A Direct Fiace Deposit ad Borrowig Method Built Upo the Web Implemeted Biddig ROSCA Model Adjuct Professor Kue-Bao (Frak) Lig, Natioal Taiwa Uiversity, Taiwa Presidet Yug-Sug Chie, SHACOM.COM INC., Taiwa Fiace Researcher Big-Lu Hsieh, SHACOM.COM INC., Taiwa Fiace Researcher Ye-Tse Che, SHACOM.COM INC., Taiwa ABSTRACT ROSCA (Rotatig Savigs ad Credit Associatio) works all over the world. I Taiwa, the biddig auctio mechaism is implemeted o the basis of ROSCA, developig ito biddig ROSCA. The characteristics of biddig ROSCA, such as direct fiace, combiatio of deposit ad borrowig, ad self-determied iterest rate are importat foudatio for future fiacial services. Yug-Sug Chie (2002) devised a ROSCA busiess method that accommodates more participats by removig the biddig leader ad by offerig credit guaratee. This study is based upo such a strategy, usig the cocept of the overall fud balace to broade the scopes of the biddig ROSCA, freeig which from the limit o participats, amout, ad time to facilitate the biddig ROSCA ad arrive at the deposit ad borrowig iterest rates that are actually determied by the market. INTRODUCTION This research will improve o the foudatio of direct fiace ad combiatio of deposit ad borrowig methods of the traditioal ROSCA (biddig ROSCA), with a aim to overcome the costraits of the operatio of the traditioal ROSCA ad expad the scope of use, as well as to utilize the computer ad Iteret techologies to create a brad ew deposit ad borrowig method. Hece, this study has three purposes: 1. To create a direct fiace method that eables the borrower ad the leder to freely choose the roles they wat to play i the system. 2. To improve the traditioal ROSCA ad break off the bouds of the origial structure to trasform it ito a direct fiace product that achieves greater efficiecy with the use of computer ad Iteret techologies. 3. To create a frictio-free tradig eviromet that allows the participats to decide the iterest rates autoomously so as to arrive at the real deposit ad borrowig iterest rates i the market. LITERTURE REVIEW This study plas to costruct a deposit ad borrowig method built upo the direct fiace model. There are two facets for discussio of our research o academic papers. The first facet is related papers o the direct fiace model, ad the other is related papers o the ROSCA model. Direct Fiace Tsai (2000) thiks that large corporatios rely more ad more o the direct fiace model for their capital eeds while mid-sized compaies ad small busiesses do ot. Tsai sees the act of busiesses borrowig from baks as a example of idirect fiace ad the act of raisig fud from issuig commercial papers i the moey market as direct fiace. As the percetage of the fud eeds of large corporatios filled by the secodary market gradually icreases at the expese of other meas of fud gatherig, it stads to reaso that large corporatios are becomig more depedet o the direct fiace model for their fiacig. Mid-size compaies ad small busiesses do ot exhibit this tred as
they are less likely to issue commercial papers o the secodary market due to the lack of credit worthiess. However, that does ot prove that mid-sized compaies ad small busiesses have less desire for direct fiace meas to satisfy their fud eeds. As this study demostrates, direct fiace will exert a icreasigly greater ifluece o the deposit ad borrowig. Biddig ROSCA Model Huag (2001) costructed a web-based system that implemeted the ROSCA model. The system shows the feasibility of a web-based ROSCA implemetatio through a set of methods that are complete, reliable, private, fair, verifiable, ad able to validate eligibility. The reliability biddig process i the web-based ROSCA implemetatio has bee verified ad prove feasible, thus there are o techical issues with the web operated system. Chie (2002) removes the biddig leader from the traditioal ROSCA ad places the fiacial istitutio i its place, utilizig computer systems ad etwork techologies to eable web-based ROSCA. With o biddig leader, the total loa amout ca be calculated i the followig two formulae depedig o whe the iterest is paid: The required fud to the bid-wier for time period 1 = (Cotributio amout x (Number of time periods 1)) premiums. The required fud to the bid-wier for the remaiig terms: A U - I N 1-1 U....