Stock-Prices-vs-Economic-Indicators

This project analyzies the Impact of COVID-19 on the Correlation Between Classic Economic Indicators and Monthly Median Stock prices.

Authors

• Chenika Bukes
• Krishna Cheemalapati
• Aryan Goel
• Dhairya Khara

Introduction

The COVID-19 pandemic has impacted the lives of people across the world. Industries from energy to healthcare have undergone drastic changes because of the pandemic. The New York Stock Exchange (NYSE) has experienced unparalleled growth during the pandemic, despite many industries being negatively impacted by the pandemic, American federal debt being higher than ever, and there being abnormally high uncertainty about everything (Bloomberg).

The US experienced grave unemployment during the first couple months of the pandemic, yet many speculate that the outflow of stimulus checks, the less busy schedules of individuals, and the ease with which one can access the stock exchange through apps like Robinhood has led to this unexpected growth. Investors have put more money into stocks in the last 5 months than in the previous 12 years combined (CNBC). The average consumer spending patterns have grown to include investments in the stock market. The S&P 500 gained more than 16 percent in 2020, a very strong return during a year of nationwide lock downs and steep job losses in the United States (Washington Post).

Altogether, so far, the stock market during the pandemic can be summarised in one word: unexpected. Classic economic indicators such as unemployment rate and number of unemployed persons per job opening have been used as predicting factors for the stock exchange throughout history (Allen). However, during the pandemic, each indicator experienced extreme swings unlike anything in recent history. This led us to question the difference in the correlation between median monthly stock prices and classic economic indicators between August 2018 to December 2019 (a period of seventeen months before COVID-19 began) versus the seventeen months during the peak of COVID, from April 2020 to August 2021.

Since the pandemic could not have been predicted, we want to know whether the median monthly NYSE price action’s correlation to unemployment rates and unemployed per job opening statistics was impacted by COVID-19. This project will attempt to answer the research question: How has the COVID-19 pandemic impacted the correlation between median monthly stock prices and the classic economic indicators of Unemployment Rates and Number of Unemployed Persons per Job Opening?

Dataset Description

Historical Stock Data
Historical stock data will be collected from Yahoo Finance using their public API. This API returns a csv file upon request. The csv file that we receive contains the ”Opening”, ”High”, ”Low”, ”Closing, Adj.”, ”Close” and ”Volume” of that stock for each date in the specified time period. Data will be collected from August 2018 to December 2019 for the analysis of the correlation before COVID-19 began and data will be collected from April 2020 to August 2021 for the analysis of the correlation during COVID-19.

Monthly Unemployment Statistics of the US
The monthly unemployment statistics of the US was found from the US Bureau of Labor Statistics (“Civilian unemployment rate”). This data is used to create a custom .xlsx file and stored in the local directory of the program. This file will have two sheets, with the first sheet (named “Before”) representing the data before COVID-19 began (August 2018 to December 2019), and the second sheet (named “COVID”) representing the data during COVID-19 (April 2020 to August 2021).

Number of Unemployed Persons per Job Opening by Month in the US
The monthly number of unemployed persons per job opening in the US was found from the US Bureau of Labor Statistics (“Number of unemployed persons...”). This data is used to create a custom .xlsx file and is stored in the local directory of the program. This file will also have two sheets, with the first sheet (named “Before”) representing the data before COVID (August 2018 to December 2019), and the second sheet (named “COVID”) representing the data during COVID (April 2020 to August 2021).

Getting Started

After cloning the repository, install all the required dependencies using pip.

pip install -r requirements.txt

Run the main.py file. Make sure to ’Run’ the file and not ’Run file in Python Console’.

Recommended calls to type into the interactive search bar are TSLA, AMZN, VGT, NFLX, VHT, and VDE or any other stock ticker of your choice.

Results and Analysis

How has the COVID-19 pandemic impacted the correlation between median monthly stock prices and the classic economic indicators of Unemployment Rates and Number of Unemployed Persons per Job Opening?

