Good news! According to equity and quant strategist Savita Subramanian's team at Bank of America Merrill Lynch (BAML), third-quarter earnings are coming in above expectations. Two weeks into the season, just over a third of the S&P 500, or 114 companies, have reported, and so far the index is killing it: 74% have beat earnings per share (EPS) expectations, 59% have beat on sales, and 49% have beat on both. Bottom-up EPS numbers are around 1.6% higher overall than analysts had forecast. BAML sees profits beating by over 2% when all is said and done.

That's good news if you were long going into the quarterly festivities, but doesn't all this good news sound familiar? On August 2, for example, the Wall Street Journal reported that analysts polled by FactSet had revised their forecast for the change in second-quarter earnings upwards by 1.5 percentage points after Facebook Inc. (FB) and Alphabet Inc. (GOOG, GOOGL) "smashed Wall Street's expectations."

Low Bar

Previous earnings seasons tell a similar story of companies blowing analysts' forecasts out of the water because companies almost always beat earnings expectations. That does not necessarily apply to individual companies: any given quarter will see a number of firms fall short on EPS, revenues or both. But overall, the majority of firms report better-than-expected results. (See also, Earnings Shocks: Stock Volatility Is Soaring.)

According to Bespoke Investment Group co-founder Paul Hickey, the average EPS "beat rate" since 1999 is 62%. The figure is even higher for the tech sector. As of October 21, the rate for the current earnings season was 68% (the report cautions that that rate is "unlikely" to hold); 58% of companies are beating on revenues this quarter.

Source: Bespoke Investment Group

BAML's historical rates, based on the period since 2000, paint a slightly less skewed picture: they have 53% of companies beating on EPS, 56% beating on revenues and 35% beating on both.

These historical averages raise an uncomfortable question: why do analysts consistently lowball EPS and revenue forecasts? If the aim is to accurately predict results, wouldn't analysts bump up forecasts a bit after a few quarters – or years even – of guessing too low? Judging by Bespoke's data, it's more common for 70% of companies to beat consensus estimates than for 50% or fewer to do so. (See also, Will 2016 Q3 Earnings Be in Line With Estimates?)

Why Prove Yourself Wrong?

Academics have looked at this question and come up with some interesting theories. Byoung-Hyoun Hwan of Cornell University and Korea University, Baixiao Liu of Florida State University, and Dong Lou of the London School of Economics and CEPR looked at the incentives that could potentially be contributing to "'Consistent' Earnings Surprises," the title of a May 2014 draft paper describing their findings.

Their hypothesis, which they say the data back up, is simple enough. Say that an analyst has a bullish rating on Super Corp. stock (to be clear, this is our scenario, not the authors'). The analyst's in-depth research and rigorous critical thinking brings them to the conclusion that Super Corp. will report EPS of $0.80 in the third quarter, a significant rise over the prior-year quarter's results of $0.65 per share. But there's always an element of uncertainty, which puts the analyst in a bind. If Super Corp.'s earnings come in at 70-something cents, the share price could fall, and the analyst has needlessly set up the bullish prediction to look inaccurate.

Shaving Your Estimate

So if you feel good about a stock, why not fudge the estimate downwards a bit? No one will blame you for expecting results that are slightly less stellar than the actual results. To the contrary, the stock price will likely rise, vindicating your bullish prognosis. (See also, Understanding Analyst Ratings.)

Here's how the paper's authors put it: "We hypothesize that analysts with a bullish stock recommendation have an interest in not being subsequently contradicted by negative firm-specific news. As a result, these analysts report downward-biased earnings forecasts so that the company is less likely to experience a negative earnings surprise." (See also, How to Evaluate the Quality of EPS.)

Interestingly, this tendency to low-ball bullish forecasts has a corollary: "analysts with a bearish recommendation report upward biased earnings forecasts so that the firm is less likely to experience a strong positive earnings surprise."

Optimistic Tendency

In general, analysts tend to be bullish. Adding up company-specific price targets earlier this month, FactSet analyst John Butters found that the overall expectation was for a 9.8% rise in the S&P 500 over 12 months. Few expect that to happen, and it is unlikely that many of the analysts who unwittingly contributed to that bullish target by setting high price expectations for individual firms believe the S&P will climb to nearly 2,400 so soon. Still, it is an interesting finding for armchair market-psychologists to chew on, demonstrating that on the whole, analysts tend to be optimistic about the firms they cover – perhaps unrealistically so.

Consistently too-high price targets (Butters found similar results looking at past months' price targets) would seem to clash with consistently too-low earnings estimates, but if you believe Hwang et al, these patterns mesh perfectly. When you're bullish on a stock, you have a counterintuitive incentive to low-ball earnings estimates. Otherwise you set the company up for a miss, which means a falling stock price, which means lower accuracy for you.

Want to learn how to invest?

Get a free 10 week email series that will teach you how to start investing.

Delivered twice a week, straight to your inbox.