(2.1) or A U N - 1 U -1 I i i 1-1..(2.2) The above method mitigates the risk of the biddig leader havig moral hazard by puttig the fiacial istitutio i charge of the operatio, it also allows the ROSCA implemetatio to become stadardized ad expad its participat base; however, it does ot remove or remedy ay of the restrictios of the ROSCA model itself. DEPOSIT AND BORROWING STRATEGY IN DIRECT FINANCE There is a iheret disadvatage with either the traditioal ROSCA model or the web ROSCA model: each biddig pool is strictly limited by several factors. All the bidders of the ROSCA must all comply with the same criteria before actual biddig may begi. These criteria iclude: first, the umber of bidders must be equal to the umber of time periods for the pool; secod, uder the same biddig pool, each bidder must use the same time frame for submittig ad wiig bids as well as payig the same cotributio amout. To mitigate the above metioed limitatios ad iefficiecies, this study will costruct a direct fiace system that is built upo the web implemeted ROSCA model ad based o the cocept of fud balace to demostrate the efficiecy improvemets uder the direct fiace model. This study utilizes the improved structure proposed by Chie (2002) that adds the cocept of fud balace to create a ew method of deposit ad borrowig withi the direct fiace model. The traditioal ROSCA is a direct fiace implemetatio with umerous major limitatios ad restrictios that has a fud balace at equilibrium. This study seeks to assimilate the differet types of ROSCA ad apply the cocept of fud balace throughout the etire trasactio. At the ed of each time period i the study, the differet eeds for fuds for each member ca be calculated (the bidders have borrowig eeds, ad the o-bidders have deposit eeds). The optimal fud balace is the fuds deposited is exactly equal to the fuds borrowed, ad therefore at zero, which is a type of direct fiace. However, the probability for the fud balace to be exactly zero is very low; thus the fud balace would likely be greater or less tha zero. If the fud balace is greater tha zero, the system would geerate iterest expese; if the fud balace is less tha zero, the system would receive iterest paymet. Thus, this study assumes that the fud balace has to be less or equal to zero. To calculate the fud balace, the system takes all the deposits from all the o-wiig participats, subtracts that umber by the total loa amout of all the wiig bids at that time period.
There are six steps i determiig the fud balace that will satisfy the eed for fuds for as may participats as possible: steps 1 ad 2 will calculate the fud balace whe there are o wiers. If the balace is less or equal to zero, the the balace for this time period is at equilibrium; otherwise, if the fud balace is greater tha zero, the system will proceed to step 3. Steps 3 ad 4 calculates the miimum umber of bid wiers that will allow the fud balace to be less or equal to zero to determie the fial outcome for this time period. Steps 5 ad 6 will take the results from the calculatios to determie the loa amout to be received by the bid wiers ad the required deposit for the owiers. The steps will be discussed i detail below. Step 1: Calculate the Fud Balace Assumig No Bid-wier First, the system calculates the curret fud balace with o wiig bids. I step 1 we oly calculate the total deposit amout from all the participats. The amout is calculated depedig o whe the iterest is realized. The three formulae below are used. Iterest paid first: U - IH.. (3.1) or U - IX (3.2) Iterest paid after: U. (3.3) IH = the maximum bid at the th time period for the specific participat IX = the required deposit amout calculated from the base iterest rate for the th time period. The base iterest rate metioed above meas a reserved bid for the auctio. It ca be provided accordig to the market term deposit rate. The three formulae used i the situatio of iterest paid first are distiguished by the participat who bids or ot. The above metioed iterest paid first meas that the total loa amout received by the wiig bid has already had the etire iterest deducted, similar to a zero coupo bod. Whereas iterest paid after refers to the wiig bidder will receive the etire loa amout at auctio close, ad the iterest paymet will be icluded ito the