Buckets, weights, and the tyranny of the marginal driver

Two arguments about macroeconomic data collide in this dossier, and both deserve a hearing. One says the aggregates — GDP, CPI, the growth composites — are conceptual fictions that give policymakers the illusion of steering. The other says that whatever the philosophical objections, the aggregates are what price assets, and the discipline lies in bucketing them, weighting them, and reading the cross-asset tape for the marginal driver. The practitioner has to inhabit both positions simultaneously. That is uncomfortable, and it is the point.

Start with the operational claim, because it is the one a portfolio can be built on. The framework advanced by Alfie Kerswell on the Pascal Hügli material is to bucket growth and inflation indicators, assign explicit weights — NFP (non-farm payrolls) at twenty percent, consumer spending at twenty-five percent — and then track both the absolute level and the speed of change across one-, three-, and six-month windows. The point of the weights is not that they are correct in some deep sense. The point is that they are fixed in advance

, which forces the analyst to be wrong in a legible way rather than wrong in a discretionary fog.

The second leg is liquidity, treated as a function of real rates rather than as a separate pillar to be nowcast from central-bank balance sheets. That is a defensible simplification: if real rates are the price of capital in inflation-adjusted terms, most of what practitioners call liquidity conditions collapses into that single variable. The third leg — cross-asset correlations, lead/lag structures, and dispersion

— is where the framework earns its keep. Kerswell's formulation is that the marginal driver of any asset is discoverable from what it is correlating with today, and that the extreme cases are already priced by the time they are visible. The edge is in the transition, not the destination.

Mispricing is pretty much me saying that my view on growth and inflation is correct and that the market's wrong.

The Austrian objection, taken seriously

Against this stands a sharper critique, aired through Quoth the Raven and voiced by Frank Shostak and, further back, Rothbard and even the BEA's own Landefeld and Parker: the aggregates are not measurements but constructions. Real GDP, as Landefeld and Parker conceded, cannot in principle be observed. You cannot add apples and oranges to obtain fruit, and you cannot add two dollars for bread to one dollar for milk and divide by two to obtain a meaningful price. Shostak's stronger claim is that these fictions are not merely imprecise but performative: policy acts on them, entrepreneurs must anticipate that policy, and the resulting feedback loop distorts the very signals the aggregates were supposed to summarise.

“Statistics are the eyes and ears of the bureaucrat, the politician, the socialistic reformer.”

— Murray Rothbard

The dossier offers no quantified forecasts to arbitrate between these views, and the reader should treat the absence honestly — this is a framework cluster, not a call cluster. Our read is that the two positions are not actually in conflict at the desk level. If the aggregates are performative, then trading them systematically is trading the policy reaction function, which is exactly what a weighted growth-and-inflation composite plus a real-rates liquidity read is designed to do. Kerswell's insistence that a bearish growth view without a time horizon can be wrong for three years is the practitioner's version of Shostak's warning: the aggregates move policy, policy moves prices, and the analyst who refuses to quantify the horizon is not being philosophically pure, only unhedged.