CREDIT CONSTRAINT AND RURAL FINANCE IN CHINA: THEORETICAL FRAMEWORK AND POLICY IMPLICATIONS. Yong Ma 1 and Shushu Li 2

Transcription

1 Pak. J. Statst Vol. 30(5), CREDIT CONSTRAINT AND RURAL FINANCE IN CHINA: THEORETICAL FRAMEWORK AND POLICY IMPLICATIONS Yong Ma 1 and Shushu L 2 1 Chna Fnancal Polcy Reseach Cente, School of Fnance Renmn Unvesty of Chna, Bejng, Chna 2 Insttute of Polcy and Management, Chnese Academy of Scences, Bejng, Chna Coespondng autho ABSTRACT In ths pape, we povde a theoetcal famewok fo undestandng the poblem of ual households cedt constant n Chna. Fou basc conclusons ae obtaned fom theoetcal and empcal analyss: (1) fnancal nsttutons have to adopt the poolng equlbum stategy to detemne cedt allocaton unde asymmetc nfomaton, whch leads to lowe effcency and hghe sk n ual fnance; (2) excessve hgh nteest ate wll be a hndance fo the sustanable development of both ual households and fnancal nsttutons snce t wll squeeze the poft magn of agcultual nvestment n the long un; (3) the development of egonal economy s mpotant as fnancal nsttutons tend to ncease the cedt n moe developed egons even n the face of asymmetc nfomaton; (4) the ncease of agcultual etun s not only helpful n attactng and stablzng fnancal esouces but also cucal fo the long-tem development of both agcultue and ual fnance. KEYWORDS Rual fnance; Cedt constant; Asymmetc Infomaton. JEL classfcaton: Q12, Q14, Q21 1. INTRODUCTION One of the coe ssues n the study of Chna s ual fnance s the cedt constant faced by ual households and how to emove the constant. Regadng ths ssue, thee ae seveal elated questons watng to be answeed: (1) How ae cedt constants fomed? (2) What factos affect the degee of cedt constants? (3) What polces should be adopted to emove the cedt constants? Theoetcally, ual cedt and ts effects on ual households have long been a majo ssue n the study of ual fnance and economcs. Despte dffeent vewponts, most studes tend to suppot the concluson that ual cedt, f popely allocated, wll sgnfcantly mpove the welfae of ual households. Khandle (1988) evaluated the effect of the ual cedt of Agcultual Development Bank of Pakstan (ADBP) and found that although moe than 69% of ual households who had no land and could baely suvve only eceved 23% of the natonal agcultual loans, ual cedt stll sgnfcantly allevated povety n the ual aea. Afte studyng the use of mcocedt of the Gameen Bank n Bangladesh, 2014 Pakstan Jounal of Statstcs 637

