When Everything is Endogenous, Econometrics Breaks Down
Inflation has fallen recently, and the standard explanation is economic weakness. Yet lower inflation should boost demand under basic economic principles, strengthening the economy.
Law Ka Chung is a Hong Kong economist and financial columnist.
Inflation has fallen recently, and the standard explanation is economic weakness. Yet lower inflation should boost demand under basic economic principles, strengthening the economy.
Fresh shocks may alter data and parameter estimates, but the core model endures — the hallmark of robust modelling. Nothing is new under the sun; resist the temptation to reinvent from scratch.
The two welfare theorems show that equilibrium implies optimality, and any optimal outcome can be achieved as equilibrium through redistribution. This supports quantity-based policies over price manipulation, yet governments often choose the latter, undermining efficiency.
When decomposing data to explain causality, the key is ensuring $\mathbf{X}'\mathbf{e} = \mathbf{0}$ — no noise in the signal, no signal in the noise. This condition underpins valid statistical inference and model credibility.
Macroeconomics builds on micro foundations but isn’t just micro writ large. Modern models, from growth to cycles, require dynamic optimisation and rational expectations—complex tools for a complex economy. Whether AI can break this decades-long mold remains an open question.
Why does capitalism work? Beyond empirical evidence, this is the formal mathematical proof: if preferences are reflected in prices, the market delivers a globally optimal allocation—no one can gain without someone else losing.
Econometrics isn’t just about proving causality—it’s about approximating truth. Through residual patterns, we test if a model holds. As with physics, it's less about certainty than refining the assumptions.