Gravity model example

Поділитися
Вставка
  • Опубліковано 1 лис 2024

КОМЕНТАРІ • 114

  • @TheKing-nx2bf
    @TheKing-nx2bf 3 роки тому +6

    This video is very educational! I'm currently doing a policy-evaluation estimation of a specific trade agreement. I have downloaded data from the CEPII database, which provides bilateral trade data from e.g., IMF, BACI and UN Comtrade, as well as numerous macro-variables. RTA variables are available and disaggregated into different trade agreements in Mario Larch's Regional Trade Agreements Database. I use Stata and regress both using either OLS (reghdfe command) and PPML (ppmlhdfe command) with exporter-year, importer-year and pair FE. All data is for 1980-2018, and FTA was implemented in 2010
    This means I run the two theoretically consistent gravity models:
    - OLS: Natural Logarithm of trade flows = beta0 + beta1*RTA + beta2*Specific FTA + Exporter Year FE + Importer Year FE + Pair FE + Error term
    - PPML: trade flows = beta0 + beta1*RTA + beta2*Specific FTA + Exporter Year FE + Importer Year FE + Pair FE + Error term
    For some reason, I continuously obtain insignificant estimates of RTA and FTA, regardless of the choice of estimation method, and regardless of the choice of year-intervals in my trade data... However, I do see that for some unknown reason, there's nearly no zero-trade flow observations in the CEPII data. Do you know of any other websites already having an assembled tradeflow data available? I seems I can only obtain for one country at a time using either IMF or UN Comtrade. That would be too time consuming... Any thoughts?
    thanks!

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому +2

      Hello! Thank you for the nice feedback :) IMF used to provide querries with numerous reference countries and partners, but that is not an option anymore. I believe the only way to compose the dataset you describe is to combine several excel sheets from IMF

  • @sthmdy8859
    @sthmdy8859 Рік тому +4

    GDP_difference can be more definite than GDP_distance which is ambiguous with distance. Thank you for this excellent lecture.

  • @user-oc8df6my5t
    @user-oc8df6my5t Рік тому +1

    This video is very informative and helpful. I would appreciate it if you could clarify how the GDP_Distance was calculated.

  • @sabagulnaz592
    @sabagulnaz592 3 роки тому +2

    How to deal with zero bilateral trade values as well as missing values of GDPs of countries? Please let us know.

  • @IdontKnow-et7qb
    @IdontKnow-et7qb 2 місяці тому

    please help me Ma'am .i am stuck off ... how to create pairs of countries and include dummy variables like FTAonein FTAbothin ..when we specically taking about trade agreements

  • @neznaika_znaet
    @neznaika_znaet 3 роки тому +2

    Hi, thank you for your videos!
    I have a question, as you said at 13min it might by a good Idea to decompose GDP/per capita in two variables. But isnt there a Problem
    of multicollinearity in this? Population and GDP have often linear relation.

  • @sihamelhachimi6965
    @sihamelhachimi6965 2 роки тому

    thank you very much for your presentation
    please help me on the calculation of the migratory flows by the radiation model

  • @amessinoubenito2592
    @amessinoubenito2592 Рік тому

    Thanks for sharing knowledge with us, please could you help me to build a gravity dataset for 15 countries from 1980 to 2020 I'm confused while trying to do it because trade values is export from country i and import from j to i. doing this for 40 years for 15 countries I dont know if I have to keep record of trade for each to each countpart for the entire period

  • @simisolashonibare733
    @simisolashonibare733 3 роки тому +1

    Thank you for this video! I'm trying to use the gravity model to find the impact on border effect for ONE industry/sector in one country of joining an FTA, and I'm not 100% sure how to adjust my variables/data for this. Instead of GDP, I was planning on using the total output in that sector, as well as exports/imports for just that sector. Is this fine? and are there any other variables that might need changing?
    Thanks for all your help!

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому +1

      Hi, it shall be fine (in terms of model setting), but please keep in mind that the scope of variables discussed in this video is not finite - in fact, that is just an example.

  • @ssrbgangimaribotan6thofthe12
    @ssrbgangimaribotan6thofthe12 6 місяців тому +1

    Hi, i loved the video as its very informational. However I have a question about the "in(TRADE)" variable in the beginning of the excel sheet, what would that be? I'm trying to input my data right now and I'm very confused on where its from. and What should I do If the data is incomplete (like halfway through they don't have the data for the %Partner_physical Capital in the world banks).
    Additionally if it don't bother you, I would like to know which formula and econometrics did you use for this scenario.