2 638 Cedt Constant and Rual Fnance n Chna Zaman (2004) advocated that mcocedt educed the faglty of famles n Bangladesh and sgnfcantly mpoved the socal welfae ndex n ual aeas. Osbone (2006) examned ual cedt sk fom a dynamc pespectve and found that ual cedt sk depends on self-poducton and consumpton featues. Theefoe, the fundamental way to lowe ual cedt sk s to enfoce self-poducton and consumpton ablty. Udy (1994) suveyed the status of ual cedt sk n noth Ngea and found that the sk of ual cedt would be detemned by nfomaton tansacton cost, socal system and othe po factos. Fo the study of Chna s ual cedt, Ln (2000) analyzed the ual cedt condtons n thee countes (ctes) and concluded that fomal cedt could affect the fxed assets fo poducton of ual households, and nfomal cedt could not eplace fomal cedt, but the volume of fomal cedt n the sample aeas was low, and thee exsted geneal cedt constants. 1 Kellee (2004) quoted Chna and Inda as examples to explcate that though the govenment ntended to epess the hgh nteest ate of pvate loans o mpove ual benefts by fomal ual cedt, the poblem of mpefect nfomaton detemned that lcensed ual fnancal nsttutons cannot eplace nfomal cedt. Smla conclusons ae found n Zhang (2005) and Zhang and Gao (2006), among othes. They advocated that the only way out fo Chna s ual fnance was to ncopoate nfomal fnance nto lcensed fnancal nsttutons, and the efom snce 2003 was an exogenous gaftng of the fnancal system whch would not succeed. Though the pevous eseaches have shed lght on vaous aspects of Chna s ual cedt constant, most of them have focused the attentons on the nadequate fnancal supply whle the demand sde s lagely gnoed. Lteatue based on the nteactve effects of demand and supply s even scace. Howeve, t s obvous that n pactce both the ual households demand fo cedt and the fnancal nsttutons supply of cedt ae dynamc and only when the demand s not satsfed would thee be a cedt constant (gap). Fo example, f the demand fo cedt s small and the aggegate supply s lage than the demand, thee would be no cedt gap even f the aggegate supply s small. In ths case, cedt gap and suppessed demand (the lack of effectve demand) should be dstngushed. Theefoe, to consde the dynamc nteacton between the cedt demand of ual households and the cedt supply of fnancal nsttutons, the two should be studed unde an ntegated famewok. Based on the above consdeatons, ths pape povdes an analytcal famewok fo undestandng the poblem of ual households cedt constant n Chna. By ntoducng an ncentve compatblty constant n the decson-makng of fnancal nsttutons unde asymmetc nfomaton, ual households cedt constant s analyzed based on the theoy of fee-dng and coss-subsdzaton effects. Seveal theoetcal popostons and polcy mplcatons ae obtaned though compaatve analyss of the model. The popostons ae elated to a wde ange of the ssues dscussed n ths pape, ncludng the mpact of nteest ate on the demand and supply of ual cedt, the elatonshp between ual cedt expanson and egonal economc development, and the multple effects of the ate of etun on both agcultue and ual fnance. Oveall, the theoetcal model and polcy mplcatons of the pape povde us wth mpotant nsghts nto the elated ssues of Chna s ual fnance. 1 Geneally speakng, ual cedt can be categozed as fomal cedt and nfomal cedt. Those ssued to ual households by lcensed fnancal nsttutons ae geneally called fomal cedt.

3 Ma and L 639 The est of the pape s oganzed as follows: secton two descbes the basc model; secton thee conducts compaatve analyss; secton fou dscusses polcy mplcatons; the fnal secton concludes. 2. THE MODEL Followng the pevous lteatue (Popkn, 1979; Gapta, 1997; Vllas-Boas and Schmdt, 1999), we assume that thee ae two types of ual households I and II n the ual fnancal maket. Type I households have a hghe capta level and ablty to poduce and nvest than type II households. Theefoe, they have a hghe magnal poduct of captal. Assume that popoton of the fnancal maket the two types of households take up ae and 1 ( 0 1) espectvely. The coespondng functons of etun on poduct ae F 1 ( L ) and F 2 ( L ), of whch L s the amount of cedt households eceve. Accodng to the law of dmnshng magnal etuns, The fst and second ode condtons of the functons should satsfy F ' ( L) 0 and F '' ( L) 0 ( 1, 2). Based on the assumpton of the magnal poduct, ' ' 1 2 F ( L) F ( L). Assume that s the maket loan ate, the cost fo ual households to boow loans fo poducton would be (1 L ). Consdeng the etun functon and cost functon of ual households, the poft functon should be F ( L) (1 ) L, of whch 1, 2. Poft functon of type I ual households: 1 F 1 ( L) (1 ) L Poft functon of type II ual households: 2 F 2 ( L) (1 ) L Based on the above functons, the optmal quantty demanded fo cedt detemned by the followng fst ode condton: L s F ' ( L) (1 ), 1, 2. (1) In ode to consde the cost fo ual households to beach the contact, we take nto account the game between them and fnancal nsttutons. In contactual elatonshps, once ual households do not abde by the contact (efuse to epay debts), fnancal nsttutons would not ssue loans to them n futue. The oppotunty cost fo the beach of contact CL ( ) can be epesented as the dscount of poft ncued by the cedt demand n the futue: C( L) F ( L) (1 ) L, 1, 2. t1 t whee ( 0 1) s the facto of the dscount on expected ncome, whch s geneally egaded as the equed ate of etun of agcultual nvestment. Rual households, accodng to the atonal man hypothess, choose to epay debts only when C ( L) (1 ) L. Theefoe, the cedt constants can be smplfed as follows: 1 F ( L) L, 1, 2. (2)

