How sensitive are estimates of the marginal propensity to consume to measurement error in survey data in South Africa Reza C. Daniels UCT

Slides:



Advertisements
Similar presentations
Graduate Methods Master Class
Advertisements

REGRESSION, IV, MATCHING Treatment effect Boualem RABTA Center for World Food Studies (SOW-VU) Vrije Universiteit - Amsterdam.
There are at least three generally recognized sources of endogeneity. (1) Model misspecification or Omitted Variables. (2) Measurement Error.
Economics 20 - Prof. Anderson1 Multiple Regression Analysis y =  0 +  1 x 1 +  2 x  k x k + u 7. Specification and Data Problems.
Economics 105: Statistics GH 24 due Wednesday. Hypothesis Tests on Several Regression Coefficients Consider the model (expanding on GH 22) Is “race” as.
Specification Error II
1Prof. Dr. Rainer Stachuletz Multiple Regression Analysis y =  0 +  1 x 1 +  2 x  k x k + u 7. Specification and Data Problems.
QUALITATIVE AND LIMITED DEPENDENT VARIABLE MODELS.
Smoking, Drinking and Obesity Hung-Hao Chang* David R. Just Biing-Hwan Lin National Taiwan University Cornell University ERS, USDA Present at National.
Prof. Dr. Rainer Stachuletz
1 MF-852 Financial Econometrics Lecture 8 Introduction to Multiple Regression Roy J. Epstein Fall 2003.
Multiple Regression Analysis
CHAPTER 4 ECONOMETRICS x x x x x Multiple Regression = more than one explanatory variable Independent variables are X 2 and X 3. Y i = B 1 + B 2 X 2i +
Multiple Regression Analysis
Met 2651 Instrument variabler (sider: 539,543, 544,545,549,550,551,559,560,561,564) Ulf H. Olsson Professor of Statistics.
Statistical Analysis SC504/HS927 Spring Term 2008 Session 7: Week 23: 7 th March 2008 Complex independent variables and regression diagnostics.
1 MF-852 Financial Econometrics Lecture 6 Linear Regression I Roy J. Epstein Fall 2003.
FINAL REPORT: OUTLINE & OVERVIEW OF SURVEY ERRORS
1 Health Status and The Retirement Decision Among the Early-Retirement-Age Population Shailesh Bhandari Economist Labor Force Statistics Branch Housing.
Correlation Question 1 This question asks you to use the Pearson correlation coefficient to measure the association between [educ4] and [empstat]. However,
1 Dealer Price Discrimination in New Car Purchases: Evidence from the Consumer Expenditure Survey Pinelopi Goldberg (JPE, 1996) Presented by Jake Gramlich.
3.1 Ch. 3 Simple Linear Regression 1.To estimate relationships among economic variables, such as y = f(x) or c = f(i) 2.To test hypotheses about these.
Tax Subsidies for Out-of-Pocket Healthcare Costs Jessica Vistnes Agency for Healthcare Research and Quality William Jack Georgetown University Arik Levinson.
  What is Econometrics? Econometrics literally means “economic measurement” It is the quantitative measurement and analysis of actual economic and business.
Multiple Regression. In the previous section, we examined simple regression, which has just one independent variable on the right side of the equation.
TIME CONSTRAINTS, DURABLE CONSUMER GOODS AND THE PREVALENCE OF OBESITY IN WESTERN EUROPE Karsten Albæk.
How do Lawyers Set fees?. Learning Objectives 1.Model i.e. “Story” or question 2.Multiple regression review 3.Omitted variables (our first failure of.
Error Component Models Methods of Economic Investigation Lecture 8 1.
Has Public Health Insurance for Older Children Reduced Disparities in Access to Care and Health Outcomes? Janet Currie, Sandra Decker, and Wanchuan Lin.
