ARDL Model. Model Two. EVIEWS

hello hello now today we shall be developing developing a are BL a our deal right ar AR deal model and here I have three variables they are GDP then export and consumption we have these three variables right and here some are some are some are I won and some some maybe I zero right but none is I to none is i2 so I can develop a our DL model and we can also call it we can also call it as bound test okay so our model actually looks them sorry our model looks like this actually it looks it looks like this all mouse re is coming up so my things I make it big in C better okay our model actually looks like this sorry okay it looks like GDP is the dependent variable that have taken GDP is the dependent variable and we have ay intercept B 1 B 2 and we have the variable Expo what is Expo export plus B 3 then our another variable is consumption right consumption is our another variable so far so this is our model right from this model I can develop the AR DL model okay so so first I have to select optimum number of blacks for this model okay so the optimum number of blacks should be I'm sorry should be okay so first I begin with del 6 lakhs right 6 lakhs then I check then I check AIC and yes I see cry to yellow okay so having 6 lakhs okay so uh so my model having 6 lakhs will be like this okay will be like this I go to quick estimate equation and so 6 lakhs right so so I put it D GDP meaning that so we take it first difference of GDP C is the constant so I put it properly now GDP black one right so make sure make sure all things are written properly then lakh to GDP lack - okay then be DDP lagg-3 then be Judy be lack for be GDP black 5 then D GDP like six okay then we start with the export be Expo that is export right export lag 1 then D Expo lag - then D Expo sorry the expo lag 3d Expo black for then D Expo black 5 then D Expo lack 6 ok so it is done then my third variable is consumption so I start with the cons what is called cons cons mais consumption okay D cons lack 1 then D cons lack - then D or should I make it all big beak on slack 3 then D cons I'm sorry lack for then P cons lack 5 then be calm relax 6 right okay so we have put all lack battle but we have to take other three variables such as such as GDP lag one right then export that is our second variable lag one and the finally comes right lakh one so here you can see our all variables are here you can see from here Khan's export and jdb right we have these three variables are here right already here and there are 350 observation in total okay here what is this what is this these three these three actually long run model so this one I have taken here whether these three variables have long run Association sheep or not so that is why here I have taken this three variable as lakh one and here you can see there is no first difference no fast difference but other variables got first difference data okay I take one by one slowly dddp dddp lag 1 dddp owes you must keep to bracket here black 3 lakh 4 lakh 5 lakh 6 then export lag one lakh to lack must have to bracket Expo should have to bracket also and he should have to bracket not one to bracket the cons because 2 because 3 because 4 to bracket and here is also to bracket B cons also to bracket ok double check again double check lag 1 lakh to lack 3 lakh 4 lakh 5 lakh 6 export lag 1 lakh to lack 3 lakh 4 lakh 5 lakh 6b cons lag 1 lakh to lack 3 lakh for lack 5 black 6 or here I've done a mistake should have a bracket right should have a fast first bracket I think second okay there is a bracket here then finally I have three variables look I think the model is set properly having 6 lakhs okay what I do I better make a copy right I keep a copy I keep a copy here right okay okay now I can run this model and here GDP is the dependent variables right and all our independent variable and and this one we can call it as a standard AR DL model right standard AR DL model having three independent variables having six lakhs each variable okay I make a copy the whole thing's right I keep a copy I keep a copy like this okay then I run the model so I press okay so the model has been estimated right this is my model GDP is the dependent variable and all our independent variable right we can see all our independent variable okay what ah what I do ah so what I do the thing I can keep it as a savings so our model actually it looks like this so here I passed it because this is my model right so I kick keep it I keep it this my model right having 6 lat okay so from here the model having 6 lakh so what is the value of blacks election are akaike criteria and what about short criteria having 6 lakh okay so III I can take a note about it I take it not Galax 6 right when does it lack 6 how much is a I see the AIC is the value is how much the IC value is minus five point nine it right from here okay so - five point nine it right and what about the second one si si criteria how much is the value so our model is we're adding she's here she's here okay si si criteria the value is minus five point seven three so I write it minus five point seven three okay so that is Lac six okay the second step I can develop the same AR DL model having four legs okay so okay uh I can show you all my data just you show all the data here I can show you just I'm showing all the data okay I can show you all the data if you want open