Volume or Order Flow? : Which contains more information in really traded Yen/Dollar Foreign Exchange Market?

Masayuki Susai (Nagasaki University)

Conclusion

Trading volume in financial asset is important data to analyse the volatility of the asset price. As for the foreign exchange market, it is difficult to obtain trading volume of each transaction. In this paper, we use transaction data. Traded foreign exchange rate, trading volume, time stamp and bid or ask information are included in our data set from ICAP, Data Mine ver.1.5. This research with real traded foreign exchange rate data, especially with real trading data might be one of the first paper in this research area.

From this tick by tick data, we construct 5 minutes and 10 minutes even spaced series. Two order flow data series are also derived from this data. One of them is the difference of the number of bid and ask in 5 and 10 minutes. This data shows which side of transaction is more in given interval. Another data is the difference of the volume traded in bid side and ask side. If this number is plus, then this number means net bid trading volume in given interval. In this case, downward pressure to foreign exchange rate occurs in this interval.

With real trading data, we explore the impact of volume and order flow on foreign exchange rate volatility. Famous hypothesis on the relations between trading volume and volatility is Mixture of Distribution Hypothesis. In line with this hypothesis, we establish the model to test whether trading volume and order flow of foreign exchange rate have significant impact on the volatility of foreign exchange rate.

The impact of trading volume on the volatility is significant in all analyses. From these results, we can confirm that trading volume is the proxy of information arrival. Especially, the impact of trading volume on GARCH effect became clear. Not so many papers found this fact in foreign exchange market to date. Information arrival may cause GARCH type volatility pattern in foreign exchange market.

Many papers are investigating the impact of order flow on the volatility of foreign exchange rate. In our research, we compare the impact of trading volume and order flow. From our main results, order flow may have some impact on the volatility. But this impact might be limited. Not on the GARCH effect but on the price volatility ((ADJ_DFX)^2), net trading volume between bid side and ask side in a given interval has clear impact.

As mentioned earlier, traded data of foreign exchange rate has not been used so far. But from now on, we can use this kind of data. There are so many puzzles in foreign exchange market. It is time to move on further to solve these puzzles with real foreign exchange trading data.

  1. See Danielsson and Payne, 2002.
  2. To our knowledge, this data has not been used as yet in any published research. Now we are going to extend the period of our data set.
  3. We estimate EGARCH model with volume in conditional variance equation. Almost same result is obtained as the result we estimate with GARCH model.
  4. See a series of papers done by Evans and Lyons in Reference of this paper. See also Frommel, Mende and Menkhoff (2007).
  5. Bauwends et al. also test the impact of the number of dealers in this equation. But we do not test this effect here because our objectives here is to explore the impact of information arrival.
  6. Same as the estimation in equation (6), diagnostic result of OLS estimation shows serious serial correlation in disturbance terms.
Search
Information

Empirical Study on Asian Financial Markets

Edited by Masayuki Susai Hiromasa Okada

本書は、2006年12月に長崎大学において開催された国際カンファレンス等で報告された東アジアの金融市場を対象とした金融および会計学における実証研究の成果をまとめたものである。