repaymet istallmets, similar to a ormal loa or mortgage. Step 2: Determie if the Fud Balace for the Previous Time Period X + The Curret Fud Balace B 0 is Less or Equal to Zero If the fud balace for the previous time period X plus the fud balace for the curret time period B 0 is less or equal to zero ( B 0 +X 0), the o oe has wo the bid at the curret time period, which meas that all the participats must deposit the required amout ito the system; coversely, if the fud balace is greater tha zero, the the system will eter step 3. Step 3: Rakig the Borrowig Iterest Rate If the fud balace calculated i step 2 is greater tha zero, the the system would rak the bidders by the correspodig borrowig iterest rate calculated through the bid. We ca use formulae below to calculate the effective borrowig iterest rate for ay participat at the th time period. First we use the formula (3.4) to calculate the average deposit iterest rate (r) for all the prior time periods; ext we use the formula (3.5) to calculate the effective borrowig iterest rate R for the curret time period. The formulae are as follows. -1 i 1-1 i 1 N-i U - I 1 r U -1 i N- -i U - I i 1 r U A j 1 (3.4) 1 1 R j.. (3.5)
N = the total umber of time periods for the auctio = the curret time period, ad N I i = the amout deposited by the participat at the i th time period i < r = the average deposit iterest rate from the 1 st time period to the i th time period i < R = the effective iterest rate A = the total loa amout for the wiig bid at the th time period (please see formulae 3.7 ad 3.8) Oce the effective borrowig iterest rates for all the participats at the curret time period are calculated, the system the raks the rates from the highest to the lowest. Step 4: Determie the Least Number of Bid-wiers that Will Allow the Fud Balace to Be Less or Equal to Zero This step relies o trial ad error to determie the umber of bidders that will allow the combied fud balace for the curret time period ad the previous time period to be less or equal to zero. Mi B +X 0.. (3.6) B = the fud balace of the curret time period whe the wiig umber of bids = X = the fud balace for the previous time period The possible total loa amout of the wiig bids ca be calculated based o iterest realizatio i the followig two formulae. A U - I N - U -1. (3.7) Iterest paid first: Iterest paid after: U N - U -1-1 A.. (3.8) I i i 1 A = the total loa amout for the wiig bid at the th time period N = the total umber of time periods for the auctio = the curret time period, ad N I = the wiig bid for the participat at the th time period I = the amout deposited by the participat at the i th time period i < i The average iterest rate for the wiig bids ca be determied. Next we ca calculate the required deposits IA for each o bid-wier (depositor). Agai, based o whe the iterest is realized, we have two formulae. N- Iterest paid first: U - IA 1 R U Iterest paid after: a...... (3.9) N- a U 1 R U IA.... (3.10) N = the total umber of time periods for the biddig auctio = the curret time period, ad N IA = the required deposit calculated from the average iterest rate for the th time period R = the average borrowig iterest rate a
We ca use IA ad the followig two formulae to calculate the amout required for deposit for the o-wiers. Iterest paid first: U - IA...... (3.11) Iterest paid after: U... (3.12) IA = the required deposit calculated from the average iterest rate for the th time period Step 5: Settlemet of Bid-wiers. For the bid wiers (borrowers), calculatig how much they will receive from the wiig bid will deped o whe the iterest is realized (Formulae (3.7) ad (3.8)). The wiig bidders will eed to repay the loa i the forthcomig time periods, the formula to calculate how much the wiers will eed to repay are as follows: Iterest paid first: U. (3.13) Iterest paid after: U I..... (3.14) I = the wiig bid for the participat at the th time period Step 6: Settlemet of No Bid-wiers If the curret time period has wiig bids, we ca use formulae (3.11) ad (3.12) to calculate the amout required for deposit for the o-wiers. If there were o wiig bids for the curret time period, but the participat did submit a bid, the required deposit amout would be calculated as the formulae (3.1), (3.2), ad (3.3). SIMULATION ANALYSIS I this sectio this study assumes that we have A, B, C, D, E, F, six participats participatig i the direct fiace model for simulatio aalysis. The biddig criteria ad biddig cash flow are listed as follows i Tables 1 ad 2. Table 1: Participat Biddig Criteria Criteria Participat Term (Year) Iterval (Moth) Total Time Periods Maximum Bid Etry Date A 1 1 12 $15,000 Jauary 1st B 1 1 12 $10,000 Jauary 15th C 5/6 0.5 20 $15,000 February 1st D 1 0.5 24 $10,000 Jauary 1st E 1 1 12 $15,000 Jauary 15th F 0.25 0.5 6 $10,000 October 1st
