28.02.2015 Views

CORRUPTION Syndromes of Corruption

CORRUPTION Syndromes of Corruption

CORRUPTION Syndromes of Corruption

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

From analysis to reform 193<br />

cannot yet track trends or assess reliability. While methodology is improving<br />

and underlying data bases are expanding, such indicators remain<br />

suggestive at best. The problems are particularly difficult with respect<br />

to changes in our cases: Mexico, I argue, has moved from Elite Cartel to<br />

Oligarch and Clan corruption, for example, while Italy, Japan, Korea, the<br />

Philippines, and Indonesia have undergone political changes <strong>of</strong> varying<br />

types and magnitudes, and in some cases it is not clear whether a given<br />

syndrome should be discussed in present or past tense. Other cases, such<br />

as Russia, might be said to be in continuing crisis, and generalizations<br />

about its politics, economy, and corruption show their age very quickly.<br />

Yet the data at hand cover only limited time spans and are gathered at<br />

intervals ranging from one to five years. I cannot yet say how the clusters<br />

have evolved, or what time lags might occur between crises and qualitative<br />

changes in corruption in specific cases. Better answers must await the<br />

publication <strong>of</strong> longer data series, the observation <strong>of</strong> more systemic<br />

changes, and continued refinement <strong>of</strong> this analysis.<br />

In this book I have tried to deal with data problems in two ways. One is to<br />

emphasize construct validity – that is, the extent to which various measures<br />

are related to each other in theoretically expected ways, and in particular to<br />

data in which we have high levels <strong>of</strong> confidence (Adcock and Collier, 2001;<br />

Babbie, 2001: 143–144). In chapter 3 the four groups are defined using a<br />

cluster analysis <strong>of</strong> six indicators and then compared in terms <strong>of</strong> a range <strong>of</strong><br />

other statistical data (Tables A– C in the Appendix). Then, in chapters 4<br />

through 7 they are compared again using qualitative accounts <strong>of</strong> corruption<br />

issues. The results are generally consistent with the basic typology and<br />

hypotheses. Second, the statistics are used to frame questions rather than<br />

answers: the cluster analysis is primarily a way to define groups for study<br />

while reducing the likelihood <strong>of</strong> pre-selecting for expected results. Those<br />

cases do suggest that an understanding <strong>of</strong> trends and contrasts in participation<br />

and institutions helps identify contrasting syndromes <strong>of</strong> corruption.<br />

The challenge now is to improve the statistical evidence, to develop models<br />

giving a clearer indication <strong>of</strong> the relationships and priority among variables,<br />

and to continue to develop richer case studies, particularly in less frequently<br />

analyzed countries. Such research might yield more or fewer<br />

categories, and will certainly change our understanding <strong>of</strong> the corruption<br />

found within each, but it would fuel the comparative debate this book is<br />

intended to encourage.<br />

Relationships among the groups<br />

A final question is both methodological and substantive. We have arrived<br />

at categories that, broadly speaking, illustrate important contrasts both

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!