Economics after the crisis - New model army

Efforts are under way to improve macroeconomic models

Jan 19th 2013 | WASHINGTON, DC |From the print edition

THE models that dismal scientists use to represent the way the economy works

are sometimes found wanting. The Depression of the 1930s and the stagflation

of the 1970s both forced rethinks. The financial crisis has sparked another.

The crisis showed that the standard macroeconomic models used by central

bankers and other policymakers, which go by the catchy name of dynamic

stochastic general equilibrium (DSGE) models, neither represent the financial

system accurately nor allow for the booms and busts observed in the real world.

A number of academics are trying to fix these failings.

Their first task is to put banks into the models. Today s mainstream macro

models contain a small number of representative agents , such as a household,

a non-financial business and the government, but no banks. They were omitted

because macroeconomists thought of them as a simple veil between savers and

borrowers, rather than profit-seeking firms that make loans opportunistically

and may themselves affect the economy.

This perspective has changed, to put it mildly. Hyun Song Shin of Princeton

University has shown that banks internal risk models make them take more and

more risk as asset prices rise, for instance. Yale s John Geanakoplos has long

argued that small changes in the willingness of creditors to lend against a

given asset can have large effects on that asset s price. Easy lending terms

allow speculators with little cash to bid up prices far above their fundamental

value. If lenders become more conservative, these marginal buyers are forced

out of the market, causing prices to tumble.

Realistically representing the financial sector would help solve the other big

problem with mainstream macro models: that they are inherently stable unless

disturbed from the outside. This feature is helpful when studying how an

economy in equilibrium responds to things like a spike in the price of

petrol, but it limits economists understanding of why economies expand and

contract in the absence of such external shocks. Highly leveraged financial

firms with portfolios of risky assets are bound to upend an economy every so

often. Having banks in models would generate shocks from within the system.

The world s big central banks are interested in these new ideas, although staff

economists are reluctant to abandon existing industry-standard models. If any

central bank is likely to experiment, however, it is the European Central Bank,

thanks to its two-pillar approach to assessing the risks of price stability.

The ECB pays as much attention to monetary analysis , which includes things

like bank lending and money creation, as to economic analysis , which is more

concerned with things like inflation and joblessness.

Improving DSGE models is the obvious way to take the lessons of the crisis on

board. But others exist too. Agent-based modelling tries to depict the

transactions that might occur in an actual economy. These models are populated

by millions of agents that gradually alter the economy as they interact with

each other. The idea was developed in the 1990s when biologists wanted to study

the behaviour of ant colonies and the flocking of birds. But modelling an

entire economy did not become practical until recently because of the sheer

number of calculations needed.

The evolutionary structure of agent-based models allows economists to study how

bubbles and crises occur over time. For example, an increase in bank lending

means more spending and therefore higher returns on existing investment, which

in turn encourages further lending. But too much lending can prompt the central

bank to raise rates if inflation starts to accelerate. Higher borrowing costs

could lead to a wave of defaults and even to a crisis if too much debt was

taken on during the boom.

The EURACE project, an initiative by a consortium of European research bodies,

has produced a sophisticated agent-based model of the EU s economy that

scholars have used to model everything from labour-market liberalisation to the

effects of quantitative easing. In Australia Steve Keen, an economist, and

Russell Standish, a computational scientist, are developing a software package

that would allow anyone to create and play with models of the economy that

incorporate some of these new ideas. Called Minsky after Hyman Minsky, an

American economist celebrated for his work on boom-and-bust financial cycles it

places the banking system at the centre of the economy.

A long road lies ahead, however. Nobody has got something so convincing that

the mainstream has to put up its hands and surrender, says Paul Ormerod, a

British economist. No model yet produces the frequent small recessions,

punctuated by rare depressions, seen in reality. But ultimately, Mr Shin

says, macro is an empirical subject. It cannot forever remain impervious to

the facts .