New Research on National Statistics and Administrative Data
Sunday, Jan. 8, 2017 1:00 PM – 3:00 PM
- Chair: Carol Corrado, Conference Board
Public Health Insurance, Labor Supply, and Employment Lock: Effects or Data Artifacts?
AbstractUsing Social Security Administration (SSA) data on employment and Census population estimates, we re-examine the labor supply effect of a large public health insurance disenrollment that took place in Tennessee in 2005. The disenrollment left about 4 percent of Tennessee's non-elderly, adult population without public coverage and was previously reported in Garthwaite, Gross, and Notowidigdo (2014) to have caused large increases in labor supply. These labor supply increases were interpreted as evidence of "employment lock," i.e. people finding employment in order to secure health insurance coverage. In contrast to the previous research conducted with the Current Population Survey (CPS), we estimate small and not statistically significant labor supply increases in SSA/Census data.<br />
We propose that the large estimates in the original analysis are biased due to data error. Thus, using CPS data we find that the observed increases in the share of workers in the CPS are driven primarily by declines of Tennessee's civilian population rather than increases in the number of people being employed. Furthermore, we find the labor supply increases estimated in the original paper mask dramatic gender differences: remarkably large labor supply increases for men and much smaller and not statistically significant for women. Finally, the alternative employment measure available in the CPS – employment during last year -- suggests considerably smaller and not statistically significant labor supply increases in Tennessee following the dis-enrollment.<br />
The paper contributes to the literature in three ways. First, it narrows the range of existing estimates of the effect of public health insurance on labor supply. Second, it uses a mostly unknown source of employment data. Third, it considers the possibility that survey data error can cause time-space-group data variation — the situation which violates standard assumptions and is often ignored in policy evaluations using difference-in-differences and triple-difference approaches.
Capturing the Productivity Impact of the ‘Free’ Apps and Other Ad-Supported Media
Abstract"Free" consumer entertainment and information from the Internet, largely supported by advertising revenues, has had a major impact on consumer behavior. Some economists believe that measured GDP growth is badly underestimated because GDP excludes online entertainment (Brynjolfsson and Oh 2012; Ito 2013; Aeppel 2015). This paper introduces an experimental GDP methodology which includes advertising-supported media in both final output and business inputs. For example, Google Maps would be counted as final output when it is used by a consumer to plan vacation driving routes. On the other hand, the same website would be counted as a business input when it is used by a pizza restaurant to plan delivery routes. <br />
Contrary to BEA's critics, including 'free' media in the input-output accounts has little impact on either GDP or TFP. Between 1998 and 2012, measured nominal GDP growth falls 0.005% per year, real GDP growth rises 0.009% per year and TFP growth rises 0.016% per year. The changes to nominal GDP, real GDP and TFP are even smaller before 1998. <br />
Our method for accounting for 'free media' is production oriented in the sense that it is a measure of the resource input into the entertainment (or other content) of the medium, rather than a measure of the consumer surplus arising from the content. BEA uses a similar production oriented approach when measured GDP. In contrast, other researchers used broader approaches to measure value (Brynjolfsson and Oh 2012), (Varian 2009) and (Bughin et. al 2011). We've also explored a more expansive accounting methodology which incorporates some of the network effects from online media. With this expansive accounting, the productivity boom from 1995 to 2000 becomes even stronger and the weak productivity growth of the past decade may be ameliorated somewhat
- C8 - Data Collection and Data Estimation Methodology; Computer Programs