainty in the case that the reference points are eogenously deterined. In epirical studies, Batean et al. (1997) tested the ipact of reference dependent preference on the echange of private goods by eperiental ethods, and Herne (1998) estiated the property of loss aversion. Peters (01), Zeisberger et al. (01), and Li and Ling (015) did recent epirical or theoretical studies on the RDP. Barberis (013) describes the review on the prospect theory. These studies assued RDP for the goods, discussed the for (preference structure) of individuals utility (value) function, and perfored epirical tests by eperiental ethods. Soe epirical studies on the visitor behavior with RDP in arkets have eained RDP on prices. The reference point of this RDP is called a reference price, and a visitor purchases an ite as the result of coparing the prices of goods with the reference prices. A fundaental discussion on visitor behavior with RDP is offered by Winer (1986). It is called the reference price odel (RPM) and is based on the assiilation contract theory (Sherif, 1963) and the adaptation level theory (Helson, 1964). A visitor gets a utility fro the difference between the actual price and the reference price. A visitor s deand is also influenced by this difference. Winer (1986), Mayhew and Winer (199), Lattin and Bucklin (1989), Greenleaf (1995), Ren et al. (014) and Kuar (014) conducted epirical studies and confired the reference price effect for purchasing a good. Putler (199) considered a visitor behavior theory for the RPM by using Kalan s (1968) utility function, which includes the prices of goods as a RDP variable (reference price) in its function. Putler (199) considered the forulation of reference price effect in utility function and analyzed the substitution and incoe effect of reference price for the Marshallian and Hicksian deand functions. Putler s (199) forulation of RPM has not been applied for bundles of goods. Thus, at least, it can be assued that a visitor s preference satisfies the transitivity for the aount of goods. Mayhew and Winer (199, 6) eplained the forulations of the internal and eternal reference prices; The internal reference price is the prices stored in eory on the basis of perceptions of actual, fair, or other price concepts, thus, people adapt to the level of past stiuli and judge new stiuli in coparison with the adaptation level. The eternal reference point is the one provided by observed stiuli in the purchase environent. For eaple, Point of purchase shelf tags that contain inforation about suggested retail price or the actual or unit price of another product against which a price can be copared. The arguent on the reference point of RPM relates to the discussion on RDP in eperiental studies in the sense that the internal or eternal reference points are fored whether they are deterined eogenously or endogenously. As the structures of a reference price in RPM, Bell and Bucklin (1999) assued the (internal) reference price as a reference price fro a visitor s previous purchase occasion; Eery (1970), Hardie et al. (1993), and Kalyanara and Little (1994) assued that the present reference point is the weighted average of the past prices of the ite and/or the weighted average between the past price and the individual s past reference point. The above arguents can be suarized as follows. The first is that although RDP has been confired in SPM or eperiental studies, few have considered revealed preference ethods (e.g., TCM) even if RDP is observed in RPM. The second is that although the reference price effect has been epirically indicated, few studies have considered the effect of RPD for the qualities of goods in a visitor s behavior v. In environental valuation studies, because a ain focus point is how changes in quality influence benefits, it is

Okuyaa and Hayashiyaa 169 useful for another analysis to consider recreation behavior with RDP. The ain hypothesis of this study is that a reference point (siilar to RDP) for an environental quality eists and relates closely to recreation behavior. Let us iagine a visitor s decision aking when he or she chooses between recreation sites A and B (e.g., river A and river B), and the sites have siilar qualities as Q a and Q b. Traditional recreation odel assuing a single trip states that if an individual prefers Q a to Q b, he or she will choose to go to site A; his or her utility is defined as uq ( a ). However, visitors can often be heard coplaining 1) This place was not as good as the last one, or, ) I have already been to that place, so let s go to a different one. These situations ean that the visitor does not judge the quality of sites in absolute ters. It is possible to consider that there is a reference point in the visitor s preference structure, e.g., the first case would relate the case that utility arises fro u( Qa Q) where Q is the quality of the previous site, and the second case would relate the case that utility arises fro u( Qa, Qb Q), where Q b is the previously visited site and Q Qb. The organization of this study is as follows. Firstly, a visitor behavior odel with the RDP is considered and analysis focuses on the relationship between RDP and deand, followed by the analysis of the benefits of RDP defined following the concept of welfare easures, and static analyses for the relationship between RDP and benefit. These analyses focus on the relationship between the position of the RDP (gain or loss) and the agnitude of the benefit because it is a fundaental consideration siilar to price or incoe in epirical welfare studies. Also, an estiation odel is considered and the total value is derived by an integrating-back approach (Larson 199, Eo and Larson 006, and von Haefen 007), and nuerical eaples are perfored to confir consistency of the estiation odel with the static analysis. Finally, the results and the unresolved issues of this study are discussed. MATERIALS AND METHODS Utility aiization proble with reference dependent preference Forulation of reference dependent preference on environental quality In this section, a visitor s recreation activity with RDP is considered by following the consuer behavior odel forulated by Putler (199). A Main difference is that RDP consists of price in Putler (199) s odel but quality in this odel. First, the forulations of RDP are considered. Let z be the aount of a coposite good and be the nuber of recreation activities for a recreation site. Respectively, the prices of these goods are p and p. Let Q be an environental quality in a site in which a political project is z assued to be ipleented. Finally, a reference point (eactly to say, reference quality) to be copared with Q be considered. In this study, let RQ be the reference quality, and the value gained fro coparing Q and RQ is the relative value. Putler (199), who also odeled the reference price effect vi, broke the visitor s judgent on RDP into three stages. In the first stage, the visitor judges the level of relative value by coparing Q and RQ. In the second stage, the visitor evaluates the level. This eans that the visitor evaluates the degree of relative value before evaluating it as his or her utility. In the third stage, the evaluated relative value is reflected in the visitor s utility function. As for the first stage, let DR, be the doain of relative values and RE be an eleent of DR. RE is a gain when RE DR, and this eans that a visitor judges RE Q RQ 0. RE is a loss when RE DR, and this eans that a visitor judges RE Q RQ 0. Finally, RE is zero when RE Q RQ 0. Let zero be included in gains for the sake of siplicity. In forulations, the notation g eans RE DR, the notation l eans RE DR 0 eans RE 0., and the notation In this study, RQ is assued to be endogenously deterined. For eaple, RQ consists of the average of all (hoogeneous) environental qualities (e.g., quality indicators of a river) that the visitor already knows. Thus, Q RQ 0 eans that the environental quality ( Q ) is judged to be relatively better than the (aggregate) qualities which this visitor has eperienced or knows about. Otherwise, Q RQ 0 eans that the environental quality ( Q ) is judged to be relatively worse than the visitor s eperience or knowledge. Equation (1) represents gain, equation () represents loss, and equation (3) represents a duy function for gain and loss because the visitor cannot eperience gain and loss at the sae tie. g I Q RQ if Q RQ 0 (1) l 1 I Q RQ if Q RQ 0 () 1 I 0 Q RQ Q RQ For the second stage, the evaluation for gain and loss, let E() be an evaluation function. The evaluation functions for gain and loss are represented as equation (4). Eg ( g) Q RQ E( g, l) 0 Q Q El () l Q RQ E ( g) 0,li E ( g) 0, E ( l) 0,li E ( l) 0 (5) g g0 g l l0 l Equation (4) iplies that the evaluation would be different for gain and loss. Equation (5) is the conditions on the liit in which case (3) (4)

170 J. Econ. Int. Finance Figure 1. Eaple of the utility function with RDP. differentiations, the notations U and U are also used for siplicity. The differentiated condition is a general one for z,, Q, and G Twice differentiable; first differentiation is positive, and second one is negative or zero i (i.e. U / z U z 0, / 0 ). The notations of differentiations on other U z U zz functions follow these notations. The property of loss ( l ) is considered by following an eaple of utility function that has the property of loss aversion (Figure 1). The dotted line is the utility function (value function) illustrated in Tversky and Kahnean (1991) and the solid line is the siplified case of the dotted line for an epirical analysis (of deand function) discussed above. In Figure 1, the first derivatives for losses in both cases are the sae as the gain ( Ul 0 ). The second derivative for losses is suarized as ull 0. Finally, differential conditions of a reference quality fro the utility function are that U / RQ u 0 and U / RQ u 0 REDR g REDR for the area of gains, and U / RQ u 0 and REDR U / RQ u ll 0 for the area of losses. Figure 1 also REDR illustrates that the increase of RQ correlates with the decrease of RE. l gg Utility aiization proble Fro equation (6), the utility aiization proble is defined as equation (7). The Marshallian deand function for recreation activity is derived as equation (8). Notice that the deand is zero if a visitor s utility is negative when RE is a loss. Thus, the utility is assued not to be negative even in the case of loss and the deand is a positive value i. Figure. Recreation deand and reference dependent preference. the utility function considered below becoes a traditional utility function that includes the absolute value of environental quality only when the relative level is equal to zero vii. Here the recreation deand is. So if a visitor uses the environental quality ties, then he or she acquires the E() aount of relative value. Let G be total gain and L be total loss defined as G E g () and L E l (). Then, a visitor s utility function is defined as equation (6) viii. Equation (6) iplies that a visitor has preferences based on both absolute evaluation ( Q ) and relative evaluation ( G, L ). Thus, the visitor s utility function becoes the traditional one fro equation (5) when GL 0. U u( z,, Q, G, L) (6) Here, n U / n denotes the n-ties differentiation for variable { z,, Q, G, L, RQ, g, l}. As for the first and second Ma u( z,, Q, G, L) s. t. y pzz p (7) z, ( p, p, y, Q, g, l) z (8) As for the Marshallian deand function, the weak copleentarity defined below is assued to hold for the environental quality ( Q ), RE (even at gains and losses), and the recreation deand. Fro the copleentarity, the Marshallian deand function increases when the environental quality or RE increases. Otherwise the deand decreases when the reference quality increases. To observe the difference between traditional deand functions and this odel, Figure illustrates a relation between the deand and the quality. Let the line fro A to D be the line in which RDP is zero, naely, Q RQ 0 for all points on the line (this deand function is equivalent to the one fro traditional econoic theory since it is equivalent to assue Q RQ 0 with the absolute value of environental quality only fro equation (5)). Net, let C be the point at which Q1 RQ 0, and let B be the other point at which the quality is less than the point C Q Q RQ ). Q0 RQ 0 iplies that RE is at loss. ( 0 1 Thus, the deand at point B is less than at point A because point A is the point at which RE is zero. Siilarly, let E be the point at which the quality is greater than point C ( 1 ). This iplies Q Q