4 640 Cedt Constant and Rual Fnance n Chna Let when L L be the uppe lmt of the cedt ange. Fom (1) and (2), t can be obtaned that L, ual households face fnancng constants. Fnancal nsttutons mplement cedt atonng. When L L, the optmal quantty demanded fo fnancng les nsde the cedt ange, thus fnancng constants ae emoved. The above analyss can be easly llustated n Fg. 1. The cedt lmt of ual households s detemned by the ntesecton of F ( L ) and 1 L, of whch those of type I and type II ual households ae L 1 and L 2, espectvely. The demands 2 L 1 and L fo cedt of type I and II ual households equal to the slope of the tangent pont of the own etun of poduct functons 1. As both coespondng cedt lmt L 1 and L 2 L 1 and constants. The fnancng gap fo type I ual households s 2 L2 L 2 ae lage than the, both types of ual households face cedt 1 L1 L and fo type II s L. Of couse, changes n the paametes of the model would change the postons of the cuves and the slopes. Once the optmal quantty demanded fo cedt falls below the cedt lmt, fnancng constants would be emoved. If thee s no nfomaton asymmety n the ual fnancal maket, nfomaton of type I and II ual households ae publc. In ths case, any fnancal nsttutons n the maket can obtan the nfomaton and thus would be able to detemne the lnes of cedt fo each household. When the type of household can be dstngushed, fnancal nsttutons would offe L 1 and L 2 as the uppe lmt of cedt anges. Cedt demand whch does not fall n the anges would lead to cedt atonng, as suggested by Stgltz and Wess (1981). In othe wods, when thee s suffcent and symmetc nfomaton, any fnancal nsttutons can employ cedt dffeentaton,.e. offeng moe cedt suppot to ual households who have moe assets and a hghe etun. Cedt dffeentaton s smla to the mcoeconomc concept of pce dscmnaton. Unde pefect nfomaton, fnancal nsttutons can devse the stategy to ssue dffeent amount of cedt to dffeent pates so as to acheve optmzaton of the cedt stuctue and etun maxmzaton. Unde asymmetc nfomaton, what cedt allocaton stategy do fnancal nsttutons adopt? Accodng to the classc game theoy, when fnancal nsttutons cannot dstngush between the two types of ual households, they wll estmate the dstbuton pobablty of the two fom expeence and fst thoughts. By dong so, they employ the stategy of poolng equlbum n cedt polcy. Based on ou model, the cost of beach of contact fo type I ual households s C 1 ( L ) and that fo type II s C 2 ( L ). The popotons fo type I and type II ual households ae and 1, espectvely. Theefoe, the cost of beach of contact of ual households expected by fnancal nsttutons s EC( L) C1( L) (1 ) C2( L). Fo ual households to epay the debts, the condton EC( L) (1 ) L must be satsfed. As a esult, the ncentve compatblty constant wll be gven by:

5 Ma and L F( L) F ( L) (1 ) F ( L) L 1 2 The magnal ate of output of the type I households s lage than that of type II,.e. ' ' ' ' ' F ( L) F ( L). Theefoe, F ( L) F ( L) F ( L). The poolng equlbum les between F 1 ( L ) and F 2 ( L ). The coespondng cedt lmt L les between L 1 and L 2 L2 L L1 (see Fg. 1).,.e. Fg. 1: Asymmetc Infomaton and Cedt Dffeentaton The above analyss shows that, to fnancal nsttutons, cedt dffeentaton unde symmetc nfomaton s pefeed to poolng equlbum stategy unde asymmetc nfomaton. The eason s that wth poolng equlbum stategy, type II ual households whch ncu hghe sk and a lowe ate of etun on nvestment take up pat of the lne of cedt of type I ual households. It equals to substtutng low-sk hgh-etun nvestment fo hgh-sk low-etun nvestment. Moeove, type I ual households ncu lowe sk and a hghe ate of etun on nvestment. Howeve, unde nfomaton asymmety and wth poolng equlbum stategy, the lne of cedt type I ual households dmnshes because of standadzaton. They now face tghte fnancng constants, whch s not favoable fo developng agcultual poducton o nvestment as type I ual households have bette condtons to tansfom o povde demonstaton. Obvously, nfomaton asymmety n the ual fnancal maket has economc sgnfcance to both ual households and fnancal nsttutons. To ual households, those who have hghe poducton and nvestment ablty (type I) face a wdened cedt gap, but those who have lowe ablty (type II) face a educed one. Effcency s sacfced n ths type of cedt allocaton usng low magnal poducton ablty to substtute low magnal poducton ablty, whch s favoable to agcultual development. Fo fnancal nsttutons, they cannot employ cedt dffeentaton to offe a easonable lne of cedt to dffeent types of ual households. Cedt allocaton usng

