FINAL EXAMINATION - INDEX 1 Mathematics Name: Prof. Pham Huu Anh Ngoe | Dr. Nguyen Phuong Anh RESEARCH METHODS IN FINANCE Duration: 100 min Head of Department of | Lecturer: Student ID: Date: 22 January, 2022 INSTRUCTIONS: 1. This is an open book examination. 2. Students sign and submit the scanned exam paper on Blackboard before 17:15. Total pages: 05 (including this page) Question 1. (20 points) Explain how we can derive a set of 95% confidence level critical values for a Dickey—Fuller Test (i.e. Unit Root Test) using Monte Carlo simulation. Question 2, (10 points) Consider the following AR() model: Ve= rt Ur As stated in chapter 5 (Classical Linear Regression Model Assumptions), the explanatory variables in a regression model are assumed to be non- stochastic, but here y;-; is stochastic. The result is that the estimator for 9 in the model will be biased in small samples. Design a simulation experiment to investigate the effect of the value of g and the sample size on the extent of the bias. Question 3. (10 points) Explain the main steps of the following program to determine MCRR (minimum capital risk requirements) for a long and short position, over a 10-day holding period using daily S&P500 data. 1]# VaR estimation using bootstrapping 2}load(" :/Programming Guide/R Guide/code/spS00.RData")3) 1ibrary (rugarch) 4 5] set. seea(12345) Glepec = ugarchspec(mean.nodel = list (armaQrder=c(0,0)),variance.zodel = 1ist( garchOrder=c(1,1) ,model="sGARCH")) TJgarchit = ugarchfit(spec,data = sp500$ret) sarchi1OfitScoof ("mu"); omega = garchi1@fitScoof ("omega") ‘= garchi10fitScoef["alphai"]; beta = garchi1efit$coef ["betai"] lh = garchitefitsvar Lresid = (garchilefitsresiduale ~ mu) 13)eres = reeid/h70.5 uf 15] = Length (a) Isfnerr = NULL 1 1s)for (m in 1:1000) ¢ 10) hfe = ret_fe = epS00_f¢ = MULL 20) random = sample(sres,size = 10,replace = 7) 21] Constructing one ste ahead forecast 22) hfe = omega + alpharresid(W]72 + beta*hLW] 28) ret_fe = mu + sqre(h_fe[1})#randos [1] 24) spS00_fc = spS00SspS00[N] *exp(ret_fc) 25| # Loop for the next 9 step ahead forecasts 26) for (i in 27] c(hfe, omega+(alphatbeta)#h_fe[i]) 2s) ret_fe = c(ret_£c,auregrt (h_fcliti])*randoa[iti]) 29) sps00_#e = c(ap500_f¢, sp500_feLi]+exp(ret_feLiti])) 30) > 31] err = rbind(mcrr, 1og(c(min(=p500_fc) ,max(sp500_fc)) /sp500$=p500.NJ)) 32) 33} 34]merrl = 1 - exp(mean(merr{,1]) - 1.645¥sd(merr[,11)) 3u|merrs = exp(mean(nerr(,2J) + 1.645+sd(merr[,2])) - 1 Question 4. (20 points) Consider the following GARCH (1,1) model y=htu,, — u,~N(0,0?) ao + en; + Bo; a) Ifyeis a daily stock return series, what range of values are likely for the coefficients j, ao, ax and f ? b) Suppose now that the researcher had estimated the above GARCH model for a series of returns on a stock index and obtained the following parameter estimates: 4t = 0.23, ao = 0.12, B= 0.38, ax = 0.12. What is the unconditional variance of y; ? Briefly prove the formula. ©) If the data are available up to time T, produce one-, two-step ahead forecasts for conditional variance of yr. Question §. (20 points) A researcher investigates the long-term relationship among 7 macroeconomic variables (including exchange