دانلود رایگان مقاله لاتین اطلاعات اندازه تجارت در بازارهای ارز خارجی از سایت الزویر
عنوان فارسی مقاله:
اطلاعات دهندگی اندازه تجارت در بازارهای ارز خارجی
عنوان انگلیسی مقاله:
Informativeness of trade size in foreign exchange markets
سال انتشار : 2017
برای دانلود رایگان مقاله اطلاعات اندازه تجارت در بازارهای ارز خارجی اینجا کلیک نمایید.
بخشی از مقاله انگلیسی:
Empirical results
To present the empirical results, we proceed as follows. First, we compare the estimatedω and φ over 145 days in our sample. Ifω > φ, then trade size conveys additional information to market participants. If, however, ω < φ, then one can conclude that traders (either informed or uninformed) do not significantly benefit from trade size. Following that, our second objective is to examine how changes in the cutoff trade size impact the estimates. Third, we investigate whether trade size responds to price volatility. Finally, we link trade size to the empirical properties of exchange rate data and particularly assess the implications for excess kurtosis. 4.1. Testing for the impact of trade size The procedure of testing for trade size effects involves comparing estimates of the restricted (ω = φ) and unrestricted (ω ̸= φ) models. The cutoff amount that differentiates large from small trades is initially set to 5000, and it is subsequently shown that this does not affect the main results.9 Table 1 lists the average estimates of αi, δi, εi, µi, ωi , and φi (i = 1, . . . , 145). The paired t-test of the equality of the means of the constrained and unconstrained models shows no significant difference for the first four parameters.10However, the difference between the two sets of estimates of ωi is statistically significant.11 Furthermore, including trade size effects (unconstrained model) significantly increases the absolute value of the log likelihood function, thus indicating that the constraint is binding. The informativeness of trade size is also confirmed by the unpaired t-test of the equality of ω¯ and φ¯ for the unconstrained model. The model concludes that ω¯ is significantly greater than φ¯. On approximately 68% of the days in the sample, ωi > φi , and the 7 By ‘‘active’’, the paper refers to traders that did not simply receive interest on their positions but placed orders during this period. The market share of these traders is approximately 86.4%. 8 This is one of the largest and most detailed tick-by-tick FX datasets ever to be used in an academic study. Please see (Gençay et al., 2015) for more information about the data. 9 Trade size is expressed in currency units of the base currency, i.e., the Euro. 10 The null hypothesis for this test is that the mean difference (d¯) between the paired observations (constrained and unconstrained) of estimated parameters is zero. The test statistic is calculated as t = d¯ √s d¯ /145 , where sd¯ is the sample standard deviation for d¯. 11 Additionally, the standard errors of ωˆi and φˆ i for the unconstrained model are consistently on the order of 10−4 and 10−5 , respectively, thus indicating statistically significant differences in the probabilities. difference in the probabilities (ωi − φi) ranges from −0.12 to 0.14 (i = 1, . . . , 145). Although there are 47 days when the probability of uninformed large trading exceeds the probability of informed large trading, this is not the typical case. We now investigate whether the findings above are robust to the choice of the cutoff amount for a ‘‘large’’ trade. Table 2 reports the results for cutoff rates of 2000, 8000 and 12 000. The focus is on the difference column from Table 1 and the mean values of the unconstrained estimates. The results indicate that in the cutoff range between 3000 and 8000, all estimates are stable, and the informed large trade size is more informative than the uninformed large trade size. The choice of cutoff values above 8000 (e.g., 12 000 in Table 2) distorts the results due to the low frequency of such large trades. Similarly, it is unreasonable to consider trades above small cutoff values (e.g., 2000 in Table 2) to be ‘‘large’’, in which case the observed effects diminish. These findings confirm the work of Chakravarty (2001) and Anand and Chakravarty (2007), who find that medium-sized trades are the most informative. This finding can also be interpreted as a ‘‘separating equilibrium’’ outcome in which informed traders submit mainly large orders (Easley and O’Hara, 1987).12 An interesting observation emerges from Table 2: the probability of both informed and uninformed large trading declines with the cutoff value. This result can be explained by the fact that increasing the cutoff value eliminates the majority of the transactions that qualify as ‘‘large’’ trades.
برای دانلود رایگان مقاله اطلاعات اندازه تجارت در بازارهای ارز خارجی اینجا کلیک نمایید.
کلمات کلیدی:
International Business Information: How to Find It, how to Use it https://books.google.com/books?isbn=1573560502 Ruth A. Pagell, Michael Halperin - 1998 - Business & Economics See Standard International Trade Classification (SITC) Size of companies, 64-68 ... 48 annual reports' informativeness in, 105 franchising in, 222 South America. High Frequency Trading and Limit Order Book Dynamics https://books.google.com/books?isbn=1317570774 Ingmar Nolte, Mark Salmon, Chris Adcock - 2016 - Business & Economics ... 229; revelation 219, 226, 230, 231; seller-initiated trades (SITs) 225–6; size 218; ... pretrade 217, 220, 232–3n informativeness: trade see trade informativeness ... 197 interest rates: foreign exchange markets 261–2 International Monetary ... IMF Staff Papers: Volume 52 - Issue 3 - Page 385 - Google Books Result https://books.google.com/books?isbn=1451955006 International Monetary Fund. Research Dept. - 2005 - Business & Economics Volume 52 International Monetary Fund. Research ... Easley, D., and M. O'Hara, 1987, “Price, Trade Size, and Information in ... 1991b, “The Summary Informativeness of Stock Trades: An Econometric Analysis,” Review ofFinancial Studies, Vol.
