Setting the budget for the coming year is a process that follows a similar structure in many organizations. It starts out with a forecasted value being extracted from the ERP-system. The individuals involved in setting the budget then gathers in order to discuss how they think the coming year will deviate from the forecasted value.
There is evidence that a combination of statistical models, such as those applied in the ERP-system, and human adjustments combined are the optimal way of handling forecasting. This is due to the model being able to pick up the general trend while the human adjustment of the forecast is able to pick up upcoming events that deviates from the historic trend. The best way of improving the budget process is thus not to change the structure but to ensure that the implication of biases are minimized.
There are two main problems in this way of handling the manual adjustments of the forecast:
1. Making the adjustment in a group will let social implications bias the adjustments.
2. Human beings are not good intuitive forecasters due to biases related to human decision-making.
So what can we do in order to improve the budget process?
Handling the social implications of the adjustments?
By handling the manual adjustment of the forecast in a group, the decision-makers opens up for a whole new array of biases related to human decision-making. By applying the Delphi-method, meaning that each individual set his/her estimation before entering the budget meeting, biases related to social implications of the forecast are eliminated. This structure could be easily implemented by each individual writing up his/her estimation on a note. Each participant then reads up his/her estimation and argues as to why that is the best forecast.
Handling the biases related to human decision-making?
There are a vast number of biases decreasing the rationality of human decision-making. However, being aware of the most common ones will make you avoid some of the most likely pitfalls.
Anchoring – Describes the human tendency of relying too heavily on the information provided when making decisions. Once a piece of information is accepted as an anchor, individuals are likely to adjust their forecast with that piece of information as a reference point. Anchoring is not wrong per see, but it is important to be aware of what information we use as anchor in our decisions.
Ease of recall bias – Explains the fact that people tend to overestimate the frequency of events that are vivid. For example, people tend to overestimate the risk of being shot or killed in a car-crash compared to the likelihood of dying from poor diet. This bias explains why many of the adjustments made tend to be over-reacting as the adjustments often relates to specific events that are more vivid than the underlying statistical trend.
Regression to the mean – Performance tend to regress towards the mean overtime. People tend to observe a trend, such as high or low sales in one department, and then expect the trend to continue for the coming year. Statistically, these extreme values tend to regress towards the mean in both over- and under performance.
Overconfidence – Means that people are likely to look for information that confirms what they already know and disregard information that is contradictory. As more information is gathered, although faulty, it is used as an anchor for validating new information. This leads to a vicious cycle of illogical confirmation.
Hindsight and the curse of knowledge – This bias refers to the human tendency of overestimating what we knew beforehand by applying later gathered information to a previous situation. This bias explains why people tend to be wise in retrospect.
The curse of knowledge – This bias refers to the assumption that the mental map of others looks the same as our own. This leads to people often failing to give important information as the listener is assumed to possess the same knowledge as the sender.