Okuyaa and Hayashiyaa 171 that RE is at gain. Thus, the deand at E is greater than the one at point D because point D is the point at which RE is zero. As a result, the deand function with RDP is the line BCE. As for the case in which Q is fied and RQ increases, the change of the deand is syetrical to the case of Figure because the increase of RQ iplies RE decreases (goes to loss) fro the definition. As a result, the first derivatives of the deand function for RE are ( )/ g 0 and ( )/ l 0, and the second derivatives ( )/ g 0 and ( )/ l 0 are supposed (See Appendi A). So, the properties of RDP in the deand function are assued to be equivalent to those of RDP in the utility function. In epirical studies, Winer (1986) and Putler (199) eployed the linear for and Suzuki et al. (001) eployed the logistic for to estiate deand functions. In addition, Suzuki et al. (001) set ( ( )/ g)/( ( )/ l) 1 to test the loss aversion in the deand function ii. Finally, the indirect utility function is derived as V u z ( ), ( ), Q, g, l v( p, p, y, Q, g, l) and the ependiture function is derived as y e( p, p, U, Q, g, l) z. Utility iniization proble and the Slutsky equation Utility iniization proble By a siilar process, the utility iniization proble is defined as equation (9) and the Hicksian deand function is derived as h h ( pz, p, U, Q, g, l) iii. Equation (10) is assued to hold h for (), (), and e(). Finally, Shephard s Lea h ( y / p ( ) ) is assued to hold. z, Min p z p s. t. U u( z,, Q, g, l) (9) z h p, p, Q, g, l, e( ) (10) z Let the first derivative of environental quality for the Marshallian deand function be equation (11), that for the Hicksian deand function be equation (1), and that for the ependiture function be equation (13). In equation (13), e( )/ Q 0 is assued (Mäler, 1974). The Slutsky equation is derived fro equation (10) as equation (14). In equation (14), the first ter is the substitute effect; the second ter is the incoe effect; and the third and fourth ters are the gain/loss effects. In this study, the gain/loss effect is assued to be positive. h I 1 I Q Q g l z (11) h ( ) ( ) e( ) ( ) ( ) e( ) I Q Q y Q g y g ( ) ( ) e( ) 1 I l y l Static analysis on the Slutsky equation (14) Net the derivatives of RE for each function are considered. As for the indirect utility function, the first derivatives are V / g 0, V / l 0 and the second derivatives assue V / g 0, V / l 0. Those lead to the first derivatives of the ependiture function as y/ g 0, y/ l 0 and to the y/ g 0, second derivatives as y/ l 0 (See Appendi B). Therefore, equation (15) holds. Equation (15) indicates that the aount of ependiture at a loss is greater than at zero, and the ependiture at zero is greater than at a gain. e, l e,0 e, g (15) Net, the derivatives of the reference quality are considered. The first derivative of the Marshallian deand function is equation (16); it is negative because ( )/ g 0 and ( )/ l 0. The first derivative of the ependiture function is equation (17); it is positive because e( )/ g 0 and e( )/ l 0. The first. derivative of the Hicksian deand function is equation (18); whether it is positive or negative depends on the gain/loss effect. If the gain/loss effect is assued to be positive, equation (18) is negative. I 1 I RQ g l y I e( ) 1 I e( ) RQ g l h ( ) ( ) e( ) I RQ g y g ( ) ( ) e( ) 1 I l y l (16) (17) (18) Finally, the total effects on the deands and the ependiture fro the increase of environental quality and the reference point are calculated by equation (19) fro equations (11), (13), (14), (16), (17) and (18). Then, it is suarized as equation (0). h h h h I 1 I Q Q g l y e I e 1 I e Q Q g l (1) (13), Q RQ Q y y e, Q RQ Q h h ( ) ( ) e( ) Q RQ Q y Q (19)

17 J. Econ. Int. Finance h h y y / Q / RQ / Q / RQ Q RQ ( )/ y (0) Corollary 1. If the increase of both the environental quality and the reference point is equivalent, then the increase of (Hicksian) deands and the ependiture are equal to the increase of both function in which assue only absolute value of Q. That is, the arginal benefit with RDP is equivalent to the arginal benefit without RDP (the traditional benefit). Equation (19) iplies that RDP does not influence the aount of deand and ependiture if the environental quality and the reference quality increase or decrease by the sae degree. Equation (0) suarizes the total effect for the arginal benefit. The values of the first and the second parentheses on the right side are equal to the value without RDP. Thus, the benefit with RDP is equivalent to the benefit without RDP. Finally, the choke price and the weak copleentarity are * considered. The choke price p is defined as equation (1), which iplies that the choke price is the price at which the Hicksian deand is zero. Notice that the Hicksian deand includes a gain and a loss. Net, the weak copleentarity is generally defined as ( ) / * 0. However, the ependiture function of this odel e Q p includes the gain and loss effect as equation (13). The odified version of the weak copleentarity is defined as equation (). p * h in p ( ) 0 (1) e( ) e( ) e( ) I 1 I 0 Q * g * l * p Definition of benefit and static analysis Definition of total value p p () It is necessary to differentiate between situations in which a project is ipleented and those in which no project is ipleented to define the benefit fro the change of an environental quality with RDP iv. Let s wo be the superscript representing the quality level at which a project is ipleented and s w be the one at which no project is ipleented. Using the notation, the utility function (and other variables) are rewritten as U s u z s, s, Q s, G s, L s. Note that a visitor does not eperience a gain and a loss at the sae tie, so RE differs based on whether or not the project is ipleented. For eaple, in one case, RE gains when s wo and it will also gain when s w ; however, in another case, RE decreases when s wo and will be at zero when s w. The benefit fro quality change is defined by equivalent variation (EV) as equation (7) and copensating variation (CV) as equation (9). EV and CV can be decoposed into three kinds of benefit fro incoe change (equations (5) and (9)), and benefit fro quality change (equations (6) and (30)). This study eaines only the benefit fro quality change v. Thus, the total value of environental quality (hereafter TV) is suarized as equation (31) vi. Equation (31) iplies EV when s w and CV when s wo. wo wo w wo wo wo wo wo wo wo wo wo EV e( p, p, U, Q, g, l ) e( p, p, U, Q, g, l ) z z (3) wo wo w wo wo wo w w w wo wo wo e( p, p, U, Q, g, l ) e( p, p, U, Q, g, l ) z z z z (4) w w w w w w wo wo wo wo wo wo e( p, p, U, Q, g, l ) e( p, p, U, Q, g, l ) z z (5) w w w wo wo wo w w w w w w e( p, p, U, Q, g, l ) e( p, p, U, Q, g, l ) (6) w w w w w w w w wo w w w CV e( p, p, U, Q, g, l ) e( p, p, U, Q, g, l ) z z wo wo wo w w w w w wo w w w e( p, p, U, Q, g, l ) e( p, p, U, Q, g, l ) z z w w w w w w wo wo wo wo wo wo e( p, p, U, Q, g, l ) e( p, p, U, Q, g, l ) z z z z (7) (8) (9) wo wo wo wo wo wo wo wo wo w w w e( p, p, U, Q, g, l ) e( p, p, U, Q, g, l ) (30) s s s wo wo wo s s s w w w TotalValue e( pz, p, U, Q, g, l ) e( pz, p, U, Q, g, l ) (31) Equation (31) includes the quality itself ( Q ),the gain ( g ), and the loss ( l ). Equation (31) is thus a coprehensive forulation wo w including the absolute evaluation (the evaluation for Q Q ) and the relative evaluation (the evaluation for wo wo w w ( g, l ) ( g, l ) ). Influence of RDP for total value Let p z, p, y, and Q (the part of absolute evaluation in utility function) be fied when s wo and s w. As for locations of RE, there are three possible areas: gain, zero, and loss. Thus, there are 3 3 patterns to deterine the value of TV (e.g., wo w TV e(, g ) e(, g ), (,0 wo w TV e ) e(, l ) ). Figure 3 shows these cases with a possible ependiture function. Let the origin of the arrow line correspond to the aount of ependiture when the wo wo in equation (31)) and project is not ipleented (i.e., e(, g, l ) the end point of the arrow line correspond to the aount of ependiture when the project is ipleented (i.e., (, w w e g, l ) ). For eaple, the third case indicates the aount of ependiture wo, and the eighth case indicates change fro e(, l ) to e(, l w ) wo w the aount of change fro e(, l ) to e(, g ). In these cases, the second and fourth, third and fifth, sith and eighth, and seventh and ninth cases ean the sae changes of TV because each condition differs only in ters of whether the RE benefits: benefit fro price change (equations (3) and (8)), goes fro low to high or fro high to low. Thus, the first, fourth, fifth, sith, and seventh cases are considered. The case of zero to zero (first case) eans there is no effect on the RE. The case of zero to gain (fourth case) eans the increase of the benefit, and the case of zero to loss (fifth case) eans the decrease of the benefit. The case of loss to zero (sith case) eans the increase of the benefit, and the case of gain to zero (seventh case) eans the decrease of

Okuyaa and Hayashiyaa 173 Table 1. Benefit and positions of RE With Without Gain Zero Loss Gain Increase Increase Increase Zero Decrease No Effect Increase Loss Decrease Decrease Decrease s w since the difference of ependiture is sallest (i.e., the ninth case in Figure 3). This case is interpreted to be a situation in which a visitor judges an objective environental quality to be better than other qualities when s wo, and judges it to be worse when s w. One reason that a visitor ight judge the quality to be worse despite environental iproveent could be that if other qualities are also iproved at sae tie and those are ore ipressive to the visitor, the reference quality would increase ore than the objective quality would increase. Another reason for such a judgeent could be a gap between the quality change achieved by the project and the quality change the visitor iagines when s wo. Therefore, the result of the project would have a negative ipact for the visitor. The benefit fro this situation is denoted by TV and defined as equation (34). GL s s s wo wo s s s w w TV e( p, p, U, Q, g ) e( p, p, U, Q, l ) (34) GL z z Figure 3. Nine cases of relative evaluation. the benefit. These results are suarized in Table 1. Let us consider the iplications of these benefits. In the cases in which the benefit increases, TV increases ost in the case in which RE changes fro a loss when s wo to a gain when s w since the difference of ependiture is biggest (that is, the eighth case in Figure 3). This case is interpreted to be a situation in which a visitor judges an objective environental quality to be worse than other qualities (which are coprehensively denoted by RQ ) when s wo, and judges it to be better when s w. In short, the ipact of environental quality change on a visitor s RDP is very big. The benefit is denoted by TV and defined as equation (3). LG s s s wo wo s s s w w TV e( p, p, U, Q, l ) e( p, p, U, Q, g ) (3) LG z z The case in which RE is zero when s wo and s w is interpreted to be a situation in which a visitor judges the quality to be equivalent to other qualities when s wo and s w because the visitor s reference quality ( RQ ) changes to the sae degree as the quality change (Corollary 1). One eaple of such a situation would be of another project being ipleented for other environental quality at the sae tie. The benefit fro this situation is denoted by TV 00 and defined as equation (33). s s s wo s s s w TV00 e( pz, p, U, Q,0) e( pz, p, U, Q,0) (33) In cases in which TV decreases, TV decreases ost in the case in which RE changes fro a gain when s wo to a loss when These results are suarized in equation (35). In addition, in the case of environental deterioration, these inequalities becae TV TV TV ). Equation (35) iplies that 1) reverse (i.e. LG 00 GL the total value that is defined only by the absolute value of quality TV ( 00 ) would be a part of the values and ) there is a possibility that the total value can be negative even if the project ais to iprove quality because there is no restriction on the reference quality change. TV TV TV (35) LG 00 GL Decoposition of use and non-use values and an interpretation of non-use value The decoposition of use and non-use value is perfored to investigate the relation between non-use value and RDP. Equation (31) decoposes TV into use value (equation (36)) and non-use value (equation (37)) by using the choke price (Neil, 1988; Larson, 199). s s s wo wo wo s s s w w w TV e( p, p, U, Q, g, l ) e( p, p, U, Q, g, l ) z z s * s w w w s s s w w w e( pz, p, U, Q, g, l ) e( pz, p, U, Q, g, l ) e p p U Q g l e p p U Q g l s * s wo wo wo s s s wo wo wo ( z,,,,, ) ( z,,,,, ) s * s wo wo wo s * s w w w ( z,,,,, ) ( z,,,,, ) (36) e p p U Q g l e p p U Q g l (37) Let us consider the definition of non-use value. Generally, the nonuse value is defined fro the properties of uniqueness and irreversibility (Krutilla, 1967). Regarding uniqueness, equation (37)

