Yahoo " (MIC) 2004 2003 B2C C2C 204 2002 36% 0.39% 1.5% 2003 125 B2C C2C 30% 70% 2004 200~250 2004 10 11 Yahoo 344 ebay 128 ( 1) 2005 1 Yahoo PDA 97,4

C2C Logstc hhchang@mcu.edu.tw Yahoo 500 Logstc 93% C2C C2C Imposng Taxes on C2C Internet Aucton the Applcaton of Logstc Regresson Abstract Accordng to taxaton prncples about Internet aucton announced by Taxaton Offce and aucton data collected from Yahoo, Ths paper apples a Logstc model to analyss, predct, and try to fnd professonal Internet auctoneers, who had to pay taxes. As a result, ths paper yelds 93% predcton probablty. The repeat of aucton tems, the tmes of aucton transacton, the rato of sellng tmes to the total of sellng and buyng tmes, and the categores of aucton tems had good explanatory power for odds, but the amount of aucton revenue dd not. So, f crtcal prncples of taxaton are based on the amount of aucton revenue and the tmes of aucton transacton only, Taxaton Offce can not dscrmnate who should pay taxes. Besdes, accordng to the results of regresson analyss, auctoneers need not to concern tax burden f they are real amateurs. Taxaton Offce should let amateur auctoneers realze t and prevent tax burden from generatng a downsde effect on C2C electronc commerce. Keywords Internet Aucton, Logstc Model, C2C Electronc Commerce - 63 -

Yahoo " (MIC) 2004 2003 B2C C2C 204 2002 36% 0.39% 1.5% 2003 125 B2C C2C 30% 70% 2004 200~250 2004 10 11 Yahoo 344 ebay 128 ( 1) 2005 1 Yahoo PDA 97,458 39,930 60,373 855,481 223,659 207,212 425,521 97,055 225,775 127,159 503,358 155,064 152,252 177,355 92,451 15 254 3,440,103 Yahoo (93 10 11 ) B2C C2C - 64 -

( ) ( ) C2C C2C 1) Yahoo ebay 2) C2C B(busness) C(customer) (2000) (2000) (2001) (2004) (2000) (2001) (2003) 2005 C2C C B Logstc ( Logt ) Probt (Neural Network) ( C B) Altman(1968) Altman and Lors(1976) Snkey(1975, 1978) Pettway and Snkey(1980) (odds) 0 1 0~1 Logstc Logstc Logstc Probt Logstc - 65 -

Martn(1977) West(1985) Whalen and Thomson(1995) Logstc Espahbod(1991) 1983 48 Logstc Logstc Ederngton(1985) Logstc Probt Probt Logstc Gentry, Newbold, and Whtford(1987) Logstc Odom and Sharda(1990) 1975~1982 65 64 Tam and Kang(1992) 1985~1987 59 Logstc Logstc Logstc (1996) (1996) (1997) (1999) (2001) (2003) Logstc Logstc (OECD) 1998 10 ( www.mof.gov.tw, 2004/11/23) ( 2004/09/09) ( 3-4 ) - 66 -

- 67 - ( 93/08/23) ( ) ( 2004/05/20). ( 2)

1 C2C C B C B ( C B) C C B 2 1. 2. 3. 6 1. 6% 2. 3. 6 20 1. 2. (1) 1% (2) 5% 3. 20 3000 1. 2. 5% 3. 6% 25% 6% 25 25% - 68 -

(1) C2C (0) C B C C C ( ) C B 1 (2004.4) 2003 Yahoo 350 ebay 17 ebay 90 2003 500 250 ( ) ( 1) 0 1 9 9 0 1 Altman(1968) (0) (1) ( ) - 69 -

3 1 A 200 Noka 50 Moto 50 50 50 A 200/4=50 2 ( 2) 3 4 A 200 20 A 200/(200+20)=0.91 1-70 -

5 Yahoo! 15 ( 1) A 200 Noka 50 Moto 50 50 50 Yahoo! A 3/15=0.6 3 / Logstc ( 2004 133) Logstc 0 1 Logstc Logstc y y ( 0) y =1 y ( ) y ( 0) y =0 ( ) y x y = α + β + ε x (1) - 71 -

P (2) ( y = 1x ) = P[( α + βx + ε ) > 0] = P[ ε > ( α βx )] (2) ε Logstc Logstc (2) P y = 1x ) = P[ ε ( α + βx )] = F( α + βx ) (3) ( F ε Logstc (3) 1 P( y = 1x ) = p[ ε ( α + βx )] = (4) ε 1+ e (4) Logstc S ε P y = 1x ) =0 ( P y = 1x ) =1 Logstc 0 1 ( Logstc Logstc ε P( y 1 = 1x ) = ( α + βx ) 1+ e (5) (5) ε ( α + βx ) x P ( y = 1x ) = p p α + βx 1 e = = (6) ( α + βx ) α + βx 1+ e 1+ e p x (7) α + βx e 1 1 p = 1 ( ) = (7) α + βx α + βx 1+ e 1+ e (odds) (8) p 1 p = e α + β x (8) p = α + βx ln 1 p (9) (9) Logstc logt form - 72 -

k (9) K p ln = + + k xk p α α β (10) 1 k= 1 p = P y = 1x, x,..., x ) ( 1 2 k p ln = α + β1repeat + β 2 AMOUNT + β 3TIMES + β 4RATIO + β 5CATEGORIES 1 p (11) p 1 p REPEAT AMOUNT TIMES RATIO CATEGORIES (11) Logstc 4 1.6 19.2 6 72 6 13 40 3~4 5 5 Levene s test t 5 6 Box s M 799.646 F 51.869 7 5 Kolmogorov- Smmov Z Logstc - 73 -

