By Arun Dahal Khatri

In recent years, the way we study and practice economics has undergone significant changes. Especially in universities, the focus has shifted away from thoroughly examining the economic question and its significance. Instead, we often jump straight into building increasingly complex models, as if the sophistication of the math is more important than the clarity of the idea behind it. This trend has created a paradox: economists are becoming more technically skilled than ever, yet less grounded in the fundamental reasoning behind their work. As an undergraduate who enjoys reading economics from the classics—Smith, Ricardo, Keynes, Friedman—to modern works in trade, development, and both macro and micro theory, I have noticed this divide. When I converse with students from different institutions—whether Ivy League, elite liberal arts, or top business schools—I often find that we can perfectly match the mathematical and statistical models. We can code the same regression, run the exact simulation, and even replicate each other’s calibration techniques. But when the discussion shifts to what the question means in economic terms, there is often a pause—a kind of intellectual gap that is harder to fill with equations.

This is not to say I have mastered the art of economic reasoning myself; far from it. But it raises a concern worth addressing: are we training economists to solve problems before we have fully defined them?

The Seduction of the Model

It is easy to see why this shift has happened. Modern economics is data-rich and computation-heavy. The rise of econometrics, big data, and machine learning has given economists a dazzling set of tools. Running a regression on millions of observations or creating a DSGE (Dynamic Stochastic General Equilibrium) model with twenty equations is now a matter of hours, not months. The problem is not with the tools themselves—they are revolutionary. The problem is that they are sometimes used to generate elegant solutions to vaguely defined or poorly understood questions. In other words, we sometimes design the “answer” before we truly grasp the “why” and “what” of the question.

This is what I call solution-first economics. And while it can yield impressive academic results, it often leaves a deeper understanding of the economic problem underdeveloped.

The Consequence: Misinterpretation of Results

One clear danger is that without deep question comprehension, results can be misinterpreted or misapplied. Consider the following example: Imagine a researcher trying to measure the effect of trade liberalization on rural incomes in a developing country. They built a sophisticated econometric model using trade volume, rural household surveys, and wage data. The regression results suggest a positive impact. However, if the researcher never thoroughly explored how rural households engage with trade—perhaps most are subsistence farmers disconnected from export markets—the positive coefficient might reflect a secondary effect, such as remittances or urban spillovers, rather than direct trade benefits.

In such a case, the model is technically correct but conceptually misleading. The “pause” before answering the theoretical question—what is the mechanism here?—was skipped.

The Classical Approach: Understanding Before Solving

The irony is that economics did not begin this way. If we look back at Adam Smith’s The Wealth of Nations, David Ricardo’s Principles of Political Economy and Taxation, or even Keynes’ General Theory, the starting point was always a careful articulation of the question. Smith pondered why nations grow rich. Ricardo examined how comparative advantage works in practice. Keynes asked what drives unemployment during recessions. The mathematical formalism in these works was minimal; the emphasis was on building a logical, coherent framework to interpret economic behavior. Modern economics, in contrast, often works backward: start with a dataset, choose a model, and then retrofit the question to match the tool. This reversal is not inherently bad—it can lead to surprising discoveries—but it risks detaching economics from the human realities it seeks to understand.

An Example from Macroeconomics

Let’s take an example from macroeconomics: the Phillips curve, which describes the relationship between inflation and unemployment.

In the mid-20th century, economists believed the Phillips curve represented a stable trade-off. Policymakers used this model to justify inflationary policies in pursuit of lower unemployment. But in the 1970s, stagflation—high inflation and high unemployment—broke the model. The issue was not the mathematics; the statistical relationship had been accurately described for decades. The problem was the question: was the trade-off stable under all economic conditions? Economists had not fully considered how expectations might shift the relationship, something later addressed by Milton Friedman and Edmund Phelps.

The lesson here is that when the question is incomplete, even the most elegant model can lead to policy mistakes.

A Conversation Between Two Students

To illustrate the gap I often see, imagine this dialogue between two economics students—one from a technical-heavy program, the other from a theory-focused program.

Technical Student: “The regression coefficient for our policy variable is significant at the 1% level. That means the policy works.”

Theory Student: “Works for what? And through what mechanism?”

Technical Student: “Well… the coefficient is positive.”

Theory Student: “Yes, but is that because the policy directly caused the outcome, or because it coincided with other factors? What’s the causal story?”

The technical student is not wrong, but without the theory, the interpretation risks becoming shallow. This is the “pause” I mentioned earlier.

Why This Gap Exists

Several factors contribute to this imbalance between understanding and modeling:

1. Curriculum Design – Many top programs emphasize quantitative skills to make graduates competitive in consulting, finance, and data-driven policy roles. The incentive is to teach the most marketable tools.
2. Research Culture – Academic publishing often rewards methodological novelty over theoretical depth. A paper with a cutting-edge model can outshine one with a beautifully argued but less technical theoretical framework.
3. Data Availability – With massive datasets now accessible, it is tempting to start with the data rather than the question.
4. Time Pressure – Students and researchers are under pressure to produce results quickly. Deep conceptual work takes time, which can feel like a luxury.

Economics as Both Science and Art

Economics is often called a science, and it is. But it is also an art—the art of framing questions about human behavior, markets, and institutions in a way that captures their complexity without losing clarity. Models are our paintbrushes, data our colors, but the question is the canvas. Without the canvas, the rest has no purpose. As a young economist, when we train economists to value precision over comprehension, we risk creating a generation of practitioners who can execute flawless technical work without truly understanding the human and institutional dynamics their work is meant to illuminate. As students, researchers, and educators, we must resist the temptation to equate complexity with insight. A beautifully complex model can be brilliant, but only if it answers a well-understood question. Likewise, a simple model can be profound if it captures the essence of the problem. In the end, economics is not about the elegance of the equation—it is about the clarity of the question. And clarity begins not with coding, calibrating, or optimizing, but with pausing long enough to ask: Do we understand what we are trying to answer?