Naslov
CAPTURING TRADING ACTIVITY DYNAMICS BY MULTI- SEASONAL PATTERNS

Autori
Arnerić, Josip

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, neobjavljeni rad, znanstveni

Skup
The 6th Dubrovnik International Economic Meeting (DIEM 2023)

Mjesto i datum
Dubrovnik, Hrvatska, 29.06.2023. - 01.07.2023

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
intraday trading ; multi-seasonality ; locally estimated smoothing

Sažetak
A great effort have been made recently in discovering multi-seasonal patterns of trading activity at stock exchanges with objective to improve the performance of trading algorithms and consequently to reduce transaction costs. In addition to hourly seasonality, daily, weekly, quarterly or even monthly seasonality can be found. Extraction of multiple seasonal components is not straightforward, i.e. many issues emerge when dealing with intraday trading data such as missing values, non-trading days, calendar variations, seasonality with different periods per cycle and outliers. Therefore, this study considers estimating the shortest seasonality first and proceeds to estimate the next seasonality from the irregular component by utilizing Loess method (locally estimated smoothing), which is repeated successively until there is no seasonality left. Application to DAX intraday trading observations, empirically demonstrates that hourly seasonality is the strongest and most persistent, indicating “U” shaped pattern. The daily seasonality exhibits moderate strength and almost linearly increasing trade pattern when approaching the end of the week as investors did not want to hold open positions over a weekend. For the same reason, trading activity was mostly intensified during the closing hours of the day. Monthly seasonality is also apparent with downturns in May, August, and December after increased trading in the previous months. Multi-seasonality patterns provide better insight into investors’ behavior, and consequently, improve trading strategies.

Izvorni jezik
Engleski

Znanstvena područja
Ekonomija

POVEZANOST RADA