6 642 Cedt Constant and Rual Fnance n Chna the poolng equlbum stategy s based towad type II ual households who have low effcency. Fnancal nsttutons face hghe sk and a lowe ate of etun on potfolo. When the expected ate of etun s smalle than the total cost, they have no choce but to ext the maket. Poposton 1: Unde asymmetc nfomaton, fnancal nsttutons can only estmate the dstbuton of sk and etun fom expeence and fst thoughts. In ths case, they have to employ the stategy of poolng equlbum n decdng cedt ssuance. Unde such stategy, ual households wth bette poducton and nvestment ablty wll face a wdened cedt gap whle fnancal nsttutons wll bea moe sk. 3. COMPARATIVE ANALYSIS AND MODEL DISCUSSIONS 3.1 The Effect of the Loan Rate on Demand and Supply In ou model, when the loan ate deceases, the slope of tangent pont 1 and the 1 constant lne l L move to the ght, causng the equlbum quantty demanded fo cedt L and the cedt lmt L to move ghtwad. When nceases, the slope of 1 tangent pont 1 and the constant lne l L move to the left, causng the equlbum quantty demanded fo cedt L and the cedt lmt L to move leftwad. Ths s to say that when the loan ate dops, ual households cedt demand nceases, and fnancal nsttutons offe a hghe cedt lmt to ual households; vce vesa. Although the change n the loan ate causes changes n the same decton n the tangent pont and the cedt lmt, the extent of the change s dffeent. Due to the paamete ( 0 1, the dscount facto of expected ncome), the change n exets an effect on the optmal quantty demanded fo cedt and the cedt lmt to dffeent extents. When dops by one unt, the slope of the tangent pont 1 dops by one 1 unt, but the slope of the constant l L lne dops by 1 tmes of one unt. In othe wods, the constant lne moves ghtwad faste than the tangent pont does. When the loan ate falls contnuously, cedt constants follow sut. Based on the above analyss, when the loan ate falls, cedt constans loosen; vce vesa. Ths s to say that the contnuous se n the loan ate mples the contnuous tghtenng of cedt constants. Ths s because when ual households face an nceasngly hgh loan ate, the poft magns of etun on nvestment would dmnsh. Smalle poftablty mples hghe cedt sk, whch suppesses the cedt ssuance of fnancal nsttutons. In exteme stuatons, even f the loan ate ses hghe than the ate of etun on agcultual nvestment, the poftablty of ual households s negatve. Fnancal nsttutons would not ssue loans to ual households and the cedt demand s entely suppessed. Poposton 2: The ncease (decease) of loan ate wll decease (ncease) both the demand and supply of ual cedt. When the loan ate ses above the ate of etun on agcultual nvestment, the poft of ual households s negatve. In ths case, fnancal

7 Ma and L 643 nsttutons would not ssue loans to ual households and the cedt demand s completely suppessed. 3.2 Cedt Expanson n the Development of Regonal Economy unde Infomaton Asymmety In the basc model, we obtaned the ncentve compatblty constant and hence 1 detemned the poolng equlbum as F( L) F1( L) (1 ) F2( L) L. As ' ' ' 1 2 F ( L) F ( L) F ( L), the poolng equlbum les between F 1 ( L ) and F 2 ( L ) cuves, and the coespondng cedt lmt L les between L 1 and L 2. Defne 1 F( L1 ) 1F1 ( L) (1 1) F2 ( L) L and 1 F( L2 ) 2F1 ( L) (1 2) F2 ( L) L Gven that 1 2, t s easy to obtan that L wll esult n an ncease n cedt ssuance. 1 L2. Ths means that a lage value of It s assumed n ou model that type I ual households account fo a popoton and type II fo 1 of total households, and type I have a bette captal foundaton and magnal poductvty. Theefoe, the hghe s, the lage the popoton type I ual households takes up, whch mples a bette foundaton and potental fo maket development. Fom the above analyss, t can be seen that the value of eflects the developmental level of a egon. A hghe value of mples a hghe developmental level of the egon, and vce vesa. The economc mplcaton of the change n s clea. Unde asymmetc nfomaton, the moe developed the egonal economy s, the geate the cedt ssuance s wth the stategy of poolng equlbum, and the smalle fnancng constants ual households wll face. Poposton 3: The development of egonal economy s cucal fo allevatng ual cedt constants caused by asymmetc nfomaton. The moe developed the egonal economy s, the lage amount of cedt s made unde the stategy of poolng equlbum, leadng to smalle fnancng constants faced by ual households. 3.3 The Multple Dvng Effect of the Rse n the Rate of Retun on Agcultue In ou model, the lne of cedt fnancal nsttutons offe to ual households s detemned by 1 L and F ( L ). When F ( L ) emans unchanged, the cedt lmt s manly detemned by the loan ate and the dscount facto. It s easy to see that when the dscount facto nceases fom to 1 ' l L moves ghtwad to l ' ', the cuve 1 L ; cetes pabus. In ths case, the cedt boundaes of both type I and type II ual households wll ncease, esultng n hghe

