PPML estimates gravity models in their
multiplicative form via Poisson Pseudo Maximum Likelihood.
PPML(y, dist, x, vce_robust = TRUE, data, ...)name (type: character) of the dependent variable in the dataset
data, e.g. trade flows.
name (type: character) of the distance variable in the dataset
data containing a measure of distance between all pairs of bilateral
partners. It is logged automatically when the function is executed.
vector of names (type: character) of those bilateral variables in
the dataset data that should be taken as the independent variables
in the estimation. If an independent variable is a dummy variable,
it should be of type numeric (0/1) in the dataset. If an independent variable
is defined as a ratio, it should be logged.
Unilateral variables such as country dummies or incomes can be added.
If unilateral metric variables such as GDPs should be used as independent
variables, those variables have to be logged first and the
logged variable can be used in x.
Interactions terms can be added.
robust (type: logic) determines whether a robust
variance-covariance matrix should be used. The default is set to TRUE.
If set TRUE the estimation results equal the Stata results for
robust estimation.
name of the dataset to be used (type: character).
To estimate gravity equations, a square gravity dataset including bilateral
flows defined by the argument y, ISO-codes of type character
(called iso_o for the country of origin and iso_d for the
destination country), a distance measure defined by the argument dist
and other potential influences given as a vector in x are required.
All dummy variables should be of type numeric (0/1). Missing trade flows as
well as incomplete rows should be excluded from the dataset.
In contrast to the other functions, trade flows equal to zero do not have
to be excluded when using PPML.
When using panel data, a variable for the time may be included in the
dataset. Note that the variable for the time dimension should be of
type: factor. See the references for more information on panel data.
additional arguments to be passed to functions used by
PPML.
The function returns the summary of the estimated gravity model as an
glm-object.
PPML is an estimation method for gravity models presented
in Santos-Silva and Tenreyro (2006) (see the references for more information).
To execute the function a square gravity dataset with all pairs of
countries, ISO-codes for the country of origin and destination, a measure of
distance between the bilateral partners as well as all
information that should be considered as dependent an independent
variables is needed.
Make sure the ISO-codes are of type "character".
Missing bilateral flows as well as incomplete rows should be
excluded from the dataset. Unlike the other functions, flows equal to zero
do not have to be excluded as the gravity equation is estimated
in its multiplicative form.
PPML estimation can be used for both, cross-sectional as well as
panel data. Nonetheless, the function is designed to be consistent with the
results from the Stata function ppml written by J. M. C.Santos-Silva
and S. Tenreyro.
The function PPML was therefore tested for cross-sectional data.
For the use with panel data no tests were performed.
Therefore, it is up to the user to ensure that the functions can be applied
to panel data. For a comprehensive overview of gravity models for panel data
see Egger and Pfaffermayr (2003), Gomez-Herrera (2013) and Head, Mayer and
Ries (2010) as well as the references therein (see also the references
included in the descriptions of the different functions). Depending on the
panel dataset and the variables - specifically the type of fixed effects -
included in the model, it may easily occur that the model is not computable.
Also, note that by including bilateral fixed effects such as country-pair
effects, the coefficients of time-invariant observables such as distance
can no longer be estimated. Depending on the specific model, the code of the
respective function may has to be changed in order to exclude the distance
variable from the estimation. At the very least, the user should take special
care with respect to the meaning of the estimated coefficients and variances
as well as the decision about which effects to include in the estimation.
When using panel data, the parameter and variance estimation of the models
may have to be changed accordingly.
For more information on the estimation of gravity equations via Poisson Pseudo maximum Likelihood see
Santos-Silva, J. M. C. and Tenreyro, S. (2006) <DOI:10.1162/rest.88.4.641>
For more information on gravity models, theoretical foundations and estimation methods in general see
Anderson, J. E. (1979) <DOI:10.12691/wjssh-2-2-5>
Anderson, J. E. (2010) <DOI:10.3386/w16576>
Anderson, J. E. and van Wincoop, E. (2003) <DOI:10.3386/w8079>
Baier, S. L. and Bergstrand, J. H. (2009) <DOI:10.1016/j.jinteco.2008.10.004>
Baier, S. L. and Bergstrand, J. H. (2010) in Van Bergeijk, P. A., & Brakman, S. (Eds.) (2010) chapter 4 <DOI:10.1111/j.1467-9396.2011.01000.x>
Head, K., Mayer, T., & Ries, J. (2010) <DOI:10.1016/j.jinteco.2010.01.002>
Head, K. and Mayer, T. (2014) <DOI:10.1016/B978-0-444-54314-1.00003-3>
and the citations therein.
See Gravity Equations: Workhorse, Toolkit, and Cookbook for gravity datasets and Stata code for estimating gravity models.
For estimating gravity equations using panel data see
Egger, P., & Pfaffermayr, M. (2003) <DOI:10.1007/s001810200146>
Gomez-Herrera, E. (2013) <DOI:10.1007/s00181-012-0576-2>
and the references therein.
# NOT RUN {
data(Gravity)
Gravity$lgdp_o <- log(Gravity$gdp_o)
Gravity$lgdp_d <- log(Gravity$gdp_d)
PPML(y="flow", dist="distw", x=c("rta", "lgdp_o", "lgdp_d"),
vce_robust=TRUE, data=Gravity)
PPML(y="flow", dist="distw", x=c("rta", "iso_o", "iso_d"),
vce_robust=FALSE, data=Gravity)
# }
# NOT RUN {
# }
Run the code above in your browser using DataLab