Beruflich Dokumente
Kultur Dokumente
Submitted to:
Chen Zhihong
Professor, SITE, UIBE
Submitted by:
1
The said fiscal stimulus was announced in February, 2011 and disbursement was made in April 2011.
2
Foreigners, constituting 40% of Singapore population, did not qualify for the growth dividend program.
3
The researchers use logistic regression to match observations and result is not included in the paper.
First, the average monthly response to the stimulus is studied using the following specification:
Yi,t= βpre X $Di X 1pre + βpost x $Di X 1post + αi + δt +ε i,i … … … (1)
where Yit is the dollar amount of total card spending, $Di is the amount of growth dividend in the
treatment group (=0 for the control group), 1pre is a binary variable equal to 1 for the months before the
announcement, 1post is a binary variable equal to 1 for the months after the announcement, δi is the year-
month dummy (to absorb seasonal variation in consumption), αi is the individual dummy (to absorb
differences in consumption preferences).
To compare the policy effect in of the post policy window into the announcement period and
disbursement period, the paper then uses the following specification-
Yi,t= βpre X $Di X 1pre + βa X $D¡ X 1announce +βd X $D¡ X 1disburse + αi + δt + εi,t. … … … (2)
where 1announce is a binary variable equal to 1 for the two months during the announcement window
and 1disburse is a binary variable equal to 1 for the months after disbursement. To study the dynamics of
the spending or debt change response the paper estimates the distributed lag model which is
specified as-
Yi,t =∑9𝑠=−4 𝛽 s X $Di X 1month s + αi + δt + εi,t … … …. (3)
Finally, the paper studies heterogeneity in the the response of growth dividend across different groups
of individuals using the following specification:
Yi,t =∑9𝑠=−4 𝛽 s X $Di X 1month +∑9𝑠=−4 𝛽g1,sX 1g1X$Di x1month s + …
+ ∑9𝑠=−4 𝛽g(N-1),s X 1g(N-1)X$Di X1month s+ αi + δt + εi,t … … … (4)
where N is the number of subgroups of consumers.
6. The summary of results:
Most of the regression coefficients are found to be significant at 5% and 1% level of significance.
The estimated results show that consumption rose significantly after the fiscal announcement: for $1
received, consumers on average spent $0.80 during the ten months after the announcement.
Consumer’s credit card debt moderately decreased during this period. The regression results also
show a strong announcement effect (19% spending increase before the cash payout). Prior to the
announcement, there is no difference in consumption trend between the treatment and control groups.
The researchers find that the consumers started spending via credit cards during the announcement
periods, and then switched to debit cards after disbursement. Consumption response is found to be
heterogeneous across spending categories and across individuals. The paper also performs tests for
robustness of its results and finds that robustness tests are qualitatively and quantitatively similar to
those in the main analysis.
7. The contribution and limitation of the paper and the possible extension:
As claimed by the researchers, the paper first documented experimentally the announcement effect
on consumption trend of stimulus program which is consistent with life-cycle theory predictions on
consumption response to unanticipated income shocks. The paper has been successful to decompose
the consumption response into different spending instruments which has likely to be underestimated
in prior work based on micro data from one single payment instrument. Moreover, the paper
documents dynamics of consumption response across different spending instruments.
It seems, however, that the paper has some limitations. If a cash dividend program is associated with
simultaneous increase in direct taxes (e.g. income tax) and indirect taxes (e.g. sales tax), the
responsiveness could be different than the estimated. The paper does not explain any such measures
taken by Singapore Government during the study period. Nevertheless, the paper has tremendous
implications on future research. Based on the methodology of this paper, the impact of fiscal stimulus
can be analyzed on the behavior of investors (assumed to be different than a rational consumer) due
to stock market collapse.
8. The publication information such as the journal, issue and year:
Researchers: Sumit Agarwal and Wenlan Qian (Business School, National University of Singapore).
Journal Information: The American Economic Review, Vol. 104, No.12 (Dec. 2014),PP-4205-4230.