COVID-19 Forecasts via Stock Market Indicators

TL;DR

This study repurposes stock-market technical indicators—candlestick patterns, MACD, and RSI—to forecast near-term COVID-19 daily cases using WHO data. It defines candlestick representations and trend signals, applies a Wilcoxon Signed-Rank test, and combines p-values across independent subsets via Fisher’s method to establish statistical significance. The authors demonstrate that certain signals, particularly Bearish MACD and some candlestick patterns, reliably predict reversals and the peaks of waves, and can indicate the onset of subsequent waves, with implications for health policy and resource allocation. The methodology is also illustrated on stock-market data, validating cross-domain predictive power for non-stationary time series and highlighting practical utility in both public health and financial domains.

Abstract

Reliable short term forecasting can provide potentially lifesaving insights into logistical planning, and in particular, into the optimal allocation of resources such as hospital staff and equipment. By reinterpreting COVID-19 daily cases in terms of candlesticks, we are able to apply some of the most popular stock market technical indicators to obtain predictive power over the course of the pandemics. By providing a quantitative assessment of MACD, RSI, and candlestick analyses, we show their statistical significance in making predictions for both stock market data and WHO COVID-19 data. In particular, we show the utility of this novel approach by considering the identification of the beginnings of subsequent waves of the pandemic. Finally, our new methods are used to assess whether current health policies are impacting the growth in new COVID-19 cases.

Figures (9)

• Figure 1: Daily Prices for S&P 500 from February 5, 2020 to March 19, 2020 with our simple polynomial regression line (shown in blue) indicates the trend. Observe that the price follows the general trend for a few unit of time, and thus a regression line can provide a useful insight.
• Figure 2: Illustrations of the construction of candlesticks. A red candlestick represents a decrease in value during the intervening period, observe that the open price is higher than the price at close. A green candlestick, conversely, indicates an increase in value. The proportions of the candlesticks are set by the open, high, low, close values over the period.
• Figure 3: Visual definitions of five candlestick patterns.
• Figure 4: Examples of accurate forecasting via candlestick patterns. The $x$-axis provides an index of time with each candle representing one time period, while the $y$-axis indicates the value of some positive-valued measurable quantity (traditionally, share price). Axes values have been omitted as they are unimportant for these illustrations. The blue line indicates the 4-day trend lines established via linear regression, confirming either an appropriate uptrend or downtrend. The light colored candle indicates the start of each candlestick pattern, observe that in all cases shown the pattern corresponds to a trend reversal.
• Figure 5: Mathematical definitions of five candlestick patterns.