Why doesn't the sum of daily percentage changes of a stock index over a given time period, equals the period percentage change?

2.1k Views Asked by At

The following is the simplified version of an actual model. It is only meant to exemplify the issue I am having.

I have a large MS-Excel sheet of 29953 rows containing the daily closing values of the Dow Jones (DJIA) since 1900/01/03. Row 1 contains the Column headers. Column A contains the Date, column B the DJIA, Column C the daily change percentages. The entire C column contains the daily closing percentage gain/loss using the following formula starting at C3 and copied all the way down: =(B3-B2)/B2.

Calculation_A. If I wish, for example, to compute the percentage gain between 1900/01/03 and now (2016-11-29), I use the following formula: =(B29953-B2)/B2 which using the actual values translates to (19121.6 - 66.6) / 66.6 = 286.068. This is, I think, the correct result.

Calculation_B. Calculation_B is the sum of the daily percentage changes (calculated in column C) corresponding, in Excel, to the formula =SUM(C2:C29953) resulting to 767.812.

Intuitively, I would have assumed that Calculation_A = Calculation_B.

My question are as follows, why (mathematically) are Calculation_A and Calculation_B yielding different results? Is there a way to make calculation_B consistent with Calculation_A?

I don't consider myself particularly optuse, but I am having a hard time solving this seemingly trivial problem. Are integrals the answer? Thanks in advance!


Update#1. Following the very good comments and answers from Brian and DougM, thank you, I was able to reconcile the two calculations on my initial PL1DAY column.

I now have another issue along the same line: I have a Excel column named PL20DAYS which calculates, for each trading day, the percentage gain/loss 20 days forward: i.e. (DAYn+20 - DAYn)/DAYn. Why would the compounding of the PL20DAYS column not equate the compounding of the PL1DAY?

1

There are 1 best solutions below

4
On BEST ANSWER

Look at a very simple example: suppose that some quantity increases by $100\%$ and then again by $100\%$ . A $100\%$ increase means that the quantity doubles. Here it doubles and then doubles again. Thus, if its initial value is $x$, it first doubles to $2x$ and then doubles again to $2\cdot 2x=4x$. It is now $400\%$ of its original value, not $100\%+100\%=200\%$.

Successive percentage increases are multiplicative, not additive. When something increases by $p$ percent, it is multiplied by $1+\frac{p}{100}$. If you now increase it by $q$ percent, you're multiplying the increased amount by $1+\frac{q}{100}$. The combined effect is to multiply the original quantity by

$$\left(1+\frac{p}{100}\right)\left(1+\frac{q}{100}\right)=1+\frac{p+q}{100}+\frac{pq}{10000}\;;$$

assuming that $p\ne 0\ne q$, this is clearly not the same as multiplying by $1+\frac{p+q}{100}$ to get an increase of $p+q\%$.