A statistical perspective on ill-posed inverse problems

A statistical perspective on ill-posed inverse problems
F O'Sullivan - Statistical Science, 1986 - jstor.org
Page 1. Statistical Science 1986, Vol. 1, No. 4, 502-527 A Statistical Perspectiveon Ill-posed Inverse Problems Finbarr O'Sullivan Abstract. ...

The ET Dialogue: A Conversation on Econometric Methodology

The ET Dialogue: A Conversation on Econometric Methodology
David F. Hendry, Edward E. Leamer, Dale J. Poirier
Econometric Theory, Vol. 6, No. 2 (Jun., 1990), pp. 171-261

Instrumental variable

Instrumental Variable Estimation of the Simple Errors-in-Variables Model
R. L. Carter, Wayne A. Fuller
Journal of the American Statistical Association, Vol. 75, No. 371 (Sep., 1980), pp. 687-692

Instrumental-Variable Estimation of an Error-Components Model
Takeshi Amemiya, Thomas E. MaCurdy
Econometrica, Vol. 54, No. 4 (Jul., 1986), pp. 869-880

Instrumental Variable Estimation in Generalized Linear Measurement Error Models
Jeffrey S. Buzas, Leonard A. Stefanski
Journal of the American Statistical Association, Vol. 91, No. 435 (Sep., 1996), pp. 999-1006

Finite Sample Properties of Instrumental Variable Estimators of Structural Coefficients
Roberto S. Mariano
Econometrica, Vol. 45, No. 2 (Mar., 1977), pp. 487-496

Does More Crime Mean More Prisoners? An Instrumental Variables Approach
Yair Listokin
Journal of Law and Economics, Vol. 46, No. 1 (Apr., 2003), pp. 181-206

Testing Identifiability and Specification in Instrumental Variable Models
John G. Cragg, Stephen G. Donald
Econometric Theory, Vol. 9, No. 2 (Jun., 1993), pp. 222-240

Consistent Estimation with a Large Number of Weak Instruments
John C. Chao, Norman R. Swanson
Econometrica, Vol. 73, No. 5 (Sep., 2005), pp. 1673-1692


Structural Equations, Treatment Effects, and Econometric Policy Evaluation Structural Equations, Treatment Effects, and Econometric Policy Evaluation
James J. Heckman, Edward Vytlacil
Econometrica, Vol. 73, No. 3 (May, 2005), pp. 669-738

Calibration

Introduction Calibration and Econometric Research: An Overview
Adrian Pagan
Journal of Applied Econometrics, Vol. 9, Supplement: Special Issue on Calibration Techniques and Eco... more
Published by: John Wiley & Sons

DEA technical efficiency score

function TE=techeff(X,Y)
%% techeff(X,Y) calculates the technical efficiency of decision making units%%% Input:% X: N x J matrix, with N inputs for each unit% Y: M x J matrix, with M outpus for each unit% J: # of decision making units%%% Output:% TE: J x 1 vetctor contains the technical efficiency of each DMU%%% Reference:% Hughes and Yaisawarng, 2004, "Sensitivity and dimensionality tests of DEA% efficiency scores," European Journal of Operational Research, 154,% p.410-422%%% Author information:% Yundong Tu% Department of Economics% University of California, Riverside% e-mail: yundong.tu@gmail.com% July 4, 2009
%check data inputs% if nargin<2% error(' Not enough input arguments to run techeff ')% end
if nargin<2 rand('twister',3); X=3.*rand(5,20); Y=5.*rand(3,20);end
%check dimension compatibilityif size(X,2)~=size(Y,2) error(' Input data are not compatible. Inputs and outputs must have the same collumns, i.e., for the same number of DMUs')end
options=optimset('Display','off');%,'MaxIter',100,'MaxFunEvals',100,'TolX',1e-2,'TolFun',1e-2);for k=1:size(X,2) [z,TE(k,1)]=linprog([zeros(size(X,2),1);1.0], [-Y,zeros(size(Y,1),1);X, -X(:,k)], [-Y(:,k);zeros(size(X,1),1.0)],... [ones(1,size(X,2)),0.0], 1.0,zeros(size(X,2)+1,1),ones(size(X,2)+1,1), 0.5.*ones(size(X,2)+1,1),options);end

Calling R from within Matlab

The COM interface allows Matlab users to call R from within Matlab. It is quite convenient for matlab users to interact with R functions. To make R project available to your matlab is very simple.

