(F) How to Better Read Earnings Reports

Earnings season will soon be upon us again. What the market will be looking for are those companies that come in with higher-than-expected earnings to reward, and for those firms that disappoint to punish.

However, there will be cases where one company beats by a small amount, say $0.05 or 2.5%, and there is a big jump in the price of a stock in reaction, but on the same day there might be another firm that beats by $0.10 or 10%, and it is treated as a relative non-event by the market. Why is that?

Well, too much of the emphasis will be one-dimensional — by what percentage or dollar amount did the actual result come in above or below the estimate? The significance of a percentage beat also varies greatly by the level of the estimate and actual. Beating by $0.02 is a much bigger deal if the consensus was looking for $0.05 than if it was looking for $5.00.

However, I would be (all else being equal) much more impressed by a company that beat by 40% if the expectation was for $5.00 and the actual was $7.00 than if the expectation were $0.05 and the actual was $0.07. While without a doubt, the size of the beat — both in dollar terms and in percentage terms — are an important metrics, it is not the only one you should be concerned about. Remember that the consensus estimate is an average (mean) of all the estimates out there. Two companies could each have say 10 estimates and a consensus estimate of $1.00.

How the Actuals Reflect the Estimates

The estimate profiles might be very different. Let us suppose that in one case, those 10 estimates range between $0.98 and $1.02. In another case the estimates range between $0.50 and $1.50. Now let us suppose that both companies report earnings of $1.05 (and we will also suppose that there were no other significant differences in the quality of earnings or one having whisper numbers that were significantly different from the “official” consensus expectations). Which company do you think would likely have the bigger reaction in the market? The one with the tighter consensus.

The posting of $1.05 would represent substantial new information about the prospects for the company. The results would have been well above what the most optimistic analyst had been looking for. For the second company, that same $0.05 and 5% beat would be essentially a non-event. It will not really force the analyst who was expecting $1.50 to become more optimistic. The analysts at the low end might be inclined to move up a bit, but the beat would most likely be considered almost random noise.

Why would there be such a difference in the spread between the two companies? Some companies are just generally easier to forecast. Generally speaking, the more leverage a company has, the harder it is to come up with an exact and accurate forecast.

Leverage Can Be Key

Leverage comes in two basic flavors — financial and operational. Companies with a lot of debt can do very, very well if things go right, but things can quickly turn disastrous if things go wrong. Equity investors get the residual, after everything else is paid off. Interest on debt has to be paid before the profit. Those costs, though, are generally fixed (unless the debt is mostly variable rate).

If a company is expected to bring in $1 billion in revenue for the quarter, but is expected have to spend $990 million (including interest costs) to generate that $1 billion in revenues, then the expected earnings are $10 million. Now suppose that it is able to generate $1.01 billion in revenues. If the costs are fixed, then that extra $10 million drops to the bottom line (OK, at least to the pre-tax line).

The result is a huge earnings surprise, 100% if we ignore taxes (or if the company has things like tax loss carry forwards). That from just a 1% better performance on the top line.

Volatility in Other Forms

Interest is not the only fixed cost. Some firms can have pristine balance sheets and no interest expense at all, but still have very volatile and unpredictable earnings. A great example of that sort of firm is the money management business.

It does not take any more brains or manpower to run $1 billion than it does to run $2 billion. The only real difference is the number of shares the fund management company buys. In the short term, the vast majority of expenses (mostly wages for analysts, portfolio managers and client contact people) are not going to change. Since the fees are generally a fixed percentage of assets under management, if those go up, earnings soar, if they go down, profits plummet.

On the other hand, a retailer will only be able to generate higher revenues if it also buys more inventory to sell. Rising revenues demand rising costs. The extent to which costs are fixed generally determines how sensitive earnings are to changes in revenues. The higher the sensitivity, the more difficult it is to forecast the earnings.

Cyclical vs. Non-Cyclical

In addition, not all companies have equally volatile revenue streams. What makes cyclical firms cyclical is that they have revenue streams that are more sensitive to the economic cycle than non-cyclical firms. If it is easy for the customer to postpone buying the company’s product when times get tough, then that company’s revenues are going to fall a great deal when the economy goes into a recession.

For example, when you are afraid that you might get laid off, you do not go out and buy a new car from Ford (F). Instead, you try to hold on to your old clunker for as long as you can.

However, by the time things get better, there are lots of people who have made the same decision and whose cars are falling apart. Demand gets pent up and then released. When that happens, revenues soar. The auto companies also tend to have a lot of leverage, both financial and operational. Not surprising, then, that earnings for auto companies tend to go from boom to bust and back again.

On the other hand, even if you are worried about your job, you are not going to stop buying toothpaste, shampoo or diapers (if you have a baby). Thus the revenue stream of a Procter & Gamble (PG) tends to be much easier to predict than that of Ford.

Thus if you look at the quarterly estimates, you find that Ford is expected to earn $0.38 per share, and P&G is expected to earn $1.14. However, the ten Ford estimates are all over the map. It has a high estimate of $0.51 and a low of just $0.19. For the full year Ford is expected to earn $1.86, but that average contains a range from $1.29 to $2.21.

In contrast, for the quarter P&G is expected to earn $1.14, but the most bearish analyst out there still expects the company to earn $1.14, while the most bullish is only looking for $1.18. For the year, P&G is expected to pull in $4.28, but the low is $4.15 and the high is $4.35.

Clearly if P&G beats by $0.05 and reports $$1.19, that will convey even more information about the company’s better performance than if Ford also beats by $0.05 and reports $0.43.  That is despite the fact that the percentage beat for Ford (13.2%) would be substantially greater than the percentage beat for P&G (4.4%).

