## Friday, April 26, 2013

## Monday, April 22, 2013

### Three Cheers for New Accounting

The U.S. is the first of its peers to adopt the 2008 System of National Accounts, an improvement on our original system (we shifted towards the SNA previously in the oughts). Most prominently, it will include R&D research and intangible assets.

This is an improvement economists have long pushed for, showing the inclusion of intangible capital (while the capital is intangible, investment in it is not) can be important in our understanding of labor productivity.

No doubt this will bring out the raving-mad conspiracy theorists. It is an exciting and long-awaited improvement to the National Income and Product Accounts.

This is an improvement economists have long pushed for, showing the inclusion of intangible capital (while the capital is intangible, investment in it is not) can be important in our understanding of labor productivity.

No doubt this will bring out the raving-mad conspiracy theorists. It is an exciting and long-awaited improvement to the National Income and Product Accounts.

## Friday, April 19, 2013

### Using the Absolute Difference Between Mean and Median to Estimate the Standard Deviation

For a normal distribution, the expectation of the mean and median is the same. However, in finite sample sizes, they will almost always be different. Given a number of observations, they will be more different the higher the standard deviation. Given the standard deviation, they will also be closer the more the observations.

Corrections presents Monte Carlo evidence that allows one to predict the standard deviation from only the absolute difference between mean and median and the number of observations, given the sample was produced using a normal distribution. Given a specific level and the number of observations, one can look up a unique value of the standard deviation.

Corrections presents Monte Carlo evidence that allows one to predict the standard deviation from only the absolute difference between mean and median and the number of observations, given the sample was produced using a normal distribution. Given a specific level and the number of observations, one can look up a unique value of the standard deviation.

Subscribe to:
Posts (Atom)