When Consensus Gets It Wrong: Measuring Financial Uncertainty with Monte Carlo Simulation

When Consensus Gets It Wrong: Measuring Financial Uncertainty with Monte Carlo Simulation

G. Monray on earnings, share prices and the hidden uncertainty behind financial forecasts

Introduction

Financial markets are built around expectations.

Analysts forecast revenues and earnings. Investors incorporate those expectations into valuations. Companies communicate guidance based on anticipated business performance. Yet a forecast is not a fact—it is an estimate surrounded by uncertainty.

That distinction was at the heart of Jorge Monray's 2024 study, Financial Uncertainties in EPS and Share Prices: A Monte Carlo Study on U.S. Business Services MNCs, published in the International Journal of Science and Research.

The study examined 30 U.S. multinational corporations in the Business Services global industry, using consensus revenue, earnings-per-share (EPS) and share-price expectations. The research applied Monte Carlo simulation with 20,000 iterations at a 99% confidence level to examine the uncertainty surrounding those expectations.

The results produced a striking contrast. For the companies analyzed, the simulation suggested that consensus revenue expectations were relatively conservative, while the consensus expectations for EPS improvement and share prices appeared considerably more optimistic. The study therefore argued for greater caution when interpreting analyst consensus forecasts.

Q: What was the central question behind this research?

G. Monray: The starting point was a very practical financial question: how certain are the earnings and share-price expectations that investors and analysts commonly use? A consensus forecast is useful because it aggregates the expectations of analysts and financial professionals. But consensus is still an expectation. I wanted to introduce a different perspective: instead of looking only at the expected number, what happens when we explicitly model the uncertainty surrounding that number? That is where Monte Carlo simulation becomes useful. It allows us to move from a single forecast to a distribution of possible outcomes.

Q: Why did you focus specifically on EPS and share prices?

G. Monray: Because EPS and share price expectations are closely connected to how investors assess corporate performance. Revenue tells us something about the scale of the business. EPS goes further because it relates corporate earnings to the number of shares outstanding. Share price, meanwhile, reflects the market's valuation of the company's expected future performance. So we were dealing with three different layers: revenue → earnings → market valuation. The interesting question was whether the level of uncertainty remained similar as we moved through those layers.

Q: What companies did you analyze?

G. Monray: The research focused on the U.S. Business Services global industry. The original universe contained 43 U.S. companies, but data sufficient for the simulation was available for 30 of them. Those 30 represented approximately 69.7% of the U.S. representation in the Forbes Global 2000 within the group identified for the study. The companies included firms such as UPS, FedEx, Capital One, Automatic Data Processing, Cintas, Moody's, Paychex, Equifax and other large U.S. business-services corporations. The objective was not to construct a model for every U.S. company. It was to examine uncertainty within a defined multinational business-services universe.

Q: Where did the underlying expectations come from?

G. Monray: The study used consensus information collected from S&P and Xignite in 2023, covering expectations for the subsequent 12-month period. The variables included consensus revenue, consensus EPS and consensus share-price expectations, together with their respective ranges. This is important because the simulation was not attempting to replace analyst expectations with an entirely independent valuation model. Instead, it took the available consensus expectations and asked: How uncertain are these expectations when we simulate a large number of possible outcomes?

Q: Why did you choose Monte Carlo simulation?

G. Monray: Because uncertainty is not well represented by a single point estimate. Suppose an analyst says that a company will generate a particular EPS next year. That number looks precise, but the underlying reality is not precise. Revenue can differ from expectations. Costs can change. Margins can move. Economic conditions can change. Market sentiment can change. Monte Carlo simulation allows us to represent these uncertainties through repeated random simulations. In this study, we ran 20,000 iterations, using a 99% confidence level. The objective was therefore not to predict one exact future number. It was to examine the range and probability of possible outcomes.

Q: What did the revenue analysis show?

G. Monray: This produced one of the most interesting results. The aggregated consensus revenue for the companies analyzed was $454.01 billion for the 2024 forecast period. The simulation produced a range of approximately $494 billion to $503 billion at the reported confidence level. Within the model, this suggested that the consensus revenue estimate was relatively conservative compared with the simulated outcomes. That is important because the conclusion was not simply that "consensus is wrong."

Rather, the simulation indicated that, under the assumptions used, the distribution of possible revenue outcomes was shifted above the consensus estimate.

Q: Were all companies contributing equally to that uncertainty?

G. Monray: No, and this is where sensitivity analysis becomes particularly useful. The analysis indicated that UPS and FedEx were the two largest contributors to the volatility in the consensus-revenue simulation, accounting for approximately 41% and 38%, respectively. This tells us something important about aggregated forecasts. An industry-level result can sometimes be driven disproportionately by a small number of companies.

Therefore, when interpreting a sector-level forecast, it is useful to identify which individual companies are contributing most strongly to the overall uncertainty.

Q: What happened when you moved from revenue to EPS?

G. Monray: The picture became considerably more uncertain. The average consensus EPS across the 30 companies was $8.03 per share. The Monte Carlo analysis indicated approximately a 61% probability of improving on those expected results under the simulation framework. In other words, the probability of exceeding the consensus expectation was far from certain.