8 644 Cedt Constant and Rual Fnance n Chna cedt lmt pe household. Ths means that the se n the dscount facto g bngs systematc mpovement to the fnancng constants. It s woth notng that the ncease n has a multple effect on the expanson of the (1 ) 1 cedt bounday. Though nceasng the multple facto of ndex f ( x) x, x a small change n can cause a lage change n the feasble egon of cedt. We have ponted out that can be egaded as the equed ate of etun on agcultual nvestment. The se n the ate of etun on agcultue would loosen cedt constants and attact moe nput of cedt esouces, whch s the pmay effect of the change n. Futhemoe, the se n mples the loweng of sk of the ente agcultual ndusty and the ncease of etuns. These cetanly wll attact moe fnancal esouces. As competton n the ual fnancal maket becomes keene, the competton between cedtos would push the loan ate down. The effect of the loweng of s multpled by, leadng to the double-multple effect whee and nteact wth each othe. As t can be seen that a se n the ate of etun on agcultue can loosen fnancng constants on ual households, ncease agcultual cedt nput, and enhance competton n the maket though the multple effect. Ths effect can pomote the smultaneous long-tem development of the agcultual ndusty and ual fnance, leadng to a vtuous cycle and a self-sustanable mechansm. Poposton 4: The ncease n the ate of etun on agcultue can allevate the cedt constants faced by ual households on the one hand, and lowe the fnancng cost by nceasng cedt ssuance and pomotng ual fnancal competton on the othe hand. The decease of fnancng cost has a multple effect, leadng to a long-tem vtuous ccle of sustanable development of agcultue and ual fnance. 4. FURTHER ANALYSIS AND POLICY IMPLICATIONS In the pevous secton, we deved seveal popostons egadng the cedt constant faced by ual households. In ths secton, we futhe dscuss these popostons wthn the specfc context of Chna s ual fnancal maket, whch may povde us wth useful polcy mplcatons to undestand the elated ssues. 4.1 Infomaton Asymmety n the Rual Fnancal Maket Ou fst poposton s that, unde symmetc nfomaton, fnancal nsttutons have to employ the stategy of poolng equlbum n cedt ssuance based on expeence and fst thoughts. Ths stategy s not pefeed even f the dstbuton of households n the maket s accuately estmated. Effcency s sacfced n ths type of cedt allocaton favong households wth low magnal poducton ablty and fnancal nsttutons face hghe sk. Hee the key questons ae: Does asymmetc nfomaton exst n the maket? If yes, how does t exst? It s obvous that fnancal nsttutons whch have been sevng ual aeas fo a long tme o ae establshed n ual aeas have bette nfomaton of ual aeas and ual households (level of assets, poducton and nvestment ablty, cedt level) than bankes n lage fnancal nsttutons. Whle these bankes ae tyng to fnd the value of to detemne the lne of cedt fo ual households, font-lne offes of

