استفاده از آنالیز جریان در ارز خارجی: شواهد اکتشافی

This paper provides questionnaire evidence on the role of flow analysis for professional traders and fund managers. This evidence suggests that besides fundamental information and technical analysis, the analysis of flows provides an independent third type of information for professionals. The view that flows can be used to learn about fundamentals is not consistent with the data. Instead, evidence indicates that flows more likely provide insight into semi-fundamental private information.

“A source of informational advantage to the traders is their access to, and trained interpretation of, the information contained in the order flow” (Goodhart, 1988: p. 456). Although this has been common market wisdom for a long time, there have been no studies that systematically examine the use of flow analysis by foreign exchange professionals. This paper provides evidence that next to fundamental and technical analysis the analysis of flows is an independent third type of information. We thus add to the recent literature by presenting empirical evidence that the analysis of flows does affect the behavior of a significant group of professional FX market participants according to their own perspective. Thus, our contribution complements other work about the informational role of flows. For example, Lyons, 1995 and Lyons, 1998 presents case study evidence on the importance of flows, Osler (1998) relies on flows to characterize exchange rate changes, Ito et al., 1998 and Covrig and Melvin, 2001 provide statistical evidence for information inherent in flows and Cai et al., 2001 and Evans, 2002 and Evans and Lyons (2002) improve exchange rate explanation by incorporating order flows. According to (Lyons (2001): p. 4), “order flow is transaction volume that is signed” (i.e. indicating purchases or sales). From an ex ante logical point of view, flow analysis may share similarities with either technical analysis or fundamentalism. This leads to three different views about flow analysis which to some extent compete with each other, and which we list in the order of their affinity to the efficient market hypothesis (EMH): Position 1. Flow analysis is an expression of limited rational behavior. In this sense flow analysis is an analogue of technical analysis (see e.g. Shleifer and Summers, 1990). This view relies heavily on the assumption of informational efficiency of markets, according to which any attempt to acquire extra information is futile or even irrational, when costly resources are invested. Position 2. Flow analysis is a manifestation of rational learning about fundamentals. This view regards flow analysis as a certain form of fundamental analysis. Conceding time constraints and informational heterogeneity, flow analysis can be viewed as a rational way of trying to detect the results of other participants’ fundamental analysis and thus parallels optimal learning from order flow such as in the seminal work of Kyle (1985). Position 3. Flow analysis provides interim information about short-run price movements but little information about fundamentals. This view is based on the assumption that the order flow can influence price paths of transactions prices in the short run. Flow analysis is understood as a separate kind of analysis if it aims at forecasting transactions prices from presently executed and planned order flows (see e.g. Ito et al., 1998 and Covrig and Melvin, 2001). According to this view, flows also contain information about short-term trading objectives or liquidity considerations of other traders that may affect short-term price movements, but that will not affect medium-term asset prices. Such information is usefully termed semi-fundamental information. Unfortunately, at present there is virtually no direct systematic information available about the importance and nature of flow analysis. Because of this lack of knowledge, it seems worthwhile to improve our understanding by conducting a questionnaire survey study among market participants. This study is organized around three questions: is flow analysis an important forecasting tool in real world markets?1 Can flow users be related to certain institutional characteristics? And finally, are there beliefs of flow users about FX markets which would reveal their motivation for applying flow analysis? This study provides several insights: it demonstrates that in addition to fundamental and technical analysis, flow analysis is indeed a major and independent third tool for FX professionals. Furthermore, the use of flow analysis is systematically related to some institutional characteristics and beliefs about the functioning of FX markets. Our results provide varying degrees of support for the three “positions” under review: the most interesting seems to be the affinity with Position 3, that is the view that flow analysis aims at exploiting semi-fundamental information. The survey approach chosen was developed as a standard methodology to establish market participants’ behavior in financial markets. The pioneering work by Shiller (1989) was first applied to foreign exchange markets by Allen and Taylor (1990). The latter thoroughly examined the use of technical analysis in the London FX market which is the reference case for our work (Taylor and Allen, 1992). The same approach was reproduced for Hong Kong based FX dealers by Lui and Mole (1998). Related studies on foreign exchange markets include Menkhoff, 1997, Menkhoff, 1998, Cheung and Wong, 2000 and Cheung and Chinn, 2001 and Cheung et al. (1999). However, to the best of our knowledge, there is no survey study explicitly examining the use of flow analysis in foreign exchange markets. Our investigation of flow analysis consists of seven steps: Section 2 reviews the literature on possible explanations for the use of flow analysis. Section 3 provides a short description of the data used. In Section 4 the evidence on the importance of flow analysis in the FX market in relation to the two established forms of fundamentalism and chartism is presented. This is followed in Section 5 by an analysis of the question whether flow users can be viewed as a distinct, coherent group of more or less rational people. Section 6 relates flow users to certain characteristics of behavior and Section 7 relates them to beliefs about FX market characteristics. We present our conclusions in Section 8. An Appendix A describes the survey data in more detail.