  • @minyarkhayat40
    @minyarkhayat40 3 роки тому +2

    Hello, thank you for the interesting video. I have a question and i hope to found an answer here. In fact, i'm using a gravity modeling to estimate the potentiel trade between 54 countries. So i need to make a data base on excel covering bilateral trade between each of the 54 counties ( i guess it will be 54*54 relationships) . In addition i'm not annalysing the overall bilateral trade between each of these countries. I focus on 20 specific products. I'm confusing how can i make a table on excel for belateral trade between 54 countries. The Destination is not only from 1 country like the exemple of poland. i have to make every one olone and for every single product. It seems to be complicated and too long. I don't know even it might be applicated (54*54*20)

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  2 роки тому

      Hi! Well, that's an interesting question. I can think only about two possible practical solutions. First, you can design different pairs of countries for different goods (such as Poland_France_agriculture, Poland_France_automotive etc.). Second, maybe, you can aggregate the values for different sectors, which, of course, will narrow the number of observations. However, I am not sure about the implications of those for your econometric analysis, as there are too many unknowns.

  • @sabrinakalinkova6435
    @sabrinakalinkova6435 5 місяців тому

    Hi, the video is very informative and helpful. It would be really helpful and great if you could share the excel panel, please.

  • @IdontKnow-et7qb
    @IdontKnow-et7qb 2 місяці тому

    please cloud you provide the excel sheet of data set for practice ... and i didnt get the points how to create pairs

  • @shivanshraman3179
    @shivanshraman3179 3 роки тому

    How did you get physical capital to GDP data? It wasn't available on the world bank data site

  • @charonaynguon1584
    @charonaynguon1584 5 місяців тому

    can you please send us the Gravity model balance panel as you showed on the video ?

  • @FaisalKhan-mn8hx
    @FaisalKhan-mn8hx 2 роки тому

    Thankyou, I m trying to put the gravity model in the trade of India with GCC can u help me in selecting the variables and how to modified the gravity model for my research

  • @erhumuagbude8806
    @erhumuagbude8806 Рік тому

    I am working on Changes in Vietnam’s Trade to the US.
    Using Price, GDP, Tariff or Trade policy as my factors that determine Changes in export supply in Vietnam. How do I get the data for this and relates it through gravity models

  • @etbedtalksAOH
    @etbedtalksAOH 10 місяців тому +1

    is it feasible to focus on land alone skipping capital and labour if analysis is on agriculture products?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  10 місяців тому +1

      Yes, you should always adjust the set of regressors to the task at hand

    • @etbedtalksAOH
      @etbedtalksAOH 10 місяців тому

      @@lazarskiopencourses4059 ty

  • @bradbullough1371
    @bradbullough1371 8 місяців тому

    Hi where did you get the specification for your remoteness variable?

  • @ilginyld
    @ilginyld 6 місяців тому +1

    Hi, thank you for this video it really helped a lot🙏🏻

  • @jeamilethhenriquez2685
    @jeamilethhenriquez2685 3 роки тому

    Hi! Thank you for this helpful video! I'm trying to use this gravity model for a new airport, but I'm not sure what variables should I use to estimate the number of trips generated by this new project. I have origin - destination matrix and distances but I guess that is not enough. Could you help me please? Do I need more variables?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому +1

      Hi! I think you might think about the overall quality of infrastructure, maybe traditional pull and push factors from migration models (if you analyze international connections). But I strongly encourage you to look though the relevant literature - I am afraid, I might not be qualified enough :)

  • @MustafaKamleh
    @MustafaKamleh 2 роки тому

    Hi, thank you for this important video. I have a question, can this model be applied to measure Trade-Based Money laundering between 10 countries? Is there a possibility to discuss it via email or a call? Thank you.

  • @geraldlimathasian2286
    @geraldlimathasian2286 3 роки тому +1

    You can show me that how to get Trades variable. Because you said it was different with trade balance. Thank You

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому

      Hi, the video starts with downloading data on imports and exports from IMF database. Then you just have to add those values and obtain the valu of bilateral trade between two counties.

  • @NkopeloMakintane
    @NkopeloMakintane 7 місяців тому +1

    Where do you get data for geographical distance...I'm sorry I did not get it

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  7 місяців тому

      Hi, you can just get it from google maps, or the database here: www.cepii.fr/CEPII/en/bdd_modele/bdd_modele_item.asp?id=6

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  7 місяців тому

      If you want it just for Europe, check this video: ua-cam.com/video/hOEwc1anN-E/v-deo.html&ab_channel=LazarskiOpenCourses

  • @amessinoubenito2592
    @amessinoubenito2592 Рік тому

    Hello this video is very educational

  • @pedrolomomenendez2587
    @pedrolomomenendez2587 3 роки тому

    Thanks for the video! I need data for bilateral trade for a lot of countries (not only one as Poland), how can I organize the data in order to estimate a panel data model? I have bilateral trade data for 140 countries among them, and I have no idea how to organize it in the form of panel data, since it is like many different panels. The same for development aid data. Thank you so much!