9 Ma and L 645 ual fnancal nsttutons (oganzatons) know clealy the detals and stuaton of each household. The gassoots ual fnancal nsttutons have an nfomaton advantage ove uban-based fomal fnancal nsttutons. These gassoots nsttutons can apply cedt dffeentaton as they have moe access to pvate nfomaton. The allocaton of cedt esouces follows moe closely the sk-etun tade-of pncple. In contast, fomal fnancal nsttutons not only face the cost of settng up the physcal netwok n ual aeas but also at an nfeo poston n tems of the method and channels fo nfomaton collecton. Theefoe, they have lttle ncentve to ente the ual fnancal maket, especally the mcocedt maket. It should be noted that although ual cedt coopeatves (RCCs) ae usually consdeed as some knd of fomal fnance, they possess an nfomaton advantage because of the wde coveage of the netwok and font-lne cedt offes. Regadless of the natue of RCCs, they ae gassoots oganzatons n tems of the opeatng methods, whch may explan why they have an edge ove fomal fnancal nsttutons n ual aeas. 4.2 Inteest Rate Polcy n the Rual Fnancal Maket Ou second poposton s that when the loan ate ses, the quantty of ual households demand fo cedt deceases, the poft magns of the ate of etun on nvestment dmnshes, and fnancal nsttutons ssue moe cedt. And when the loan ate ses above the ate of etun on agcultual nvestment, fnancal nsttutons wll ext the maket and ual households cedt demand wll be completely suppessed. In ths case, a loan ate whch s too hgh s unfavoable to both ual households and fnancal nsttutons. The above concluson can be easly llustated by the followng model. Assume a boowe wth medum sk needs to boow captal L to fnance a poject. Followng Stgltz and Wess (1981), the pobablty of success of the poject s p and the coespondng ate of etun s R ; the pobablty of falue s 1 p and the coespondng etun s 0. All pojects have the same expected ate of etun,.e. p R p R R, j. The densty functon of the pobablty of success of the poject j j s f( p ) and the coespondng dstbuton functon s F( p ). The loan ate fo the boowe to boow fom the bank s. If the nvestment poject succeeds, the boowe has to pay the bank L(1 ) ; f the nvestment poject fals, the boowe has to pay the bank a popoton of the pncpal, whee 0 1. If the bank, wth medum sk, s n a compettve cedt maket, the supply of cedt captal s an nceasng functon of etun on loan pe unt,.e. S S( ), S 0, of whch s the etun on loan pe unt. The expected poft of the boowe s E pl( R 1 ) (1 p) L Only when the expected poft s lage than 0 wll the boowe apply fo loans. Assume ˆp to be the ctcal success ate of the poject at evey nteest ate, whch means ˆp s the success ate when 0 E s satsfed. Fom the above equaton, the followng can be obtaned:

10 646 Cedt Constant and Rual Fnance n Chna R pˆ (1 ) Because de / dp L(1 ) 0, only when p pˆ, wll the boowe apply fo loans. As pˆ 0, pˆ R (1 ) 2 0 The above fomula shows that when the loan ate ses, the ctcal success ate of the nvestment poject wll fall. Because the oveall aveage success ate of the pojects s pˆ p P pf ( p) dp / F( pˆ ), we can obtan 0. Theefoe, as the loan ate ses, the 0 qualty of pojects fnanced by the loan wll dop. Ths concluson mples that smply asng the loan ate may change the sk status of the cedt maket. As moe and moe hgh-sk boowes ente the maket, the accumulated sk may lead to the collapse of the cedt maket. Evdence fom Chna s ual fnancal maket also suggest that n the stuaton n whch the ate of etun on agcultue s geneally low, smply asng the loan ate to cove the cost of fnancal nsttutons can hadly ncease the loan amount ssued to ual households. On the contay, t dscouages ual households fom obtanng loans. Accodng to ou suvey on the feld staff (each membe was esponsble fo 2,700 ual households on aveage) of seven RCCs n two povnces n westen Chna, f the loan ate ses to hghe than 12%, the ual households demand fo cedt dops sgnfcantly. If the loan ate ses to hghe than 15%, ual households dae not apply fo loans. We gave ou questonnae to 70 households n fou towns n the two countes. The answe to the queston Can you bea a loan ate of 15% fo poducton o consumpton? was 100% negatve. The study of the eseach goup of The Chnese Banke (2007) gave a smla concluson: whle ual households can bea a wde ange of nteest ates, they have lttle ntenton to bea hgh ones. Of the 330 qualfed samples, 70% households answeed that they could only bea a loan ate lowe than 10% and only 30% could bea a loan ate hghe than 15% (see Table 1). Table 1 The Affodablty of Rual Households towads Loan Rates The Hghest Loan Rate whch Rual Households can bea (X) Numbe of Households Popoton (%) X 5% % < X 7% % < X 10% % < X 15% X > 15% Souce: Reseach goup of The Chnese Banke (2007).