rate, oil price, commodity prices) using Johansen approach. a) Afier checking that the variables are integrated of order 1, the appropriate lag length of the VECM model is determined from thefollowing table of information criteria. What is your suggestion about the optimal lag length? Table 3. Lag selection criteria for cointegration Lag LogL LR FPE AIC sc HQ ° -1937.576 NA 211094 21.1476 21.2873 21.2082 1 -709.640 (2335.748 6.77E-07 8.4961 9.7541 9.0060* 2 ~635.807, 134.023 6.10¢-07" 8.3892" 10.765 9.3523 3 16308 68264 © SO4E-07 8.6555 12.1500 10.0718 4 546.295 82086 «9535-07 «8.8076 134203. 10.6771 5 503.422 66.641 1236-06 9.0372 14.7682 11.3600 6 444.481 86.488 1.36E-06 9.0922 15.9414 11.8683 7 -378.602 90.942" 1A2E-06 9.0718 17.0392 12.3010 8 -335.731 55.452. L9SE-06 9.3014 18.3871 12.9840 “indivates lag der selected bj the criterion, LR: sequential modified TR test sais (each test af 3% level), FPE: Final Prediction error, AIC: Akaike information criterion, SC: Schwarz information criterion, HQ: Hannan-Quinn information criterion. b) The following table gives results for the Johansen test about the number of cointegrating relationships r using A-trace and A-max tests. What is your conclusion about the number of cointegrating relationships r? Trace Maximum igen Ho: Heater Figenvatue E! 0.08 0.05 Ranker TNT ‘senate Rao Sta Critical Statstle Critheal Value Value ° 5 o2sae 15628 Dora? 159530 GaaTo 52308 1 7 02307" 783.2241 MHO00-12SG1S 49.828 46.251 2 6 01727283122 9417695754 3591340078 3 5 012g -710.3557 «58263. 9$19. 26021 33.877 4 4 0109 -697338 3224047856 20400 27.584 5 3 ois -687.1391 183029797 9262132 6 2 00128 -682.5078 2568 S495 2a 7 1 0.0007 __-681.2870 0126 38h 0.126 38H Note: =** (4) denotes rejection of hypothesis at 1% (5%) significance level, LR. test shows two conmlegrating {equations at 5% significance level, ©) From the following table of results, write down the r long-term relationship equations (the value of r is obtained from the previous question) and write down the (potential) VECM model. Colntegrating vectors 3" and adjustinent coefficient a Maize Rice Soybean Beans Cocoa Crude Oil Exchange aX Ox Bod 0 0017 0523 0.019 0.000 2339 Maize (a1) 0.049 0.032 ho 10295 0.759 -0.039 0.001 1.636 Rice (ax) 0173 0.180 Soybean (as) -0.304 0.084 Beans(a.) 0.612 0.001 Cocoa(as) 0.106 0212 Crude Oil (ae) 1.822 7159 Exchange (a7) __0.000 0.003Question 6. (20 points) a) Discuss the following graph of a series of daily stock index returns from 2001 to 2012. b) Discuss the following table of results for stationarity test and ARCH- effect test of the return series using p-values. ADF stat Pp -34.1348*** 0.0000 ARCH(1)LM Stat P 203.6634*** 0.0000 ©) Discuss the following tables of results for GARCH(1,1) model (5), GIR model (6) and EGARCH model (7). of =a, +S aul, +TB%, © = a oP =a, + Dau?, +7024, +82, © where d,_ is a dummy variable that is: _f[lifu,, <0, bad news ~ | Oifu,,20, goodnews at: 4 logla?)