Trading Rules Profitability in the Emerging FX Market: Danger of Data Snooping∗

We study the profitability of technical trading rules (hereafter "TTR") for the emerging foreign exchange markets, which are usually perceived to be less efficient. We find little sup-portive evidence for TTR once we control for data snooping bias, as opposed to conventional wisdom that technical analysis is more likely to be useful in less efficient market. Data snoop-ing bias is intrinsic in searching for profitable trading rules by researchers and practitioners. We quantify the danger of data snooping bias using formal test with data snooping check. If we ignore the data snooping bias, we do find thousands of profitable rules for every considered emerging currency exchange rates. The profitability disappears in most cases, however, when data snooping bias is taken into account. We use an universe of 25,988 trading strategies for 10 emerging market currencies, consisting of simple rules, charting rules with kernel smoothing and complex trading rules. Most of these strategies are popular in the trading practice, but have not yet been studied for emerging market currencies in the literature. We also find that in the charting rules with kernel smoothing, those based on the under-smoothed spot series perform better than over-smoothed and original spot series, confirming that under-smoothing mimics better the eyeball smoothing of professional traders