# Reasons for negative autocorrelation of forward prices

I am working on each trade day's forward prices of gasoline. I noticed that the autocorrelation at lag 6 is significantly negative. I know how to interpret negative autocorrelation in a statistical way, but how can I interpret it in an economic way? On term structure, the forward prices of gasoline should go up or go down together. For example, if one large oil field in Saudi Arabia gets attacked, prices all go up together. So how come the autocorrelation is negative here.

• A very wild guess: Do you have 6 trade dates per week (Mon thru Sat), i.e. could this be a day-of-week-effect? Jun 22 at 21:01
• Thx for replying! I have 5 trade dates per week. And on each trade date, the forward contracts have monthly term dates. Jun 22 at 21:19
• Are you looking at returns/changes or raw prices? Jun 22 at 22:51
• You said that the autocorrelation was negative but you did not give a value. It is one thing if the autocorrelation was -0.01 and another if it is -0.5. Can you tell us what the autocorrelation value is? Also if you use a lag of 5 is it positive? Is it positive at 7?
– Bob
Jun 23 at 1:50
• I'm looking at raw prices. The lags of 5 and 7 are also negative, but they are insignificant around -0.15. The lag of 6 is -0.75. Jun 23 at 13:07