東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Can the STAR or the EGARCH-M model o...

Chen, Yung-Cheng.

FindBook

Google Book

Amazon

博客來

Can the STAR or the EGARCH-M model outperform the random walk model for short-run exchange rate forecasts? The case of Taiwan and Japan.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Can the STAR or the EGARCH-M model outperform the random walk model for short-run exchange rate forecasts? The case of Taiwan and Japan./
作者:	Chen, Yung-Cheng.
面頁冊數:	136 p.
附註:	Source: Dissertation Abstracts International, Volume: 60-11, Section: A, page: 4100.
Contained By:	Dissertation Abstracts International60-11A.
標題:	Economics. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=9949960
ISBN:	9780599517882

Can the STAR or the EGARCH-M model outperform the random walk model for short-run exchange rate forecasts? The case of Taiwan and Japan.
Chen, Yung-Cheng.

Can the STAR or the EGARCH-M model outperform the random walk model for short-run exchange rate forecasts? The case of Taiwan and Japan. - 136 p.

Source: Dissertation Abstracts International, Volume: 60-11, Section: A, page: 4100.

Thesis (Ph.D.)--The University of Memphis, 1999.

This item must not be sold to any third party vendors.

This dissertation focuses on finding a superior exchange rate forecasting model, which can outperform, in terms of the mean square error, the random walk at short-run horizons. The reason is twofold. First, short-run forecasting is very useful since most market participants have to make decisions frequently. An accurate forecast can reduce operational costs and even make "excess returns." Hence, this study focuses on monthly, weekly, and daily horizons. Second, short-run forecasting is very challenging since many structural models attempting to out-predict the random walk at short- and medium-run have failed, according to Frankel and Rose (1995). The hypothesis that exchange rates follow a random walk has very important implications. Pollock (1990) suggests that, if exchange rates follow a random walk, then the Efficient Market Hypothesis (EMH) implies that market participants couldn't systematically beat the market. This questions the use of forecasting models based on econometrics, time series or technical analysis. Therefore, the random walk model will be used as the benchmark for performance.

ISBN: 9780599517882Subjects--Topical Terms:

517137
Economics.

Can the STAR or the EGARCH-M model outperform the random walk model for short-run exchange rate forecasts? The case of Taiwan and Japan.
LDR:04436nmm a2200373 4500 001 2060166
005 20150917133432.5
008 170521s1999 ||||||||||||||||| ||eng d
020 $a 9780599517882
035 $a (MiAaPQ)AAI9949960
035 $a AAI9949960
040 $a MiAaPQ $c MiAaPQ
100 1 $a Chen, Yung-Cheng. $3 1929055
245 1 0 $a Can the STAR or the EGARCH-M model outperform the random walk model for short-run exchange rate forecasts? The case of Taiwan and Japan.
300 $a 136 p.
500 $a Source: Dissertation Abstracts International, Volume: 60-11, Section: A, page: 4100.
500 $a Major Professor: David M. Kemme.
502 $a Thesis (Ph.D.)--The University of Memphis, 1999.
506 $a This item must not be sold to any third party vendors.
506 $a This item must not be added to any third party search indexes.
520 $a This dissertation focuses on finding a superior exchange rate forecasting model, which can outperform, in terms of the mean square error, the random walk at short-run horizons. The reason is twofold. First, short-run forecasting is very useful since most market participants have to make decisions frequently. An accurate forecast can reduce operational costs and even make "excess returns." Hence, this study focuses on monthly, weekly, and daily horizons. Second, short-run forecasting is very challenging since many structural models attempting to out-predict the random walk at short- and medium-run have failed, according to Frankel and Rose (1995). The hypothesis that exchange rates follow a random walk has very important implications. Pollock (1990) suggests that, if exchange rates follow a random walk, then the Efficient Market Hypothesis (EMH) implies that market participants couldn't systematically beat the market. This questions the use of forecasting models based on econometrics, time series or technical analysis. Therefore, the random walk model will be used as the benchmark for performance.
520 $a This study adopts the Smooth Transition Autoregressive (STAR) and Exponential Generalized Autoregressive Conditional Heteroscedastic in Mean (EGARCH-M) model, which incorporate non-linearity in the model. In particular, the STAR model allows the exchange rate movements between troughs and peaks to be smooth rather than discrete, as in Terasvirta (1990). On the other hand, the EGARCH-M model allows fatter-tails (leptokurtic), instead of normal, distribution of the data and permits the variance of the regression to change over time. In addition, this model, which can incorporate exogenous variables, such as interest rate differentials, further improve the forecasting accuracy for exchange rates.
520 $a To increase the forecasting accuracy, we update the model to include the information from all previous periods. After estimating the random walk, the STAR model, and the EGARCH-M model using Taiwan's and Japan's daily, weekly, and monthly exchange rate changes in natural logarithm form, this study finds:
520 $a First, the normality test indicates that all time series under investigation do not follow the normal distribution (at about 0.01% level), except for Japan's monthly data.
520 $a Second, the in-sample mean square error of the EGARCH-M is lower than that of the OLS (Ordinary Least Square) estimates, except for Japan's monthly and Taiwan's weekly data.
520 $a Third, economic fundamentals can be used to improve the forecasting accuracy for both Taiwan's and Japan's (monthly) cases. For example, money, industrial output (as a proxy for income), and interest rate differentials are all highly significant (at about 0.01% level) for Taiwan. Although, they are not statistically significant for Japan, these variables still improve the out-of-sample forecasting accuracy.
520 $a Finally, the STAR model has the lowest out-of-sample mean square error in every case under investigation. The EGARCH-M model has the second lowest also in every case. In other words, both the STAR and the EGARCH-M model have outperformed the random walk. This gives support to Hsieh's (1988a) view that the logarithmic growth rates do not follow a random walk.
590 $a School code: 1194.
650 4 $a Economics. $3 517137
650 4 $a Commerce-Business. $3 3168423
690 $a 0501
690 $a 0505
710 2 $a The University of Memphis. $3 1025952
773 0 $t Dissertation Abstracts International $g 60-11A.
790 $a 1194
791 $a Ph.D.
792 $a 1999
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=9949960