Logit Regression on Overnight S&P500 Performance and Its Impact on KLCI

In a previous article, we demonstrated the capabilities of logistic regression (“logit”) on predicting potential loan defaulters (Read more here).  This week, we intend to illustrate another example of logit model application.

We often hear that overnight Wall Street’s performance might have an impact on the next day’s KLCI results, but how shall we quantify this?  To answer the question, we could use logit model to study the probability of the KLCI to close positively or negatively, given the performance of overnight S&P500.

In a logit model, if KLCI gains on the next day, it will be labelled as “1”, while it will be labelled as “0” if KLCI has loss on the next day.  X is the overnight S&P500 performance while b₀ and b₁ are the coefficients.  The coefficients are then estimated using Maximum Likelihood Estimation (MLE)

Chart 1 is the probability of KLCI Gain/Loss on next day given overnight S&P500 performance based on the logit model.  The probability distribution shows that if overnight S&P500 gained 5%, it is probably 99% sure that KLCI will gain on the following day.  If overnight S&P500 gained 1%, the chances for KLCI to gain on the next day is around 70%.  What if overnight S&P500 loss is 2%?  Then the probability of KLCI to close positively on the following day would be around 15%.

Chart 1

Table 1 shows the first 5 rows of the data and their respective equations while Table 2 is the summary of the logit with the estimated coefficients. 

Table 1

Table 2

Reference:

CFA Program Level II Reading Assignment by Sanjiv R. Das, PhD, Richard A. DeFusco, PhD, CFA, Dennis W. Mcleavey, CFA, Jerald E. Pinto, PhD, CFA, and David E. Runkle, PhD, CFA