4 (0) 1.103.147 0.001 (1) 2.327 3.898 0.176 0) 16090.749 5537.080 2265.524 (1) 60951.415 10061.414 19736.319 (0) 13.113 1.862.841 (1) 40.451 7.124 9.611 / (0).628.028.018 (1).947.006.0.005 (0).282.018.009 (1).127.0092.005 5 Levene s test for equalty of varances T test for equalty of means F Sg. t Sg.( ) 94.972.000-8.973 498.000-8.973 250.306.000 30.540.000-4.865 498.000-4.865 255.561.000 110.984.000-9.645 498.000-9.645 252.811.000 / 439.567.000-22.153 498.000-22.153 288.005.000 74.179.000 12.832 498.000 12.832 361.925.000 =5% 6 Box s M 2520.982 F 249.919 10 1185676.5.000 7 / Kolmogorov-Smmov Z 8.452 8.050 8.452 8.810 5.769.000.000.000.000.000 =5% - 74 -

8 Omnbus Test (11) 2534.238 Cox & Snell R 2 Nagelkerke R 2 2 R 1 (11) Cox & Snell R 2 Nagelkerke R 2 0.656 0.875 5 Hosmer & Lemeshow Test (12) 7.823 P 0.451( 0.05) ( 2004 442) 8 Omnbus Test Cox & Snell R 2 Nagelkerke R 2 Hosmer & Lemeshow Test 534.238(.000) 0.656 0.875 7.822(0.451) P 9 (11) Wald 95% 9 Wald ( ) 6.249 1.005 35.059.000.000.000.000.997.032.010 9.600.002 / 6.972 1.524 20.918.000-6.669 1.998 11.139.011-13.325 2.085 40.832.000 =5% 10 (0) (1) 17 6.8% ( C B) (1) (0) 18 7.2% - 75 -

(11) 93% 5 (odds) 10 0 1 0 233 17 93.2 1 18 232 92.8 93 0.5 Yahoo C2C ( C B) Logstc 250 5 93% 0( ) 1( ) ( C B) 1( ) 0( ) C2C - 76 -

1. - 2004 2. 48 1 1997 3. 1753 89 4 4. 25 3 92 5. logt 4 1999 85-104 6. --- 32 3 89 5 7. 40 4 92 12 8. 4 6 89 6 9. ---Delph APH 60 93 3 10. 32 6 90 1 11. Logt 20 1 2001 12. 13 2 1996 13. 8 3 1996 14. 34 2 91 3 15. Altman, E. I., Fnancal Ratos (1968), Dscrmnant Analyss and the Predcton of Corporate Bankruptcy, the Journal of Fnance, Vol.23, No.4, p598-609. 16. and B. Lors (1976), A Fnancal Early Warnng System for Over-The-Counter Broker-Dealers, the Journal of Fnance, Vol.31, No.4, p1201-1217. 17. Collns, R. A. (1980), an Emprcal Comparson of Bankruptcy Predcton Models, Fnancal Management, summer, p52-87. 18. Ederngton, L. H. (1985), Classfcaton Models and Bond Ratngs, The Fnancal Revew, Vol.20, No.4, 1985, p237-262. 19. Espahbod, P. (1991), Identfcaton of Problem Banks and Bnary Choce Models, Journal of Bankng and Fnance, vol.15, no.1, p53-71. 20. Gentry, J. A., P. Newbold and D. T. Whtford (1985), Classfyng Bankrupt Frms wth Funds Flow Components, Journal of Accountng Research, Vol.23, p146-160. 21. Martn, D. (1977), Early Warnng of Bank Falure: Logt Regresson Approach, the Journal of Bankng and Fnance, vol.1, no.3, p249-276. - 77 -

22. Odom, D. M., and Sharda, R. (1990), A Neural Network Model for Bankruptcy P redcton, I J CNN-9011. 23. Ohlson, J. (1980), Fnancal Ratos and the Probablstc Predcton of Bankruptcy, Journal of Accountng Research, Vol.18, no.1, p109-131. 24. Pettway, R. H., and J. F. JR. Snkey (1980), Establshng On-Ste Bank Examnaton Prortes : An Early-Warnng System Usng Accountng and Market nformaton, Journal of Fnance, vol.35, no.1, p137-151. 25. Snkey, J. F. JR. (1975), a Multvarte Statstcal Analyss of the Characterstcs of Problem Banks, Journal of Fnance, vol.30, no.1, p21-34. 26. (1978), Identfyng Problem banks How Do the Bankng Authortes Measure a Bank s Rsk Exposure?, Journal of Money, Credt, and Bankng, vol.10, no.2, p185-193. 27. Tam, K. Y., and Y. Kang (1992), Manageral Applcaton of Neurnal Networks: the Case of Bank Falure Predcton, Management Scence, 38(7), p926-946. 28. West, R. C. (1985), A Factor Analytc Approach to Bank Condton, Journal of Bankng and Fnance, vol.15, no.1, p253-266. 29. Whalen, G. W., and J. B. Thomson (1998), Usng Fnancal Data to Identfy Changes n Bank Condton, Economc Revew, Federal Reserve Bank of Cleveland, 2 nd Quarter, p17-26. - 78 -