1 MF-852 Financial Econometrics Lecture 10 Serial Correlation and Heteroscedasticity Roy J. Epstein Fall 2003.
The Demand Side: Consumption & Saving. Created By: Reem M. Al-Hajji.
Random Regressors and Moment Based Estimation Prepared by Vera Tabakova, East Carolina University.
Bureau of Economic Research, University of Dhaka The Role of Credit in Food Production, Food Security & Dietary Diversity in Bangladesh Authors Dr. Sayema.
Lecture 7: What is Regression Analysis? BUEC 333 Summer 2009 Simon Woodcock.
Instrumental Variables: Introduction Methods of Economic Investigation Lecture 14.
Correlation. Up Until Now T Tests, Anova: Categories Predicting a Continuous Dependent Variable Correlation: Very different way of thinking about variables.
Overview of Regression Analysis. Conditional Mean We all know what a mean or average is. E.g. The mean annual earnings for year old working males.
5. Consistency We cannot always achieve unbiasedness of estimators. -For example, σhat is not an unbiased estimator of σ -It is only consistent -Where.
Using microsimulation model to get things right: a wage equation for Poland Leszek Morawski, University of Warsaw Michał Myck, DIW - Berlin Anna Nicińska,
Randomized Assignment Difference-in-Differences
Lecturer: Ing. Martina Hanová, PhD. Business Modeling.
1 1 Using administrative registers to evaluate the effects of proxy interviews in the Norwegian Labour Force Survey Øyvin Kleven, Ib Thomsen and Ole Villund.
AUTOCORRELATION 1 Assumption B.5 states that the values of the disturbance term in the observations in the sample are generated independently of each other.
The Instrumental Variables Estimator The instrumental variables (IV) estimator is an alternative to Ordinary Least Squares (OLS) which generates consistent.
Effects of migration and remittances on poverty and inequality A comparison between Burkina Faso, Kenya, Nigeria, Senegal, South Africa, and Uganda Y.
Income Convergence in South Africa: Fact or Measurement Error? Tobias Lechtenfeld & Asmus Zoch.
1/25 Introduction to Econometrics. 2/25 Econometrics Econometrics – „economic measurement“ „May be defined as the quantitative analysis of actual economic.
SECTION 1 TEST OF A SINGLE PROPORTION
Lecturer: Ing. Martina Hanová, PhD. Business Modeling.
Analysis of Mismeasured Data David Yanez Department of Biostatistics University of Washington July 5, 2005 Biost/Stat 579.
Chapter 4. The Normality Assumption: CLassical Normal Linear Regression Model (CNLRM)
AN ANALYSIS OF HOUSEHOLD EXPENDITURE AND INCOME DATA
Analysis of results.
REGRESSION DIAGNOSTIC II: HETEROSCEDASTICITY
Instrumental Variable (IV) Regression
More on Specification and Data Issues
STOCHASTIC REGRESSORS AND THE METHOD OF INSTRUMENTAL VARIABLES
More on Specification and Data Issues
Instrumental Variables and Two Stage Least Squares
Chapter 6: MULTIPLE REGRESSION ANALYSIS
How sensitive are estimates of the marginal propensity to consume to measurement error in survey data in South Africa Reza C. Daniels UCT
Introduction to Microeconometrics
Instrumental Variables and Two Stage Least Squares
Some issues in multivariate regression
Instrumental Variables and Two Stage Least Squares
Tutorial 1: Misspecification
Instrumental Variables Estimation and Two Stage Least Squares
Financial Econometrics Fin. 505
Chapter 2: Steps of Econometric Analysis
More on Specification and Data Issues
Presentation transcript:

