COVID-19 forecasts by Hybrid Intelligence: end of the pandemic, vaccine date, GDP impact

The ongoing coronavirus pandemic is creating unprecedented uncertainty. When will the lockdown end? How many people will be affected? What will the impact on the economy be? Expert views vary widely, so Cindicator applied the predictive power of Hybrid Intelligence to try to look into the future.

How Cindicator’s Hybrid Intelligence works

Cindicator’s predictive analytics combines the collective intelligence of 140,000 registered analysts with machine learning. Some users are professional financial analysts, others are amateur market watchers, and all of them use crypto.

First, analysts make forecasts by answering questions in the Cindicator app. Analysts win or lose points for correct and incorrect forecasts and at the end of the month those with the most points receive rewards in CND tokens.

After the forecasts are collected, machine learning models apply different weightings to individual answers to come up with a final indicator. These indicators are sent to holders of CND tokens, most of whom are traders. Since Q4 2017, Cindicator’s accuracy has remained at above 60%.

Questions about the coronavirus crisis

When it comes to forecasting the course of the pandemic and its effects, there are simply too many unknowns.

To narrow the scope of forecasts, we first asked the Cindicator community which questions they would ask.

After reviewing all the suggestions, we posted the following questions on the Collective Intelligence platform:

• How many days have passed since you were in public?
• By what date will a coronavirus vaccine be announced?
• By what date will the WHO declare the pandemic over?
• By how much do you think your country's GDP will fall in 2020, according to IMF data?
• By how much will global GDP fall in 2020, according to IMF data?
• How many coronavirus cases will be officially confirmed by the end of 2020, according to worldometers.info?

Since most of Cindicator’s users are traders or investors, we also asked two questions about the markets:

• Will E-mini S&P 500 Futures trade below the Q1 low (USD 2,174) at any point before the end of 2020?
• Will BTC trade below the Q1 low (USDT 3,782) at any point before the end of 2020?

In their forecasts, analysts estimated the probability of these events giving predictions between 0% (not happening) and 100% (will definitely happen).

COVID-19 predictions

The questions were posted for three days (3–⁠5 April) and we collected thousands of forecasts.

Since there is no precedent for forecasting the pandemic and its impact on the economy, we didn’t apply machine learning models in assigning weightings to individual answers. Instead, we used the median answer as an estimation as suggested by the wisdom of the crowd theory.

So let’s review the answers.

How many days have passed since you were in public?

To understand their reference points, we asked how long respondents have been staying at home.

On average, Cindicator’s users have been self-isolating for 9 days. The median answer was 11.

The median answers for Russia and the United States – two of the most widely represented countries on the Cindicator platform – were 6 days. However, most users did not indicate their country.

By what date will a coronavirus vaccine be announced?

Analysts on average expect the vaccine to be ready by mid-September to mid-October as represented by a median of 162 days and a mean of 192 days from 6 April.

By what date will the WHO declare the pandemic over?

On average, Cindicator’s analysts think the World Health Organisation will declare the end of the pandemic by early September or October (mean: 184 days; median: 148 days).

By how much do you think your country's GDP will fall in 2020, according to IMF data?

Analysts think that their country’s GDP will decline by 10%. However, some analysts appear overly pessimistic and the median answer suggests a less dramatic decline of 4.8%.

Answers came from users from over 60 countries, yet not all of them are equally represented on the Cindicator platform. Furthermore, some analysts did not state their country.

Here are the median forecasts from the top 10 economies that are also well represented on the platform:

• United States -8%
• Germany -4.5%
• India -2.8%
• UK -5%
• Canada -20%

By how much will global GDP fall in 2020, according to IMF data?

Cindicator’s respondents expect global GDP to fall by 5% (the median answer).

Over a quarter of analysts forecast a global GDP contraction of over 20%. For comparison, the world’s GDP fell by an estimated 15% during the Great Depression (between 1929 and 1932), and anywhere from 0.1% (IMF’s estimate) to 1.96% (World Bank) in 2009. While potentially unrealistic, this shows that many expect the coronavirus pandemic to have an extremely severe impact on the global economy.

How many coronavirus cases will be officially confirmed by the end of 2020?

Cindicator’s analysts expect 5 million confirmed cases (median answer) by the end of the year.

Coronavirus impact on S&P 500 and Bitcoin price

The big question on every investor’s mind is whether we have already seen the bottom in the markets, considering that the worst for both global health and the economy are yet to come.

Since 2017, Cindicator has collected over 15 million forecasts, mostly in financial markets. For the questions below, machine learning models have assigned weightings to individual forecasts based on track records in forecasting equity and crypto markets respectively.

Hybrid Intelligence indicators based on binary questions suggest that the price of Bitcoin will not retest the Q1 low, while the future of the S&P 500 is less clear.

Will E-mini S&P 500 Futures trade below the Q1 low (USD 2,174) at any point before the end of 2020?

• Indicator: 46.67%
• Interpretation: analysts are uncertain about the outcome and are slightly bearish.

Will BTC trade below the Q1 low (USDT 3,782) at any point before the end of 2020?

• Indicator: 13.3%
• Interpretation: analysts believe there is a very low probability of hitting new lows this year.

Conclusions

While the above predictions are grim, we don’t have a crystal ball. The reality may be better or worse, but forecasts at least help us to reduce uncertainty by showing expectations in numerical form.

It is helpful to understand what others believe, as they will make decisions based on these beliefs whether they turn out to be right or not.

If you are interested in collaborating with Cindicator’s Hybrid Intelligence, feel free to suggest any ideas you have by sending us an email.