- J.C. Penney was downgraded on Monday by Wells Fargo with a price target revision of $5.00 which was down from $6.00.
- But how often do the professional analysts get it right with respect to J.C. Penney? Can we predict future stock prices based on the analyst price targets?
- We put together a regression model based on 137 analyst price targets from the last 2 years for J.C. Penney which can be used to predict the future stock price.
- In the case of J.C. Penney, analyst price targets in our regression model can be used to explain 57% of the change in subsequent stock prices.
J.C. Penney (NYSE:JCP) was recently downgraded by Wells Fargo, which also changed their price target from $6.00 to $5.00. This caused the stock to fall from the initial increase seen on Friday after the earnings report on Thursday. However, what do various price targets created by analysts indicate concerning future performance? Sometimes the various price targets can lead to conflicting information for investors. For example, while Wells Fargo reduced its price target for JCP on Monday, based presumably on the recently released earnings report, Deutsche Bank increased its price target from $6.00 to $10.00 on Friday (reiterating a Hold status). How do analysts at two different firms take similar price targets and go in two different directions, with the latter effectively doubling the former? Can we create a regression model to predict future performance based on an aggregate of analyst price targets?
It would appear that we can create a regression model; however, the accuracy is far from perfect. We will also re-address the Wells Fargo revised price target of $5.00 to discover what it really predicts with respect to the future stock price.
There has been some research in the past conducted on the effectiveness of using analyst price targets to determine future stock prices. Generally, they have not been concerned with a single stock, but with analyst recommendations in general. Brav & Lehavy have conducted analysis showing how price targets have impacted stock price over the short term (subsequent 5 days) and long term (subsequent year). With respect to the subsequent 5-day reaction, the following was found:
Price target revisions are accompanied by a mean five-day abnormal stock return of -3.9% around downward revision announcements and +3.2% for upward revisions (Brav and Lehavy 2003).
Asquith, Mikhail, and Au found the following:
Earnings forecast revisions and recommendation revisions are significantly and positively associated with the market's reaction at the time a security analyst report is released.
The argument for this is that an analyst releases new information into the market even though prior information is used by the analyst in the creation of the report. However, due to the regurgitation of information previously known by the market, the researchers indicate that their R^2 is fairly high compared to other types of research. We will address this in more detail later under the discussion of limitations, but there will most likely always be an issue with multicollinearity or the interdependence of data sets which will result in regression models showing higher levels of accuracy that might simply be explained by the analyst price targets.
With that disclaimer stated, here is the analysis.
Brief Summary of Our Regression Model:
We utilized two sets of data for the model, we used the Analyst Ratings Network data for JCP. We also utilized historical price data gathered using Google Finance. We merged the two data sets to compare the analyst ratings by date against the subsequent average for the business week, two weeks, month, two months, and three months. There were 137 different analyst price targets that were usable from the last 2 years. We then created five regression models, setting in each the predictive or independent variable as the analyst price target which explains the future or subsequent average stock price.
Here are the summary statistics on each different model:
|Time Frame||12 Week||8 Week||4 Week||2 Week||1 Week|
|Adjusted R Square||0.57||0.60||0.61||0.63||0.64|
R^2 is the test of accuracy for the regression model. It indicates how much of our dependent variable is explained by our independent variable. We therefore have an accuracy of 57%, 60%, 62%, 63%, and 64% which shows us that in the short term, analyst price targets are fairly good at indicating the future average stock price.
However, when we state that the price target of analysts is a strong indicator of the future average stock price, it does not mean that the price target of $5.00 as in the case of Wells Fargo analysts would indicate that the price will actually be $5.00. In this case, the analysts are really quite inaccurate with respect to creating their price targets. However, based on some changes, we can create a strong future value of the price based on the analyst recommendations.
We therefore consider the coefficients that I have listed above as well. The predicted or expected value based on any given model would be given by the equation y = mx + b, or stated in another way, the Intercept plus the Price Target Coefficient times the Analyst Price Target. So within the next 3 months, if we utilized the Wells Fargo price target, we would have the following $6.43+(0.51*5.00) = $8.97 which is roughly where JCP price currently falls. If we considered the Deutsche Bank analyst price target, we would take $6.43+(0.51*10) = $11.53 which is probably a bit high (but if we consider the after hours trading on Thursday reaching a high of $10.50, not entirely implausible).
We went back and applied the equation for each time period, including 1 week of analyst price targets for the 1 week category, and 2 weeks for the 2 week category, etc. We then took the average of the resulting predicted prices. This is what we found:
|Predicted JCP Stock Price Based on Aggregate Analyst Price Targets|
Here are the predicted numbers based on only the Wells Fargo price target of $5.00:
|Predicted JCP Price Based on Wells Fargo Analyst Price Target|
As you can see, this is approximately where the price of JCP currently falls, this is to be expected based on the previous two years of market response to the analyst estimate.
Discussion of Assumptions & Limitations:
Generally, I am not the biggest fan of conducting regression analysis to explain the behavior of stocks because of the interdependence of the data sets. As an example, we would certainly not wish to use the P/E ratio to explain changes in stock price as the P/E ratio contains actual price data and would therefore correlate very well with changes in price, but at the same time not be a good independent variable as it would hardly be independent of the dependent variable.
We are therefore making the assumption that investment firms are creating target prices based solely on actual intrinsic company information as opposed to the current stock price. To the extent that this assumption does not hold true, there will be error in our analysis. Conceptually, the idea of price targets are meant to be created based on company performance which is independent of stock price and while there is some research to suggest that they do not do this, that would be beyond our current discussion.
So how can you utilize all of this information?
It depends if your other analysis causes you to be bullish or bearish with respect to JCP. If bearish, this analysis might cause one to be hesitant to short JCP because of the $5.00 price target set by Wells Fargo, since when taken with all other price targets within the last 2 years, there is a 64% chance that the stock will not go below $8.46 within the subsequent 5 days of the Wells Fargo downgrade. However, had one the ability to short early on Monday then that would probably have been the correct decision. If we see the stock pop up significantly in the next few days, then JCP will probably be a good stock to short once again. All of this would also seem to be in line with prior research mentioned earlier.
If other analysis indicated a generally bullish sentiment with respect to JCP, then I would simply wait for the stock to drop a little further and then purchase more shares. Right now, with respect to the Wells Fargo analysis, however, it would appear that the stock is roughly where it ought to be (even though it is far way from $5.00). Consequently, it is difficult to suggest either a long or short position in the stock simply based off of this analysis and at this current point in time.
While there may not be immediate use with respect to 'buying' or 'selling' based on this form of analysis, it is hoped that this will help the reader by establishing benchmarks for future shorting or long strategies based on the analyst price targets. You can also download the data and the analysis as an excel file right here.