Preliminary Abstracts for Economics 551 Term Papers
From: Seunghyun Hong
Dear Prof. Rust;
About the paper topic, I did not decide the title yet. But what I have in
mind now is analyzing evolutionary property of time series data, like
monthly production data. I think there must be some property changes in
data series which is usually analyzed through structural break. What I
want to do is analyze the property chages in spectarl domain rather than
time domain.(Jointwork with TA Woocheol Kim)
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
Hong, SeungHyun ( LUKE )
- Department of Economics, Yale Univ.
- Tel/Fax : (203)436-2787
- e-mail : seunghyun.hong@yale.edu
- Address : P.O.Box 200054
Yale Station
New Haven CT 06520
U.S.A.
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
____________________________________________________________________
Econometrics II Paper Brief Proposal
"Export Concentration and Earnings Instability in Developing Countries"
by Garth Frazer
For Professor J. Rust
18 February 1998
Export diversification in developing countries has been studied for
its links to reducing export instability, and increasing export and
overall growth. Theoretically, export diversification is assumed to
decrease export instability, as producing a variety of products should
offset the adverse consequences of short-run fluctuations in demand for
specific export products. However, early cross-section empirical
studies (in the 1970s and early 1980s) found the relationship to be
weak. Slightly more recent studies (as late as 1990), using time-
series, have found that export diversification reduces export
instability, as expected. I hope to analyze this relationship
between export instability and export diversification, or its
opposite, export concentration, beginning with countries in Africa.
The number of countries in my study is not yet finalized, as it will
partly depend on obtaining adequate data. Comparable export earnings
data is available, at least on paper, if not in electronic form, for
most countries. The export concentration indices (Gini-
Hirschmann indices) will require more work to construct, using
three-digit SITC (Standard International Trade Classification)
categories.
____________________________________________________________________
From: Halla Qaddumi
Abstract:
I plan to estimate domestic (household) demand for water across various
consumption and housing categories in Hyderabad (Andhra Pradesh), India.
The data that will be used for empirical estimation of water demand comes
from a survey covering 862 households commissioned by the Hyderabad
Metropolitan Water Supply and Sewerage Board in 1991-92. The objective of
the study is to determine the price elasticity for water, as well as to
investigate the welfare effects of alternative water prices. The
rationale for the study is to demonstrate the perverse effect of the
existing water tariff structure on level and pattern of water consumption
and to test the hypothesis that marginal cost (efficient) pricing of water
has positive conservation and welfare implications.
____________________________________________________________________
From: David Love
Professor Rust-
A dry title for my paper might be, "Transportation and Jobs in the
Inner City." I'll work on the flare. I want to examine the effects of
transportation availability on employment opportunities in the United
State's cities (New York esp.) This can be done several ways. One
posssibiblity, if the
data are available, would be to look at the impact of a new subway or
train line on employment. Actually, I'm foggy as to the right approach,
but I have the blind confidence to try anyhow. I'm guessing that my paper
will develop as I collect data.
When I narrow my search a bit, I'd like to talk to you and see
what you think.
____________________________________________________________________
From: Ling Hu
Dear professor John Rust:
I have chosen the topic for the paper: income distribution in China urban
area, 1986 and 1994. I am sorry that this message is hardly an abstract as
I have not seen the data yet. Doctor Yu Xue Jun in Growth Center has some
first hand data on this topic. We had a brief talk days ago. But as he has
a presentation today, he can only meet me and explain the data to me in
detail tomorrow.
I choose this topic because I always want to do some research on Chinese
economic issues. In the past 15 years, China experienced profound
transformation. Many characteristic need to be described and analysised.
Although income determination is a quite minor aspect, it is quite
intersting and instructive.
For example, as many have observed, the difference between rich and poor
was enlarged dramatically in China in recent years. In this paper, I would
like to try to see to what degree this difference is "FAIR", or can be
explained by education, skill, location, and so on. Then the residual can
be regarded as luck, and deficiency of the institution.