To answer this research question we will investigate three stock indexes. These indexes are the Vangaurd Information Technology Index Fund ETF (VGT), Vangaurd Healthcare Index Fund ETF (VHT), and Vangaurd Energy Index Fund ETF (VDE).

Figure 1: VGT Index

From the graphs above we can see that the unemployment rate vs the median monthly stock price before COVID19 has a correlation coefficient of 0.77 which indicates a moderately strong positive correlation compared to 0.85 for the R² value of the unemployment rate during COVID-19 vs the median monthly stock price before COVID-19. This means COVID-19 did not have a particularly meaningful impact on the strength of the correlation between technology stocks and unemployment rates. As for the correlation between unemployed per job opening vs median monthly stock prices, before COVID-19 there was a 0.02 R² correlation and during COVID-19 there was an R² value of 0.85. This means COVID-19 heavily impacted the correlation between unemployed per job opening and median monthly stock prices for technology stocks.

Figure 2: VHT Index

From the graphs above we can see that the unemployment rate vs the median monthly stock price before COVID19 has a correlation coefficient of 0.25 which indicates a weak positive correlation compared to 0.74 for the R² value of the unemployment rate during COVID-19 vs the median monthly stock price before COVID-19. That means COVID-19 had an impact on the strength of the correlation between healthcare stocks and unemployment rates. As for the correlation between unemployed per job opening vs median monthly stock prices, before COVID-19 there was a 0.04 R² correlation and during COVID-19 there was 0.75 R² value. This means COVID-19 heavily impacted the correlation between unemployed per job opening and median monthly stock prices for healthcare stocks.

Figure 3: VDE Index

From the graphs above we can see that the unemployment rate vs the median monthly stock price before COVID-19 has a correlation coefficient of 0.13 which indicates a weak positive correlation compared to 0.35 for the R² value of the unemployment rate during COVID-19 vs the median monthly stock price before COVID-19. That means COVID-19 did have an impact on the strength of the correlation between energy stocks and unemployment rates. However, both of these correlations are weak and hence we can deduce from the data that in general unemployment rates does not impact energy stocks in the same way it impacts technology and healthcare stocks. As for the correlation between unemployed per job opening vs median monthly stock prices, before COVID-19 there was a 0.01 R² correlation and during COVID-19 there was 0.4 R² value. Once again, this means COVID-19 had an impact on the correlation between median monthly energy stock prices and unemployed per job openings, but the correlation is still not statistically significant.

Limitations

The primary limitation that we encountered with the data set, was that data for all of our independent variables was only available based on months, while the data with regards to the stock price is only available on a daily basis. Thus in order to solve this problem, we converted the daily stock data into monthly stock data by taking the median stock price of each month. The primary limitation with this is that the median stock price cannot accurately consider all the fluctuations that occur during a month, and cannot represent the month as a whole on an accurate basis.

Conclusions

Altogether, COVID-19 most greatly impacted the correlation between unemployed per job opening statistics and the median monthly stock price for all three indices of stocks: technology, healthcare and energy. The index which showed the least impact by COVID-19 was the VDE energy index. This might be because energy companies’ stocks are more impacted by factors such as government policies and price elasticity of demand. Ultimately, there is a clear impact of COVID-19 on the correlation between unemployment rates and median monthly stock prices, as well as, unemployed per job opening and median monthly stock prices.

Further Investigation

In regards to further exploration, we could identify more classic economic indicators and not only compare the correlation between before and during COVID, but we could also compare the classic economic indicators against each other to identify which indicator correlates the best with the different categories of stocks (technology, healthcare, education etc.) before and during COVID. Furthermore, our investigation focused on the United States, using stocks only listed on the NASDAQ, therefore we could further our investigation by getting stock information from other stock exchanges to include statistics from other countries. Then we could analyse the correlation by country as well to possibly uncover other trends in the impact of COVID-19 on the correlations between stock price predictors and stock prices.