TO DO LIST:
1. Install Matlab: http://www.mathworks.com/
2. Install R: Open http://www.r-project.org/ , choose download R and then select a mirror.
3. Install the R package "rscproxy" in R. The package is available at: http://cran.r-project.org/. Go to package and find rscproxy. This can also be simply done by typing "install.packages('rscproxy') in R prompt. Then choose a mirror. Wait for 2 seconds and it is done.
4. Download R-(D)COM Interface by clicking R_Scilab_DCOM3.0-1B5.exe .
5. Download MATLAB R-link from: http://www.mathworks.com/matlabcentral/fileexchange/5051
6. Unzip the downloaded MATLAB R-link.zip into your current directory in matlab.
7. You can enjoy the convenience of having Matlab to call R now. Test if it is working by type "Rdemo" in Matlab commond window. If you are interested in nonparametric denstiy estimation, run the following MatlabcallRnp.m file

function MatlabcallRnp()
%% MatlabcallRnp() illustrates how to use Matlab to call the %% NP package in R. %% Note: all .m files in the Matlab_RLINK folder should be %% copied into the current directory in Matlab. You do not %% need run R when it is called.% Author Infor.% Yundong Tu% Department of Economics,% Univeristy of California, Riverside% e-mail: yundong.tu@email.ucr.edu
%% Connect to an R SessionopenR
%% Push data into R with putRdata()a = randn(200,1);putRdata('a',a)%%{%% Call the density (R function to estimate density) %% command into Matlab with evalR() %get the evaluation points for the density estimator xx= evalR('density(a)$x'); %get the corresponding evaluation of density fx= evalR('density(a)$y'); %figure subplot(2,1,1) plot(xx,fx); %plot the density function% %the density plot can be also called by R as evalR('plot(density(a),mfrow=c(2,1))');%}
%% Call the npudens (R function to estimate density) %% command into Matlab with evalR() %get the evaluation points for the density estimator evalR('data=data.frame(a)'); evalR('library(np)')
evalR('xx=npudens(data)$eval'); %dataframe xx=evalR('xx[[1]]') %matrix %get the corresponding evaluation of density fx= evalR('npudens(data)$dens') %figure %xy=eda; [xx,ind]=sort(xx); fx=fx(ind); subplot(2,1,2) plot(xx,fx); %plot the density function %the npudens plot can be also called by R as evalR('plot(npudens(a),add=FALSE)');%}%%{%% Run a series of commands and grab %% the result using getRdata()%save time to call density() just once, also get the bandwidthevalR('den=density(a)'); xx=evalR('den$x'); % the same as: evalR('xx=den$x'); xx=getRdata('xx');fx=evalR('den$y'); bw=evalR('den$bw');
figureplot(xx,fx);hold ontitle({['R-output DENSITY'];['Bandwidth=',num2str(bw)]});hold off%}%% Close the connectioncloseR

Other related links:
http://www.cs.ubc.ca/~murphyk/Software/callingRfromMatlab.html
http://webscripts.softpedia.com/script/Scientific-Engineering-Ruby/Statistics-and-Probability/MATLAB-R-link-35485.html

http://neurochannels.blogspot.com/2010/05/how-to-run-r-code-in-matlab.html

It will be a new world

Looking back to my econometric life for the past 5 years, many important persons have come to my life to help me make a difference. And my world has been changing all the time, indeed. There have been lights, heros, mentors, partners, as well as friends. And each experience is so precious and exciting. I feel deeply thankful for all those people there, adding new elements into my life and guiding me to move forward.

It is indeed not easy to stay in academia since there are way many things to learn, Econometric theory, computing techniques, economic applications and new stats elements as well. Deep love of all these rock and sand is sure the way to success in research life and the road to the happiness I have imagined years ago. It would be impossible to know everything, but good to know that there are things that need to learn in each stage as time goes by. To think deeper the questions at hand and the papers in progress, to learn the social skills to get to know the life of other professionals, to get things done more carefully, more precisely, more concisely, more intuitively and more regorously etc, are key steps to achieve academic accomplishment and recognizations. It will be a new world for sure and keep on marching forward.

