An index of political corruption

An index of political corruption

The government doesn't publish statistics on the "volume" of political corruption ("...the loss to the public from the subversion of formal, or implicit, rules of government behavior."). Trends in its level have to be inferred from trends in related things.

Edward Glaeser and Claudia Goldin ("Corruption and Reform: An Introduction." NBER Working Paper 10775. September 2004) have created indices of newspaper reporting on corruption, and, on the theory that where there's smoke there's fire, use them as a general indices of changes in levels of US political corruption from the early 19th Century to about 1970. (The authors are aware of the potential shortcomings of the index and mention them in the article.)

They searched out instances of the use of variants of the words "corruption" and "fraud" in an electronic data base of the New York Times going back to 1851, and a data base of small town papers going back to 1820, and counted the number of articles or pages in which the words appeared. Key word counts were divided by series of word counts for "political," variants of politic, and "January." The division was done to "deflate" the series to account for changes in the size of the newspapers ("January"), or for changes in attention given to political stories ("political" or "politic").

Three series are reported, all moving fairly consistently. These show:

"The first great boom in corruption reporting occurs around the 1840 election. Stories of corruption during this period focused on Tammany Hall and also Martin Van Buren, the first President who owed his success to a political machine (for a description of Van Buren�s activities in the banking sector, see the Bodenhorn essay). The next peak in corruption reporting occurred between 1857 and 1861 and focused on voting irregularities in Kansas. There is a global peak in the 1870s during the Grant administration. Top stories concerned Cr�dit Mobilier, the Whiskey Ring, and Reconstruction and the Ku Klux Klan. Finally, there is a small local peak in the late 1920s during the era of prohibition and the Teapot Dome scandal..." p 13

In general there appears to be an increase in corruption in the early 19th Century, high levels between about 1840 and 1870, declining levels from 1870 through the 1940s and relatively low levels thereafter. The authors note that the evidence for an increase going into 1840 is not much more than suggestive.

The authors explain the changes in levels of corruption by looking at the changing patterns of the benefits and costs of corrupt activity. In the early 19th Century, the increasing size and scope of government reduced the cost of corrupt actions:

"...these factors suggest that the benefits of corruption will rise with the size and discretion of the government and the amount of social and economic regulation. Benefits from corruption will also rise when the size of assets or damages involved in property rights disputes increases (as in Glaeser and Shleifer 2003). As we will discuss later, the late nineteenth century was a period of increasingly larger governments, more valuable public assets, more aggressive regulation (as discussed by Novak�s essay) and bigger stakes litigation. The potential benefits from corruption rose during the period along almost every dimension." p8

Corruption didn't decrease because of decreases in the size of government. But changing laws and mores changed the costs of corrupt activity:

"But the decline in corruption between the mid-1870s and 1920 was not associated with declining returns to corruption. The size of the government continued to rise and the returns from corruption in the judiciary increased as well. The big change over the twentieth century has been in the costs facing corrupt politicians. In 1900, many actions we would now prosecute were legal. Governments rarely prosecuted themselves, and the higher levels of government were sufficiently weak that they could not provide a check on local corruption. Newspapers had long provided exposure of corrupt practices, but in many smaller cities the news media was sufficiently tied to the political establishment that it was unlikely to trumpet information unfavorable to that establishment." p19

"By the early twentieth century, the full apparatus of modern checks on corruption were in place. Rules had generally replaced discretion in many areas such as patronage. Different levels of government more effectively patrolled each other. Greater competition and political independence in the news media meant that corrupt activities and charges of corruption were more likely to be reported everywhere in America, not just in the big cities. Finally, voter expectations about corrupt behavior had changed and revealed corruption was more likely to lead to political defeat." p 19

This article is going to be an introduction to a collection of papers on U.S. corruption to be published by National Bureau of Economic Research (NBER) under the title, Corruption and Reform: Lessons From Americas History. The papers were presented at a conference this past July. Glaeser and Goldin's paper is available from the NBER web site: "Corruption and Reform: An Introduction".