The literature I am reading now is the book by J. Mincer: Schooling,
experience and earnings. There are also many articals on this topic.
However, as China is in both development and transformation, it is quite
special. And Chinese literature on this topic is very rare, as this is a
phenomenon only from last decade. As late as early 80s, major urban income
are salaries and most of the salaries are fixed. Difference in salaries is
very small.
I am sorry that I have not finished the proposal by now. But I will try my
best to catch up in the spring break. As this is far from well prepared, I
think it is not helpful to be on the web.
Best regards!
Ling Hu
____________________________________________________________________
From: Nancy Epling
Professor Rust:
Here is my paper topic and abstract. Please let me know if you have any
questions or concerns.
Thanks,
Nancy
"Consumer Benefits from New Goods Produced in an Oligopolistic Market"
This paper will develop a model of the US Automobile Industry and estimate
the valuation of a given new product into the market. A discrete choice
model will be used in conjunction with population weights to produce an
estimated demand function. Utilizing this estimated demand function, the
paper will estimate consumer benefit from the addition of a given new
product. A "virtual" price that sets demand equal to zero will be
evaluated using similar methods to those in Hausman (1994). Applications
from this study will include implications for the evaluation of the
consumer price index in regard to goods produced in oligopolistic markets.
Micro data from the Consumer Expenditure Survey conducted by the Bureau of
Labor Statistics as well as the Automotive News Market Data Book will be
used.
____________________________________________________________________
From: Liang Peng
professor:
I would like to test some theorem about households' consumption,
especially the Random-Walk Hypothesis. I think it is important to
understand households' consumption decision for us to study many economy
phenomenon such as growth and fluctuations.
I would organize my paper as following:
1), describe the Random-Walk Hypothesis and try to make its implication
about households' consumption.
2), Try to identify some variables trough which households would make
prediction of their income. Simulate the "households' income predicting
process".
3), Test if Random Walk Hypothesis is successful in describing households'
consumption.
4), Try to give a explaination of the possible deviation from hypothesis
to reality.
The main methods of economerics would be IV and OLS. Other methods may
be proved to be also important when I go on.
I would start to do my paper immediately. Thank you very much.
Best Wishes!
Your Sincerely
Liang Peng
____________________________________________________________________
From: Alok Kumar
Experiments with a Symbolic Framework for Time Series Prediction and
Visualization
Alok Kumar, Yale School of Management
alok.kumar@yale.edu
http://pantheon.yale.edu/~ak237
A symbolic framework for time series prediction is proposed. First of
all, an enhanced dataset is created by replacing each datum
by a vector (known as a feature vector or FEVA) which is a snapshot of
the “shape” of the local neighborhood of the datum. The
vector attempts to capture how a given datum is embedded in its
surrounding. Preprocessing is followed by a clustering
procedure that groups together data points that are embedded in
similar surroundings. A symbolic representation of the given time
series is obtained by representing each datum by an appropriate
cluster number (a “symbol”) and then two separate modeling
paradigms are used to approximate the dynamics of the underlying
process: (a) an Elman recurrent neural network model and (b)
a second order Markov Chain. In a separate step, we use a string
matching algorithm and several simple visualization models to
identify (hidden) structure in the given time series. The approximate
pattern matching technique gives a reduced form
representation of the time series in terms of a set of strings of
symbols (words). Efforts are currently underway to develop
techniques that would exploit this structural information to improve
the prediction performance. The symbolic framework is tested
on two real datasets and the preliminary results are encouraging.
Keywords: Time Series Prediction, Feature Vector Analysis (FEVA),
Symbolic Representation, Clustering, Recurrent Neural
Networks, Markov Chains, String Matching, Data Visualization.
A current version of the abstract (and the working paper) is available
at http://pantheon.yale.edu/~ak237/symbol1.html
____________________________________________________________________
From: Talia Bar
Dear Professor Rust,
I am interested in doing a health economics realted paper.
