Skip to main content

Voter ID Laws and Voter Turnout

A Correction to this article was published on 14 June 2019

This article has been updated


In recent years, many states have enacted laws imposing strict identification requirements for voting. Proponents contend such laws are necessary to prevent voter fraud while opponents claim the laws disenfranchise legitimate voters, particularly black and Hispanic voters. This paper uses data from 2000 to 2014 federal elections to examine whether these new identification laws reduce voter turnout, either overall or among minority groups. The results provide no evidence that strict identification laws affect overall turnout or black turnout. However, the results do indicate a small reduction in Hispanic turnout, but this effect is statistically significant only if state fixed effects are not included in the estimation.

This is a preview of subscription content, access via your institution.

Change history

  • 14 June 2019

    The name of the third author was incorrect in the initial online publication. The original article has been corrected.


  1. Several other states, including North Carolina, Texas, and Wisconsin, enacted strict ID laws between 2000 and 2014, but the laws in these states faced court challenges and were amended to be non-strict ID laws or had implementation delayed until after 2014.

  2. There were 376 observations in the black turnout regressions; the missing observations were Idaho (2000, 2004, 2006, 2012, and 2014), Montana (2000, 2002, 2004, 2006, 2012, and 2014), New Hampshire (2002), New Mexico (2010), North Dakota (2006 and 2010), South Dakota (2004, 2006, and 2012), Vermont (2000, 2004, 2006, and 2014), and Wyoming (2002 and 2006). There were 392 observations in the Hispanic turnout regressions; the missing observations were Alabama (2002 and 2012), Maine (2006), Mississippi (2006), North Dakota (2000 and 2010), South Dakota (2014), and West Virginia (2006).

  3. For elections with no statewide offices on the ballot, we averaged the margin of the House of Representatives elections.

  4. At the suggestion of a referee, we also report results that use time fixed effects instead of the unemployment rate to allow for the possibility that the unemployment rate does adequately capture year-specific effects. For example, 2008 had very high turnout, likely because of the economic crisis and/or the historic candidacy of Barack Obama, but the unemployment rate at the time of the election was 6.8%, a figure lower than the subsequent two elections. In any event, the results were similar using both approaches.

  5. Bailey (2017, pp. 253–258) provided an explanation for the use of fixed effects as a way of dealing with endogeneity.


Download references

Author information

Authors and Affiliations


Corresponding author

Correspondence to E. Frank Stephenson.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

The original version of this article was revised: The name of the third author has been corrected.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Heller, L.R., Miller, J. & Stephenson, E.F. Voter ID Laws and Voter Turnout. Atl Econ J 47, 147–157 (2019).

Download citation

  • Published:

  • Issue Date:

  • DOI:


  • Voter identification
  • Turnout
  • Voter fraud
  • Disenfranchisement


  • K16
  • D72
  • H8