174 J. Econ. Int. Finance Table. Paraeters. Q g l p y wo Q -0.9534-0.0163 0.035 0.003 0.007 0.056 9.91.384 14 17 w Q includes the reference point so that the environental quality is copared with others. There is a possibility that the uniqueness will not hold. However, Krutilla (1967) stated that uniqueness is not necessary condition for his arguent. One reason is that there is a possibility that a siilar environental quality eists in another arket which is difficult to access. Thus, a visitor can copare the objective environental quality viii of one place with others even if the other places are out of reach. If the eistence value defined only by absolute evaluation (in this study, it is the case in which RE 0 ) is the pure eistence value, equation (37) would interpret the ipure eistence value. Integrating-back approach Von Haefen (007) presented three ethods to estiate the total value of environental quality by arket data. In this paper, the integrating-back approach is eployed viii. This approach is useful for obtaining the total value fro a deand function. The central idea is to derive the quasi-ependiture function, which was developed by LaFrance (1985). Recently, Eo and Larson (006) presented an estiation odel based on the integrating-back approach. This ethod can calculate the use and non-use value fro arket data such as the forulations derived below. Let the deand function be equation (38). Here Q Q I g (1 I) l. As for the estiation, visit Q g l nuber ( ), travel cost ( p ), incoe ( y ), and environental quality ( Q ) are observed in the recreation arket. In addition, other environental qualities ust be accounted for to deterine the reference quality. It is necessary to research the data of visitors knowledge about other qualities, or their eperience of sites they have visited. Then the reference quality is constructed as shown above and the data of gain ( g ) or loss ( l ) are calculated i. ln[ ( )] p y Q (38) The quasi-ependiture function is equation (39), where the constant of integration is Uep Q. The (indirect) utility function is equation (40). Let the price and the incoe be fied for siplicity, and the notation s be oitted in these variables. Then the total value is derived as equation (41), where s s ep Q p y Q. TV consists of the deand s s s ( ( p, y, Q ) ) and the quality ( s Q ). Finally, TV is decoposed into non-use value (NUV) as in equation (4) and use value (UV) as in equation (43). / ep p Q UepQ 1 e p, U, Q ln (39) U (1/ )ep( y) (1/ )ep( p Q) ep( Q) w wo 1 Q 1 ln Q TV w wo ep Q Q w wo NUV Q Q (40) (41) (4) 1 ln w 1 wo UV TV UV Q Q Paraeters for benefit calculation (43) A project concerning quality iproveent is assued. The siulation is focused only on the total value because non-use value is defined as erely the difference of quality change i. Paraeters are specified in Table. The paraeters,, Q,, p, and y are the sae as those in Tables 1 and of Eo and Larson (006) (the results of estiation odel for non-use value). The variables g and l are originally designed so as to satisfy the property of RDP of deand function ( ( / g)/( / l) 1)). For the quality level, Eo and Larson (006) used biocheical oygen deand (BOD) ii for the wo w estiation; Q 14 and Q 17 are designed as these levels. Since positive utility does not arise for values ( Q ) under 10 in this odel, the quality levels and reference qualities are set at values over than 10. Therefore, this foration cannot be used for arbitrary values of paraeters. The siulation is thus perfored for reference quality ( RQ ) ranging fro 10 to 0 in one-point increases. RESULTS AND DISCUSSION The siulations were perfored by each functions (equations (38) to (41)). The ain focus of the discussion is to eaine the differences between the traditional benefit calculation odel and the odel of this paper by analyzing the relations between RQ and TV. Deand function Figure 4 shows the change of the deand corresponding

Okuyaa and Hayashiyaa 175 Deand levels 0.4 0.4 0.38 0.36 Utility levels 3.6 3..8.4 0.34 10 1 14 16 18 0 Reference point Q=14 Q=17 Figure 4. Reference quality and deand function. 10 1 14 16 18 0 Reference quality Figure 6. Reference quality and utility function. Ependiture levels 4 3 1 0 10 1 14 16 18 0 Reference quality Figure 5. Reference quality and ependiture function. to each reference point. The gray line is the deand wo function when the quality level is Q and the black line is w the deand function when the quality level is Q. The point 14 for the gray line and the point 17 for the black line are the points at which RE 0 (i.e., the inflection points for each function). The inflection points are not discussed in previous section. However, this forulation is eployed because it is coonly used in RPM odels (for the forulation of deand function) and eperiental studies (for the forulation of value function). In previous section, the relation between the deand and RE was analyzed. The deand decreases when RE becoes negative in Figure. Figure 4 shows both deand levels decrease corresponding to the increase of RQ (the increase of RQ eans that RE becoes negative). This feature is the sae in Figure. Ependiture and utility function The ependiture function (equation (39)) is shown in Figure 5. In previous section, the condition y / RQ 0 is discussed (equation (17)) and it is reflected in Figure 5. Here, the U in the ependiture function is the utility at point 14 in Figure 6. As for the property of RDP, the gradient at the loss is greater than the gradient at the gain. Siilarly, the condition U / RQ 0 for the (indirect) utility function is reflected in Figure 6. Especially, the properties of loss aversion (li g0ug)/(li l0ul) 1 are observed. That is, the properties of loss aversion in the deand function (e.g., g 0.03 and l 0.07 ) are also reflected in the utility function. Benefit calculation Figure 7 shows the change of TV corresponding to each level of reference point. The black line shows the case in which RE in the second ter of equation (31) is fied at zero and RE in the first ter of equation (31) changes (eactly to say, RQ in the first ter of equation (31) changes) iii. As a result, the total value decreases

176 J. Econ. Int. Finance Total value Figure 7. Reference quality and total value. Table 3. Eaple for Table 1 fro Figure 5. With Without Gain Zero Loss Gain [8,10] [7,10] [5,10] Zero (1,3) (3,3)[7,7] [5,7] Loss (1,6) (3,6) (6,9) Total Value 3.5 1.5 1 0.5 0 3.5 1.5 1 0.5 0 10 1 14 16 18 0 Reference quality RE=0(without) RE=0(with) 1 3 Cases Figure 8. Relative evaluations and size of total value. following the increase of RQ. This case corresponds to the fourth case in Figure 3. The increase of RQ in the first ter eans that the value (the aount of ependiture) of the first ter of equation (31) decreases, so the difference between the first and second ters of equation (31) is a decrease. The black line also shows the relation between RE and the total value in Table 1. Let (row, colun) be the eleent in the atri corresponding to the row and colun of Table 1. (e.g., (zero, loss) eans decrease in Table 1). In Figure 5, for eaple, the difference between the value at point 3 and that at point 6 (hereafter abbreviated as (3, 6)) corresponds to (zero, loss) if the wo value at 3 is the one when Q Q and the value at 6 is w the one when Q Q. Siilarly, other situations can be considered (e.g., (6, 9) corresponds to (loss, loss)). Other eaples are listed in Table 3. The gray line in Figure 7 shows that RE in the first ter of equation (31) is fied and RE in the second ter of equation (31) changes. As a result, the total value increases following the decrease of RQ. This case corresponds to the sith case in Figure 3. The increase of RQ in the second ter of equation (31) eans that the value of the second ter decreases, so the difference between the first and second ters of equation (31) is an increase (the total value increases). The gray line also shows the relation in Table 1. Let iv [row, colun] be the eleent in the atri corresponding to the row and colun of Table 1 (e.g., [zero, gain] eans increase fro Table 1). In Figure 5, for eaple, the difference between the values at point 7 and point 10 (abbreviated as [7, 10] below) corresponds to [zero, gain] if the value (the aount of ependiture) at wo 7 is the one when Q Q and the value at 10 is the one wo when Q Q. Siilarly, other situations can be considered (e.g., [8, 10] corresponds to [gain, gain]). Other eaples are listed in Table 3. Figure 8 shows the relation in equation (35). The situations (the aount of deand and environental wo w qualities when Q Q and Q Q ) are given in Figure 7. Case 1 indicates the value of TV LG when the first ter of equation (3) is set as point 0 on the gray line and the second ter of equation (3) is set as point 10 on the black line in Figure 7. Siilarly, Case indicates the value of TV 00 when the first ter of equation (33) is set as point 14 on the gray line and the second ter of equation (33) is set as point 17 on the black line. Case 3 indicates the value of TV GL when the first ter of equation (34) is set as point 0 on the gray line and the second ter of equation (34) is set as point 10 on the black line. The result of equation (35) is confired. Figure 9 shows the price change and the total value. The black line is the sae as in Figure 7, the gray line is the line when the price of the black line set 0, and the dotted line is the line when the price of the black line set 40. The total value decreases when the price increases.