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому +1

      Hi, unfortunately, it is not possible to specify several reference counties in IMF DOTS anymore. So, the easiest way is to download 140 datasets for 140 reference counties and then use R for binding them in a single dataset and selecting the unique pairs of counties. Oh, and please keep in mind that those values might depend on the reporting counties (for Turkey-Germany trade, for instance, they will depend on whether you took Turkey or Germany as a reference country).

  • @IdontKnow-et7qb
    @IdontKnow-et7qb 2 місяці тому

    how to design data set please explain it in detail

  • @sthmdy8859
    @sthmdy8859 Рік тому

    HI teacher! I do the same procedure for the variable, REMOT but I got the figures minus although all are positive figure. Please advise how can I handle this?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  Рік тому +2

      Hi, if I understood correctly, you got a negative coefficient next to the Remoteness variable. It makes a perfect sense, since this is just a modified distance variable

  • @tramdo917
    @tramdo917 3 роки тому +1

    the distance in formula REMOT in your file how to calculate, i did not see the formula in excel file, can you explain?

    • @vitaly_p
      @vitaly_p 3 роки тому +1

      Distance*GDP(Partner)/GDP Global

    • @tramothithanh1864
      @tramothithanh1864 3 роки тому

      @@vitaly_p thank you, i read the other video to explain, distance is geographical km, REMOT not mentioned but thank you, i got it ReMot on file, stupid question mentioned distance not REMOT...distance is not variable x factor it is constant. Thank you!

  • @economicst.v5192
    @economicst.v5192 3 роки тому +1

    Thank you. It's really helpful. I am going to test gravity model for Pakistan with 10 trading parteners. For this purpose I have selected GDP of Pakistan, Gdp of trade partner, distance, Tariff, Real exchange rate and Trade agreement. I m confused in making excel file for estimation. Will u please help me in formatting excel file of these variabled

    • @dabiresgustave4526
      @dabiresgustave4526 9 місяців тому

      Salut, comment vous allez? Sur quel site on peut avoir les données sur le taux de change réel?

  • @Orange1117
    @Orange1117 3 роки тому

    Thank you very much, this video has been a huge help. I wanted to ask, how can we incorporate time invariant dummies in a fixed effect model? Also, could you provide a source(s) for the _remoteness_ variable in 9:46? Thank you so much!

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому +2

      Hi, technically, there should be no time-invariant dummies - I am not sure, but probably Gretl allowed for including them because of a bug or smth. Please check this paper: Wei S. (1996), Intra-National Versus International Trade: How Stubborn Are Nations in Global Integration, NBER Working Paper, No. 5531.

  • @tokofora8783
    @tokofora8783 3 роки тому

    I did my gravity model to analyze bilateraltrade. I modified it with some variables using gretl as you did. But i'm confuse on how to interpret the OLS result. Can you help me?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому

      Hi, please try this video and the comments section:
      ua-cam.com/video/hOEwc1anN-E/v-deo.html&ab_channel=LUDepartmentofEconometrics

  • @waqasnaseem556
    @waqasnaseem556 3 роки тому

    Thank you very much,this video help me alot to do my research but as you mention i will upload this research and please let me know how i will find yours research?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому +1

      Hi, link to the research you can find in the description of this video: ua-cam.com/video/hOEwc1anN-E/v-deo.html&ab_channel=LUDepartmentofEconometrics

    • @waqasnaseem556
      @waqasnaseem556 3 роки тому +1

      @@lazarskiopencourses4059 well thank you for response i checked there but i need this research paper the poland one.thank you

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому +1

      @@waqasnaseem556 hello, unfortunately, there is no paper, just an example I used for my students

  • @tridishabayan9243
    @tridishabayan9243 2 роки тому +1

    Can I use the same techniques in case of trade between GCC and India also?