as a group this is my data right all are here right 350 data okay okay now now I shall develop the ARD l model have been having four legs right okay I go to quick and estimate equation and here my models are six lakhs so here I make four lakhs so I delete this two from here this two from here and this 2 from here also and I delete this one from here okay this one is here okay now as before here GDP is that dependent variable and all are in depth and variable having four lakhs right all our remain unchanged right all our remain unchanged all the remain unchanged here right all our remain unchanged right already main unchanged here okay now I I run this model having four legs right okay so I proceed I think it is alright i double-checked flag go on like to like four leg one leg two left for lag on and for press okay so it is my second a RDL model you can see D GDP is the dependent variable having four lakhs right okay here you can see the akaike information criterion 6.01 short information criteria - five point eight point okay so what i can do i I can keep a record I can keep your record of this these two right so one is five point six point zero one and five point eight four I keep a record here okay lakh okay now I take lakh oh sorry sorry sorry lacked for so what is the value of K I see the value of AIC has in - six point zero one and value of s IC has been - should - five point eight four right - 6.01 si si five point eight four and make sure lower the value of AIC and is si si better the model all the time okay now I take model having two lakhs right then we check about the value of AI Si and si si okay as before I click on the quick estimate equation okay this is my model or here I take only lakh to write each independent variable so I am deleting and making it - ha ha okay I have it okay I did it more idle it more export one expert - I delete all these things by deleting all these things then come up T cons the cons to the country they conspire I did it all except I keep all the three variables right which may have which three which three may have long run association ship okay so this is my model having two lakhs cutting it is fine set I can click OK to run the model okay so this is my model the new model having lack 2 and D GDP is the dependent variable as before and you can see Lac to write all variables have blacked to and here as before how much is the is the value of akaike the value of a KY KZ becomes six point zero two and value of schwate become five point nine one okay so what I do I can keep a record of it k lack when there is a lack - what is the value of AI see it becomes a - six point zero two and si si become - five point nine one okay okay now you can see from here lack six lakh for and latter okay out of this three model which model has lower AIC and si si criteria so after checking all these three models here lack two has the lowest value of AIC and the lowest value of Si Si right you see it is the lowest and it is the lowest also so model having lack two is the best model right out of all the three model okay now we can concentrate on lack to model to proceed further okay so our model is here already lack to model lack to model they already here right this is my lap to model you can say I have the lat two models again okay now I have to check whether this lack to model has serial correlation or not and I have to check whether this model is stable or not right okay how to proceed okay okay first I check whether this AR DL model has serial correlation or not how to check I go to view then residual diagnostic then serial correlation LM test and here I keep black to press okay and you can see the observed R square value so here the probability value is six point one zero percent meaning that we cannot reject null hypothesis rather we accept null hypothesis meaning that this model has no serial correlation so we are happy about this model okay then we check whether this model is stable or not as before I click on the view then stability diagnostic then I click on the recursive estimate okay then I click on the cushion test right cushion test I press ok and here you can see this line this line is within this two red line this line is within this two red line meaning that our model is stable right so our model does not have any serial correlation and our model is stable so we are happy about this model okay now now we can go for bound testing right and and we can check whether our variables have long run association ship or not okay here here I told you I told you so our model is looks like this I'm sorry I the model okay our model is like this right okay now now we can check whether GDP export and consumption all these three variables have long run Association shape or not we can check it now and okay now the question is that how to check it how to test it and we call it bound testing okay okay here you can see this one here is the model having lack two and the hand and this model is the best because AIC and si si value is the smallest okay in this model this one is c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 so it is c8 c9 and sidon okay now now we can apply the wall statistics right we can apply what wall statistics to check it yeah we shall apply all set okay it is c c8 c9 sitter okay as before I go to view and coefficient diagnostic and I go to wall test right okay wall test