THANDO NGOZO: Navigating bias and noise for optimal fiscal policy

Bias and noise can markedly influence the decision-making process in public policy arenas, notably in fiscal policy matters. This can lead to weak economic expansion, increased disparities in wealth distribution between the rich and the poor, and eroded adherence to authority. For fiscal policy to cultivate awareness, equitability and utility, policymakers must guard against bias and noise. 

Our ability to predict events is extremely limited and we are not very good at it. However, this does not stop us from making forecasts and convincing ourselves they are accurate. We get a pleasant feeling of certainty by telling a coherent story about why events occurred. When our narrative is simple and compelling, we become blind to alternative explanations. Because of hindsight biases, we are constantly rewriting history and persuading ourselves that we knew all along what would happen. 

Most of us are taken aback by the revelation that our capacity to foresee future events is generally inadequate compared with the predictive accuracy of a simple rule or formula. In fact, sophisticated algorithms can make better predictions than us with access to less information than the average person. The critical advantage of rules and models is that they are noise-free. 

Judgment is delicate and intricate in how we perceive it. It is not easy to identify the subtlety as noise. Despite this well-established truth we find it hard to believe that mindless compliance with simple rules will often achieve more accuracy than we can. We readily accept that psychological biases result from systematic error, but we find it hard to admit that it also causes randomness. 

Public policy decisions can be influenced in many ways. In SA fiscal policy decisions are influenced by factors including ideological, political, cognitive and informational biases. These biases can lead policymakers to advocate for specific perspectives and interests, potentially resulting in favouritism towards certain groups, neglect of other communities, and adherence to outdated economic models and assumptions that do not align with economic reality. 

Biases can cause public resources to be distributed to projects and programmes that profit particular groups or regions at the expense of others. This misallocation of resources aggravates prevailing income disparities. Political connections influence fiscal policy decision-making, resulting in partisan dealings with individuals or businesses closely linked to government officials. It can easily culminate in government contracts, subsidies and other benefits being distributed to politically well-connected elites at the expense of fair competition and equitable distribution of resources. 

Stimulus packages

Biases may also be evident in tax policies that favour affluent individuals or corporations. Due to the various tax benefits and privileges they receive, their tax rates can be substantially lower. This can fuel income disparities and erode social and economic justice efforts. 

Biases can influence how economic stimulus packages are conceived and executed, leaning towards politically powerful industries or areas instead of those that truly require assistance. Some sections of the economy or specific sectors can gain advantages through special subsidies and exemptions from the rules or other types of fiscal support. This special treatment alters markets, stifles competition and prevents new and potentially innovative sectors. 

Noise is the random variations or inaccuracies that could distort the decision-making process and outcomes. In the context of fiscal policy, noise results from inaccurate forecast models, inadequate data, significant market uncertainties, behavioural biases and even psychological factors such as groupthink or herd mentality among policymakers. Noise creates widespread damage by leading to irrational and suboptimal policy choices. 

When most of our mistakes in a series of verdicts go in one direction, we have bias. Bias represents the average of errors, such as when budgets consistently favour certain sectors over others. However, removing bias from judgments, while desirable, is insufficient to eliminate errors. The remaining errors are not shared in the judgments among the differing individuals, but are individually idiosyncratic. Such remaining errors after removing bias constitute the unwanted variability in judgments about the same thing — they are noise.

Noise can occur when government funding for infrastructure projects is unevenly distributed. Though efforts are made to remove bias from project selection, individual decision-makers may introduce noise by favouring projects not for objective reasons but based on personal preference. Though the tax policy is designed to be fair and impartial, variations in how different tax officials apply tax incentives can lead to noise, resulting in inconsistency and inefficiency in the fiscal system.

Policymakers should adopt a comprehensive plan to tackle bias and noise in fiscal policy decisions in SA. This plan should prioritise disclosure, a sense of responsibility, a wide range of views, evidence-based analysis closely related to policy practice, and commitments to the hand-in-hand relationship between policymakers and their principals, the so-called stakeholders. Some key strategies that can help mitigate bias and noise in public policy include:

• Policymakers should nurture a climate of variety and incorporation within fiscal policy execution to enable the consideration of an extensive spectrum of perspectives, narratives and abilities. They should promote and support multidisciplinary collaboration and involve various stakeholders to dispute presumptions and reduce deficiencies. 
• Policy should be formulated and analysed using strong evidence, data analytics and impact assessments. Investment in research and programme-building initiatives should be prioritised to improve the quality of economic analysis and forecasting in government. 
• Resilient strategies should be established for scenario planning and risk management, including uncertainties, potential biases and alternatives in decision-making. Extreme conditions and subjects of interest should also be tested to see how these strategies react to distressing exertion. 

By implementing these strategies in the policy-making process SA may potentially uplift the calibre, legitimacy and realisation of fiscal policies, which in turn may contribute to sustaining stability, social unity and economic prosperity.

• Ngozo, a research manager at the Financial & Fiscal Commission SA, writes in his personal capacity.