Adjusted vs Unadjusted Close price

Question

I have written some code that retrieves the "Close" and "Adjusted Close" price for all the funds available to me via my employers 401k. The code grabs the prices monthly all the way back to 2000 (or whenever the fund was created). The goal was to backtest various investing strategies over that time period. Here's what I don't understand... In an attempt to sanity test the data I retrieved I noticed that different sites have different adjusted prices, why?

For the sake of this discussion, let's use the example of PIMCO Real Estate Real Return (PRRSX). The code I wrote collects data from a website called Alphavantage. For sanity purposes, I wanted to compare the value from Alphavantange with three other sites: PIMCO's site, Yahoo and MarketWatch.

First, note that PIMCO lists only a NAV, nothing about adjusted vs unadjusted. Second, Yahoo and Alphavantage use the term "adjusted" while MarketWatch used "unadjusted". With respect to what Yahoo just calls the "Close", the 3 sites match. Only the "adjusted" and "unadjusted" differ. The picture below summarizes what each website returns for a value on 9/28/2012.

Answer 1

It's pretty frequent for historical data of mutual funds to be a little wonkey, Google's data is pretty notorious for missing dividends and/or capital distributions.

PIMCO Real Estate Real Return (PRRSX) had a 1:2 split on 8/7/2015.

So it looks like MarketWatch and PIMCO itself are adjusting only for the reverse split not dividends or other distributions. The close NAV of $10.76 divided by two, because there is now one share for every two shares that existed on 9/28/12, is $5.38. MarketWatch probably calls this "unadjusted" as the data is not adjusted for dividends or capital distributions and it doesn't consider a split to be an "adjustment" event. $5.38 can also be verified on Google's chart which shows unadjusted data and very clearly shows the price change on 8/7/2015.

When I pull the Yahoo data it looks like the split is simply built in to the historical data, as evidenced by the close price on 8/6/2015 is listed as $7.72 rather than the true historical close of $3.86. I suspect Alphavantage is also only accounting for dividends and/or capital distributions.

Total distributions have been $3.1195 since 9/28/2015 according to data from Seeking Alpha. Yahoo appears to be using a 53.31% adjustment factor by that date. You adjust by subtracting the percentage the distribution represents of the NAV on the day of the distribution. So for a $10 unit, a $0.50 dividend represents 5% meaning you adjust everything past that date by to 95%. When I place the distributions against the close data I come to a 67.83% adjustment by that date for a share price of $7.30, which is actually pretty close to the Alphavantage adjusted share price of $7.50.

It's hard to pinpoint the exact cause for the deviation in the adjusted amount; though some of the deviation in adjustment likely involves a difference in the number of digits being captured in the various math being done.

The moral of the story is data and adjustment methodologies are not always consistent and data collection isn't infallible .