11 Ma and L Regonal Economc Development and ts Impact on Rual Cedt Ou thd poposton s that the development of egonal economy s cucal fo allevatng cedt constant caused by asymmetc nfomaton: In egons wth a bette economy, fnancal nsttutons tend to ssue moe cedt to ual households even unde asymmetc nfomaton. Ths s because when fnancal nsttutons could not dentfy the types of ual households, they manly efe to the oveall development of the egonal economy to detemne the cedt ssued. In othe wods, the developmental level of the egonal economy has a sgnalng effect. It s woth notng that uban-based lage commecal banks ae the man fnancal nsttutons that ae toubled by asymmetc nfomaton. They ae eluctant to ente ual cedt maket because of hgh nfomaton costs. Howeve, when the egonal economy sends out encouagng sgnals, these fnancal nsttutons wll expand cedt ssuance as moe type II ual households tansfom to type I. Theefoe, pomotng the development of the egonal economy s an mpotant channel fo attactng extenal fnancal esouces. As the ual economy develops, uban-based fnancal nsttutons wll natually e-ente the maket and ncease the ual cedt ssuance as well. Ths means that the ate of etun on agcultue should be contnuously stmulated and development of the ual economy should be pomoted n ode to attact the nflow of extenal fnancal esouces nto the ual fnancal maket based on effcency. In the end, the mutual development of agcultue and ual fnance wll be acheved. The elatonshp between ual cedt and egonal economc development can be easly llustated by Fg. 7 and Fg. 8, whee the total amount of ual cedt ssued s plotted togethe wth dffeent levels of egonal economc development based on cosspovsonal data. Fom Fg. 7 we can see that ual cedt s postvely coelated wth egonal GDP and one pecent ncease n GDP wll esult n 0.93 pecent of ual cedt. The same concluson can be obtaned fom Fg. 8, whee captal fomaton s used as a poxy fo egonal economc development. In a ecent study, L and Ma (2013) also show that the development egonal economy can sgnfcantly allevate the poblem of asymmetc nfomaton n the agcultual loan maket and educe the tansacton costs between maket patcpants as well. Theefoe, n ode to encouage moe ual cedt, the qualty of egonal economy should be mpoved contnuously. Fg. 7: GDP and Total Amount of Rual Cedt

12 648 Cedt Constant and Rual Fnance n Chna Fg. 8: Captal Fomaton and Total Amount of Rual Cedt 4.4 The Long-Tem Sustanable Development of Rual Fnance Ou fouth poposton s that an ncease n the ate of etun on agcultue can loosen the cedt constants on ual households, as well as loweng the fnancng cost by nceasng cedt ssuance and pomotng ual fnancal competton, leadng to a vtuous ccle of sustanable development of both agcultue and ual fnance. As s well known, fnance s an ntemeday ndusty. In ode to guaantee commecal sustanablty, commecal fnancal nsttutons wll not ente ual cedt maket unless the loan etun could compensate fo the expected loss. On the contay, f the sk-adjusted ate of etun on agcultual loans s hghe than the aveage poftablty of fnancal nsttutons, the poft-seekng motves would natually lead them nto enteng the maket and expandng the busness n cedt. Theefoe, the se n the ate of etun on agcultue causes not only an ncease n the nflow of fnancal esouces; t also mpoves the sustanablty of exstng fnancal esouces. Only a contnuous se n the ate of etun can lead to a wn-wn stuaton of long-tem development of agcultue and ual aeas. A futhe poposton s that the allocaton of fnancal esouces should taget an mpovement n the ate of etun on agcultue. Any fnancal supply whch cannot do so s neffectve (o even pedatoy). Theefoe, beng effcency-oented s the bass fo the long-tem development of both agcultue and ual fnance. Though n ou model cedt constants on ual households ae closely elated to the ate of etun on agcultue and the development of egonal economy, othe factos may also play an mpotant ole n allevatng ual households cedt constants. In ode to llustate ths, egesson analyss s conducted based on suvey data fom 1,100 ual households n a westen cty of Chna. The empcal esults n Table 2 shows that vaous factos affect the amount of cedt that ual households obtan, ncludng egonal economc development, total famly assets, land pe capta, famly ncome ( both agcultual and nonual), famly labo foce ato, educatonal level, agcultual technology acqued, and socal captal. These esults tend to suggest that the mutual development of agcultue and ual fnance s not an easy task and depends on many thngs. In ode to emove the cedt constant faced by ual households, almost evey aspect of ual socety needs to be taken nto consdeaton. Fo nstance, n most cases, povety s elated to the lmted contact wth the maket, netwok, and technques ual households have. Ths mples that non-fnancal sevces such as techncal tanng can mpove the ate of etun on nvestment of the poo, whch wll n tun mpove the envonment of ual fnance.

