@Dao’s comment got me thinking about how we might investigate the effect of a share going ex-div. We can do this by looking at two ETFs which track the same index, but have different dividend dates. Looking at the ratio of prices will factor out any effects that are due to sentiment.
Below is a table of closing prices for VUKE and ISF, two FTSE 100 trackers. VUKE went ex-div on 26 September, 2019. ISF had surrounding ex-div dates of 12 September and 12 December.
Note how the share price ratio of 100*VUKE/ISF drops to 4.47 on 26-Sep when VUKE goes ex-div. The VUKE share price drops to 32.39, but rises back to 32.72 on 27 September, more than it was on the 25th, but it is still depressed when compared to the ISF price.
I suppose it would be possible, by switching back and forth between VUKE and ISF, to collect no dividends. Instead, one would accumulate greater capital gains. That could be beneficial when capital gains tax rate is 20% and dividend tax rate is 32.5% (for UK higher rate taxpayer). I wonder if the trading costs due to bid/offer spread would wipe out any savings on tax. Roughly speaking if the yield is 4.8%, then on £10000 of shares we save (.325-.20)*480 = £60 in tax, but spend 0.05%, 8 times a year in switching costs, amounting to £40. So it appears only marginally worthwhile.
However, the savings can be greater in the case of an accumulating ETF such as SWDA. If you hold this share on the distribution date of 31 December you are liable to dividend tax on the nondistributed, but reportable, income for the year (eg $1.0685 per share in 2019). By selling just before 31 December you avoid this tax liability. Prices of accumulating ETFs tend not to drop at their distribution dates because dividends are continuously being reinvested, not just paid out on 4 specific dates each year.