The question that interests me is whether an increase in the
percentage of women that follow the regular path of checkups
recommended for an early detection of breast cancer can reduce
costs of treatment, thus making it worthwhile for health
insurance organizations to give incentives for women to
follow that path.
I have not yet located a data set that can help me test this
question. I will continue the search and may decide to change
the topic if I do not locate such a data set.
Thank you,
Talia Bar
____________________________________________________________________
From: Ana Fernandes
Prof. Rust
The topic for my econometrics II paper will be the following:
(note that slight changes may occur related to data gathering and
processing problems)
The paper will attempt to relate economic growth with openness of
economies.
The study will be performed for developing countries and the methods of
estimation used will be OLS and Granger Causality tests.
GDP growth will be regressed on all traditional factors (resulting of a
generalized production function) and a new one a measure of openness of
economies. Several measures of openness will be used.
Robustness of results will be tested by introducing other macropolicy
related variables that could dismiss the effects of openness on growth.
The causality issue between openess and growth will also be studied.
Ana Margarida Fernandes
____________________________________________________________________
A Cross Sectional Study of the Relative-Income Hypothesis
Michael McBride
ECON 551, Spring 1998
Abstract
Past empirical work has suggested that a person's subjective well-being is
based in large part on relative and not absolute income. This conclusion,
called the "Relative-Income Hypothesis" casts doubts on the viability of
many current economic theories and policies and thus warrants a closer
look. This paper will examine this relationship between relative income
and SWB using cross sectional data and a simple model. Emphasis will be
placed on identifying the norms and reference groups which are used in
making relative income comparisons. Generational effects will also be
examined.
Planned Econometric Methods
First, the truncated income measures will be dealt with. Second, the "norm
income" which is used in making income comparisions should be predicted
based on relevant demographic information (Clark & Oswald, 1996).
Different norms and reference groups will be examined. Third, (probably
using a probit model) subjective well-being will be regressed on the
predicted norm income and other relevant variables for each of the
different reference groups. Fourth, significance level wills be tested.
The strength of the conclusions will be discussed.
Data Source
General Social Survey at the National Opinion Research Center located on
the Internet at http://www.icpsr.umich.edu/gss/
Some Data Problems
1) Identification: SWB could be influenced by health, absolute income,
relative income, family, etc., or SWB could cause one of those. We will
assume that SWB is the dependent variable, and there is support for this
assumption. There could also be generation effects, ie., each new
generation has higher aspirations, and as the older generations die then
the younger generations preferences alter the overall SWB measures.
2) Simultaneous Equations:
A) Race and environment may affect absolute/relative income.
B) Relative income may affect marriage status.
3) Income measurement is truncated at a high income level.
4) Preference falsification (Kuran 1995) is probably not too important, but
should be noted. Other concerns about the meaningfulness of the data is
also to be recognized.
Selected References
Clark, Andrew, and Andrew Oswald. 1996. "Satisfaction and Comparison
Income." Journal of Public Economics 61: 359-381.
Diener, Ed. 1984. "Subjective Well-Being." Psychological Bulletin 95:
542-575.
Easterlin, Richard. 1974. "Does Economic Growth Improve the Human Lot?
Some Empirical Evidence," in Paul A. David and Melvin W. Reder, eds.,
Nations and Households in Economic Growth: Essays in Honor of Moses
Abramovitz. New York: Academic Press, Inc.
Easterlin, Richard. 1995. "Will Increasing the Incomes of All Increase
the Happiness of All?" Journal of Economic Behavior and Organization
27:35-47.
Kuran, Timur. 1995. Private Truths, Public Lies: The Social Consequences
of Preference Falsification. Cambridge, Massachusetts: Harvard University
Press.
McBride, Michael. 1997. The Relative-Income Hypothesis and Income
Redistribution. M.A. thesis, University of Southern California.
McDonald, Maurice and Robin A. Douthitt. 1992. "Consumption Theories and
Consumers Assessments of Subjective Well-Being." Journal of Consumer
Affairs 26: 243-261.
