Losing sleep at the market: an empirical note on the daylight saving anomaly in Australia

Economic Papers (Economic Society of Australia), Dec, 2003 by Andrew C. Worthington

The purpose of the present paper is to add to this small but intriguing body of work the results of an analysis of the Australian equity market. To the author's knowledge this is the first of its kind in Australia, and adds significantly to the nascent literature concerning the economic benefits and costs of daylight saving. The paper itself is divided into four main areas. Section 2 discusses the starting and ending dates of DST in Australia since 1979/80. Section 3 explains the empirical methodology and data collection employed in the analysis. Section 4 presents the results. The paper ends with a brief conclusion.

2 Daylight Saving in Australia

Two dates are relevant for the implementation of DST in Australia: (i) a Sunday a.m. starting date, usually in late October, and (ii) a Sunday a.m. ending date, normally in late March in the following year. However, one complication that does exist in Australia is that not all states and territories currently use DST, and of those that do the timing and adherence to the usual starting and ending dates has varied over the sample period. As shown in Table I, of Australia's six states and two territories only New South Wales (NSW), Victoria (VIC), South Australia (SA), Tasmania (TAS) and the Australian Capital Territory (ACT) have implemented DST annually since 1979/80, while Queensland (QLD) and Western Australia (WA) adopted DST spasmodically, from 1989/90 to 1991/92 and in 1983/84 and 1991/92, respectively. The Northern Territory (not shown) has never used DST. In order to reflect these differences in the current analysis, two sets of starting dates and three sets of ending dates are identified.

The first set of starting dates is where the most populous states of NSW and VIC both shift to DST at the same time. These account for 24 cases in the 24 years in the sample and on most occasions, include many of the remaining states and territories. The second set is when a state or territory other than NSW or VIC commences DST on some other date. This represents 11 cases in 24 years. The other set of dates is the DST ending dates and these comprise three categories: (i) occasions when NSW and VIC shifted from DST at the same times (19 cases), (ii) dates when NSW and VIC moved out of DST at different times (4 cases in each respect) and (iii) dates when a state or territory other than NSW and VIC shifted out of DST at a different time to NSW and VIC (8 cases). As before, at least some other states and territories may also have moved out of DST at these times. Accordingly, in the twenty-four year sample there have been thirty-five weekends in Australia where DST started at 2:00 a.m. on Sunday and thirty-five weekends where DST ended on Sunday at 3:00 a.m.

3 Research Method and Data

To test for the economic effect of daylight saving time change we look at the first trading day following the transition to or from DST. The data employed in the study are market returns drawn from the Australian Stock Exchange (ASX) over the period 2 January 1980 to 5 May 2003. Both the All Ordinaries Price and Accumulation (including dividends and capitalisation changes) Indices are used. Following Kamstra et al. (2000, 2002) and Pinegar (2002) a series of daily raw returns are calculated of which three categories are identified. These are: (i) a weekend return calculated from the Friday and Monday closing prices for weekends that do not include a DST starting or ending Sunday, (ii) a DST starting and ending weekend return calculated using the Friday and Monday closing prices for weekends including a DST starting or ending Sunday, and (iii) other days being those other than DST and non-DST weekends. The DST starting and ending weekends are further categorised according to the different sets of starting and ending dates discussed above.