question archive Econometrics/Fall 17 Homework Set #5 1
Econometrics/Fall 17 Homework Set #5 1
Subject:BusinessPrice: Bought3
Econometrics/Fall 17 Homework Set #5
1. Explain the nature, cause and consequences of heteroscedasticity in your own words.
Detect possible heteroskedasticity in your equation--with four variables (in the previous assignment). Justify your answer (explain why there is/is not heteroskedasticity). Suggest corrective remedies (even if serial correlation is not detected).
2. In a metropolitan center with more than four million residences and 51 municipalities, housing prices increase across the board above the national average yet the municipalities show uneven housing-price changes. You are a young economist hired to help understand the nature of the region’s housing affordability. Statistical analyses are needed to provide policy measures to keep the housing prices at or below the national growth rate. You are given access to a local library. Quarterly sales and prices and other necessary records are available for the last twenty-five years in 40 different metropolitan areas. Your job is to design and execute a regression analysis. After doing a brief literature search, design an econometric study to determine the determinants of housing price growth. Prepare a proposal including a model, its justification, expectations, possible tests for robustness and any other creative elements to prove yourself.
3. After a brief research, come up with two models (your own): one could convincingly benefit from a distributed lag variable and another again convincingly dynamic (autoregressive) scheme. Discuss the pros and cons of these versions specific to your models. And if there are any possible issues how to correct them.
4. Forecast, with the four-independent-variable model you used in the previous assignment, the poverty rate for:
a. average values; and b. two years past the most recent period after extrapolating dependent variables. Explain your reasoning and method of extrapolation
c. also estimate the 95% confidence interval for both estimates
d. also explain factors that could further improve your forecasts.