Smith, Tom. 1990. Context Effects in the General Social Survey. See link
for details: http://www.icpsr.umich.edu/gss/report/m-report/meth55.htm
van de Stadt, huib, Arie Kapteyn, and Sara van de Geer. 1985. "The
Relativity of Utility: Evidence from Panel Data." The Review of Economics
and Statistics 67: 179-187.
____________________________________________________________________
From: Michael Vaney
To: John Rust
Subject: Paper topic for Michael Vaney
Tentative Title: Estimating the Cost of Capital for Canadian Natural
Gas
Pipelines, a Comparison of Methodologies.
In this paper I propose to review the manner in which the National
Energy
Board, which regulates interprovincial gas pipelines in Canada,
determines
the cost of equity capital for the pipelines under their jurisdiction.
At
this point in time I'm not even sure of which method they use, though I
suspect it is a variant of Discounted Cash Flow. I will try to
determine
a method based on CAPM or APT might lead to different results. This is
dependent on the National Energy Board being able to provide me with
data
for this (I am awaiting their reply).
Michael Vaney
____________________________________________________________________
Amil Dasgupta
February 17, 1998
Proposal:
A Non-Structural Investigation of Currency Contagion
Joint term paper for Econometrics II and The Emerging Markets of Latin
America.
In this paper I shall present econometric evidence for contagion
in foreign exchange markets. For our purpose, contagion can be defined as
the phenomenon by which successful speculative attacks on a currency leads
to (or increases the probability of) speculative attacks on other
currencies.
The recent crises in East Asia have drawn attention to the issue
of contagion. Large International Monetary Fund aid packages for affected
countries have typically been justified by the argument that crises in one
currency tend to spill over into other currencies which leads to global
volatility and dislocation. This raises important questions. Does
contagion exist? What causes it? To what extent is it a neighborhood
phenomenon defined by measurable parameters (e.g., comparative
fundamentals)?
Economic literature on the subject is scant. A plethora of models
address the issue of currency crises in general (e.g. Krugman 1979,
Obstfeld 1986 and 1995, Ozkan and Sutherland 1995). Very few deal
analytically with contagion. The only recent theoretical candidates are
an (N+1)-country center-periphery model by Buiter, Corsetti, and Pesenti
(1998) and an unpublished model by Goldfajn and Valdes (1995) emphasizing
the role of financial intermediation. Empirical literature on the subject
is sparser. Structural econometric studies do not exist. There have been
no established models to test. The best recent non-structural study is
Eichengreen, Rose, and Wyplosz (1996). These authors use data from OECD
countries from 1959 to 1992 to estimate the extent to which speculative
currency attacks are temporally correlated. They estimate a binary probit
model to conclude that "the existence of a currency crisis elsewhere in
the worldraises the probability of an attack on the domestic currency by 8
percent" (Eichengreen et al, 1996, p. 463).
In the absence of established structural models, I shall limit
this initial exercise to non-structural estimation. The results contained
in Eichengreen et al (1996) need to be modified to include the Tequila
crisis and the East Asian crisis. I shall replicate Eichengreen, Rose,
and Wyplosz's estimation, and extend it to include data from Latin America
and East Asia since 1993. I shall also modify their model to include some
more structure by adding "economic proximity weights" (e.g., the extent of
mutual trade) to determine whether contagion is a neighborhood phenomenon.
This paper shall, therefore, help to provide empirical answers to the
first and third of the three questions posed above.
Preliminary Bibliography
1. Buiter, W., G. Corsetti, and P. Pesenti 1998, "Interpreting the ERM
Crisis: Country-Specific and Systemic Issues," forthcoming in Princeton
Studies in International Finance, 1998.
2. Eichengreen, B., A. Rose, and C. Wyplosz 1996, "Contagious Currency
Crises: First Tests," Scandinavian Journal of Economics, v. 98, n. 4, p.
463-84.
3. Goldfajn, I., and R. Valdes 1995, "Balance of payments crises and
capital flows: The rose of liquidity," Unpublished manuscript, MIT.