This illustrates an important principle in financial forecasting: even when revenue expectations appear relatively achievable, translating revenue into earnings introduces another layer of uncertainty. Margins, operating costs, financing costs and other factors intervene between revenue and EPS.

Q: Which companies had the greatest influence on EPS volatility?

G. Monray: The sensitivity analysis identified Capital One as the largest contributor, at approximately 42%, followed by FedEx at nearly 19%. That finding reinforces the value of sensitivity analysis. A model can tell you that uncertainty exists. Sensitivity analysis helps tell you where that uncertainty is coming from. For management and investors, that can be more actionable than simply knowing the final probability distribution.

Q: And what happened with share prices?

G. Monray: This was arguably the most striking part of the research. The estimated average share price for the 30 companies was $198.98. After the simulation, the study reported only a 15% probability that the simulated result would be at or above the expected level. The interpretation in the paper was that the expert consensus on share prices appeared comparatively optimistic.

This result should be interpreted carefully. Share prices are influenced by considerably more than corporate earnings. Investor expectations, interest rates, risk premiums, market sentiment and valuation multiples can all affect prices.

Therefore, the finding should not be read as a claim that the market must fall. It indicates that, within the study's simulation framework, the consensus share-price expectation carried substantial uncertainty.

Q: Why does uncertainty increase as we move from revenue to share price?

G. Monray: Because each stage introduces additional assumptions. Revenue is relatively close to the underlying operating activity of a company. EPS depends on revenue but also on costs, margins, financing and the number of shares. Share price introduces another layer: how the market values those expected earnings.

Conceptually, we can think of the process as: Revenue uncertainty → earnings uncertainty → valuation uncertainty.

The further we move away from the underlying operational activity, the more assumptions enter the analysis. That is why financial forecasts should not be treated as equally certain simply because they are presented as precise numbers.

Q: Is this a criticism of financial analysts?

G. Monray: Not necessarily. Analyst consensus serves an important function. It aggregates information and provides investors with a reference point. The purpose of the research was not to suggest that analysts are systematically incapable of forecasting. The point is that consensus itself has a probability distribution around it. A consensus figure of $8.03 EPS, for example, should not psychologically become an $8.03 certainty. Monte Carlo simulation provides a way of making that uncertainty visible.

Q: What is the biggest danger when managers use consensus forecasts?

G. Monray: The danger is false precision. When a financial model produces a number with two decimal places, it can create the impression that the future is known with great accuracy. But the precision of the number does not necessarily reflect the precision of the underlying information. The more appropriate approach is to ask:

• What assumptions produced the forecast?

• What is the range of possible outcomes?

• How sensitive is the result to individual companies or variables?

• What is the probability of exceeding or falling below the consensus?

• Which assumptions contribute most to the uncertainty?

That turns forecasting from a single-number exercise into a risk-analysis exercise.

Q: What does this mean for financial managers?

G. Monray: It means that forecasting should be connected with probability and sensitivity analysis. A CFO should not only ask, "What is our expected EPS?" The better questions are:

• "What is the probability distribution around that EPS?"

• "Which assumptions have the greatest influence?"

• "What happens if revenue is different from expectations?"

And perhaps most importantly:

"How much confidence should we actually place in the forecast?" Monte Carlo simulation provides one practical way of answering those questions.

Q: What is the broader contribution of this study?

G. Monray: The broader contribution is the application of uncertainty analysis to a very practical financial problem. We often talk about forecasting as though the objective were to discover the correct number. But the future does not provide us with one predetermined number that we simply need to find. There is a range of possible outcomes. The value of Monte Carlo analysis is that it allows financial analysts and managers to work explicitly with that uncertainty rather than hiding it behind a single forecast.

Q: How does this research connect with your earlier work?

G. Monray: There is a strong methodological continuity. Earlier research examined corporate risk using Monte Carlo simulation and sensitivity analysis, including the relationship between customer concentration and financial uncertainty. Here, the same general philosophy is applied to a different financial problem. The question moves from "How does the structure of a company's customer base affect risk?" to "How uncertain are the financial expectations surrounding large multinational companies?"

In both cases, the objective is similar: turn uncertainty into something that can be measured, simulated and analyzed. The most important message of the research is not that consensus forecasts are necessarily wrong. It is that a consensus forecast is still a forecast.

For the 30 U.S. Business Services multinationals examined, the Monte Carlo analysis produced materially different probability profiles for revenue, EPS and share-price expectations. Revenue simulations were above the aggregated consensus estimate, while the probability of improving on the EPS consensus was reported at 61%, and the probability of reaching or exceeding the average share-price expectation was reported at only 15%.

These results illustrate how uncertainty can change as we move from operating performance to earnings and ultimately to market valuation.

The study therefore advocates a more cautious approach to financial forecasting: not because forecasts are useless, but because the probability surrounding the forecast can be as important as the forecast itself.

Publication

Monray, J. (2024). Financial Uncertainties in EPS and Share Prices: A Monte Carlo Study on U.S. Business Services MNCs. International Journal of Science and Research, 13(9), 1591–1596. DOI: 10.21275/SR24926085653. The article was published in September 2024.

Methodological note: the findings reflect the companies, consensus data and assumptions used in the 2024 study. They should not be interpreted as a current 2026 forecast for these companies or for the U.S. Business Services sector.