时间、成交量、持仓量都可以用来参考换月。

谢谢你啊 ，我后来也看了很多论文，有些也都提到了这个方法，可是现在是这篇已发表的论文跟我用同样方法取得同一产品的数据，最后按道理应该是用同样数据的论文，我们连均值都明显不一样。。。我觉得这个很糟糕，我只是写个论文，不想追求准度，只想追求不犯错。。。。

First, you can roll the contract on either the first business day of expriation month or the last trading day of that contract. In most cases the rolling date has no signficant impact on results.
Second, to compute returns you must use prices from the same contract instead of two different contracts. For example, attached are futures prices for WTI crude oil. The last trading day of December contract 'CL2013Z' is 11/20/2013. The return on 11/20/2013 should be ln(93.33)-ln(93.34). However, the return on 11/21/2013 is NOT ln(95.44)-ln(93.33). It should be ln(95.44)-ln(93.85), i.e., both prices need come from the January contract. By doing this, you can avoid big jumps across contracts.

This is the typical way to calculate futures returns in academic research. See appandix in Singleton (2014).

The reason is that returns based on prices across contracts are actually unattainable. See Sanders and Irwin (2012).
By the way, some papers did in this way but they didn't say it.

Sanders, D. R. and Irwin, S. H. (2012). A reappraisal of investing in commodity futuresmarkets. Applied Economic Perspectives and Policy, 34(3):515–530.                                                                                                 

Singleton, K. J. (2014). Investor flows and the 2008 boom/bust in oilprices. Management Science,60(2):300-318.

                                
                        
               
                                
                        
               

看起来还是不错的

想知道这个问题楼主解决了么？最近查了一点资料发现还是不知所云，有知道的大神解释一下吗？？网址见下面。https://quant.stackexchange.com/questions/419/is-there-a-standard-method-for-getting-a-continuous-time-series-from-futures-dat