Okuyaa and Hayashiyaa 177 Total value Figure 9. Price change and total value. Total value 3.5 1.5 1 0.5 0.5 1.5 1 0.5 0-0.5-1 10 1 14 16 18 0 reference quality TV(without=0,p=9.91) TV(without=0,p=0) 10 14 18 6 30 Reference quality TV(without=0) Figure 10. Etree case of reference quality. Finally, a proble on calculation is discussed. It is called the etree oving to loss. Figure 10 shows the case (by using the black line fro Figure 7) in which the reference quality changes etreely to loss. In this case, the value goes to negative such as at point 30 despite the project s ai of environental iproveent. This happens because there is no restriction on the value of reference quality for s wo and s w (therefore, this case occurs in the case of the gray line, e.g., if RQ goes below 10). Thus, a boundary condition should be set for the reference quality or preference structures. Estiation of deand function Final section described how to estiate the deand function represented as equation (38). In estiating, an independent variable is the visit nuber ( ); dependent variables are the travel cost ( p ) and household incoe ( y ) observed in the recreation arket (or collected by a survey). The data on environental quality ( Q ) in recreational sites are also used for a dependent variable. Here, the reference quality data ( RQ ) would be collected by asking respondents; For eaple, how uch quality level ( RQ ) of the environent do you need? If RQ Q, the respondents would be categorized as gainrespondents. In the inverse case, loss-respondents. Researchers would arbitrarily decide how to categorize the case of RQ Q ; either include it in gain-respondents or loss-respondents. Whether respondents preferences are the absolute or relative valuation would be eained by coparing the deand functions; ( p, p, y, Q) and ( p, p, y, Q, g, l). z Conclusion Most environental valuation studies have eployed absolute evaluation assuption for valuation ethods. However, soe eperiental studies in econoics and psychology pointed out that an individual s decision aking is influenced by reference dependent preference, naely relative evaluation. The purposes of this paper are 1) to forulate a recreation behavior odel with RDP for environental quality and ) to forulate an application odel for benefit calculation. As for the base odel, the travel cost ethod is eployed as discussed below. First, the odeling of RDP in a visitor s utility function and the properties of RDP on deand and ependiture functions were eained. The analysis for the deand function revealed that the siultaneous changes of the environental quality and the reference quality are equivalent to the condition in which only the absolute evaluation is considered. This iplies that soe effects (e.g., a fraing effect) of RDP for the value of the benefit arise only when the quality and the reference point change in different directions. Second, the definitions of the benefits and static analysis were considered. The finding was that the total value defined only by absolute evaluation is one of the z

178 J. Econ. Int. Finance benefits that include RDP. Thus, it is necessary to deterine whether the benefit (or, in ters of this study, visitor behavior) includes RDP or not. If the benefit includes RDP, then the benefit can change depending on the reference points. This paper copare the benefits 1) relative evaluation is loss before ipleenting a project and relative evaluation is gain after ipleenting a project, ) relative evaluation is zero before ipleenting a project and relative evaluation is zero after ipleenting a project, 3) relative evaluation is gain before ipleenting a project and relative evaluation is loss after ipleenting a project. As a result, the benefit of first case is bigger that the second case and the second is bigger than the third case (equation (35)). Third, this property was confired though siulations. Since the coputable forulation reflects the theoretical findings, this odel can be used to estiate the recreation deand function with RDP and to calculate the benefit. Finally, soe probles for epirical study should be entioned. First, the recreation deand for a single site is assued in this paper. However, the structure of reference quality needs the aggregation of other qualities that a visitor has already eperienced or knows about. Thus, it is natural to assue there are ultiple sites for recreation deand. One solution could be to use the ultiple-site trip for travel cost ethod, e.g., the Kuhn- Tucker Model. Second, the loss effect cannot be estiated if there is no reservation utility (discussed in theoretical analysis). Since the loss effect eans that the utility is at a negative value, it is possible for a visitor not to select such recreation sites. Third, if the influence of RDP is confired, there is a possibility that the benefit will change over a long period of tie due to the change of the reference quality. Thus it ay be necessary to consider the structures of RDP for a dynaic odel. Notes i. The fraing effect is a phenoenon in which an individual s preference changes depending on how options are presented (fraed) in a questionnaire (Tversky and Kahnean, 1991). ii. The status quo bias is a phenoenon in which an individual tends to prefer to reain at the status quo due to an aversion toward loss (Kahnean et al., 1991). iii. The endowent effect is a phenoenon in which an individual feels that a good has a higher value once he or she has becoe the owner of the good. This effect has been eplained as being equivalent to status quo bias (Kahnean et al., 1991). However, since it is not clear in these studies whether or not the property right is the ain coponent of the status quo, these two biases are eplained separately. iv. These biases are discussed etensively in the proble of the disparity between willingness to pay and willingness to accept. Mitchell and Carson (1989) present cases of this disparity, and RDP is one of the. Haneann (1991) deonstrated the cause theoretically without RDP. However, recent studies indicate RDP as the ain cause of endowent effect (Horowitz and McConnell, 00; Plott and Zeiler, 005; Brow, 005). v. Suzuki et al. (001) estiated the deand function with RDP for services qualities in the airline arket. vi. Putler (199) odeled the reference price effect as ref follows: let p be the price of a good, p be the reference ref price of the good. Then the gain is p p 0, and the ref loss is p p 0. This iplies that a visitor gains a utility if the price is ore inepensive than the reference point. vii. Putler (199) set E( ) 0. However, it is generally assued that Eg ( g) 0 and El ( l) 0 in the utility function with RDP. Thus, those conditions are eployed in such studies as Munro and Sugden (003). viii. In addition, since G and L are directly involved in the utility function, it ay be useful to construct a utility function to assue functions F and F such as G F ( ) E ( g) and L F ( ) E ( l). An eaple of the g g Utility function is, where l l g F E g g and 1/ g( ), g( ) 0.5 1/ Fl( ), El( l) l. i. As for the notation, the first derivatives are denoted as U z u, the second derivatives are denoted as / z U / z uzz U / z uz, and the cross derivatives are denoted as. Note that u gl is not defined.. Foral conditions are based on a odified version of Bowan et al. (1999). The first is that U is strictly increasing in RE. The second is to epress the relation between the arginal utility of a loss and the arginal utility of a gain, defined as U ( RE) U ( RE) U( RE) U( RE) for 0 RE RE. The third is to represent an assuption of diinishing arginal sensitivity defined as RE is strictly concave in RE 0 and RE is strictly conve in RE 0. The fourth is to represent that a person can evaluate losses even when coparing very sall losses to very sall gains, given that there eists a value M s.t. li 0( u ) l RE RE REDR /( ure REDR ) M. In addition, if RE is the linear for, a siple condition to epress loss aversion is li g0ug/lil0ul 1. i. It is natural to think there is a value of RE at which the value of the utility becoes positive if RE eceeds the value ( RE RE U 0 ). This iplies that it is necessary to assue a reservation utility for a recreation activity ( ) if a recreation activity occurs even in the case that RE is at loss. This study assues that the absolute value of Q in the utility function will perfor the role.

Okuyaa and Hayashiyaa 179 Thus, u(, Q, L) 0 for RE RE is assued. ii. Let be each paraeter. For eaples, a linear for is p p y y QQ gg ll and a logistic for is 1/{1 ep( p y Q g l)} p y Q g l in the notation of this paper. The condition to epress the loss aversion as / 1 for the linear for and g l ( ) ( ) / 1 for the logistic for. This g l g l iplies that these deand functions ust reflect the properties of utility function (although those would not be reflected eactly). iii. In addition, the ependiture function is defined as h h y pzz p by using the Hicksian deand function. iv. Whether the project is for environental iproveent or for environental deterioration does not atter. The difference corresponds to the definition of willingness to pay and willingness to accept. However, the project is ainly assued to be an environental iproveent project in this study, as discussed below. v. An additional decoposition is shown in Appendi C. vi. In epirical studies, equation (31) would be rewritten wo Q astv e( ) / Qd w. It would be necessary to Q consider the kink at RE 0 if the doain of RE changes between s wo and s w. Thus, equation (31) is the definition when e() is differentiable at RE 0. vii. He also cannot access to the objective environental quality in the definition of equation (37). The reason why he cannot access is due to eploying the choke price to define the equation. See equation (1). The choke price is defined as the price at which (Hicksian) deand is zero. This interpreted as being a situation in which a visitor cannot visit the recreation site because of high travel costs. viii. The related papers concerning this ethod are Neil (1988) and Bullock and Minot (006). The probles are 1) there is no guarantee to derive the choke price fro arbitrary functional for to deterine the use value and non-use value, and ) it is difficult to apply the ethod to the case of ulti-site trips. i. For eaple, assue that the reference quality coes fro the average of the qualities that a visitor has known. If visitor A who plans to go to site 1, which has quality= knows site s quality=3, site 3 s quality=1, and site 4 s quality=5, then the visitor s reference quality is 11/4=.75 and RE =-.75=-0.75. Thus, visitor A s RDP is at loss. If visitor B, who plans to go to site 5, has RE =5-.75=.5, the RDP is at gain. For a data set, it is siple to set if a visitor s RE is at gain, then his or her data of loss is zero (Suzuki et al., 001).. Eo and Larson (006) assued the constant of Uep Q. is used to estiate the integration as eistence of non-use value. Thus, the odel in this study assues 1 and adds RE to their odel. i. In addition, the result of analysis for eistence value are siilar one for the total value since the difference of both definition (total value and the non-use value) are only the definition of price and the prices are fied in each definition (See equation (31) and equation (37)). ii. BOD is an indicator of water pollution levels and it is generally assued to have a negative effect on recreation deand. However, Eo and Larson (006) used negative values of BOD for estiation. Thus, the estiated paraeter is positive and is directly used for in this study. Q iii. In this case, the absolute value of the environental quality is the value at point 14 because RE of the second ter is fied at zero. Siilarly, the absolute value of the environental quality is the value at point 17 in the gray bars. iv. Note that parentheses ( ) is used for black line and brackets [ ] is used for gray line. Conflict of Interests The authors have not declared any conflict of interest. REFERENCES Barberis NC (013). 'Thirty Years of Prospect Theory in Econoics: A Review and Assessent', J. Econ Perspect. 7(1):173-196. Batean I, Munro A, Rhodes B (1997). 'A Test of the Theory of Reference-Dependent Preferences', Q J Econ. 11():479-505. Bell DR, Bucklin RE (1999). 'The Role of Internal Reference Points in the Category Purchase Decision', J Consu Res. 6():18-143. Bockstael NE, McConnell KE (007). 'Environental and Resource Valuation with Revealed Preferences: A Theoretical Guide to Epirical Models', Springer. New York. Bowan D, Minehart D, Rabin M (1999). 'Loss Aversion in a Consuption Savings Model', J. Econ. Behav. Organ. 38():55-178. Brow TC (005). 'Loss Aversion without the Endowent Effect, and Other Eplanations for the WTA WTP Disparity', J. Econ. Behav. Organ. 57(3):367-379. Bullock DS, Minot N (006). 'On Measuring the Value of a Nonarket Good Using Market Data', A. J. Agr. Econ. 88(4):961-973. Eery F (1970). 'Soe Psychological Aspects of Price', In Taylor B, Wills G (eds.) Pricing Strategy. Brandon/Syste Press pp.98-111. Eo YS, Larson DM (006). 'Iproving Environental Valuation Estiates through Consistent Use of Revealed and Stated Preference Inforation', J. Environ. Econ. Manage. 5(1):501-516. Freean III AM, Herriges JA, Kling CL (014). 'The Measureent of Environental and Resource Values: Theory and Methods (third edition)', Resource for the Future Press. Washington, DC. Greenleaf EA (1995). 'The Ipact of Reference Effect on the Profitability of Price Prootions', Market Sci. 14(1):8-104. Haneann WH (1991). 'Willingness to Pay and Willingness to Accept: How uch can They Differ?', A. Econ. Rev. 81(3):635-647. Hardie BGS, Johnson EJ, Fader PS (1993). 'Modeling Loss Aversion and Reference Dependence Effects on Brand Choice', Market Sci. 1(4):378-394. Helson H (1964). 'Adaptation-Level Theory', Harper and Row. New York. Herne K (1998). 'Testing the Reference-Dependent Model: An Eperient on Asyetrically Doinated Reference Points', J. Behav. Decis. Making. 11(3):181-19. Herriges JA, Kling CL, Phaneuf DJ (004). What s the use? Welfare estiates fro revealed preference odels when weak copleentarity doesn t hold. J Environ Econ Manage. 47: 55-70.

