My Framework

Tyler Cowen, Scott Sumner, and Arnold Kling have all given their macro frameworks. It’s a good exercise, even if I’m a little late to the party. So, here are the basic propositions with which I approach the world.

A few notes beforehand: mine starts a bit further back than the examples so far. I’ve tried to be more systematic; the list starts with epistemology and ends with macroeconomics. It might seem like a mismatch of momentousness, but since I’m a specialist to some extent, I’ve tried to organize the points to reflect an approximately equal amount of mental energy spent on each major point, with a broad trajectory from general to specific.

1. The universe is intelligible.

• On its face, this is a statement about the mind, not about the universe.
• Intelligibility is defined as exhibiting order, or patterns. Nevertheless, I don’t think order is an entirely subjective concept, so I take the statement to imply something about the universe as well.

2. The language faculty – i.e. the recursive manipulation of arbitrary symbols – is the decisive difference between human and animal consciousness.

• It is not, however, a “finely tuned” symbolic logic system like a computer; rather, a bundle of heuristics which act “as if” they were a logic machine.
• Recursive logic is unique to human cognition, but not necessarily the dominant feature all the time. Heuristics can be exploited, though in high stakes games they can be deliberately overridden (see below, “demand curves slope downward”)

3. The fact-value distinction is irreducible.

• Perception is filtered and structured by pre-conscious judgements about the significance of various aspects. This judgement (“theory”) is not essentially different from value judgements which operate on the conscious level.
• Raw factual data, therefore, never speak for themselves. They must be interpreted – i.e. significance must be imputed to them. It is better to do this explicitly than implicitly.

4. Variation and selection are necessary and sufficient to explain complex order.

• Value judgements about any particular order cannot be coherently made without 1) a value judgement on its selection mechanism (what it selects for and how strongly), and 2) a specification of the alternative.
• The rules which govern a complex order can themselves arise within another complex order. Biology ultimately reduces to chemistry – at least in principle; it may be impossible in practice to interpret an order on the basis of its “parent” order. Still, in terms of the patterns of particular elements, the containing order must necessarily be more complex than the contained. This is why conscious design – the result of several levels of complex order – is severely limited in its potential complexity compared to higher levels of complex order.

5. Wherever possible, we should prefer to rely on known ordering forces rather than conscious planning for social arrangements.

• Markets select strongly for the satisfaction of consumer wants. The deliberate abrogation of market outcomes, on the other hand, sets in motion an ordering process which is poorly understood, highly unpredictable, and fails to select for desirable results.
• In other words, apparatus cost is the decisive argument against most policy, regardless of its particular merits.

6. Demand curves slope downward; supply curves slope upward.

• In other words, “Incentives matter”. An increase in the cost of some activity results in a marginal reduction in that activity; an increase in the return to some activity results in a marginal increase in that activity.
• In economics, “cost” and “return” must be understood subjectively. However, these notions can be used in an “as if” manner to illuminate the workings of selection mechanisms in general, even without intentionality on the part of the things being selected for (e.g. why game theory works in evolutionary biology).

7. The frictions in a selection process are just as analytically important as the selection process itself.

• For example, the concrete forms of economic order are shaped by the form of transaction costs. This makes sense if the selection mechanism operates on the basis of cost, inclusive of frictions – as it usually does, in some form or another (see #6).
• Indeed, the very idea of frictions as distinct from constraints, or transaction costs as distinct from costs more generally, requires the analyst to hold the structure of the higher-level order fixed (see #4). This may or may not be a legitimate assumption, but again it is better to state the relevant alternative explicitly rather than implicitly.

8. Economic organization, not (just) technical progress, is the key to economic growth and development.

• Advanced technology requires an extremely fine division of labor.
• All the technological knowledge in the world will not save a civilization from collapse if it should fail to organize itself along low-transaction-cost lines – hence the “dark ages” following the collapse of Rome’s money economy (see the next point).

9. The two most significant transaction-cost-reducing innovations in human history so far have been liberalism and monetary exchange.

• A well-functioning, stable, and standardized monetary system substantially reduces haggling costs and the necessity for trust in making anonymous exchange.
• Liberalism – both as a policy norm (the rule of law) and a set of personal norms (the commercial virtues; respect for others’ rights to life, liberty, and property) – expands the pool of trustworthy trading partners to a historically unprecedented level.
• Modern civilization is organized at such a degree of complexity that both of these elements are likely necessary to keep transaction costs low enough to sustain it.

10. Social norms are generally not rationally justifiable; they must be accepted either tout court or on the basis of a mythology.

• Social norms allow social selection mechanisms to operate with a minimum of friction by legitimizing their results. Liberalism, for example, reduces frictions by legitimizing the market as a selection mechanism.
• The mythology of liberalism (natural rights) being now widely recognized as a mythology, the market continues, albeit with increasing frictions. The greatest practical threat is rent-seeking (an alternative selection mechanism) and leftism in its various strands, broadly speaking (a set of social and policy norms with their own mythologies, which legitimizes certain forms of rent-seeking)

11. Increasing returns are the primary explanandum in economic development and international trade.

• Hence the uselessness of the Solow model. The division of labor is the primary driver of increasing returns (see #8).
• Increasing returns take the form of mutual feedback between two elements (for example, the volume of trade and the division of labor). They therefore do not imply the feasibility of a “big push”: if one element proceeds too far out of lockstep with the other, the result will most likely be waste.
• Institutional path dependence – being stuck at a local optimum, lacking some aspect of liberalism or low-friction monetary exchange – is a huge, and perhaps the main, obstacle to development worldwide.

12. Say’s Law works.

• Economic progress is driven by increases in the supply of goods; not the demand for goods (which in turn is driven by #8). An economic theory which gets this backward will be hopelessly confused on nearly every important macroeconomic issue.
• Prices do not adjust instantly, so there is some room for short-run demand-driven fluctuations. To the extent an economy is characterized by liberalism and low-friction monetary exchange however, the short-run will be tolerably short.

13. Short-run demand fluctuations depend only on the total volume of spending.

• These probably characterize most depressions and bubbles.
• If these are to be stabilized (i.e. monetary pressure on prices avoided), monetary policy can do so far more effectively and with a lower apparatus cost than fiscal policy. The former can engineer a “pure” short-run demand boost; the latter is primarily redistributive. Consumption spending is not an important economic magnitude for nearly any question.
• NGDP targeting is probably a modest improvement over the current monetary regime in this respect. Gross Output targeting is probably even better, to the extent that it targets total spending rather than final spending – at least if its reliability (expected variation around the “true” value) is not significantly worse than NGDP’s. Targeting a Divisia aggregate would probably be substantially similar to an NGDP target except in the case of large exogenous changes in the demand for the services of money.
• However, #5 suggests something like Free Banking would be more stable and robust than even an ideal monetary policy target.
  

14. Banking is important.

• On the asset side of the balance sheet, intermediation ensures a more efficient distribution of changes in the money stock.
• On the liability side, banks in a competitive banking system – even under a central bank – face market pressures to adjust their quantity of circulating liabilities in accordance with the demand for money. Where there is a central bank, this mechanism reduces its epistemic burden: monetary policy does not need to work as hard (see again #5).
• The best argument for central banking is perverse elasticity in the face of changes in the demand for liquidity. The best argument against it, naturally, is apparatus cost.
• In the equation of exchange, M must be understood broadly as the total services of money which exist in the economy, the overwhelming majority of which is serviced by private liabilities. Plugging in the quantity of a single asset (like M0) or a simple addition of heterogeneous classes of assets (like M2) will result in misleading conclusions.