my dissertation proposal
DESCRIPTION
Slides for my dissertation proposal, presented to my panel on February 22, 2010. http://brianbelen.blogspot.comTRANSCRIPT
Dissertation Proposal
(An Econometric Analysis)Remittances
PropertiesofPhilippine
PresentationThis
will touch upon:
The Motivation for the research.
A description of the data Treatment.
Initial departures and preliminary Results.
Issues and Directions to explore.
MotivationThe
for the research.
are particularly significant Remittances
Philippines for the
and for Filipinos .
The Filipino Diaspora and Workers’ Remittances
0
5
10
15
2000 2001 2002 2003 2004 2005 2006 2007Overseas Filipinos Workers’ Remittances(in Millions of Filipinos) (in Nominal US$ Billions)
Sources: International Labor Organisation (2009) and Bangko Sentral ng Pilipinas (2008)
The Seven Largest Remittance Recipient Countries
Germany
Spain
France
Philippines
Mexico
China
India
0 15 30 45 60
Remittances % of GDP
Source: The World Bank Migration and Remittances Data (November 2009)
(in Nominal US$ Billions) (2008 GDP)
SOME
Observations
about the
literature
(ON REMITTANCES)
Many studies have approached the topic from a development perspective.
Of the approximately 102,000 articles on remittances available on Google Scholar,not a single one was written by me!
“Stylized facts”about remittances contrast with findings from individual country cases.
Some approaches have yet to be applied to the Philippine case.
Why Not...?
Apply new approaches to Philippine Data.
Inquire into the cyclicality and other properties of remittances to the Philippines.
Approach the topic from a macroeconomic perspective.
Develop the research like an open-source project.-
TreatmentThe
applied to the data
Data for Overseas Filipino Workers’ (OFW) remittances and other macroeconomic variables were obtained for 1989Q1 to 2008Q4.
Each time series (v) was assumed to have the following components:
v = trend + cyclical + seasonal + et t t t t
Three techniques were tried to remove the seasonal component of the data.
Two business cycle filters were considered to remove the time trend and isolate the cyclical component of the data.
ln(y) (Raw Data)
Quarterly Frequency, 1989Q1-2004Q4
y
1990 1995 2000 2005
6.6
6.8
7.0
7.2
7.4
ln(r) (Raw Data)
Quarterly Frequency, 1989Q1-2004Q4
r
1990 1995 2000 2005
3.0
3.5
4.0
4.5
5.0
1989Q1-2008Q4 1989Q1-2008Q4
Percentages are Normalized to 1
Quarterly Frequency: 1989Q1 to 2008Q4
Gro
wth
Ra
te (
Pe
rce
nt)
1990 1995 2000 2005
-0.04
-0.02
0.00
0.02
0.04
by LOESS
by Moving Averages
by Dummy Variables
HP Filtered y by Seasonal Adjustment Technique
Percentages are Normalized to 1
Quarterly Frequency: 1989Q1 to 2008Q4
Gro
wth
Ra
te (
Pe
rce
nt)
1990 1995 2000 2005
-0.04
-0.02
0.00
0.02
by LOESS
by Moving Averages
by Dummy Variables
CF Filtered y by Seasonal Adjustment Technique
Percentages are Normalized to 1
Quarterly Frequency: 1989Q1 to 2008Q4
Gro
wth
Ra
te (
Pe
rce
nt)
1990 1995 2000 2005
-0.4
-0.2
0.0
0.2
0.4
0.6
by LOESS
by Moving Averages
by Dummy Variables
HP-Filtered r by Seasonal Adjustment Technique
Percentages are Normalized to 1
Quarterly Frequency: 1989Q1 to 2008Q4
Gro
wth
Ra
te (
Pe
rce
nt)
1990 1995 2000 2005
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
by LOESS
by Moving Averages
by Dummy Variables
CF Filtered r by Seasonal Adjustment Technique
ResultsPreliminary
from work on the topic thus far
In keeping with the literature, the idea was to develop regression models using variables relevant to remittance sending:
GDP Past remittance Exchange Rater = f ( , , )
First, the research explores the cyclicality of workers’ remittances to the Philippines.
Perhaps the different remittance-sending behavior described by the models is due to the type of OFW
deployed in each particular country.
Conjecture:
Vector autoregressions were employed with the end in view of understanding remittances’ macroeconomic impact.
A lag length of one was used to prevent overfitting using VARs.
Impulse-response functions were also employed to model the effect of shocks to the variables.
IRFS USING HP-FILTERED DATA
xy$xhp.ya
0.00
0.05
0.10
0.15
xy$x
hp.ra
0.00
0.05
0.10
0.15
0 1 2 3 4 5 6 7 8 9 10
Orthogonal Impulse Response from hp.ra
95 % Bootstrap CI, 1000 runs
xy$x
hp.ya
-0.02
0.00
0.02
xy$x
hp.ra
-0.02
0.00
0.02
0 1 2 3 4 5 6 7 8 9 10
Orthogonal Impulse Response from hp.ya
95 % Bootstrap CI, 1000 runs
IRFS USING CF-FILTERED DATA
xy$x
cf.ya
0.00
0.01
0.02
xy$x
cf.ra
0.00
0.01
0.02
0 1 2 3 4 5 6 7 8 9 10
Orthogonal Impulse Response from cf.ya
95 % Bootstrap CI, 1000 runs
xy$xcf.ya
-0.02
0.00
0.02
0.04
0.06
xy$x
cf.ra
-0.02
0.00
0.02
0.04
0.06
0 1 2 3 4 5 6 7 8 9 10
Orthogonal Impulse Response from cf.ra
95 % Bootstrap CI, 1000 runs
DirectionsIssues
to explore
and
Why are there different results between different filtering techniques?
Can the exchange rate effects be accounted for in the analysis?
How to handle structural breaks in the data?
How do remittances measure up against other inflows like FDI?
What other macroeconomic variables would be relevant to the analysis?
Brian L. Belen
brianbelen.blogspot.com
@brianbelen
END OF PRESENTATIONT H A N K Y O U