  • @bakhtiyorjonbokhodirov6827
    @bakhtiyorjonbokhodirov6827 Рік тому +1

    hello, where can I find data for physical capital?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  Рік тому

      Hi, from PennWorld Table, the link is here: www.rug.nl/ggdc/productivity/pwt/?lang=en

  • @Hahdjxkdmx
    @Hahdjxkdmx 2 роки тому

    Cool video :) working on something similar. What would you do if you had a lot of missing values for Exports and Imports for certain countries?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  2 роки тому +1

      Hi, yeah, there are many missing values for the less developed countries. Then you can switch the reporting country (like data on Iraqi imports from Germany = data on German exports to Iraq). Definitely adds to the workload, but might be helpful

  • @howhow8604
    @howhow8604 3 роки тому +1

    Can you show how to take the data such as GDP_pol, GDP partner, Pol_GDP_per_capita Partner_GDP_per_capita, %Pol_physical capital %Partner_physical capital, Pol_land Partner_land Pol_labor Parner_labor Pol_cap_ratio Partner_cap_ratio Pol_land_ratio Partner_land_ratio.
    I don't know how to collect them

  • @vandananautiyal7848
    @vandananautiyal7848 3 роки тому

    Hello can you please tell us the name of the research paper which you used for this particular analysis

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому +1

      Hi, when it comes to the relative endowment (i.e., the difference between capital-to-worker and land-to-worker rations) and GDP distance, please check Baxter and Kouparitsas (2005) [Baxter, M., and Kouparitsas, M.A. (2005). What determines bilateral trade flows? Working Paper Series WP-05-11, Federal Reserve Bank of Chicago]. For the adjusted measure of geographical distance, please check Wei (1996) [Wei S. (1996), Intra-National Versus International Trade: How Stubborn Are Nations in Global Integration, NBER Working Paper, No. 5531] and/or Trotignon (2010) [Trotignon J. (2010). Does Regional Integration Promote the Multilateralization of Trade Flows?: a Gravity Model Using Panel Data. Journal of Economic Integration, 25(2), pp. 223-251]. And when it comes to SCALE, I cannot find the original source, unfortunately.

    • @vandananautiyal7848
      @vandananautiyal7848 2 роки тому

      @@lazarskiopencourses4059 thank you very much

  • @theeducation2510
    @theeducation2510 3 роки тому

    You have awesome style ..I want to work on same topic relevant with OPEC ..
    Plz tell me How I can access this site freely
    "Direction of Trade, IMF"..
    I am trying but this site is not available freely

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому

      Thank you so much! Sure, here is the link data.imf.org/?sk=9D6028D4-F14A-464C-A2F2-59B2CD424B85&sId=1390030341854

    • @theeducation2510
      @theeducation2510 3 роки тому

      @@lazarskiopencourses4059 still not opening , any proxy ?

    • @theeducation2510
      @theeducation2510 3 роки тому

      How we find out GDP World ,this is not available in world bank site ,plz help

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому

      @@theeducation2510 hi, you can download the excel sheet from the World Bank - world GDP will appear at the end of the list

    • @theeducation2510
      @theeducation2510 3 роки тому

      Thank you so much I got it

  • @lusinemzikyan2883
    @lusinemzikyan2883 2 роки тому +1

    HELLO, PLEASE TELL ME HOW CAN I DOWNLOAD WORLD GDP? THANK YOU

  • @theeducation2510
    @theeducation2510 3 роки тому

    Hello mam you can provide references of research articles from which you have designed all these variables ,i.e REMOT , SCALE, GDP dist , Labor-endow , cap-endow ..

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому

      Hi, when it comes to the relative endowment (i.e., the difference between capital-to-worker and land-to-worker rations) and GDP distance, please check Baxter and Kouparitsas (2005) [Baxter, M., and Kouparitsas, M.A. (2005). What determines bilateral trade flows? Working Paper Series WP-05-11, Federal Reserve Bank of Chicago]. For the adjusted measure of geographical distance, please check Wei (1996) [Wei S. (1996), Intra-National Versus International Trade: How Stubborn Are Nations in Global Integration, NBER Working Paper, No. 5531] and/or Trotignon (2010) [Trotignon J. (2010). Does Regional Integration Promote the Multilateralization of Trade Flows?: a Gravity Model Using Panel Data. Journal of Economic Integration, 25(2), pp. 223-251]. And when it comes to SCALE, I cannot find the original source, unfortunately.

  • @МахмудЯллавач
    @МахмудЯллавач 3 роки тому

    Can you please 🙏 tell what we should do if there is heteroskedestity in the model?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому

      The easiest way to overcome this issue is to use heteroscedasticity consistent covariance matrix. Most popular choices are White and Newey-West (this one also accounts for autocorrelation)

    • @МахмудЯллавач
      @МахмудЯллавач 3 роки тому

      Can you recommend any practical examples so I learn how to do that ?