180 J. Econ. Int. Finance Horowitz JK, McConnell KE (00). 'A Review of WTA / WTP Studies', J. Environ. Econ. Manage. 44(3):46-447. Kahnean D, Tversky A (1979). 'Prospect Theory an Analysis of Decision under Risk', Econoetrica 47():63-91. Kahnean D, Knetsch JL, Thaler RH (1991). 'Anoalies: The Endowent Effect, Loss Aversion, and Status Quo Bias', J. Econ. Perspect. 5(1):193-06. Kalan PJ (1968). 'Theory of Visitor Behavior When Prices Enter the Utility Function', Econoetrica 36(3/4): 497-510. Kalyanara G, Little JDC (1994). 'An Epirical Analysis of Latitude of Price Acceptance in Visitor Package Goods', J. Consu. Res. 1(3):408-418. Kőszegi B, Rabin M (006). 'A Model of Reference-Dependent Preferences', Q. J. Econ. 11(4):1133-1165. Krutilla JV (1967). 'Conservation Reconsidered', A. Econ. Rev. 57(4):777-786. Kuar MPS (014). 'Brand Preference and Buying Decision: A Study with Reference to Organized Indian Edible Oil Brands', Afr. J. Mark. Manage. 6():17-6. LaFrance JT (1985). 'Linear Deand Functions in Theory and Practice', J. Econ. Theory. 37(1):147-166. Larson DM (199). 'Further Results on Willingness to Pay for Nonarket Goods', J. Environ. Econ. Manage. 3():101-1. Lattin JM, Bucklin RE (1989). 'Reference Effects of Price and Prootion on Brand Choice', J. Market. Res. 6(3):99-310. Li X, Ling W (015). 'How Fraing Effect Ipact on Decision Making on Internet Shopping', Open J. Bus. Manage. 3:96-108. Mayhew GE, Winer RS (199). 'An Epirical Analysis of Internal and Eternal Reference Price using Scanner Data', J. Consu. Res. 19(1): 6-70. Mäler KG (1974). 'Environental Econoics: A Theoretical Inquiry', John Hopkins University Press; Baltiore. Mitchell RC, Carson RT (1989). Using Surveys to Value Public Goods: The Contingent Valuation Method, Resource for the Future Press, Washington. Munro A, Sugden R (003). On the Theory of Reference Dependent Preference, J. Econ. Behav. Organ. 50(4):407-48. Neil JR (1988). Another Theore on Using Market Deands to Deterine Willingness to Pay for Non-traded Goods, J. Environ. Econ. Manage. 15():4-3. Peters H (01). A Preference Foundation for Constant Loss Aversion, J. Math. Econ. 48(1):1-5. Phaneuf DJ, Siderelis C (003) An Application of the Kuhn-Tucker Model to the Deand for Water Trail Trips in North Carolina, Mar Resour Econ. 18(1): 1-14. Plott CR, Zeiler Y (005). The Willingness to Pay Willingness to Accept Gap, the Endowent Effect, Subject Misconceptions, and Eperiental Procedures for Eliciting Valuations, A. Econ. Rev. 95(3):530-545. Putler D (199). Incorporating Reference Price Effects into a Theory of Visitor Choice, Market. Sci. 11(3):87-309. Ren J, Zhao DM, Zhang XH (014). Online Buying Decision Preference and Prospect Theory, Adv. Mat. Res. 96-930:3850-3853. Sherif CW (1963). Social Categorization as a Function of Latitude of Acceptance and Series Range, J. Abnor. Psychol. 67():148-156. Shrestha RK, Seidl AF, Moraes AS (00). Value of Recreational Fishing in the Brazilian Pantanal: A Travel Cost Analysis Using Count Data Models, Ecol. Econ. 4(1-):89-99. Suzuki Y, Tyworth JE, Novack RA (001). Airline Market Share and Custoer Service Quality: A Reference Dependent Model, Transport Res A-Pol. 35(9):773-788. Tversky A, Kahnean D (1991). Loss Aversion in Riskless Choice: A Reference Dependent Model, Q. J. Econ. 106(4):1039-1061. Von Haefen RH (007). Epirical Strategies for Incorporating Weak Copleentarity into Visitor Deand Models, J. Environ. Econ. Manage. 54(1):15-31. Winer RS (1986). A Reference Price Model of Brand Choice for Frequently Purchased Products, J. Consu. Res. 13():50-56. Whitehead JC, Groothuis PA, Southwick R, Foster-Turley P (009). Measuring the econoic benefits of Saginaw Bay coastal arsh with revealed and stated preference ethods, J. Great Lakes Res. 35(3):430-437. Zeisberger S, Vrecko D, Langer T (01). Measuring the Tie Stability of Prospect Theory Preferences, Theory Decis. 7(3):359-386.

Okuyaa and Hayashiyaa 181 Appendi A. Deand function Let be the undeterined ultiplier and the Lagrange equation be defined by equation (A.1). The first derivatives of equation (A.1) are suarized as equation (A.). Let dp, dp, dy, and dq be zero, and p be equal to 1. Then, the z total derivative of the deand function is derived as equation (A.3) by using Craer s rule. Here, A is the deterinant of the atri (the first parenthesis of the left side of equation (A.)). u( z,, Q, g, l) ( pzz p y) (A.1) uzz uz pz dz dpz ( uzqdq uzgdg uzldl) uz u u Q d dp ( uqdq ugdg uldl) pz p 0 d dy ( zdpz dp ) (A.) 1 d {( ugdg uldl ) uq ( uzgdg uzldl )} A (A.3) The first derivatives are equation (A.4) fro equation (A.3). It is necessary to hold u u u 0 for gains and u z g Q zg u u 0 for losses to satisfy the first differential condition discussed above. The second derivatives are equations l Q zl (A.5) and (A.6). It is necessary to hold equation (A.5) as non-positive for gains and equation (A.6) as non-negative for losses to satisfy the second differential condition. These conditions are suarized as equation (A.7). d u u u d u u u, g dg A l dl A g Q zg l Q zl { ( ug uquzg ) / g} A ( ug uquzg )( A / g) g A { ul uquzl / l} A ( ul uquzl )( A / l) l A { ( u g u Q u zg )/ g} ( u )( / ), l u Q u zl A l A A ( u u u )( A / g) { ( u u u )/ l} g Q zg l Q zl (A.4) (A.5) (A.6) (A.7) Finally, an eaple of deand function and the derivatives are illustrated by using the eaple of utility function in 1/ 1/ footnote 8; U z ( Q 0.5 g l). Fro the first derivatives of the Lagrange equation, the deand function is derived as equation (A.8), and the first and second derivatives for gains and losses are equations (A.9) and (A.10), where B p C p 0 and C { Q 0.5 g l} 0. C { Q l} 0 is assued to ensure the utility is positive even if RE is at loss and the deand would be positive(see footnote 10). The first derivatives are satisfied in both gains 1 and losses; however, the second derivatives need the conditions p B C 1 1.5 0 for gains and 4p B C 3 0 for losses. This iplies that the deand function satisfies the conditions when 1 3/ 4 p B C 9/ 4 holds. y p ( Q 0.5 g l) p g l 3 0.5 p yb C 0 3 4p yb C 0,, 0.15 p y p B C 1.5B C g 1 4 8p y 4 p B C 3A B l 1 4 (A.8) (A.9) (A.10)

18 J. Econ. Int. Finance Appendi B. Indirect utility and ependiture functions The derivations of indirect utility function and ependiture function are considered. For siplicity, let the total gain and total loss be decoposed into ( g, ) and ( l, ), respectively, in the utility function as in the eaple above. Let V be defined as equation (B.1),where * indicates the deands and the undeterined ultiplier are at optial level. V u z Q g l y p z p (B.1) * * * * * (,,,, ) ( z ) Then, V/ g and V / l are equivalent to u g and u l by the envelope theore. Thus, the second derivatives of the indirect utility function for losses also follow the condition of the utility function. Net, let us consider the ependiture function. As for the total differential equation of V, let the variables ecept for l and y be fied. Then, * dy dl u is derived due to u 0, where the arginal utility of incoe ( ) is positive. The second derivative is * / l / 0 l * d( dy / dl)/ dl ull / 0 ll * d( dy / dg) / dg ugg / 0 second derivative is due to u 0. By a siilar process, the first derivative for gains is dy / dg 0 and the due to u 0. gg Appendi C. Decoposition of total value The decoposition of the benefits fro RE and Q is considered by using equation (31). Equation (31) is decoposed s s s w wo wo into equation (C.1) and equation (C.) by using additional ependiture e( p, p, U, Q, g, l ) for (C.1) and s s s w w w e( pz, p, U, Q, g, l ) for (C.). The first ter of each equation is defined as the benefit fro environental quality change since other variables are constant. Siilarly, the second ter of each equation is defined as the benefit fro RE change. There are two ways to decopose TV. z s s s wo wo wo s s s w wo wo e( pz, p, U, Q, g, l ) e( pz, p, U, Q, g, l ) TotalValue s s s w wo wo s s s w w w e( pz, p, U, Q, g, l ) e( pz, p, U, Q, g, l ) s s s wo w w s s s wo wo wo e( pz, p, U, Q, g, l ) e( pz, p, U, Q, g, l ) TotalValue s s s w w w s s s wo w w e( pz, p, U, Q, g, l ) e( pz, p, U, Q, g, l ) (C.1.) (C.) The conditions are that the first ter of (C.1) is equal to the first ter of (C.) and the second ter of (C.1) is equal to the second ter of (C.) is e( ) / Q e( ) / Q. As for non-use value, this condition holds due to the choke wo RE price. For eaple, the decoposition is perfored in equation (4) as the following equation. wo wo wo w I g g (1 I) l l NUV Q Q Q Q w w I g g (1 I) l l w RE

Vol. 7(8), pp. 183-191, August, 015 DOI: 10.5897/JEIF014.0636 Article Nuber: 9706F654746 ISSN 141-667 Copyright 015 Author(s) retain the copyright of this article http://www.acadeicjournals.org/jeif Journal of Econoics and International Finance Full Length Research Paper Banking sector refors and output growth of anufacturing sector in Nigeria (1970-011) Olanrewaju, Oluwagbenga Gideon 1*, Areo, Adeleke Gabriel and Aiyegbusi Oluwole Oladipo 1 Departent of Econoics, Babcock University, Ilisan, Reo, Ogun State, Nigeria. Departent of Econoics, Obafei Awolowo University, Ile-Ife Nigeria. Received 7 Deceber, 014; Accepted 15 July, 015 The study investigated epirically the effect of banking sector refors on the output of anufacturing sector in the Nigerian econoy between 1970 and 011 with a view to eaining the etent of the ipact of banking sector refors on the anufacturing sector. The study eployed annual secondary tie series data fro 1970-011, sourced fro the Central Bank of Nigeria s statistical bulletin and annual report and stateent of accounts, National Bureau of Statistics final accounts and IMF International Financial Statistics (IFS) using the ethodology of Cointegration analysis and Error Correction Mechanis (ECM). The epirical results showed that the effects of Bank assets, Lending rate, Echange rate and real rate of interest on anufacturing output were positively significant but with very low ipact. On the other hand, the financial deepening and interest rate spread negatively and significantly ipacted on the output growth of anufacturing sector in Nigeria. Overall, the conclusion that eerged fro the findings suggests that the effects of banking sector refors on the output growth of anufacturing sector were significantly low in the Nigerian econoy. However, the findings indicated that the ipacts of the various banking refors could vary widely on the econoy depending on the tie lags involved. Consequently, the policyakers ust be prepared to initiate proper countercyclical banking refors that will serve as buffer easures to lessen or abort the negative ipacts of any banking refors on the anufacturing output growth. Thus a fleible accoodating banking refor regie is advocated for Nigeria. Key words: Banking sector refors, Error Correction Mechanis (ECM), anufacturing sector, and Cointegration analysis. INTRODUCTION The banking sector represents the nucleus of the financial syste in an econoy; thus it constitutes a potent and frontline econoic policy tool in the hand of the onetary authorities for realizing key acroeconoic policy objectives. A well-functioning banking sector is sufficient to jup-start a sustainable econoic growth. Hailton (1781) opined that banks were the surest engines that ever were invented for spurring econoic growth. The recent eperience in Eastern Europe and Asia has shown that countries that oved quickly to fi *Corresponding author. E-ail: golarenwaju@yahoo.co Authors agree that this article reain peranently open access under the ters of the Creative Coons Attribution License 4.0 International License

