I Try to Predict the S&P 500 and the Results Are Surprising

I always ask myself the eternal question. Can we predict stock markets? Is there a reliable way to do this for peasants such as me?

Big banks and investment companies have abundant resources to project the stock markets and individual companies’ share prices. Ordinary investors like you and me don’t. And we always suffer as small-time retail investors.

But what if we can create simple models to do what the big boys are doing? Models that use indicators and numbers that we can get from regular websites?

That is what I will try to find out here. Can we project the stock markets on a shoestring budget and match the big boys of finance?

We sure can. I have already prepared the model and you can access it from here!

Let’s get to the basics

The stock market reflects what’s happening to the whole country. Anything can affect it, and the prices will essentially show what investors think about the economy.

So, it’s a game of ‘sentiment’. And sentiments are very subjective. Luckily for us, a couple of indicators are proxies to these.

In my head, three indicators come to mind.

1. CBOE volatility index (VIX): This is how ‘fearful’ or ‘optimistic’ investors are on the stock market. The higher it is, the more fearful they are and vice versa.
2. Services Purchasing Manager Index: A survey is sent out every month to companies in the services industries to examine how ‘positive’ or ‘negative’ they are in their respective businesses. 50 and above indicates positive sentiment, while below 50 indicates negative sentiment.
3. Manufacturing Purchasing Manager Index: ?Similar to the services purchasing manager index but for manufacturing companies.

These 3 indicators are easy to obtain from investing.com where you can also download the Excel files for them.

The best form of learning is to copy from other people what they do. I read this recent paper in 2021 by Gaspariene, Remeikiene, Sosidko, and Vebratie and they identified three indicators that are related to S&P500 performances.

1. 3-Month Treasury Bills: Returns or yields on 3-month government bonds issued by the U.S.
2. WTI Crude Oil Price: Crude oil price for West Texas Intermediary
3. Personal Savings Rate: How much is saved by consumers in the U.S.

Thus, I set out to investigate whether these variables can project the S&P 500 levels for the 3 months ahead.

Here is How Effective the Models Are

I am a firm believer in making things simple. Hence, I am just using a simple regression to regress the variables here. It’s the simplest form of ordinary least squares regression.

The equation is the following:

Surprisingly, 2 variables are not significant or related to the S&P 500. They are Services PMI and WTI crude oil prices. And the intercept is too high at -6,254.

I made two key changes to the model. Firstly, I removed the two variables that are not significant. Secondly, I set the intercept to 0.

The model came out to be highly effective and significant with a fit (R-squared) of 95%. This means the model can explain 95% of the S&P 500. All variables are significant at the 5% level.

To project forward, I would need to utilize lags in the model. In simple terms, it means that data for the variables from December 2023 to February 2024 should be able to predict the S&P 500 level in March, April and May 2024.

Hence, I also made models that have one to three months lag in projecting the S&P level. This means I will have a forecast for March, April and May 2024.

Here are how effective the models are, and their estimates for the variables.

And the Projections

Hence, all these models yielded these projections for the next 3 months.

Here are some explanations. Manufacturing PMI has been increasing from 48.2 in December 2023 to 50.7 in February 2024, which indicates improvements in manufacturing companies’ sentiments.

And CBOE volatility index has also increased from 12.45 to 15.42 which indicates that risk sentiments are increasing. This is what’s interesting. It seems like an increase in risks pushes the S&P 500 level. Here’s my interpretation. Investors in the S&P 500 are riding some sort of risky trends. For me, this is the AI trend.

The 3-month Treasury bills also have the same story as the CBOE Volatility Index. By right, the higher the treasury bills, the more it shows that risks are increasing in the economy and the S&P 500 should decrease. But no, it seems that investors like risk at the current juncture and the S&P 500 is enjoying that.

So, what can you do? And how does this add up?

Now, firstly, these are just models. And models are biased. I did not test for stationarity or heteroskedasticity, as I felt they would be too complicated for the regular person to wrap their heads around.

The models that I have created are just meant to be a guide in the myriad of tools out there. And it’s also because I dislike the tendencies of big players to manipulate the market with their views and models to their advantage.

How many times we have seen big banks and investment firms buy in bulk some companies and ask other smaller investors to do so also? And right when they are at the top, they sell them brutally at the expense of us. Their reasoning is always, “Valuations are unsustainable now and too high”. No shit, who was the one who bought them in bulk in the first place?

My aim here is to provide us smaller investors with some foresight and not let big players have their way every time.

I believe in education and the motto of Go Research and Invest!