How sensitive are estimates of the marginal propensity to consume to measurement error in survey data in South Africa Reza C. Daniels UCT Vimal Ranchhod UCT

Outline Context Question Econometric problem Proposed solution Data Results Caveats Conclusion

Context Best micro level data in SA for incomes and expenditures comes from StatsSA income and expenditure surveys (ies 95, 00 and 05) From here, we can estimate the marginal propensity to consume (MPC), i.e. what proportion of every rand of disposable income do households spend. This can be broken up into various categories of expenditure. The marginal propensity to save (MPS) is defined as 1 – MPC, with corresponding definition.

Context (2) A common problem in survey data is measurement error in responses. i.e. The data captured might not truly reflect the financial reality being measured. In the `classical measurement error’ case, this leads to attenuation bias. Estimated relationships are weaker than the true relationships. For other types of measurement error, even being able to sign this bias may not be possible.

Question How sensitive are estimates of the MPC to measurement error (m.e.) in the IES data. – We propose to estimate this sensitivity using an instrumental variables approach, with wage data from the same households, but from a different survey, namely the LFS 2000:2, as our candidate instrument.

Econometric Problem Suppose the “true” relationship is: – Y i = B 0 + B 1 X i * + u i, where: Y is the outcome variable of interest, X * is the ‘true’ value of the dependent variable, And u is a mean zero error term. The subscript i refers to person i, where i=1, …, n – It can be shown that, asymptotically, the OLS estimator of B1 obtained from a regression of Y on X * will be consistent if and only if: Cov(X *, u) = 0

OLS estimator with measurement error Suppose that we observe X instead of X*, – Where X = X* + e, E[e]=0, cov(X *, e) = 0 and cov(u, e)=0 By regressing Y on X, we can show that the probability limit of our estimate of B 1 from an OLS regression = B 1 + cov(X, u-B 1 e)/Var(X) = B 1 – B 1 (Var(e)/[Var(X * ) + Var(e)]) This is known as attenuation bias. – In essence, regardless of the type of m.e. we are considering, the crucial question to ask is whether or not the covariance between the observed X and the composite error term is zero.

M.E: An IV solution Suppose we had another variable, Z, which is: – Correlated with our observed X, and – Uncorrelated with the composite error term, v=(u-B 1 e) Then Z would provide a valid instrument for the endogenous regressor X. In particular, another noisy measure of X *, eg. Z=X * + k would suffice, if k is uncorrelated with both u and e.

Asymptotically plimB hat1,IV = B 1 + (corr(Z,v)/corr(Z,X 1 ))*(var(v)/var(X 1 )) 0.5 (Recall that v=(u-B 1 e)) Now, var(v) ≠0, var(X 1 )≠0 and corr(Z,X 1 )≠0, therefore the IV works iff corr(Z,v)=0. i.e corr(X 1 * + k, u-B 1 e)=0, So the crucial requirement is that k and e are uncorrelated.

Implementing the solution We match data on individuals from the IES 2000 and LFS 2000:2. The data were obtained in October and September respectively. IES contains more detailed information on multiple sources of income and categories of expenditure, expenditure at HH level. LFS contains information on employment status and wage income.

Summarizing the Data Table 1: Summary statistics of sample Variable # of individuals # of HHs25964 Mean HH size3.96 % of sampleMean HH Yd p.a. Mean HH Exp p.a. Ratio of exp/Yd African Coloured Indian White Total Male Female Notes: 1. Means do not include sampling weights. 2. Means are obtained by race or gender of HH head.

Mean Expenditure by Category CategoryMean ExpenditureAs % of Yd alcohol & cigs beverages clothes durable goods education food health housing insurance other non durable own production recreation hh services transport utilities TOTAL EXPENDITURE Total Disposable Income32058 Notes: 1. Proportion in sub-category do not necessarily add to 100, due to omitted categories such as debt servicing.

OLS and IV coefficients OLSIV COEFFICIENTyds.eyds.e alcohol & cigs *** *** beverages ***-6.5E *** clothes *** *** durable goods0.0142*** *** education0.0153*** *** food0.0355*** *** health0.0274*** *** housing0.0747*** *** insurance0.217*** *** other non durable0.0828*** own production0.315*** *** recreation *** *** hh services0.0317*** *** transport0.0639*** *** utilities0.0119*** *** TOTAL0.981*** ***-0.026

Caveats Sensitive to imputation method of wage data from categories. Conceptual difficulties on how to treat debt financing and dissaving or borrowing. IVs are fairly weak, many HH’s get income from grants. If m.e. correlated with income levels, eg. Rich HHs always understate, then its not clear that the solution is valid. Non-response also not accounted for.

Conclusion Promising avenue of investigation IV’s do have significant first stages Lots more to be done … – Do by income quintile, – And by age of HH head, or some form of HH composition.