184 J. Econ. Int. Finance their banking industry were able to achieve a sustainable growth rate than those that did not. This indicates that there is a causation that runs fro banking sector developent through the anufacturing sector to econoic growth. Banks originate and facilitate financing which are fundaentally necessary for carrying out productive investents (Akingbola, 001; Bukhari, 005; Carbo; 007) A large body of the literature on developent reveals that the services provided by well developed financial interediaries such as banks are essential drivers for innovation and growth (Bencivenga and Sith, 1991; King and Levine, 1993; Levine, 1997; Levine and Zervos, 1998; Loayza et al., 000; Graff and Karann, 001) aong others. In contrast, eplanation initiated by Robinson (195) posited that finance does not eert a causal ipact on growth and that the findings of causations are soewhat isleading. Rather than this, banking developent ay follow real sector growth as a result of higher deand for financial services; hence the financial syste grows in response to econoic epansion. As econoic activities grow, there will be ore deand for both physical and liquid capital, such that the growth in anufacturing sector induces the financial sector to epand, and thereby increasing copetition and efficiency in the financial interediaries. Based on the above controversies related to the causal ipact of banking refors on growth, the study eained the ipact of banking refors on anufacturing sector growth considered as a crucial subsector of the real sector. The need to eaine this ipact is borne out of the following reasons: First, it will enable the policy akers to assess the contribution of the etant banking refors to real sector growth and to overall growth of the econoy. Second, it enables the policy akers to predict when best to reshuffle and substitute any unproductive banking refors strategies that are not eeting the desired goal of achieving the pre-deterined output growth in the econoy. Third, there is the likelihood that variants of refor strategies aking up a enu of refors ay not likely have the sae effects on real sector output growth. The present study therefore contributes to the literature by eaining the individual effects of each of the banking sector refor instruents to anufacturing output growth. The reaining presentation is organized into four sections. The theoretical and epirical issues are eained in section. The ethodology is contained in section 3. Section 4 provides the epirical results while the last section contains the concluding rearks and policy iplications THEORETICAL AND EMPIRICAL ISSUES The evolutionary theory of debt-interediation clais that bankers were actually big erchants or firs which evolved fro loaning their ecess funds to other erchants or firs, to discounting bills of echange and to issuing their own bills. These erchant bankers relied on their own equity to perfor banking operations. They do not only enable the econoic transactions but also anticipate the by creating their own bills for the future econoic transactions. Thus, as ore equity capital is built via retained profits, the erchant bankers begin to lend and then learnt the skills of screening, selecting and onitoring. They start obilizing deposits and accuulating increasingly diversified loan portfolios as their quality skills and capabilities of screening, selection and onitoring iproved (Winkler, 1998). The popular view underlying these theories is that in a world of inforation asyetry, high transaction costs and oral hazard (which result in adverse selection), banks not only have to obilize and allocate savings but they also have to eert corporate control anage risks to lower the cost of researching potential investents (Levine, 1997). Faa (1980) applied the Modigliani-Miller (MM) theore of irrelevance pure financing decision to banking sector. He found that portfolio anageent activity of banks under strong MM theore is irrelevant to econoic activities. However, the role of a copetitive banking sector in a general equilibriu is passive. Johnson (1986) in a siilar study observed the sae line of arguent by assuring that a copetitive banking syste would be under constant incentive to epand the noinal oney supply and thereby initiating inflation. Thus if finance is oney, and oney is a veil financial developent is a neutral factor in real econoic developent since increase in banking operations leads to increases in oney supply, and so, inflationary prices. By iplication, increasingly better resource allocation depresses saving rates such that growth is retarded (Levine et al., 000). In a response to the question does finance cause econoic growth? King and Levine (1993) eplored the Schupeter s stateent that banker authorizes people in the nae of society as it were to innovate. They used various easures of financial developent in 1 regression equations and found that all the indicators of interediation developent are strongly associated with real per capita GDP growth, the rate of physical capital accuulation and iproveents in the efficiency with which econoies eploy physical capital. They also show that coercial banks advance credits better than any other financial institutions and this is due to the risk sharing inforation services provided by coercial banks. However, their findings are not tantaount to the conclusion that finance causes growth; but it ay be that finance is only a leading factor. Levine and Zervos (1998) etend the work of King and Levine to include the independent ipact of stock arkets, as well as banks, on real econoic growth. They found that stock arket liquidity and banking sector

Olanrewaju et al. 185 developent are independently and positively correlated with both current and future rates of capital accuulation and econoic growth. Siilarly, Bencivenga and Sith (1991) assert that the introduction of financial developent in any econoy shifts the coposition of savings towards capital, causing interediation to be growth prooting. However, the ajor objection to the views of King and Levine (1993) and Levine and Zervos (1998) is that of unobserved differences in industrial coposition across countries which tend to eplain both the variance observed in financial developent and the variance observed in growth. Greenwood and Javanovic (1990) in their study eployed an endogenous growth odel to deonstrate that there is a positive two-way causal relationship between output growth and financial sector developent. They opined that, the process of growth stiulate higher borrowing requireents for working capital and investents, thereby necessitating the entry and epansion of ore banking institutions. While the process of financial interediation by banks, encourages investent projects to be financed ore efficiently, thereby stiulating investent and output growth. In eplaining the causality evidence whether finance is an engine of growth or not Jayaratne and Strahan (1996) observed that rates of real per capital growth in incoe and output increased significantly following interstate bank refors in USA. They also note that iproveents in the quality of bank lending via branch banking and not increased volue of bank lending are responsible for growth changes. However, the causality direction sees to depend on the studied countries. Jayaratne and Strahan (00) investigated the ethods to estiate the ipact of changes in laws governing bank copetition on entrepreneurial behaviour. They found that countries with ore concentrated local banking arkets have lower rates of incorporation, and when these countries opened their banking arkets to eternal copetition, the rate of incorporation increased. Thus, the reoval of regulatory barriers increased bank copetition, which in turn, caused higher rates of business incorporation. Beck et al. (000) also take advantage on the questions of unobserved heterogeneity and spurious casuality; though without considering industries as well as countries, but applying novel econoetric techniques. They use a dynaic Generalized-Method-of Moent (GMM) panel estiator that allows siultaneously the eploitation of tie series variation in the data to account for unobserved country-specific effects for the inclusion of lagged variables as regressors, and controls for endogeneity of all the eplanatory variables. They also use an instruental variable estiator in order to etract the eogenous coponent of financial interediary developent and found a positive effect of the financial developent. In Nigeria, several epirical attepts have been ade to assess ore generally the relationship between financial liberalization and econoic perforance (Soyibo and Adekanye, 199; Ikhide and Alawode, 001; Akpan, 004). There eist other studies which eaine the perforance of financial sector refors on econoy across soe sub Sahara African countries (Soyibo, 1994; 1997; Eenuga, 1998; Aryeetey and Sebnet 1998; Aryeetey, 000 aong others). For instance, Soyibo and Adekanye (199) eaine the links between interest rates, savings, investent and oney supply in Nigeria. They found that there eists positive relationship between returns on financial assets and the rate of savings. They also showed that bank deposits are iportant in the level of productive investent in Nigeria. However, they cautiously noted that the general epectation in ters of the link between savings, investent and econoic growth is abiguous due to structural iperfections such as inforation asyetries, oral hazards, and the likes. Pradhan et al. (013) eained the causal neus between econoic growth, banking sector developent, stock arket developent, and other acroeconoic variables in ASEAN countries between 1961 and 01 using panel vector auto-repression. The study showed that banking sector developent Granger-causes econoic growth. This result confors with findings of Mezioghu and Walde-Rafael (014) and Bojanic (01) and Chaiechi (01). Also, Dwyer et al. [013] eained the relationship aong banking crises, econoic growth and recession covering 1 countries. They found evidence of ied effects of either negative or positive on econoic growth during the econoic crisis but evidence of ied results after the crisis. METHODOLOGY Model specification Based on the arguents presented in the theoretical fraework and the intuition fro the reviewed literature, the odel adopted in this study is the Schupeterian Circular flow of creditary production (1934).The popular view underlying this theory is that a sectoral output of the entrepreneur will depend on banking refor easures, lending capacity of the banking syste, and other conditioning variables that are capable of influencing the productivity of capital (A). Therefore, the relationship between banking sector refors and output growth of the anufacturing sector via investible funds can be epressed as: MFGO t=ƒ(bf t,lc t,a t,). (1) Where MFGO is the anufacturing sector output growth; BF is the easure of banking refors that is proied by real interest rates (RR); interest rate spread (IRS); lending rates (LR); ratio of broad oney (M ) to noinal GDP as a easure Deposit Money Banks liquid liabilities (M /GDP). The lending capacity of the banking syste is easured by ratio of Deposit Money Banks assets to total banking assets (BA). A, in the odel one represents those

