MEASURING FOREIGN EXCHANGE PRESSURE: A TEXT MINNING APPROACH

ABSTRACT : This study investigates the effectiveness of sentiment analysis, using text-mining approach within the context of big-data analytics, in measuring foreign exchange pressure in Nigeria. It begins with the construction of two sentiment-based index of foreign exchange pressure; the first, labelled EMP, constructed by text-mining public sentiments about foreign exchange management in Nigeria, within the platform of twitter, while the second, labelled EMP_Trend, constructed from Google Trend as an index of sampled search of related words around foreign management in Nigeria. Thereafter, the study tested the effectiveness of both indices in signaling movement in
exchange rate (IEW) in Nigeria relative to existing traditional measures of foreign exchange pressure in Nigeria. The Predictive Regression Model (PRM) and Clark and West (2007) frameworks were employed. Findings from the study suggest that foreign exchange market pressure index using Sentiment Analysis may hold sufficient information in predicting and signaling movement in exchange rate (IEW) in Nigeria. Specifically, EMP_Trend and EMP were found to improve the forecast of IEW, as their estimated Clark and West coefficients were both positive and statistically significant at 5 per cent. The study recommends that monetary authorities leverage sentiment analysis to monitor future direction in exchange rate, with a view to implementing policies that would moderate the prevailing instability in the foreign exchange market in the Nigeria.
KEYWORDS: Foreign Exchange Pressure, Sentiment Analysis, Text-mining, leading indicator, predictive regression model.