Google (GOOG) is due to report its Q3 earnings at 4:30PM Eastern Time today, much to the anticipation of the market as a whole. In general, I agree with previous articles written about the upcoming report which can be found here, here, and here. I also like the implied volatility study combined with the earnings whispers found here. As a disclosure, I am bullish on Google from a financial analysis standpoint, but wanted to conduct this study in relation to data analysis.
What none of the previous articles addressed was the in-depth history of Google’s earnings data and how that data may actually give the average investor insight into the announcement later today. In analyzing Google’s earning history, this article will focus on three main points:
- Google has a solid history of surprising analysts
- A study of earnings seasonality is useful in Google’s case
- Investors should be cautious when trading Google's earnings today
This study does not attempt to provide a solid trading rule, but should be another layer of analysis after completing a standard qualitative and quantitative analysis. We will be looking at Google’s last 22 quarters (since Q1 2006) for earnings data. I realize the amazing amount of volatility that the market has seen since 2006, but have no guarantees that the future will be any different.
As well, for this study, I will be looking at alpha returns (compared to the SPY), so the return studies will not be as skewed by the economic environment of the specific quarters.
Google Continuously Surprises Analysts
If you look at the histogram below, it is easy to discern that Google has a history of beating analyst estimates. Granted, it is difficult to quantify exactly where the search engine giant is going, but on average, Google beats estimates by 5.6%, and has surprised 17 out of 22 quarters (77.27%).
Although earnings surprises seem to be the norm with Google, the reaction to these surprises is anything but normal. What tends to happen is that when the actual earnings number comes out, the market takes some time to digest the underlying numbers (margins, additional data, etc.) before reacting. In fact, if you look at a scatter plot with the surprises on the X-axis and Google’s alpha the day after the announcement on the Y-axis, it shows that the actual surprise percentage is not the only determining factor in the next days’ trading alpha (see chart below).
If you are not familiar with scatter plots, basically, the horizontal X-axis shows the percentage surprise, while the vertical Y-axis shows the alpha percentage on the day after the announcement. The equation is interesting because if interpreted, basically says that if x=0 (average analyst target was correct), the stock will drop approximately 4.59%. By using the equation given by a linear regression, for there to be any type of positive alpha the day after the earnings announcement, Google would have to surprise by more than approximately 5%. However, this approximation is definitely not a hard rule, as can be seen from the scatter plot above. What we can conclude is that although there is a relationship between surprises and positive alpha, I wouldn't bet my portfolio on it.
I also congregated the third quarter EPS data for each year and analyzed that series (in red on the scatter plot). Initially, it looks like Q3 shows some abnormally high alphas compared to the surprise percentages. However, on further examination, it appears that two of the data points are skewed because Google surprised to the negative side on the quarters preceding the Q3’s under examination.
Google’s Earnings Seasonality
When I set up a financial model, I always check every line for seasonality. I do this because in many cases, there are unique trends that emerge. For instance, when looking at a microcap manufacturer, they may not have energy hedges set up, and so their energy costs swing according to the season.
I also like to look at earnings seasonality. In Google’s case, if you take an average of the last few years, an earnings seasonality trend is present. I took an average of the percentage of revenues that each quarter is responsible for over the fiscal year, and got the following table.
So, by using the data table, along with the first and second quarters’ EPS data, I was able to extrapolate an annual EPS and calculate a “naive” Q3 EPS range of $8.24 to $10.42.
Another way of looking at it was to look at past years and calculate a Q3 EPS range by calculating Q3 EPS as a percent of the sum of the first and second quarter EPS numbers. Doing this gave me a range of $8.50 to $9.98.
Making Money by Trading Google Today
Once again, my goal is not to put forth an EPS projection, but rather to empower the investor with a data framework to layer onto his or her own financial projections. Just as a reference, I have messily compiled this chart with accompanying table showing my projected ranges, the analysts projected range and average target, as well as the average surprise number.
The only thing that I really will conclude with this study is that the data does not look good for Google’s earnings projections. Especially considering that even on the high side, not a single analyst is above that average surprise number. And since the average surprise is almost built in to the price, according to the data, unless we see more than a 5% surprise, GOOG may not see a positive alpha tomorrow.
Once again, I am bullish on Google, and believe that they will probably do just fine today and tomorrow. However, a deep dive into data analysis can be useful when layered onto fundamental analysis.
I am interested if anyone found this study to be informative or useful. Please let me know in the comments.
Disclosure: I have no positions in any stocks mentioned, but may initiate a long position in GOOG over the next 72 hours.