RIZAL TECHNOLOGICAL UNIVERSITY

City of Mandaluyong and Pasig City

Economic Forecasting

Submitted by:
Ortilla, Mischelline Star M.

Submitted to:
Engr. Danfort Raven Inocencio

BSECE/CEIT-29-501E

1st Semester A.Y. 2024 – 2025

Economic forecasting is a critical tool for policymakers, businesses, and analysts to

predict future economic conditions based on historical data and current economic indicators. The

purpose of forecasting is to inform decisions regarding fiscal policy, business strategies, and

investment planning. Forecasting methodologies vary widely, from simple models based on past

trends to more complex approaches involving econometric models and artificial intelligence

(AI). The accuracy of these predictions plays a significant role in mitigating risks and optimizing

resource allocation.

This research provides an overview of economic forecasting methods, the challenges

faced in the field, and emerging trends in predictive techniques.

PROCESS

Economic forecasting involves several steps, each contributing to the overall accuracy

and relevance of the predictions:

1. Data Collection:

Economic forecasts are based on a variety of data sources, including government

statistics, business surveys, financial indicators, and international economic reports. Key

indicators such as GDP growth, unemployment rates, inflation, interest rates, and

consumer confidence are central to forecasting models.

2. Model Selection:

There are different types of forecasting models used in economics. These include:
 ○ Time Series Models: These models, such as ARIMA (Auto-Regressive

Integrated Moving Average), rely on historical data to predict future trends.

○ Econometric Models: These models consider the relationships between various

economic variables. They use statistical methods to model the impact of one

variable on another, for example, how changes in interest rates can affect inflation

or GDP growth.

○ Machine Learning Models: With the rise of AI, machine learning models are

gaining popularity for their ability to analyze large datasets and uncover complex

patterns not easily discernible by traditional models.

3. Model Validation:

After developing a model, forecasters test its predictive power by comparing its output

against real-world outcomes. This validation process is crucial for determining whether a

model is reliable enough to be used for decision-making.

4. Forecasting and Monitoring:

Once validated, the model is used to generate forecasts, which are regularly updated as

new data becomes available. Continuous monitoring and recalibration of models are

necessary to account for economic shifts and unexpected events, such as recessions or

financial crises.
OBSERVATION

Several studies and articles have highlighted the strengths and weaknesses of different

economic forecasting methods:

Accuracy of Traditional Models:

While time series and econometric models have been staples of economic forecasting for

decades, they are often criticized for their inability to account for sudden, exogenous shocks,

such as the COVID-19 pandemic. A study by Stock and Watson (2007) found that while these

models can be reliable in stable economic environments, they tend to perform poorly during

times of volatility.

Machine Learning in Forecasting:

In recent years, machine learning has shown promise in improving forecasting accuracy. Models

like neural networks and support vector machines can process vast amounts of data and identify

non-linear patterns that traditional models may miss. A study by Candelon and Lütkepohl (2021)

demonstrated that machine learning techniques, particularly deep learning, were able to

outperform classical models in predicting macroeconomic variables like GDP and inflation.

The Role of Big Data:

The increasing availability of big data has expanded the tools available for forecasting. Real-time

data from sources like social media, credit card transactions, and satellite imagery can provide

more granular insights into economic activity. For example, a study by Choi and Varian (2012)

showed that Google search data could be used to predict economic indicators like retail sales and

unemployment claims with a high degree of accuracy.

Despite advancements, forecasting remains a highly uncertain exercise. Models often struggle

with predicting rare, extreme events (black swan events), and their accuracy declines when there

are significant changes in the underlying economic structure. Additionally, forecasters often face

challenges in interpreting complex data due to factors like human bias, data noise, and imperfect

information.

CONCLUSION

Economic forecasting plays an essential role in decision-making, but it is an inherently

uncertain process. While traditional models have their merits, the growing availability of big data

and advancements in machine learning provide opportunities for more accurate predictions.

However, even with these new tools, there is no foolproof method for forecasting, as unforeseen

events and complex variables can lead to large errors. The ability to dynamically adjust forecasts

and maintain flexibility in response to changing economic conditions is key to improving the

effectiveness of economic predictions.

RECOMMENDATION

To improve economic forecasting:

1. Adopt Hybrid Models:

Combining traditional econometric models with machine learning techniques could yield
 more accurate predictions by leveraging the strengths of both approaches. Hybrid models

are increasingly used in practice and have shown promise in improving forecast accuracy.

2. Incorporate Real-Time Data:

The use of real-time data, including big data from alternative sources, could enhance the

responsiveness of forecasts to sudden economic changes. Forecasting models that

incorporate non-traditional data, such as online activity, can provide more timely insights.

3. Develop More Robust Uncertainty Quantification:

Forecasters should focus on improving the ability of models to quantify uncertainty. By

providing a range of possible outcomes and their probabilities, models can help decision-

makers understand and manage the risks associated with economic uncertainty.

4. Improve Model Transparency and Explainability:

As machine learning and AI models become more integrated into economic forecasting,

it is crucial that these models are interpretable. Policymakers and businesses need to

understand the rationale behind predictions to make informed decisions.

-Stock, J. H., & Watson, M. W. (2007). "Why has US inflation become harder to forecast?"

Journal of Money, Credit, and Banking, 39(1), 3-33.

-Candelon, B., & Lütkepohl, H. (2021). "Machine learning techniques in macroeconomic

forecasting: A review." Journal of Economic Surveys, 35(2), 429-461.

-Choi, H., & Varian, H. (2012). "Predicting the present with Google Trends." Economic Record,

88, 2-9.

-Hendry, D. F., & Clements, M. P. (2004). "Economic forecasting: A retrospective and

prospective view." Journal of Applied Econometrics, 19(3), 235-254.