=a +78, lox{o?,)+ Yo, « cr isl ike GARCH(,1) GJR-GARCH EGARCH Value Pp Value Pp Value | p Mean Equation Gy 0.0009 0.0000 | 0.0006 0.0023 | 0.0005 | 0.0055 Variation Equation Gy 4.69 x10 | 0.0000 | 6.07 x10-* | 0.0000 | -0.5147 | 0.0000 a 0.1318, 0.0000 | 0.0727 0.0000 | 0.2531 0.0000 Y - - 0.1036, 0.0000 | -0.0687 | 0.0000 B 0.8496 0.0000 | 0.8441 0.0000 | 0.9636 0.0000APPENDIX Table A2.2 Critical values of Student's tdistribution for different probability levels, a and degrees of freedom, » @ 04 025 018 01 005 0.025 O01 0005 0001 0.0005 1 08249 10000 1.9626 3.0777 63138 127062 318205 62.6567 318.2087 626.6189 2 02887 08165 13862 18856 29200 42027 69616 99248 223271 31.5991 3 02767 0.7649 12498 16277 23524 3.1824 45407 58409 10.245 129240 4 02707 07407 11896 15392 2198 2764 37469 4.6041 7792.60 5 02672 07267 11558 14759 20150 25706 3.649 4.0321 5.8934 GAGES 6 02648 0.7176 112 14298 19492 24469 31427 3.7074 5.2076 .95KR 7 02632 a7IM 1192 14149 1896 29546 29989 34905 47852 5.4079 $ 02619 0.7064 11081 13968 18595 23060 24965 23554 45008 5.0813, 9 02610 0.7027 10997 13890 18x91 22622 28214 32498 4.2968 4.7809 10 02602 0998 osm 14722 18125 22281 27628 416N3 4147 SHO 11 02596 0.5974 1.0877 12634 1.7959 2201027181108 4.0287 «4.4370 12 02590 0.6955 10832 13562 17823 21788 26810 3065 3.9296 4.178, 13 02586 0.6938 10795 13502 17709 21604 2.6503 30123 3.8520 4.2203 14 0.2582 0.6924 10763 13450 17613 21448 26245 29768 3.7874 4.1405 15 02579 06912 10735 13406 75H 2134 2.6025 29467 4.7328 4.0728 1G 02576 06901 LO711 13368 17159 21199-25835 2.9208 3.6862 4.0150 17 02573 0.6802 10690 13324 17396 24008 2.5669 2.8082 «2.6458 3.9651 18 02571 0884 1.0672 13304 17341 21009 25524 28784 3.6103. 39216 19 02569 06876 1.0655 13277 1.7291 20920 25395 28609 35794 3.884 20 02567 0.6870 1o6d0 13253 1.7247 Dosed 2.5280 28459 4SSI8 3.8495 21 02566 0.6K 10627 14212 17207 20796 251% BNI 45272 -BIOD 22 02564 0.6858 LO6l4 13212 17171 20729 25082 28188 2.5050 3.7021 23 02562 0.6853 10603. 13195 17139 20587 24999 28073 3.4850 3.7676 24 0.2562 0.6848 1.0593 13178 1.7109 2.0639 24922 2.7969 3.4668 3.7454 25 02561 0.6844 Lose 13163 17081 20595 24Ns1 27874 «94502-37251, 26 0.2560 0.6840 10575 1.3150 1.7056 20583 247K6 2.7787 «1.4350. 7066 27 02559 06837 0567 1.3137 17083 QosIs 24727-27707 420.6896 28 02558 06824 10560 13125 17011 2oss4 24671 2.7633 «3.4082 3.6799 29 02557 06830 1.0553 13114 16991 20452 24620 27364 3.3962 3.6594 0 02556 0.6828 1.0347 13104 16979 20422 24572-27300 3.3852 3.6460 35 02553 0.6816 10520 1.3062 16896 20301 24377-27238 3.2400 35911 40° 02550 0.6807 10S00 12021 16819 2021 24233-27045 2.3069 -BS5IO 45 02549 0.6800 10485 12006 16704 Jo 24121-26895 «3.2815 3.5203 50 02547 0.6794 10473 12987 16759 20086 24022 26778 3.264 3.4960 60 0.2545 0.6786 1.0455 1.2958 1.6706 2.0003 23901 2.6603 3.2317 3.4602 70 02543 0.6780 10442 12038 16669 19944 23808 26479 4.2108 3.4350 No 02512 0676 10432 12022 166h 19901 22739-26187 49S3-BAIG 90 02541 0.6772 10424 12010 16620 19867 22685 2636 34a22 34019, 100 0.2540 0.6770 L048 1.2901 1.6602 19840 23642 26259 3.1737 3.3903 120 0.2539 0.5765 1.009 12886 16577 19799 23578 261M 3.1505 3.3735 150 02538 0.6761 10400 12872 1655119759 2as15 26090 3.4455 .3566 200 0.2537 0.6757 1.0391 12858 16525 19719 22451 260062815 3.8398, 300 0.2596 06752 10282 12844 16499 19679 2938825029 ITS «3.7 be 02593 06745 10264 12816 1649 19600 2326225758 3.0902 3.2903,