4. Krugman, P. 1979, "A model of balance of payments crises," Journal of
Money, Credit, and Banking, v. 11, p. 311-25.
5. Obstfeld, M. 1995, "International currency experience: New lessons and
lessons relearned," Brookings Papers on Economic Activity, v. 2, p.
119-220.
6. Obstfeld, M. 1986, "Rational and self-fulfilling balance of payments
crises," American Economic Review, v.76, p. 72-81.
7. Ozkan, G., and A. Sutherland 1995, "Policy measures to avoid currency
crisis," Economic Journal, v. 105, p. 510-19.
___________________________________________________________
Economics 551b
Econometrics Project -- Economics 551b
David Yoon
Title: Economic Voting and Panel Survey Data
Abstract:
That economic conditions influence presidential votes is empirically well
established. But whether voters base their choices on the performance of the
national economy (sociotropic voting) or changes in personal finances
(pocketbook voting) is less clear.
Investigators have used either cross-sectional survey data, which does not
allow for the effects of changing economic conditions, or aggregated
time-series data, which does not allow for
the differentiation between sociotropic and pocketbook voting. By moving to
panel data -- provided by the Inter-university
Consortium for Political and Social Research (ICPSR) --
for the years 1992, 1993, 1994, and
1996 (four waves), I hope to flush out the effects of changes in both the
national economy and personal finances of individual respondents on presidential
approval (a proxy for the vote).
A dynamic fixed effects model will be used initially. Individual respondents
will be assumed to have fixed effects (intercepts) that influence their
approval or disapproval of the sitting president that are time-invariant for
the duration of the study -- such influences as race, party
identification, age group effects, social standing, geographic effects, that
are thought to affect the vote. One desirable consequence of using the fixed
effects model is that since these attributes are generally time-invariant,
they may be omitted without biasing our results.
In order to account for the effects of measurement error in the personal
finance measure, the 1992 waves responses will be used as instruments --
since the 1992 survey was taken during the Bush administration while 93 -96
surveys are all during Clinton's tenure, the study will essentially be using
a three wave panel. Whether there would be advantages to including fixed time
effects (economic or political shocks) will also be investigated.
___________________________________________________________
Professor Rust,
I am thinking of writing a paper that is related to Kelly Brownell's
(professor in Dept. of Psychology at Yale) proposal for a tax on
unhealthy foods.
Dr. Brownell is concerned with fighting diet-related illness on a national
level. He claims that "We live in a toxic food environment where
high-calorie and high-fat foods are available at low cost". Basically, he
proposes taxing foods that are high in calories and low in nutritional
value(e.g.Twinkies) and subsidizing foods high in nutritional value(e.g.fruits).
Whether the "fat tax" will be effective depends on price elasticities of
the targeted healthy and unhealthy foods, which I would like to estimate
in my paper. My main concern right now is about the availability of data.
Three states(California, Maryland and Maine) have levied this kind of tax
on the sale of junk food in recent years but no data were collected before
the laws were repealed (The tax is still in effect in Maine,though).
I'll contact prof. Brownell and his students to find out if they have the
data I need.
I would greatly appreciate any suggestions you might have regarding this topic.
Yan Li
___________________________________________________________
From: Lingfeng Li
Professor Rust,
I am interested in doing an empirical test on the Cox, Ingersoll and
Ross (1985) model of term structure of interest rates.
Theory:
Problem with testing affine term structure models is that the original
models are hard to fit in any realistic econometric test and extensions
from these models impose doubts on the relevance of results. Given my
limited knowledge on this topic and time constraint, I will basically
follow the one-factor extension of this model done by Brown and Debvig
(1986). Brown and Debvig did the first direct test of CIR model with US
data from 1952-1983, which did not yield encouraging results. With the
implementations of this model and increasing size of data in the last
ten years, I suspect there could be something more exciting.