Table 2: Biddig Cash Flows Jauary 1st Jauary 15th February 1st February 15th March 1st March 15th April 1st April 15th May 1st May 15th Jue 1st Jue 15th Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow A 800 (156,200) 15,000 15,000 15,000 15,000 15,000 B 300 9,700 300 9,700 400 9,600 400 9,600 400 9,600 400 9,600 C 700 14,300 800 14,200 900 (269,700) 15,000 15,000 15,000 15,000 15,000 15,000 15,000 D 800 8,092 600 9,400 600 9,400 600 9,400 500 7,880 500 9,500 400 9,600 400 9,600 400 9,600 400 9,600 400 (224,800) 10,000 E 700 14,300 700 14,300 500 14,500 500 14,500 500 14,500 500 14,500 F Fud Balace (148,108) 33,400 38,700 47,600 (246,820) 48,600 39,600 48,700 39,600 48,700 (194,800) 49,100 Accumulated Fud Balace (148,108) (114,708) (76,008) (28,408) (275,228) (226,628) (187,028) (138,328) (98,728) (50,028) (244,828) (195,728) Auctio Result A wis Noe Noe Noe C wis Noe Noe Noe Noe Noe D wis Noe July 1st July 15th August 1st August 15th September 1st September 15th October 1st October 15th November 1st November 15th December 1st December 15th Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow Bid Cash Flow A 15,000 15,000 15,000 15,000 15,000 15,000 B 300 9,700 400 (108,400) 10,000 10,000 10,000 10,000 C 15,000 15,000 15,000 15,000 15,000 15,000 15,000 15,000 15,000 15,000 D 10,000 10,000 10,000 10,000 10,000 10,000 10,000 10,000 10,000 10,000 10,000 10,000 E 500 14,500 500 14,796 500 9,500 500 (164,000) 15,000 15,000 F 200 (49,000) 10,000 10,000 10,000 10,000 10,000 Fud Balace 40,000 49,200 40,000 (68,604) 40,000 44,500 (9,000) (119,000) 50,000 60,000 35,000 45,000 Accumulated Fud Balace (155,728) (106,528) (66,528) (135,132) (95,132) (50,632) (59,632) (178,632) (128,632) (68,632) (33,632) 11,368 Auctio Result Noe Noe Noe B wis Noe Noe F wis E wis Noe Noe Noe Noe As Table 2 shows, o Jauary 1 st, the first time period, there were 2 participats participatig i the biddig, A ad D. Each submitted a bid of $800. The fud balace would be $23,400 ($14,200 + $9,200 = $23,400) if there were o wiers. As this umber is greater tha zero, the system would eed a wier for the auctio. Through calculatio of (3.4) ad (3.5), the effective iterest rate for participat A is higher tha the effective iterest rate for participat D, ad the system will award the bid to participat A. The total loa amout is $156,200, ad participat A would eed to repay $15,000 o the first of every moth util the last time period, December 1 st. Because participat A wo the bid, participat D is ow a depositor. Usig participat A s effective mothly iterest rate of 0.9248%, the required deposit amout for participat D is calculated to be $8,092. The fud balace for this time period is ($148,108), which is $8,092 - $156,200. The above is summarized i the followig Table 3. Participat Table 3: Compariso of Borrowig Iterest Rates (Jauary 1 st ) Criteria Calculatios Effective Total Iterval Maximum Iterest Total Loa Time (Moth) Bid Rate amout Periods ( mothly ) Term (Year) Required Deposit Amout A 1 1 12 $15,000 0.9248% 156,200 D 1 0.5 24 $10,000 0.7064% 8,092 Regardig Jauary 15 th, the secod time period, there were three participats who submitted bids, B, D, ad E, with bids of $300, $600, ad $700 respectively. Suppose all three participats fail to wi the bid, the fud balace for this time period would be $33,400 ($9,700+$9,400+$14,300=$33,400). The add this umber to the fud balace for the previous time period of -$148,108, the accumulated fud balace for this time period comes to $114,708. As the accumulated fud balace is less tha zero, o participat will wi the bid. Similarly, the accumulated fud balace cotiues to be less tha zero for the third ad fourth time periods o February 1 st ad February 15 th, o oe would wi the bid for those two time periods, either.
