Archive for September, 2015

Histogram vs Lorenz

Here is a question that might be worth of some thinking. Suppose one has a random sample of say 10000 observations on income of individuals and wants to record just two summary statistics for each of say 20 groups made from this data. What summary statistics should be recorded? Continue reading

Another Look at Melbourne’s Climate

QMelb

Source : See here page 4 by Roger Koenker

The plot illustrates 10 years of daily maximum temperature data for Melbourne. The curves are estimated conditional quantile functions for the quantiles {0.05,0.1,….,0.9,0.95}.

Histogram vs Lorenz Curve

I will put an updated post later

Grouped Data

Grouped data are not uncommon in income and wealth distribution studies. Inference about distributions with grouped data is one of my research areas (joint with several colleagues). However I am occasionally asked ,in seminars or privately, whether at the age of “big data” there is any point in working with grouped data. I think grouped data are still interesting and at present probably unavoidable in global income distribution studies for following reasons: Continue reading

Theory vs Measurement

I have long held the view that in many applications improved models or econometric techniques are of less importance compared to better measurements or higher data quality. I was glad then to find that I am not alone. The following 60-second lecture (a 2013 video) by one of the leading econmetricians seems to hold a similar view.  Continue reading

Influence function

One of the useful tools in asymptotic and robust statistics often not discussed in econometric textbooks (even the advanced ones) is the influence function. In this post, I try to explain the concept and its applications.

Continue reading

Lorenz Curves-1

I am going to write about Lorenz curves and inference about them in a series of posts. This first post tries to explain why there are so much interests in Lorenz curves.

CDF and PDF are the first things come to mind when thinking of a distribution but there is another representation namely the Lorenz curve that is as widely used and estimated (if not more) in some contexts such as income distribution analyses. Here are the reasons that I can think for popularity of Lorenz curves: Continue reading