so whether C 8 equals c 9 equals C 10 oh sorry Satan whether they are equal to zero or not so actually this one is our null hypothesis whether whether GDP lag 1 export lag 1 and consumption lag 1 whether they are 0 or not jointly okay so so I press ok so I press ok now so I press ok now to check it ok so press ok already okay then after that okay so this is the world test result F statistics right ok how much is the F statistics and this is the probability right but here this is the here we shall take the critical value from the bound testing or okay here I am writing how much is the F statistics twenty three point five five right that is my F statistics so here our F status of sorry here F statistics is how much abdullah mr. haire society is at the value of f statistics has come up just one minute oh sorry why this coming here okay how do studies takes value just I take it up okay I take it up a little bit so that we can again okay our F statistic value F statistic value has been I become twenty three point five five and this value should be compared should be compared with should be compared with pisarra press pasaron critical value right value at 5% level right s around critical values at 5% level and here our model is what our model is unrestricted unrestricted intercept our model is unrestricted intercept and our model has no trend right so now we can check the critical value when the model has unrestricted intercept and no trained so that is our model that we have developed now we can take that critical value at 5% level okay so from the purpose our our table right from the special table our lower bound value lower bound value is a three point seven nine and upper bound value upper bound value has been 4.85 right so so we have this two value from the pasaron table okay then what is the guideline the guideline is sorry the guideline is when de yep statistics right is a more than upper bound value upper bound up our upper bound value right we can reject the null hypothesis right now hypothesis right that is the guideline so here our f statistics is how much you can see twenty three point five five right that is our F statistics twenty three point five five and this one is more than upper bound value four point eight five right so so we can reject the null hypothesis and can accept the alternative hypothesis okay what is our null hypothesis that we have said our estimate our null half that we said already here you can see our null hypothesis is c8 c9 and C then equal to zero that we have just received so we can reject we can reject this null hypothesis meaning that c8 c9 and C are not 0 jointly it means that the three variables such as three variables are which are three variables such a I am showing again three variables such as GDP export and consumption have long run association ship right that means all the three variables move together in the long run okay so so there so all the three variables can move together in the long run so now wow we can develop the model further using short run and long run issue we can set it properly okay and okay now now now first we check what is our long run model that I showed you already last time it looks like this that is our original model that I showed you this original model and here all the three variables have long run association shape right okay now from this model I shall be taking the residual of this long run model right okay then okay so first I did I I develop this model right this develop model I develop it how as before how to develop it I go to quick estimate equation so GDP is the dependent variable then C then expo then consumption so and so this model we can develop right so so so I proceed okay okay so so this one is my long-run model okay now here GDP is the dependent variable export and consumption our independent variable now from this model I shall derive the residual of this model right how you can see the residual actually already here it has been already estimated right residual of the model so I take a copy to copy and I can paste it again right past it and the I give it as a name ect error correction term right press okay you can see on variable has been created so this one is the erode correction term actually this variable and this variable is the same this variable and this variable is the same variable right okay now I can develop our model further having Lac - okay how to proceed I proceed as before okay as before so this is my model right no not this one okay so so our best model having lack two that we know already okay I run it again estimated equation D GDP then I put C B GDP having lag one D GDP having lack to then D export having lag 1b Expo having black - sorry then D consumption right having like 1 then D comes having like two okay now so the this one we have already done okay now I am adding one more variable that is called ECT lack one right just I have added this variable one period lack of the residual from the long run model okay so now the this is my model so I am double checking because lack to is the best model that is why I have taken it like this I think it is it is it is fine I take a coffee copy if you need it okay I think we have done it properly there is no problem so far so far I see so I can proceed if you allow me press ok okay this is my model you can see from here okay so this my dependent variable right variables