Just look at an earlier blog about things econometricians do at
http://toyond.blogspot.com/2009/05/behave-as-econometrician.html

What’s New in Econometrics: Time Series, 2008

NATIONAL BUREAU OF ECONOMIC RESEARCH, INC.

SUMMER INSTITUTE 2008

What’s New in Econometrics: Time Series

James H. Stock and Mark W. Watson, Organizers

The role of Economic constraints in Econometrics

NONPARAMETRIC KERNEL REGRESSIONSUBJECT TO MONOTONICITY CONSTRAINTS by Peter Hall and Li-shan Huang, 2001, AS

NONPARAMETRIC ESTIMATION WHEN DATA ONDERIVATIVES ARE AVAILABLE by Peter Hall and Adonis Yatchew, 2007, AS

CONSTRAINED NONPARAMETRIC KERNEL REGRESSION:ESTIMATION AND INFERENCE by JEFFREY S. RACINE, CHRISTOPHER F. PARMETER, AND PANG DU

Econometrics___Bruce E. Hansen

Here is the link to the pdf version of Professor Hansen's graduate level text book:
Econometrics. Although it is still in its infancy, it is one of the best books for beginners. His way of understanding and presenting the material is not only deep, but easier for readers.

Visual Understanding of Higher-Order Kernels

Visual Understanding of Higher-Order Kernels
Author(s): J. S. MarronSource: Journal of Computational and Graphical Statistics, Vol. 3, No. 4 (Dec., 1994), pp. 447-458
Published by: American Statistical Association, Institute of Mathematical Statistics, andInterface Foundation of America
Stable URL: http://www.jstor.org/stable/1390905

Computationally Efficient Classes of Higher-Order Kernel Functions

Computationally Efficient Classes of Higher-Order Kernel Functions
Author(s): Belkacem Abdous
Source: The Canadian Journal of Statistics / La Revue Canadienne de Statistique, Vol. 23, No. 1(Mar., 1995), pp. 21-27
Published by: Statistical Society of CanadaStable URL: http://www.jstor.org/stable/3315548

Finite sample theory: A discussion

9:59 PM Amy: about Prof. Ullah's paper last Friday, what's the main point? Have he and Yongbao developed some estimation even when the error term is nonnormal?

Yundong Tu to Amy show details 10:18 PM (19 hours ago) Reply

Their papers are not to develope estimation result but to provide finite sample approximation for higher order moments of the estimators ( estimators, say beta hat, are taken as given, which can be MLE, GMM, IV, LS, etc.). The other paper is about expectation of quadradict form. They also provide finite sample approximation for these terms. Finite sample approximation differs when the error terms are nonnormally distributed from normally distributed case. These approximation results, however, could be used to study the properties of some other estimators, for example, the estimator of rho in the spatial autoregressive model, or the estimators of the coefficients in the MA or AR models.