  • @mahabubrahman1229
    @mahabubrahman1229 2 роки тому

    how you calculate Poland capital and land ratio and partner country's land and capital ratio?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  2 роки тому

      Hi, you can get the data from the World Bank. Land is measured in acres (arable land), and capital is calculated as the product of gross capital formation (% of GDP) and GDP

  • @arielcheang1627
    @arielcheang1627 3 роки тому

    Could you please why do you multiply two countries' populations? What's the that mean?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому +1

      Hi, traditionally, economic weight of the countries is assessed based on the product of GDP. However, GDP might be biased because it incorporates i) information about the size of the economy; ii) information about income/well-being, etc. Therefore, by using the product of populations, you can avoid this effect (especially if you include GDP per capita variables as well). Both methods are pretty common, it's up to you. But usually my results look nicer when using product of populations as the indicator of economic weight, especially when using simple computational techniques

  • @dinobrown5956
    @dinobrown5956 3 роки тому

    Thank you for this wonderful tutorial. Is it appropriate to select more countries as partner countries. eg. Instead of Poland alone you add two more countries as partner countries?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому

      Hi! Sure, it will make no difference.

    • @dinobrown5956
      @dinobrown5956 3 роки тому

      I realized the IMF- DOTs platform does not permit selecting more than one country as counterpart countries? Kindly comment on this

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому

      @@dinobrown5956 unfortunately, that's true - adding more countries would require merging several tables in Excel or R

  • @lifeofadepressedgirl.9134
    @lifeofadepressedgirl.9134 2 роки тому

    excuse me but can you explain more clearly about the remot? thank you very much, I really need to understand it urgently

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  2 роки тому

      Hi, please check these papers: Wei (1996) [Wei S. (1996), Intra-National Versus International Trade: How Stubborn Are Nations in Global Integration, NBER Working Paper, No. 5531] and/or Trotignon (2010) [Trotignon J. (2010). Does Regional Integration Promote the Multilateralization of Trade Flows?: a Gravity Model Using Panel Data. Journal of Economic Integration, 25(2), pp. 223-251]. The underlying idea is that the effect of geographical distance varies depending on the size/wealth. Also, I suspect that it's a smart strategy to incorporate distance to the fixed effects estimation (i.e., to transform the time-invariant variable to the time-variant variable).

  • @aoghaly
    @aoghaly 3 роки тому

    Could you please share the excel sheet and if the full course is available for download?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому +2

      Hi, the course is not available for download, but the entire playlist is on UA-cam to stay (ua-cam.com/video/f0ZVKY360aI/v-deo.html&ab_channel=LUDepartmentofEconometrics). The data file is available here: www.researchgate.net/publication/344452737_Replication_data_for_Beck_K_2020_What_drives_international_trade_Robust_analysis_for_the_European_Union_Journal_of_International_Studies_133_68-84_doi_10142542071-8330202013-35

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому +4

      Hello, you can download the file from my Google Drive: drive.google.com/file/d/1oPMfswfGbWF3fUd11vslQa9mqD7hciWv/view?usp=sharing

    • @aoghaly
      @aoghaly 3 роки тому +1

      Thank you so much for your help

    • @alammir2666
      @alammir2666 3 роки тому

      Your video is quite helpful ..I also completed my mphil thesis in trade using gravity model

    • @sonyvebiyanto8667
      @sonyvebiyanto8667 2 роки тому

      @@lazarskiopencourses4059 thx so much

  • @sharpaquos2806
    @sharpaquos2806 3 роки тому

    why you put 6 cross-sectional units instead of 9 and number of time periods are 22 instead of 21 (1995-2016)?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  3 роки тому +1

      Hi! It covers all years from 1995 till 2016 (including these years). So, there 22 years in total

    • @sharpaquos2806
      @sharpaquos2806 3 роки тому

      @@lazarskiopencourses4059 : the ols output indicates that your capital endowment and remoteness have negative coefficient. Does it mean for 1 percent remoteness increases with the trade partner, the volume of trade will decrase 63 cents ? with decrase being 25 cents for dissimilar in factor endowments is that correct? lastly, for land endowment increase by 0.007 cents if the bilateral trade (or volume of trade) predominated among country that is dissimilar of it?

  • @caraye379
    @caraye379 2 роки тому

    Asc, Madama I need most export somalia

  • @icarus6087
    @icarus6087 2 роки тому

    Is it possible to share this file?

  • @fabricenzepang3026
    @fabricenzepang3026 3 роки тому

    please share the programme you use

  • @mdmamunurrashid2076
    @mdmamunurrashid2076 2 роки тому

    why my landendownment is negetive?

    • @lazarskiopencourses4059
      @lazarskiopencourses4059  2 роки тому

      Hi, the designed variable reflects the degree of similarity of the land endowment. So, you can interpret it in terms of the HO theory.