the details of the formulas

I quick answer is I’m not going to reveal exactly what the formulas are. But I will reveal the methodology used, and the components of the TSPinvestor Basic and Plus formulas.

After setting sights on this quest, I decided upon regression analysis to solve the problem.  Certainly this would reveal which factors were important – and it did.  But it required some math and creativity to get to the best solution.

First I downloaded the monthly Thrift Savings Plan returns for each fund from the inception of the S and I funds — from January 2003 until present (September 2015). The Basic uses only the Thrift Savings Plan funds – G, F, C, S, and their own historical values.

I tried numerous combinations of these to determine the best predictor model (formula) of the next month’s return for each fund. I then used scaled combinations of these formulas to get an allocation percentage for each fund. Once I get the formula, I run the historical data back through the model and get monthly allocations from the formula and can determine over time the month-to-month returns, and the annual returns.

The initial annual return from 2005-2014 turned out to be about 11%. This was an answer I could not believe – it was beating the individual fund’s best performer (S fund) by more than 2% annually. I sought to disprove the data by investigating the spreadsheet thoroughly and looking intently at the monthly data for inconsistencies in logic.  I couldn’t find anything other than a few administrative errors that once corrected made very little difference in the error.

But I looked at the process and determined that my regression analysis did not optimize the reduction in independent variables. So I ran the regression analysis again and optimized the variables.  After running to data through again, the annual return was now at about 12%. So I locked this formula down (called the Basic).

I was not at all content with the answer – I wanted a better return – a better formula, what I later called the Plus. The question I asked, “Can economic information help better predict the returns?”  I investigated the economic landscape and settled on three monthly indicators of economic performance – which I will not reveal – that are diverse and common. I downloaded the monthly data which overlapped the other data (2003-2015).

After determining the formula and running historical data through, the annual return was nearly 19% from 2005-2014.

Again, 19% — a number that is simply unbelievable. The data backs the number.

If there is one pervasive observation of the month-to-month data – the TSPinvestor Basic and Plus are both effective at minimizing the large loss months. The Plus is more effective at getting the large returns after the big losing months.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s