Interpreation of tests for autocorrelation

Within my sample, I have different financing round (about 300) of several companies (about 200) between 2010 and 2017. To control for the time the company exists, I have included the variable "age" (in number of years - with comma values depending on the month, e.g. 3.5 for three and a half years).
I am pretty new into statistics and I did tests for autocorrelation in R (Durbin-Watson and Breusch-Godfrey).
As far as I have understood:
- the DW test should be between 1.5 and 2.5 (my results are: DW = 1.7737, p-value = 0.01029)
- the BG test should have a p-value lower than 0.05 (my results are: LM test = 3.4041, df = 1, p-value = 0.06503)

I have following questions:
1) On which order do I have to conduct the BG-test (based on the years or months - 1 or 12)?
2) Is my understanding of the test interpretation right, or did I make a mistake somewhere?
3) The p-value of the BG test is higher than 0.05, so there should be autocorrelation, right? How can I remove this autocorrelation?


Less is more. Stay pure. Stay poor.
Side note, there isn't anything magical about 0.05. If you were going to get a surgery and they said the pvalue related to dying as a result of the procedure was 0.056, would you say well it is larger than 0.05, so I should be fine. Probably not. Pvalues can be influenced by sample sizes and among other things. It has been awhile since I used those tests, so I am going to opt not to provide any direct interpretations, in an attempt to not accidently mislead you.

Good Luck