186 J. Econ. Int. Finance conditioning variables which could also deterine the productivity of the invested capital. These variables include: power infrastructure (ENG); anufacturing capacity utilisation (MCU); trade openness as the degree to which the banking syste is able to interediate funds across borders, easured by ratio of iports plus eports to noinal GDP (TO) and echange rate (EXR). Therefore, in order to assess the effects of the banking refors on the output growth of anufacturing sector in Nigeria, the following relationship was investigated: MFGO t =ƒ(lr t, RR t, IRS t, BA t, TO t, ENG t, M /GDP t, EXR t, MCU t -------------------------------------------------------------------------- () This technique allows different easures of banking developent to be epressed in ters of a single inde (Stock and Watson, 00 a, b). Thus, the new proy for the banking developent, as denoted by BF, is able to capture ost of the inforation fro the original dataset which consists of the three financial developent proies. In order to test the links between the output growth of anufacturing sector and banking variables as well as those conditioning variables, we partially log-transfored equation () to have: log(mfgo t) = β 0 + β 1 LR t + β RR t + β 3 IRS t + β 4 log BA t + β 5 log TO t + β 6 log ENG t + β 7 log M /GDP t + β 8 EXR t + β 9 MCU t + t. (3) Where: log MFGO is denoted as LnMFGO, easured by the inde of anufacturing output; BF is proied as a banking refors variable; and is subdivided into: LR, LnBA, LnM /GDP, EXR, IRS which is the spread between deposit and lending rates; and RR, which is the real deposit interest rate less the rate of inflation easured by GDP deflator, or lending rate less the interest rate spread; Log TO is the log su of iports and eports easured as a share of GDP; Log ENG is the log of physical infrastructure proied as industrial energy consuption inde; EXR is the log of real effective echange rate inde; Log MCU is the log of anufacturing capacity utilization easured by the average capacity utilisation rates of the anufacturing sector. The a priori epectations of the odel in equation 3 is that we epect the banking variables LR, RR, IRS and EXR with an eception of M /GDP, to have inverse relationship with the output growth of anufacturing sector. But on the other hand, M /GDP and all other conditioning variables such as BA, ENG, and MCU and TO in the odel should have a positive relationship with the dependent variable, MFGO. β 0, is the constant and β1 to β 9 are the coefficients, while ε t is the stochastic error ter. Thus, β 1, β, β 3 and β 8 <0; β 4, β 5, β 6, β 7, and β 9 >0. However, the intercept values (β 0,) could either be positive or negative. Sources of data In order to investigate the relationship between banking refors and output growth in anufacturing sector in the period 1970 to 011, the study ade use of secondary data to obtain values for the variables in the odel. Data were sourced fro the publications of Central Bank of Nigeria Statistical Bulletin and Annual Report and Stateent of Accounts (various issues). Data on anufacturing output were obtained fro the publications of National Bureau of Statistics (NBS) and International Financial Statistics (IFS). Techniques of data analysis Engle and Granger (1987), deonstrated that co-integration variables ust have an error correction representation in which an error correction ter (ECM) ust be incorporated into the odel as in equation 4: Ln MFGO t = λ 0 t-1 + λ1 Ln BF t t + λ Ln TO t + λ 3 Ln ENG t + λ4 EXR t + λ 5 Ln MCU t + ε t. (4) Here, denotes first difference operator. t-1 is the one period lagged value of the residual fro the co-integration regression. The λ 0 coefficient of the error correction ter captures the adjustent towards the long-run equation. ε t represents white noise with usual assued zero ean and constant variance. Thus, Model 4 becoes the Error Correction Model. EMPIRICAL RESULTS It is pertinent to eaine the statistical properties of the data series. All the tie series data eployed in the analysis were subjected to stationarity test. The two conventional tests eployed are Augented Dickey Fuller (ADF) and Philips-Perron (PP) tests. The results of the two tests are reported in Table 1. As shown in the table, we use both the arguented Dickey-Fuller (ADF) and the Phillips Perron Statistics. It was found that all the variables ecept the interest rate (RR) ehibit non stationarity at their levels but stationary at their first differences. The ADF and Phillips Perron tests are carried out against the null hypothesis that there is unit root, that is, I (1), non stationary of the series. With a saple size of 39, the critical values for the ADF and Phillips Perron (without trend) at 1% and 5% significance levels are 3.6 and.94 respectively. Absolute values of ADF and Phillips Perron values are less than critical values indicate a rejection of the null hypothesis. The results of the test as reported above indicates that both the ADF and Phillips Perron test Statistics confired that differencing the variables once will guarantee their stationarity. Since the series were integrated of order one or I (1) in the terinology of Engel and Granger (1987), and their first differences were stationary, consequently the presence of significant cointegrating relationships aong the variables were thereby deterined. Test of Cointegration Cointegration analysis helps to clarify the long ter relationships between the integrated variables. A ajor defect of the unit root test is that it cannot discriinate between true and near true rando walks processes. Thus, it becae necessary to perfor additional tests to show that the variables of the odel are cointegrated. The establishent of long ter convergence, i.e, long run equilibriu between the variables enables us to carry out estiations using cointegration and error correction odel

Olanrewaju et al. 187 Table 1. Unit Root test- of the series -without trend (1970-008). At Levels At First Differences Variable ADF Philips-Peron ADF Philips Peron Order of Integration LMFGO -0.957334-0.798496-3.68888-6.363431 I(1) LR -1.41418-1.818578-6.573497-9.413967 I(1) LBA -.174196 -.5030-4.86379-5.69675 I(1) LENG 0.048759 0.15184-4.3474-6.073509 I(1) IRS -1.38686-1.791809-6.5341-9.649741 I(1) MCU -1.637037-1.450511-5.757-3.703099 I(1) LM /GDP -.17890-1.980600-4.01854-5.768 I(1) RR -4.071365-3.6731 - - I(0) EXR -1.3479-1.367673-3.67031-5.33076 I(1) LTO -1.733006 -.455093-5.04045-9.38813 I(1) Source: Copiled by the author. Table. Johansen cointegration test assuption of linear deterinistic trend in the data series : EXR, LBA, LENG, IRS. LR, LM/GDP, MCU, LMFGO, LTO, RR. Eigen value Likelihood ratio 5% critical value 1% critical value Hypothesized No. of CE (s) 0.986403 479.9538 33.13 47.18 None** 0.934444 39.570 19.89 04.95 At ost 1** 0.794641 34.1573 156.00 168.36 At ost ** 0.77961 178.755 14.4 133.57 At ost 3** 0.651693 133.18991 94.15 103.18 At ost 4** 0.63561 96.7568 68.5 76.07 At ost 5** 0.59476 60.94195 47.1 54.46 At ost 6** 0.418618 34.55514 9.68 35.65 At ost 7 * 0.333348 15.5799 15.41 0.04 At ost 8 0.038687 1.380936 3.76 6.65 At ost 9 *(**) denotes rejection of the hypothesis at 5% (1%) significance level. Likelihood ratio test indicates 9 co-integrating equation(s) at 5 per cent significance level. (ECM), estiate our odels. The Johansen Maiu Likelihood procedure was therefore eployed and the results are presented in Table. As stated in Table, the result shows that 9 co- integrating equations are found to eist aong the variables. By iplication, there is a long run relationship aong the variables because there is at least the presence of one cointegrating vector, which suffices to confir cointegrating relations. Specifically, the results of the cointegration test suggest that banking refor and conditioning variables at first differences converge in the long run. It was iportant to note that the eistence of cointegration vectors aong a group of variables ight not iply that there was causal influence or relationship between pairs of variables included in the odel involving cointegrated variables. Consequently, the eistence of equilibriu between a group of variables should not be interpreted to ean that equilibriu eist between all pairs of variables in the odel. Thus, the changes in the banking refor variables ight not have had significant ipact on anufacturing sector s output growth which perhaps ight have been induced by other variables included in the odels which ight be responsible for the possible long run relationship. The result fro the noralized equation with respect to MFGO is: LMFGO = 7.849* LR + 0.056 *RR 0.738 * IRS + 1.0350* LBA + 10.4408 * LTO 0.643 * LENG 8.4888* LM/GDP + 0.0345 *EXR 0.1015 * MCU None of the banking variables has the apriori signs ecept interest rate spread (IRS). A decrease of 1 per cent in the interest rate spread (IRS) leads to an increase

188 J. Econ. Int. Finance in anufacturing output (MFGO) by about 0.7 per cent. But on the other hand, a decrease of 1 per cent in the real rate of interest will result in 0. per cent decrease in anufacturing output growth. Siilarly, a reduction in lending interest rate (LR) by 1 per cent will also lead to a decrease in anufacturing output (MFGO) by about 8 per cent. Moreover, a 1 per cent increase in the financial is deepening indicator (M/GDP) leads to a decrease in MFGO by about 8 per cent. The effects of echange rate (EXR) on MFGO also do not confor to the apriori negative signs. This is an indication of weak effects of banking sector refors eperience in Nigeria. The lending capacity of the banking syste easured by the ratio of DMBs assets to aggregate banking assets (BA) does not have the epected positive sign. It has coefficient value of 1.03. This iplies that an increase of 1 per cent in banks assets leads to 1 per cent increase in MFGO. This is not to say that lending capacity of banks is not iportant in deterining the quantity of bank credits to the anufacturing sector, but it probably reflects a case of credit diversion fro the real sector to consuer credit arkets. In the sae vein, the signs of the coefficients of ost of the conditioning variables on anufacturing sector output are negative and thus do not confor to the apriori epectations. A 1 per cent increase in the ratio of net eports to GDP (TO) results in about 10 per cent increase in MFGO. The effects of power infrastructure on output of anufacturing sector which shows a negative sign does not confor to the epected positive sign. This confirs the evidence that in Nigeria, over the years, power infrastructure plays insignificant role in bringing about increased output. Thus, an increase of 10 per cent in anufacturing capacity utilization will brings about an increase of only 0.1 per cent in output growth of anufacturing sector. The focus in this section is to present the over paraeterised version of the error correction odel as well as the parsionious version. The conteporaneous as well as the lag variables presented are in their loglinear for which iplies that the coefficient estiates in all the odels are elasticities showing the percentage changes in the eogenous variables that condition the percentage change in the endogenous variable (anufacturing output growth). The over paraeterised banking sector refor odel is siplified until theory-consistent and data-coherent results are achieved by one by one deleting of the insignificant variables. Both Schwarz and Akaike inforation criteria show that the parsionious odel is a better odel than the over paraeterised odel because of the reduction in their values fro -1.738979 and -3.190138 to 0.737941 and -0.305079. This is an indicative of the robustness of the parsionious odel. To ensure the validity of the estiates of the parsionious odel in table 3, tests to verify the etent of the violations of the assuptions of Ordinary Least Squares estiates were carried out. The first test is the Breusch-Godfrey Serial Correlation LM test presented in Table 4. The probably of F-statistic at 5 per cent significance level shows that the null hypothesis of no serial correlation cannot be rejected. The second test carried out is the Autoregressive Conditional Heteroskedasticity (ARCH) test to verify the presence of Heteroskedasticity in the error ter. This is presented in Table 5. The ARCH result in Table 5 shows that the null hypothesis of the absence of ARCH effects cannot be rejected at 5 per cent significance level. On the basis of the two tests presented in tables 4 and 5, the parsionious odel results could largely be relied upon. A critical look at the parsionious odel results presented in table 3 shows that the past values of anufacturing output were positively related to its current value, and equally significant. As a result a 1 per cent increase in a year lagged period value of anufacturing output will bring about 44 per cent increase in the current value of anufacturing output. Siilarly,1 per cent increase in the two year lagged periods will equally result in 69 per cent increase in the current value of anufacturing output. While the a priori epectation of the signs was et in a year lagged period values of lending rate and interest rate spread, other variables like echange rate, current bank assets, two-year lagged value of interest rate spread, two-year lagged value of power infrastructure, financial depth indicator at level, past and current values of interest rates were not properly signed even though significant. In fact, the positive signs on lending rate and real interest rate contradict the theoretical prediction according to which a higher cost of capital would discourage productive activity by entrepreneurs. However, the current value of power infrastructure, a year lagged period value of interest rate spread and two-year lagged period value of financial depth indicator et the apriori epectations of the sign. The Duy variable introduced into our regression odel in order to capture the effect of the regie shift fro financial regulation to deregulation of the banking sector in 1986 was found to be positive but not significant. However, the Duy variable was still left in the parsinious odel even though its effect on the robustness of our analysis was arginal. The overall regression is significant. All the variables, besides current interest rate are significant at 5 per cent significance level. The results also show that the error correction ter (ECM) is negative and at the sae tie significant with a very low probability value of 0.09. In other words, the negative coefficient of ECM signifies that there is an adjustent in the syste in case of any disequilibriu. Thus, about 3 per cent of the

