September 16, 2020

Covid-19: Modelled scenarios for Covid-19 in Victoria and New South Wales

Our goal is to share regular updates to modelled scenarios for the potential development of Covid-19 in Australia (specifically, Victoria and New South Wales).

  • Last updated: 16th September 2020
  • Next update: 23rd September 2020


Our data analytics teams operate a multi-state model to generate scenarios for the potential development of the Covid-19 pandemic across several countries in Asia Pacific, as well as South Africa.

Having modelled the first-wave developments in Australia, Matthew Tiong has continued to apply and adapt these models to the second wave of infections spreading in Victoria and New South Wales. While Matt has had the benefit of feedback from the Covid-19 working group of the Institute of Actuaries of Australia, the model outputs and opinions expressed here are his own.

Victoria: Modelled scenarios as at 16th September 2020

Quick transition to step 3 for regional Victoria, metro Melbourne on track to meet step 2 threshold by 28th September, but step 3 threshold on 26th October is emerging as a stretch target

On Sunday night, metropolitan Melbourne and regional Victoria entered steps 1 and 2 of the Andrews’ government’s 5-step plan respectively. Given the progression to subsequent steps of the plan for both regions requires a daily new case threshold to be met, our modelling provides an indication of the likely achievability of these targets.

Comparing the dark blue dotted line with the grey bars and the medium blue line illustrates that the state-wide transmission rate has remained stable (for the third week in a row now), and that the 7-day rolling average for newly reported cases has moved largely in line with what we expected last week (ie the 7-day average fell to 36 on 16th September versus our estimate last week for 16th September of 40).

Regional Victoria

Regional Victoria can enter step 3 when the 14-day average for new cases reduces below 5 provided no cases with an unknown source are reported over the same period. The good news is that regional Victoria achieved this threshold on Tuesday, triggering the transition to step 3 with a significant easing in restrictions (effective from 11:59pm on 16th September):

  • No restrictions on leaving home
  • Public gatherings up to 10 people outdoors
  • Visitors allowed at home from 1 other household (up to 5 people)
  • Hospitality open for predominantly outdoor seated service only
  • Outdoor contact and non-contact sport for kids and non-contact sport for adults to resume
  • Intrastate travel allowed across 3rd step area (ie regional Victoria until metro Melbourne also reaches step 3 on 26th October at the earliest)
  • Outdoor venues and events permitted, subject to approval
  • Accommodation and camping to reopen

Given that the progression to step 4 requires no new cases for 14 days state-wide (and no earlier than 23rd November), those living in regional Victoria should expect these rules to last for at least 2 months, and quite likely longer. As we asserted last week, if the recent experience in NSW is indicative, having no new cases reported state-wide over a 14-day period will be a difficult threshold to achieve, even with effective suppression and contact tracing.

Metro Melbourne

As with last week, metro Melbourne is highly likely to achieve the threshold required for step 2 as our model is now projecting a 14-day average of ~26 cases per day (state-wide) on 28th September.

The achievability of step 3 remains uncertain however, as similar to a week ago our model is projecting a 14-day average of ~6 cases per day on 26th October, which is slightly above the required threshold. In order to achieve a 14-day average of fewer than 5 cases per day by 26th October, our modelling suggests that the average transmission rate between now and 26th October would need to reduce by a further 10 – 15% compared to the current prevailing rate. This would seem unlikely should metro Melbourne move to step 2 on 28th September as planned, and given that regional Victoria is already transitioning to step 3 as mentioned earlier.

In saying this, a key learning from the regular recalibration of our model is how sensitive the predictions are to small changes in the transmission rate, which we can reasonably expect to fluctuate. It also remains true that the low number of daily new cases required to achieve an average of <5 per day is more heavily impacted by day-to-day volatility (rather than a stable average transmission rate). Achievement of the step 3 threshold by 26 October is genuinely uncertain, and arguably an odd choice of a threshold.

Given the experience we have seen in NSW and elsewhere, the risk is that transmission may increase again in VIC as a result of the anticipated easing of restrictions, and it is more than fair to question whether targeting what is effectively zero for step 3 (and actually zero for steps 4 and 5) is the right strategy.

The appropriateness of a chosen recovery plan is logically a function of the underlying level of infection and explains (at least in part) the variant approaches between countries. Very low levels of underlying infection in Australia (ie <1%) make an eradication strategy worth considering. However, an article published this week by the Sydney Morning Herald referenced research suggesting that by July (ie before the peak of the second wave), the total number of cases (including unreported and asymptomatic cases) in the Australian population was about 70k, which is about 60k more than the figure for confirmed cases at the time. Our model also produces estimates for the total number of cases, so it will be interesting to compare the outputs from our model against this research – however, more work is required to calibrate our model for total cases, so we’ll save this topic for next week.

New South Wales: Emerging experience as at 16th September 2020

7-day average for new cases trends back downward, stable pattern in daily new cases, sensible relaxing of NSW/QLD border restrictions is surely appropriate

This week, NSW saw the number of cases continue to trend downwards, with the 7-day rolling average falling from 5 last week to 4 this week. Daily new cases continue to remain at a low level and given that the majority of cases reported over the past week relate to known clusters/sources, it appears that transmission is under control. The pattern that we identified last week also appears to be continuing, with about two thirds of days since mid-August having observed a daily new case figure in the 3 to 8 range, and the remaining third being roughly evenly split between above and below this range.

Some have suggested that testing rates have been reducing and have rightfully pointed out that if this is the case, this might be at least partially responsible for the lower number of new cases seen this past week. A quick look at the testing statistics suggests that there might be some merit in this suggestion.

While there does appear to be a downwards trend in the number of tests conducted over the past couple of days, the 7-day rolling average does appear to follow a cycle, and it is too early to say whether the recent downwards trend represents a departure from that cycle. Nonetheless, it is worth noting that the proportion of tests returning a positive result (measured as the 7-day average for newly reported cases divided by the 7-day average for number of tests conducted) despite some reasonably significant volatility, has remained at around the same level since the number of daily cases started to show a pattern in mid-August (as discussed earlier), and hence we might expect the number of newly reported cases to increase again if testing picks up over the next couple of days. By extension, this would suggest that there are Covid-19 positive cases in Greater Sydney that aren’t reported. It’s an important reminder and levels of diligence need to remain high.

When will the NSW/QLD border reopen? The rate of transmission in Queensland has remained low with a total of 6 new cases being reported over the past 7 days. With yet another week of experience suggesting that NSW is capable of keeping transmission under control, and with Queensland continuing at current levels, our view remains that it is unreasonable keep the NSW/QLD border closed, and that a sensible relaxing of restrictions is appropriate.

Related Knowledge

Scroll to Top