Data:
This is my major concern about this project. I have not started looking
at data yet. Although it is not hard to find monthly records of T-bill,
note and bonds, it may be a pain for me to decide the applicability of
these data.
Methods:
Maximum likelihood methods and basic time series methods are most likely
to get involved. Some nonlinear methods will also be used. But I will
try to avoid tackling any technical stuff beyond my knowledge.
I am currently working on the algebraic part and hope to get it done
before Spring Break.
Regards
Lingfeng Li
(203)782-0781
___________________________________________________________
From: Richard Akresh
"Agricultural Production, Extension, and Gender Issues"
Given the importance of the agricultural sector in sub-Saharan
Africa in terms of both percentages of the population involved and share
of economic output and the relative stagnation of this sector, it is a
pressing need for most governments to try to improve agricultural
productivity and growth. Agricultural extension programs are one way to
narrow the gap between the best available techniques and those presently
being used. Historically, extension programs have produced minimal results
in Africa especially when compared to the agricultural productivity gains
in Asia during the Green Revolution.
During the last ten years, efforts have been made to introduce a
new extension program called the Training and Visit(T&V) management
system. Presently there is some controversy regarding its effectiveness
and ability to increase crop productivity. There have been several World
Bank discussion papers that tried to quantify the effects of the new
extension programs in Burkina Faso and Kenya. While the papers provided
some positive support for T&V effectiveness, there was no attempt to study
the differential gender effects of T&V.
Since women are becoming increasingly important as farm managers
in Africa, it is crucial to make sure their agricultural productivity is
rising as well. I want to analyze whether men are more likely than women
to adopt new technologies and follow extension advice and whether they are
more efficient in production. Also does the extension advice have a
different effect on the productivity increases for men compared to the
productivity increases for women? Given the relative efficiencies of
production, it might make more sense to put a larger share of agricultural
extension services at the disposal of that group.
___________________________________________________________
Preliminary Proposal for Econometrics paper
I.P. Singh
Some Issues in Measuring Price and Promotional
Elasticities from Store-Level Scanner
Data in
Comparison to Market Data
Over the last few years, availability of store-level and
household-level scanner data has brought new insights in the household
purchase behavior. Yet managers continue to use market level data to
estimate the price and promotional elasticities, resulting from variety
of marketing strategies. It would be interesting to see how the
store-level elasticities differ from those estimated from market data. A
previous work suggests that market level data leads to biased estimates
in comparison to store-level estimates. I plan to use the recently
released A.C. Nielson store-level data. To do the required analysis, the
store-level data could be aggregated to get the market data.
___________________________________________________________
I am currently interested in econometric study of the transition and
development economies. Particularly, I feel myself familiar with
Ukrainian economy.
Thus, I suggest that my future research papers deals with econometric
analysis and modelling of Ukrainian economy, described by several
macroeconomic indicators. Its main purpose will be to study and compare
several possible methods of construction of a simple model of the Ukrainian
economy. This model will be based on monthly time series statistics, and
will be able to serve for short-time forecasts.
Some of the variables which relationships I am going to explore are real
and nominal GDP, wholesale and consumer prices inflation, some money
aggregates and the exchange rate. I am going to use official data on
these indicators, available from Ukrainian sources. All the statistics will be
as current as possible. However, it may be necessary to make some additional
study of the series, and, possibly, to transform them or to construct some
new series. All the preliminary analysis of the data will be described.
I am going to explore several possible methods of estimation. Some of
them can be:
Simple multiple LS regression
Vector autoregression
Error correction methods
ARIMA methods
Exponential smoothing and other simple extrapolation methods
I am not yet sure which of these or other possible methods or their
combinations will be found suitable to the data, thus I do not claim
that all of them appear in the analysis. But the study will include several
most appropriate methods, and a quantitative comparison of the results.
The paper will include an analysis of single equations composing the
constructed model in terms of the coefficients values and significance,
and other useful statistics.
------
Yours truly, Andrei Jirnyi
E-mail: aj56@pantheon.yale.edu
Phone: (203) 865-4339