Unbalanced Panel Data

1
Unbalanced Panel Data and Stata

• Kuan-Pin LinPortland State University andWISE,
  Xiamen University

2
Panel Data Anslysis

• Panel Data Models
• Fixed Effects Model
• Random Effects Model

• Panel Data Definition
• Unbalanced Panel
• Balanced Panel
• Short Panel
• Long Panel

3
Panel Data Analysis

• Benefits
• More Data Variability
• Less Colinearity among Variables
• Control for Unobserved Heterogeneity
• Limitations
• Time Serial and Cross Sectional Correlation
• Missing Data
• Randomly and Non-randomly
• Sample Selection Bias

4
Unbalanced Panel Data

• Incomplete panels are more likely to be the norm
  in typical economic empirical studies
• Problems with non-response and measurement errors
• Gaps in Time Series
• Holes in Cross Sections

5
Panel Data Analysis Using Stata

• Declare panel data and variables
• xtset
• Panel data analysis xt commands
• xtdes
• xtsum
• xtdata
• xtline
• Panel data regression
• xtreg

6
Returns to Schooling Koops and Tobias (2004)

• Study the relationship between wages and
  education, ability, and family characteristics.
• Data is available in two parts. The first file
  contains the panel of 17,919 observations on the
  Person ID and 4 time-varying variables. The
  second file contains time invariant variables for
  the individual or the 2,178 households. See the
  article for details on the empirical model and
  data construction (data achieve and demo program
  1 2).
• Data source U.S. National Longitudinal Survey of
  Youth (NLSY), U.S. Dept. of Labor, Bureau of
  Labor Statistics.

7
Returns to Schooling Koops and Tobias (2004)

• Part 1
• Column 1 Person id (ranging from 1 to 2178),
• Column 2 Education,
• Column 3 Log of hourly wage,
• Column 4 Potential experience,
• Column 5 Time trend.
• Part 2
• Column 1 Time invariant ability,
• Column 2 Mother's education,
• Column 3 Father's education,
• Column 4 Dummy variable for residence in a
  broken home,
• Column 5 Number of siblings.

8
Housing Prices in China Wang (2011)

• Study the effects on housing consumption and
  prices from the privatization of state-owned
  housing reform began in 1994 (data file and demo
  program 1 2)
• Based on standard regression analysis, the
  removal of price distortions allowed households
  to increase their consumption of housing and led
  to an increase in equilibrium housing prices.
• Using CHNS, data is obtained for 31677
  individuals from about 2900 households in
  counties and cities of 9 provinces over 6 years
  (1989, 1991, 1993,1997, 2000, 2004).

9
References

• B.H. Baltagi, and S.H. Song, Unbalanced Panel
  Data A Survey, Statistical Papers 47, 493-523,
  2006.
• B.H. Baltagi, Chapter 9 Unbalanced Panel Data
  Models, Econometric Analysis of Panel Data, 4th
  ed., John Wiley, New York, 2008.
• G. Koops, and J.L. Tobias, Learning About
  Heterogeneity in Returns to Schooling, Journal
  of Applied Econometrics 19, 827-849, 2004.
• S-Y. Wang, State Misallocation and Housing
  Prices Theory and Evidence from China, American
  Economic Review 101, 2081-2107, 2011.