13 Ma and L 649 Regonal economc development Total famly assets Land pe capta Agcultual ncome Non-ual ncome Labo foce ato Educatonal level Agcultual technology Socal captal Constant Table 2 Regesson Analyss of Cedt Obtaned by Rual Households Independent Vaable: Cedt Obtaned by Rual Households (3.964) (8.144) (3.356) (2.291) (3.828) (-1.055) (-0.732) (3.952) (7.630) (3.406) (2.189) (3.962) (-1.204) (4.721) (-1.557) (3.836) (7.473) (2.872) (2.135) (3.745) (-1.216) (3.541) (1.936) (-0.797) (3.839) (7.216) (2.862) (2.155) (3.713) (-1.217) (3.629) (1.988) (4.461) (-0.888) Obsevatons Adjusted 2 R Notes: 1. Fo ths analyss, ove 1,300 questonnaes wee dstbuted among 11 countes of the cty, of whch 1,100 wee etuned and qualfed fo use. 2. The backeted value s the t-value. 3. at 10% confdence level, at 5% confdence level, at 1% confdence level. 5. CONCLUSIONS How to solve the poblem of ual households cedt constant s one of the coe ssues n Chna s ual fnance. Ths pape povdes a theoetcal famewok fo undestandng ths poblem and seveal theoetcal popostons as well as polcy mplcatons ae obtaned though compaatve analyss. The basc conclusons of ths pape ae as follows: (1) fnancal nsttutons have to adopt the poolng equlbum stategy to detemne cedt allocaton unde asymmetc nfomaton, whch leads to lowe effcency and hghe sk n ual fnance; (2) excessve hgh nteest ate wll be a hndance fo the sustanable development of both ual households and fnancal nsttutons snce t wll squeeze the poft magn of agcultual nvestment n the long un; (3) the development of egonal economy s

14 650 Cedt Constant and Rual Fnance n Chna mpotant as fnancal nsttutons tend to ncease the cedt n moe developed egons even n the face of asymmetc nfomaton; (4) the ncease of agcultual etun s not only helpful n attactng and stablzng fnancal esouces but also cucal fo the longtem development of both agcultue and ual fnance. ACKNOWLEDGEMENTS The eseach s suppoted by Natonal Natual Scence Foundaton of Chna (Gant No ), Natonal Natual Scence Foundaton of Chna (Gant No ) and Natonal Socal Scence Foundaton of Chna (Gant No. 12&ZD089). REFERENCES 1. Gupta, M. and Chaudhu, S. (1997). Fomal Cedt, Coupton and the Infomal Cedt Maket n Agcultue: A theoetcal Analyss. Economca, 64(254), Kellee, T. (2004). Impefect Substtutes: The Local Poltcal Economy of Infomal Fnance and Mcofnance n Rual Chna and Inda. Wold Development, 32(9), Khandle, S. (1988). Fghtng Povety. London: Routledge. 4. L, S. and Ma, Y. (2013). Fnancal Ecology, Rual Income and Agcultual Development: Evdence fom Chna. Pak. J. Statst., 29(6), Ln, Y. (2000). Re-dscuss the Insttuton, Technology and Chna s Agcultual Development (n Chnese). Bejng: Pekng Unvesty Pess. 6. Osbone, T. (2006). Cedt and Rsk n Rual Developng Economes. Jounal of Economc Dynamcs and Contol, 30(4), Popkn, S. (1979). The Ratonal Peasant: the Poltcal Economy of Rual Socety n Vetnam. Bekeley: Unvesty of Calfona Pess. 8. Reseach goup of The Chnese Banke (2007). Fnancal demand n tadtonal ual aeas and the dstbuton of nsttutons. The Chnese Banke, 7, Stgltz, J. and Wess, A. (1981). Cedt Ratonng n Makets wth Impefect Competton. Amecan Economc Revew, 71(3), Udy, C. (1994). Rsk and Insuance n a Rual Cedt Maket: A Empcal Investgaton n Nothen Ngea. Revew of Economc Studes, 61(3), Vllas-Boas, J. and Schmdt-Moh, U. (1999). Olgopoly wth Asymmetc Infomaton: Dffeentaton n Cedt Makets. The RAND Jounal of Economcs, 30(3), Zaman, H. (2004). The Scalng-Up of Mcofnance n Bangladesh: Detemnants, Impact, and Lessons. Wokng Pape. 13. Zhang, J. (2005). Fnancal efom and development of the ual fnancal maket-the case of Wenzhou. Chna Development (n Chnese), 2(1), Zhang, J. and Gao, X. (2006). Game n captal njecton and efom of ual cedt coopeatves. Jounal of Fnancal Reseach (n Chnese), 309(3),