AN ANALYSIS OF SUPERBOWL VOLATILITY Sean Puneky 10
AN ANALYSIS OF SUPERBOWL VOLATILITY Sean Puneky 10 February 2009
Initial Theory A study in Behavioral Economics Traders trade what they know Theory: If traders witness a commercial from a major add campaign, they are more likely to develop trading ideas for the stock, increasing the volatility of the stock Volatility should reflect fundamentals, not trader behavior
A Testable Hypothesis Test uses data from a major event like the Superbowl where many companies release their best commercials Companies that advertise in the Superbowl should see the Realized Volatility of their share prices increase on days following the Superbowl
Structure of the Test A one sided t-test is used comparing the volatility before the Superbowl to the volatility after Variables for t-test: Equation:
Structure of Data sets are means of the Realized Variances from an 8 minute sampling interval on the five trading days before and after the Superbowl Log(mean) from before the Superbowl subtracted from the log(mean) from after the Superbowl for each year T-test performed with null hypothesis that difference is equal to zero
Stocks Chosen Working with three stocks: PEP (Pepsi), KO (Coke), and FDX (Fed Ex) Both Pepsi and Fed Ex advertised in the Superbowl over the testing period while Coke advertised only before the Superbowl in 19982006
Volatility Signature Plot of PEP
PEP Histogram
Volatility Signature Plot of KO
KO Histogram
Volatility Signature Plot of FDX
FDX Histogram
T-Test Results (9 df) PEP T-value of 2. 27714421054183 Significant at 2. 5% level (t>2. 26216) KO T-value of -0. 250909814238321 Not significant FDX T-value of -0. 645862272357625 Not significant
Conclusions Although PEP and KO show promising results, FDX is inconsistent with theory PEP volatility increases without any change in the fundamental value of PEP
Extensions Investigate the FDX dilemma Examine the effects of weighting Superbowls by viewership or price of air time Expand analysis to other ad campaigns
- Slides: 15