Fun with Standard Deviations

There is a handy measure to gauge the spread of estimates, without relying on the high and low outliers. It is known as the standard deviation, which involves taking the average of the squared differences between the individual estimates and the average. Most of the time, two thirds of the individual estimates will be within one standard deviation of the average, and 95% will be within two standard deviations. The quarterly standard deviation is $0.10 for Ford, but only $0.03 for P&G. The beat would be just 0.5 standard deviations for Ford but 1.66 standard deviations for P&G.

However, even that understates things a bit, since the higher the average the higher the standard deviation. For example, if all the estimates for Ford were ten dollars higher than they are now the standard deviation would still be $0.10. However, that would represent a much tighter consensus than we currently see.

To really be able to compare the spread between two companies one has to adjust for that. The best way of doing that is to look at the standard deviation as a percentage of the mean. This is known as the coefficient of variation (CV). For the quarter for Ford, it is 26.3%, for P&G it is just 2.63% — a ten-fold difference. For the full year, the numbers are somewhat lower, with Ford having a CV of 13.4% and P&G having a CV of just 1.2%, but note that it is still more than a tenfold difference. The lower the standard deviation, the larger the likely price move from a given percentage beat.

Even though tight consensus firms tend to be more sensitive to earnings beats (or misses) than loose consensus firms, as a general rule, they tend to be less risky than loose consensus firms. A wide range of estimates indicates that the companies business model is one that is exposed to a significant amount of variation, that it has a lot of leverage or an unstable earnings stream. Because of the inherent unpredictability of their earnings, those companies tend to trade at lower multiples (holding growth expectations constant) than do tight consensus firms.

If fact, the CV can be a better measure of risk than beta, the most widely used measure of a stock’s riskiness. Even though beta is the cornerstone of modern portfolio theory, it is entirely a backward-looking measure (how much has the stock’s price tended to fluctuate on a daily basis in the past relative to how much the S&P 500 fluctuates that day).

That is more than a bit ironic when the key message of MPT is that one cannot use historical data to outperform the market, since the market will already incorporate that information into stock prices (in more formal terms, the market is efficient with respect to past price moves). The standard deviation of analysts’ estimates, and thus the CV, is an entirely forward looking measure. The lower the coefficient of variation, the more certain the market is about a company’s earnings.

Longer-Term Trends Matter

If a company has consistently grown at a steady pace, say 10% per year, then it generally means that it has a steady source of revenues (say, a high proportion of recurring revenues) and a cost structure that is mostly variable. Under those conditions, analysts are likely to come to similar conclusions about what the company will earn this quarter or this year. The market will generally be more comfortable extrapolating those earnings into the future.

Sometimes that will prove to be a mistake, but it is something the market generally does. To go back to our Ford versus P&G example, I am pretty confident that P&G will be in the black in, say, 2016 and that its earnings that year will be north of $5 per share. Quite frankly, I have no idea what Ford will earn in 2016 — it could just as easily be earning over $8 per share (meaning it is trading for less than 2x 2016 earnings) or it could be bleeding red ink as if it had a severed artery.

Thus between now and 2016, P&G is likely to be a much safer stock than Ford. If Ford does earn $8 per share in 2016, it will far outperform P&G between now and then, if it goes back to bleeding cash, it will far underperform. With P&G you will sit back and collect a nice dividend that will probably grow over time, but not see either massive capital appreciation nor a price collapse. Its price five years from now could well be just around where it is today ($64.75), although I’m a bit more bullish than that on the company. I think it is highly unlikely that Ford will still be going for its current $14.10. It could be going for $7 or it could be going for $30.

More Estimates Help Accuracy

The usefulness of the CV as a measure also depends on having enough estimates in the consensus. If there are three or fewer individual estimates for a quarter or year, then the standard deviation or CV is not going to be all that useful in evaluating a company’s earnings surprise, or for that matter its long-term riskiness. However, for larger companies with more than seven or eight estimates, it is a very important tool for evaluating just how good or bad a company’s earnings report actually is.

Longer-term investors should also pay attention to the spread of the estimates, since in the world of investing, risk is just as important as return. There are several aspects to risk, but the variability of a company’s earnings stream going forward has to be near the top of the list in terms of importance.

If a company has a history of steady earnings growth, or even just of steady earnings and there is not a tight consensus about what the company is going to earn this quarter or this year, that is an important clue that something has significantly changed at the company. So it is important to look not just at how the company’s earnings have fluctuated in the past, but also at the degree of harmony between the analysts covering the stock.

When earnings start to roll in next week — the semi-official start of earnings season is on July 11th when Alcoa (AA) reports — make sure to look at not just how the earnings came in relative to the average estimate, but look at it with the spread around that average as well. The tighter the spread, the more significant any given surprise will be.

ALCOA INC (AA): Free Stock Analysis Report

FORD MOTOR CO (F): Free Stock Analysis Report

PROCTER & GAMBL (PG): Free Stock Analysis Report

Zacks Investment Research

About vitalstocks

This is a sample profile field. Vitalstocks is the operating company for Stockbloghub. This will place the picture of the author or company in the profile. Here is another extra line of information.


Powered by Facebook Comments

Similar Posts: | | | | | | | | Auto Manufacturers - Major | Consumer Goods

RSS feeds: AA | Alcoa Inc. | Central Vermont Public Service Corporation | CV | F | Ford Motor Company | PG | Procter & Gamble Company | Auto Manufacturers - Major | Consumer Goods |

Other Posts by | RSS Feed for this author