on this model has lacked - okay okay this model you can see is these are all what short-run coefficient right right these are right short trunk option and this one is the speed of adjustment ECT is the speed of adjustment here it is how much sixty four point nine four percent the speed of adjustment toward long-run equilibrium is sixty four point nine four percent and here what is the guideline it should be negative and it should be significant it is negative and it is significant meaning that the whole system whole system can get back to equilibrium long-run equilibrium at the speed of sixty four point nine four percent okay okay now I check whether this model has serial correlation or not whether this model has stability or not I go to view as before then residual diagnostic then serial correlation LM test and here I take lack two as before I can see F statistics right so we can reject the null hypothesis meaning that this model has serial correlation that is not desirable this model has serial correlation okay now we check whether this model has stability or not using Kusum test as before I go to view their stability diagnostic then records ship estimate as therefore then cushion test as before this lime is within these two red line meaning that this model has is stable so we are happy about this model but the problem is that this model has serial correlation we cannot accept this model okay so what I should do now okay so what do you do I can I can delete one variable or I can drop one variable to so that the so that we can get a better result okay what I do from our model I delete this variable from this model right I delete I delete D GDP lag 1 from this model and other remain unchanged other remain unchanged so now this my model then press ok ok this is my new model you can see D GDP lag 1 is not here okay in this model I check whether there is serial correlation or not I go to view ratio diagnostic serial correlation LM test I take 2 you can see we cannot reject null hypothesis meaning that there is no serial correlation we are happy about this model okay then we take whether this model is stable or not I click on the view then stability diagnostic then record sieve estimates cushion test press ok you can say this line is within this to standard deviation right sorry you within this to red line meaning that our model is stable so our model this model is stable we are happy about this model right because no serial correlation and the model is stable okay now from here we can check speed of adjustment and also also the short run causality from here we can check okay so first I talk about short run causality from export right okay it is C 1 C 2 C 3 C 4 right C 3 C 4 right c3 c4 and c5 c6 c7 okay so first I check c3 and c4 right right so what I do I click on the view then coefficient diagnostic work test c3 required to c4 we're doing wall test it is our null hypothesis C 3 C 4 is 0 or not I click on the ok wall test it is our null hypothesis they are 0 we cannot reject null hypothesis meaning that we accept null hypothesis meaning that they are zero meaning that export meaning that export export lag 1 export lag to jointly cannot cause GDP so we call it short run causality meaning that there is no short run causality running from export to GDP ok now we take one more whether this to consumption lag 1 and lakh to c1 c2 c3 c4 c5 c6 right c5 and c6 I do it as before diagnostic wall test c5 equal to 6x equal to 0 right this man all hypothesis press okay this is the null hypothesis we cannot reject null hypothesis we accept null hypothesis meaning that they are 0 that means that means there is no there is no short-run causality running from consumption lag 1 and consumption lack two to our dependent variable which is GDP right ok so here in our model there is no short-run causality from da from the independent variable to dependent variable now now we can check the speed of adjustment that is ECT and the coefficient is seventy five point six two percent it is negative and also significant because probability is less than five percent as a result we can say speed of adjustment towards long run equilibrium or we can say other way this system this whole system is getting an adjusted at the speed of at the speed of seventy five point six two percent towards long run equilibrium okay I can write this one if you if you ask me again I write it this one this way I can write it the system is getting adjusted towards long run equilibrium equilibrium at the speed of at the speed of seventy five point six two percent so so that is the outcome right of this a RDL model having lack two so model having lack two is the best model because it has no serial correlation the model is stable and also the model is getting getting to long-run equilibrium at the speed of at the speed of seventy five point six two percent only problem anot problem in this model there is no Truong causality running from independent variable to dependent variable I think we are done thank you very much for being with me for a while

Channel: Sayed Hossain

Views: 115,738

Rating: 4.7238493 out of 5

Keywords: Hossain, Academy, Data, Analysis, Econometrics, Statistics, EVIEWS, STATA, Sayed Hossain, Time series, ARDL, bound test, pesaran, Autoregressive distributed lag model

Id: mTa7n1KY9iQ

Length: 50min 15sec (3015 seconds)

Published: Tue Dec 30 2014