Olanrewaju et al. 189 Table 3.The parsionious error correction odel of DLMFGO. Variable Coefficient Std.Error t-statistics Prob. DLMFGO(-1) 1.63013 0.37758 3.853496 0.003 DLBA 0.768793 0.316803.4673 0.0356 DLTO 0.73391 0.77698.64880 0.046 DLTO(-1) 1.09550 0.45477.4114 0.0360 DLTO(-) 1.633461 0.44411 3.67717 0.0043 DIRS(-1) -0.347 0.069048-3.35835 0.0089 DIRS(-) 0.1167 0.060301 1.67436 0.088 DLENG 0.9457 0.088106 3.339816 0.0075 DLENG(-) -0.171697 0.110078-1.559775 0.1499 DLM/GDP -3.366537 0.69431-4.849304 0.0007 DLM/GDP(-) -1.456169 0.37484-3.890546 0.0030 DLR(-1) 0.184419 0.051895 3.55376 0.005 DLR(-) -0.105693 0.051011 -.07196 0.0651 DLR 0.16617 0.056040.966060 0.0141 DRR(-) 0.0400 0.006413 3.49748 0.0058 DIRS -0.59540 0.07888-3.9040 0.0080 DEXR(-1) 0.08514 0.006099 4.674815 0.0009 DEXR(-) 0.011534 0.003461 3.33300 0.0076 DUM 0.179034 0.177 1.40708 0.1897 ECM(-1) -0.394704 0.891977-0.684714 0.09 R =0.84; S.E=0.19; F =.3503 DW=1.915; RSSl=0.3533; S.C=0.7379; AK INFO=-0.3050. Source: Copiled by the authors Table 4. Breusch-Godfrey Serial Correlation LM Test. F-statistic 1.30974 Probability 0.83576 Obs*R-squared 3.385340 Probability 0.18407 Table 5. Autoregressive conditional Heteroskedasticity (ARCH) result. F-statistic 1.415501 Probability 0.53710 Obs*R-squared 5.553867 Probability 0.35035 disequilibriu in the output growth of anufacturing sector in the previous year is autoatically corrected in the current year. This also appears significant deonstrating the fact that doestic and endogenous factors go a long way in eplaining output growth of anufacturing sector in Nigeria. The iplication of the above analysis is that the conteporaneous values of the banking variables used in the study appear not to ipact positively on the output growth of anufacturing sector during the period under review. Thus, it becoes evident that the banking industry is inefficient and uncopetitive given the negligible ipact the sector had on the anufacturing activity. However, the results show that the past (lags 1 ) values of both banking variables and anufacturing output, as well as the values of other conditioning variables were responsible for the growth of anufacturing output in Nigeria during the period. The finding contradicts the position of Ikhide and Alowode (001) which supported the positive role of financial developent in banking sector growth in Nigeria. CONCLUSION AND POLICY IMPLICATIONS Overall, the study found that the epectations of

190 J. Econ. Int. Finance increased output growth of the anufacturing sector in Nigeria in the wake of banking refors are still far fro being realized. The coefficients of banking developent indicators show a negative ipact of banking refors on output growth of anufacturing sector. The results of the Nigeria-specific study confired that the Nigerian banking industry has not been efficient in aeliorating inforational asyetries, reducing transaction costs and allocating resources to the anufacturing sector. This indicates that the Nigerian banking sector has not considerably played its legitiate role of channeling financial credits to the anufacturing sector regarded as an engine of growth. It is therefore pertinent to suggest that the etant banking sector refors in Nigeria need to be carefully reviewed and closely onitored. This calls for sound institutional and legal fraework and sound corporate governance aied at reducing structural lapses and corruption level. This steps will create an enabling environent for banking sector to operate and ensure that the sector positively ipact on the anufacturing sector output growth. In addition, the policyakers ust be prepared to initiate proper alternative banking refors that will serve as counter easures to lessen or abort the negative ipacts on the anufacturing output growth caused by any banking refors. Thus a fleible accoodating banking refor regie is advocated for Nigeria. Conflict of Interests The authors have not declared any conflict of interest. REFERENCES Akingbola BO (001). Strategies for Effective Manageent of Financial Institutions in the New Millenniu J the Chartered institute of Bankers of Nigerian (CIBN), Lagos. Akpan DB (004). Financial Liberalization and Endogenous Growth: The case of Nigeria. AIDEP Publication 3():1-4. Aryeetey E, Sebnet L (1998). Essential Financial Market Refors in African Background Paper for Can Africa Clai the 1 st Century? World Bank. Aryeetey E (000). Recent Developent in African Financial Market: Agenda for Further Research. Paper presented at the plenary Session of the Biannual Research Workshop of the AERC, Nairobi, Kenya Deceber -7, pp.111-15. Beck T, Levine R, Loayza N (000). Finance and the sources of Growth J Financ. Econ. 58:61-300 Bencivenga VR, Sith BD (1991). Financial Interediation and Endogenous Growth. Review of Econ. Stud. 58:195-09 Bojanic A.M. (01). The Ipact of Financial developent and Trade on the Econoic Growth for Bolivia. J. Apppl. Econ. 115(j):51-70. Bukhari MS (005).``Banking and Econoic Growth: Case of the Gabia Inaugural-Dissertation Christian-Albrechts-Universtitaet, zu Kiel 005. Carbo S (007). Financial Innovations in Banking; Ipact on Regional Growth Departent of Econoic, University of Granada, Spain. CBN (008). Statistical Bulletin, Central Bank of Nigeria, Abuja. Chaiechi T (01). Financial Developent shocks and Conteporaries feedback effort on key acroeconoic indicators: A post Keynesian Tie Series analysis, Econ. Model. 9():487-501. Dickey D, Fuller WA (1981). Likelihood ratio statistics for Autoregressive Tie Series with a Unit Root, Econoetrical, 49: 1057-107. Dwyer GP, Devereu J, Baler S, Taura R (013). Recession, Growth and Banking Crises. J Int. Money Financ. 38:18-40. Engle RF, Hendry DF, Richard J (1987). Eogeneity. Econoetrica 51: 77-304. Eenuga C (1998). The outcoe of the Financial Sector Refors in West Africa In Ediz (ECLS) Econoic Refors, Sub-Sahara, Africa IDRC Ottavia. Faa EF (1980). Banking in the Theory of Finance J. Monet. Econ. 6:39-57. Graff MA, Karann A (001). Does Financial Activity Cause Econoic No. 1/01, Techno Growth? Dresden Discussion paper Series in Econoics No. 1/01, Technische Universitat Dresden. Greenwood J, Jovanovic B (1990). Financial Developent, Growth and the distribution of Incoe, J Polit. Econ. 98: 1076-1108. Hailton A (1781). http:/www.worldbank.org/research/conference/ finance. U.S.treasury/financeDept./Version3.pdf. Ikhide SI, Alawode AA (001). Financial Score Refors, Macro Econoic Instability and the order of Econoic Liberalization: The Evidence for Nigeria AERC Research Paper 11, Noveber. Jayaratne J, Strathan PE (1996). Financial Growth Neus: Evidence fro Bank Branch Deregulation, Quart. J Econo. III, Pp. 693-670. Jayaratne J, Strathan PE (00). Financial Growth Neus: Evidence fro Bank Branch Deregulation, Q. J. Econ. 3:693-670. Johansen S, Juselius K (199). Soe structural Hypothesis in a ultivariate Co-integration Analysis of Purchasing Power Parity and Uncovered Interest Parity for the U.K. J. Econ. 53:11-44. King RG, Levine R (1993a). Financial and Growth: Schupeter Might be Right Quart. J Econ. 108(3):717-737. King RG, Levine R (1993b). Finance, Entrepreneurship, and Growth: Theory and Evidence, J Monetary Econ., 3: 513-54 Levine R, Loayza N, Beck T (00). Financial Interediation and Growth; Causality and Causes J. Econ. Lit., 46:31-77. Levine R (1997). Financial Developent and Econoic Growth: View and Agenda. J. Econ. Lit. 35:688-76. Loayza NK, Schidt-Hebbel LS (000). What Drives private Savings across the World? Rev. Econ. Stat. 8(). Mezioghu K, Walde-Rufael Y (014). Financial developent, trade operations and econoic growth in African countries: New Insight for Panel Casualty Appraisal, Econ. Productivity 37():386-394. Pradhan RP, Air MB, Hall JH, Bahoni (014). Causal Neus Between Econoic Growth, Banking Sector Developent, Stock Market Developent And Other Macroeconoic Variables The Case Of AESAN Countries. Rev. Financ. Econ. 3:155-173. Robinson J (195). The Generalization of the General Theory, in The Rate of Interest and other Essays, Macillan, London, pp.67-14 Schupeter JA (1934). The Theory of Econoic Developent, Oford University Press. Soyibo A, Adekanye F (199). Financial Syste Regulation, regulation and Savings Mobilization in Nigeria AERC Research paper. Soyibo A (1994). Financial Liberalization and Bank Restructuring in Sub-Saharan Africa: Soe Lessons for Sequencing and Policy Design Plenary Paper, AERC Deceber, pp.4-9. Soyibo A (1997). Banking Sector Refors in Africa, Effects on Savings, Investent and Financial Developent Afr. Dev. Rev. 9(1). Stock JH, Watson MW (00a). Macroeconoic Forecasting Using Diffusion Indees. J. Bus. Econ. Stat. 0:147-163. Stock JH, Watson MW (00b). Forecasting Using Principal Coponents fro Large Nuber of Predictors. J. A. Stat. Assoc. 97:1167-1179. Winkler A (1998). Financial Developent, Econoic Growth and Coorporate Governance Johann Wolfgang Goethe Universitaet, Frankfurt, AM Main, Working paper Series: Finance No 1.

Olanrewaju et al. 191 Appendi A: 30000 0000 10000 10000 5000 0 0-10000 -5000 70 75 80 85 90 95 00 05 Res idual Ac tual Fitted Figure 1. Actual and Fitted graphs of MFGO odel 8000 6000 4000 000 0-000 -4000-6000 70 75 80 85 90 95 00 05 MFGO Residuals Figure. The residual graph of the MFGO odel.

Journal of Econoics and International Finance Related Journals Published by Acadeic Journals Journal of Hospitality Manageent and Touris African Journal of Business Manageent Journal of Accounting and Taation International Journal of Sociology and Anthropology Journal of Public Adinistration and Policy Research African Journal of Marketing Manageent