10:29 PM Amy: so when error tems are nonnormal, we can still estimate the coefficients such as in VAR models. What 's the difference between this way and other methods approximating nonnormal errors to a normal distribution?
10:31 PM me: yes, you still can estimate using the same method as if the error term is normal
Amy: in Ullah's way?
me: no the classical way he is sillent about the estimation approach he is only concerned with the moments of the estimators
10:32 PM which is not quite a concern in macro, i think
Amy: but you say we can still estimate the coefficients me: yes
Amy: That's what I am considering
me: but we do not know the higher oder moments the properties of that is provided by Ullah and Bao
10:33 PM Amy: when the error term is nonnormal, could we use some methods of finite sample to estimate> Since they mention the MLE
me: finite sample is not to estimate the coefficients but to approximate higher order moments, say skewness and kurtosis of the estimators
10:34 PM Amy: I see. I am not familiar with finte sample
me: yes, you can still use MLE, GMM, IV and LS, etc. for the estimation purpose but once you get these estimators, you might be interested in its higher order properties the estimator you get would be not normally distributed
10:35 PM especially when the error term is not normally distributed and when the sample is small or even moderate large not even close to normal
Amy: I c.
me: finite sample approach is one method to tell how far is your estimator from a normal random variable
10:36 PM typical way to examine this is to check the property of skewness and kurtosis and see how far they are from those of normal distribution
10:37 PM in large sample, everything (estimator) is normally distributed, so there is no difference but what we have usually is not large sample observations
10:38 PM this leaves a room for finite sample theory to improve upon the large sample theory to get more accurate properties of the estimators we derived
Amy: But what will we do if we find the estimator is not normal distribution?
10:39 PM me: we impose finite sample corrections then
10:41 PM you know that for normally distributed random variable, skewness is zero and kurtosis is 3. if it is not normally distributed Ullah and Bao provide formula for those, which should also be used when sample size is small
10:42 PM Amy: I know that. But how to corret them?
10:43 PM me: use the fomula in Ullah and Bao for Skewness and Kurtosis
Amy: and then?
me: In Ullah's book, there should be fomula for mean and variance
10:45 PM Amy: I see. Maybe I should read the book firstly. But is there any way to use this correction way for estimation?
me: no it is not for estimation of the coefficient in econometric models like y=xb+e
10:46 PM their method is silent about the estimation approach, as i said earlier it can be used for a general class of estimators called extremum estimators
10:47 PM Amy: I c. Thank you. me: their approximation result depends on the following assumption sqrt(n)(bhat-b) converges to a normal distribution it's a pleasure
10:48 PM Amy: see. Thanks a million~~ I am looking at your pics in picasa
10:49 PM me: :)
10:51 PM Amy: ok, night~~ good dream.
me: u2

A Future Role for the Econometric Society in International Statistics

Charles F. Roos
Econometrica, Vol. 16, No. 2 (Apr., 1948), pp. 127-134
Published by: The Econometric Society A Future Role for the Econometric Society in International Statistics

Finite-Sample Asymptotics

Small Sample Asymptotics
Author(s): Michael A. FlignerSource: Journal of Educational Statistics, Vol. 13, No. 1 (Spring, 1988), pp. 53-61

Asymmetric Loss Functions

Optimal prediction under asymmetric loss: Christoffersen, FX Diebold - Econometric Theory, 1997 - jstor.org
Further results on forecasting and model selection under asymmetric loss: Christoffersen, FX Diebold - Journal of Applied Econometrics, 1996 - jstor.org
Financial Asset Returns, Direction-of-Change Forecasting, and ... :Christoffersen, FX Diebold - Management Science, 2006

R-np package

The following are from Racine's webpage:

Software
The np package (current version 0.30-1) for R (www.r-project.org) Obtaining: Available directly from the Comprehensive R Archive Network (cran.r-project.org)
Direct link to the np package on CRAN
Announcement on the R-packages mailing list
October 2007 Rnews article (pdf)
Manual (pdf)Vignette (pdf)FAQ (pdf) (html)

Nonparametric Econometrics: A Primer

Racine, J. S. (2008), "Nonparametric Econometrics: A Primer," Foundations and Trends in Econometrics: Vol. 3: No 1, pp 1-88. http://dx.doi.org/10.1561/0800000009

Incidental Prameter Problem

Neyman and Scott (1948) have noticed this problem in statistics and econometrics more than 50 years ago and their paper titled 'Consistent estimates based on partially consistent observations' was published in Econometrica. It incites researches over this topic in both economics and other natural science subjects, besides Bayesian and Conditional frequentist approach developed in Statistics literature. Lancaster reviewed the development up to 2000 and cast new insight into this problem.


Consistent Estimates Based on Partially Consistent Observations
J. Neyman and Elizabeth L. Scott
Econometrica, Vol. 16, No. 1 (Jan., 1948), pp. 1-32

The incidental parameter problem since 1948
Tony Lancaster
Journal of EconometricsVolume 95, Issue 2, April 2000, Pages 391-413

Asymptotic Theory and Econometric Practice with Large Dimension of Parameters

Asymptotic Theory and Econometric Practice
Author(s): Roger Koenker
Source: Journal of Applied Econometrics, Vol. 3, No. 2 (Apr., 1988), pp. 139-147

Koenker addressed the issue of estimation of models with increasing dimensionality in parameters when the number of observations are increasing, which is first noticed by Huber,
Robust Regression: Asymptotics, Conjectures and Monte Carlo
Peter J. Huber
The Annals of Statistics, Vol. 1, No. 5 (Sep., 1973), pp. 799-821