Scientists from the Higher School of Economics (HSE) and University of Valladolid have developed a neural network prediction model of corruption based on economic and political factors. The results of the research were published in Social Indicators Research.
Researchers contend that corruption must be detected as soon as possible in order to take corrective and preventive measures. Because public resources for combating corruption are limited, efforts should focus on areas most likely to be involved in corruption cases. They use a unique database that brings together the main cases of political corruption in Spain. Then they propose an early warning corruption model to predict whether corruption cases are likely to emerge in Spanish regions given certain macroeconomic and political determinants.
The model provides different profiles of corruption risk depending on the economic conditions of a region conditional on the timing of the prediction.
Scientists of HSE and University of Valladolid have used self-organizing maps (SOMs), a neural network approach, to predict corruption cases in different time horizons. SOMs are a kind of artificial neural network that aim to mimic brain functions. SOMs have the ability to extract patterns from large data sets without an explicit understanding of the underlying relationships. They convert nonlinear relations among high dimensional data into simple geometric connections. These properties have made SOMs a useful tool to detect patterns and obtain visual representations of large amounts of data. Consequently, predicting corruption is a field in which SOMs can become a powerful tool.
The results show that economic factors prove to be relevant predictors of corruption. Researchers find that the taxation of real estate, economic growth, increased house prices, and the growing number of deposit institutions and non-financial firms may induce public corruption. They also find that the same ruling party remaining in power too long is positively related to public corruption.
Depending on the characteristics of each region, the probability of corrupt cases emerging over a period of up three years can be estimated. Then the different patterns of corruption antecedents were detected. Whereas in some cases, corruption cases can be predicted well before they occur and thus allow preventive measures to be implemented, in other cases the prediction period is much shorter and urgent corrective political measures are required. The method consists of a sophisticated algorithm with multiple non-linear relations according to which the determinants of the propensity to corruption change throughout the time.
“Our research develops a novel approach with three differential characteristics. First, unlike previous research, which is mainly based on the perception of corruption, we use data on actual cases of corruption,” said one of the authors of the research Félix J. López-Iturriaga, Leading Research Fellow at the International Laboratory of Intangible-driven Economy of HSE. “Second, we use the neural network approach, a particularly suitable method since it does not make assumptions about data distribution. Neural networks are quite powerful and flexible modeling devices that do not make restrictive assumptions on the data-generating process or the statistical laws concerning the relevant variables. Third, we report the probability of corruption cases on different time scenarios, so that anti-corruption measures can be tailored depending on the immediacy of such corrupt practices. Our model allows patterns of corruption to be identified on different time horizons.”
Since corruption remains a widespread global concern, a key issue in the research is the generalizability of the model and the proposed actions. Scientists have used fairly common macroeconomic and political variables that are widely available from public sources in many countries. In turn, the model can be applied to other regions and countries as well. Of course, it could be improved if country or region-specific factors were taken into account.
The approach in the research is interesting both for academia and public authorities. For academia, scientists provide an innovative way to predict public corruption using neural networks. These methods have often been used to predict corporate financial distress and other economic events, but no studies have yet attempted to use neural networks to predict public corruption. Consequently, the researchers extend the domain of neural network application.
For public authorities, they provide a model that improves the efficiency of the measures aimed at fighting corruption. Because the resources available to combat corruption are limited, authorities can use the early corruption warning system, which categorizes each province according to its corruption profile, in order to narrow their focus and better implement preventive and corrective policies. In addition, this model predicts corruption cases long before they are discovered, which enhances anticipatory measures. The model can be especially relevant in countries suffering the severest corruption problems. In fact, European Union authorities are highly concerned about widespread corruption in certain countries and can use this approach to prevent corruption.