As for the fifth time period, March 1 st, participats C ad D submitted bids for $900 ad $500, participat A has already wo a bid, ad thus is ieligible. Agai, assumig participats C ad D do ot wi the bid, the fud balace for this time period is $38,600 ($14,100+$9,500+$15,000=$38,600). Addig this umber to the fud balace of the previous time period of -$28,408, ad the accumulated fud balace for March 1 st comes to $10,192. As the accumulated fud balace is greater tha zero, a bid wier is eeded for this time period. Through calculatios of formulae (3.4) ad (3.5), the effective iterest rate offered by participat C is higher tha the effective iterest rate offered by participat D, ad the system awards the bid to participat C, with the total loa amout of $269,700 ad repaymet of $15,000 o the first of the moth begiig i April util the time period eds i December. For participat D, the required deposit amout is calculated through the effective iterest rate offered by participat C (1.2618% mothly), ad it comes to $7,880. The fud balace for this time period ca the be calculated. The required deposit amout for participat D $7,880 the total loa amout for participat C $269,700 + the repaymet for participat A $15,000, ad the fud balace comes to -$246,820. Table 4 lists the required deposit amout ad the effective iterest rates for the bids submitted by each participat. Participat Table 4: Compariso of Borrowig Iterest Rates (March 1 st ) Criteria Calculatios Effective Total Iterval Maximum Iterest Total Loa Time (Moth) Bid Rate amout Periods ( mothly ) Term (Year) Required Deposit Amout C 5/6 0.5 20 $15,000 1.2618% 269,700 D 1 0.5 24 $10,000 0.3889% 7,880 CONCLUSION I a world leders ad borrowers still rely o the idirect fiace model, the advatages of the direct fiace model are udeiable. This study utilizes the traditioal ROSCA model to create a ew direct fiace method. Usig the accumulated fud balace cocept to determie the auctio outcome at a particular term, the method improves upo the ROSCA model by removig the limitatios o the umber of participats, time, ad cotributio amout, ad closely approaches the realizatio of a truly direct fiace model. Furthermore, this study also provides the ROSCA model a ew way to calculate iterest rates, allowig the participats to obtai a fair ad mutual beeficial iterest rate uder varyig coditios. Below lists the mai accomplishmets of this study. 1. Direct fiace deposit ad borrowig method: this study creates a ew method by usig the priciples of the traditioal ROSCA model. It allows the borrowers ad the leders to iteract without a itermediary, thus miimizig cotrols o iterest rates. 2. Itroduces the cocept of accumulated fud balace. The traditioal ROSCA model is severely limited by the umber of participats, time, ad cotributio amout, makig implemetatio extremely difficult. This study combies differet types of ROSCA alog with participats of varyig requiremets together, ad automatically determies the outcome of the auctio based upo the accumulated fud balace at each term. Oce the outcome is determied, the system the calculates the total fud wo by the wier, ad the required deposit amout for the o bid-wiers. This ot oly removes the limitatios of the traditioal ROSCA model, it also greatly improves the efficiecy for the platform operators that implemet this method. 3. Reflectio of market rates. As a explicit beefit of the direct fiace model, both the leder ad the borrower obtai the most favorable iterest rates. As such, the iterest rates determied through the method ca best reflect the iterest rates for the curret market coditios. Ad, as this method combies all ROSCA with differet coditios, it elimiates the probability for those illogical bids that adversely affect the traditioal ROSCA biddig to affect the ew method.
REFERENCES Chie, Y. S. (2002, August). A Maagemet Method for Fiacig with Credit Cotrol (Patet publicatio No. 00497059). Taipei: Itellectual Property Office, Miistry of Ecoomic Affairs, Taiwa, R.O.C. Chuag, W. (1997). Study of fiacial itermediatio -Aalysis of direct fiace ad idirect fiace. Upublished master s thesis, Soochow Uiversity, Taipei. Ho, Y. H. (2004). Game theory aalysis of ROSCA. Upublished master s thesis, Natioal Cheg Kug Uiversity, Taia. Huag, K. H. (2001). A practical protocol of etwork fellowship auctio. Upublished master s thesis, Natioal Chaghua Uiversity of Educatio, Chaghua. Huag, R. J. (1998). The strategy of savigs cooperatio. Taipei: Ha Yu Publicatio Compay. Lee, M. G. (2007). Determiats i the growth of direct fiace/idirect fiace. Upublished master s thesis, Natioal Taiwa Uiversity, Taipei. Tsai, H. W. (2000). Moetary policies, direct fiace ad idirect fiace. Upublished master s thesis, Natioal Taipei Uiversity, Taipei.