Flow analysis in foreign exchange markets has not been a subject of systematic examination so far. In this respect it shares the fate of technical analysis, which was also quite neglected until a few years ago when Allen and Taylor (1990) conducted their survey. Compared to the wide area of anecdotal and accidental information, our questionnaire establishes better substantiated knowledge in two fields, i.e. regarding on the one hand the importance of flow analysis and on the other, the appropriate understanding of the nature of flow analysis as reflected by three competing positions. The importance of flows may be highlighted by two facts: • First, it becomes obvious that there is a third form of analysis in the market besides fundamental and technical analysis. In the group of respondents, about every second FX dealer and every third fund manager allocated 25% or more of information used to flow analysis. • Second, the relationship with the other two forms of market analysis shows that flow analysis is neither closely related to a preference for “fundamentalism” or “chartism”, nor is it a substitute for either of them: it rather represents an independent third form of analysis relevant for professionals. The survey results have also shed some light on our understanding of the role of flow analysis in foreign exchange markets. Several hypotheses have been tested revealing evidence on the explanatory power of three competing views. The results are compiled in summarized form in Table 5. They provide a clear picture of the explanatory power of three competing positions: • Flow analysis does not seem to be basically used as a tool to learn about the fundamental information of others, as claimed in Position 2. • Moreover, the use of flow analysis does not appear to be clearly related to indicators of less rational behavior, thus slightly opposing the view of efficient markets, as stated in Position 1. • However, the evidence seems to accord best with Position 3. This is the view that flow analysis aims at exploiting semi-fundamental private information. Table 6. Coverage and response rates Institutions Addressed Addressed Response Response rate Inst. Quest. Inst. Quest. Inst. (%) Quest. (%) Banks Directly 35 112 28 72 80.0 64.3 Banks via Association 15 99 14 73 93.3 73.7 Fund manag. companies Directly 57 180 29 58 50.9 32.2 Total All 107 391 71 203 66.4 51.9 (Subtotal) Banks only 50 211 42 145 84.0 68.7 (Subtotal) Directly only 92 292 57 130 62.0 44.5 Note: 7 banks which were directly addressed and received 22 questionnaires are located in Austria. Table options Table 7. Items with significant change over time Questions or relations See Table Response in 1992 Response in 2001 Impact on positions Importance of flow analysis 1 17.9% 23.5% Pos. 1: ↓ Flow use and higher degree of education 2 Negative correlation for all, dealers and fund managers Positive correlation for fund managers Pos. 1: ↓ Pos. 3: ↑ Flow use and higher age 2 – Negative correlation for dealers Pos. 1: ↑ Flow use and superior position 2 Positive correlation for dealers – (Pos. 3: ↓)a Flow use and institution size 2 − Positive correlation for fund managers Pos. 2: ↓ Pos. 3: ↑ Flow use and importance of psychology 4 Positive correlation for all – Pos. 1: ↓ Flow use and longer time of information processing 4 – Positive correlation for fund managers Pos. 2: ↑ Pos. 3: ↑ Note: “–” indicates no significant relation. a The disconnection of flow use and superior position may be influenced by the younger dealers relying stronger on the more modern technique of “flow analysis” (see Section 5). Table options Table 5. Evidence from hypothesis tests regarding Position 1, Position 2 and Position 3 No. Hypotheses Position 1 Position 2 Position 3 description Flow analysis as sign of non-optimal behavior Flow analysis as instrument to learn about fundamental news Flow analysis as instrument to exploit semifundamental private information 1 Absolute importance − 2 Relative importance − 3 Subgroup importance + 4 Correlation with other information types − − + 5 Relation to education − + 6 Relation to age + 7 Relation to position 8 Relation to proprietary dealers 9 Relation to share of customer business 10 Relation to company size − + 11 Relation to forecasting horizon + (−) + 12 Relation to sources of information − − + 13 Belief in psychological factors 14 Belief in longer time for information processing + + 15 Belief in importance of large market participants + “+” indicates supporting and “−” refusing evidence from the test.