Opening remarks by
Ms. Sinikka Salo, Member of the Board

It is a great pleasure for me to open this conference. Its theme "Expectations and Business Cycle Dynamics" is time-honoured but still topical.

Most if not all theories of business cycle dynamics attach great importance to expectations. A general theory of business cycles, if there will ever be such a theory, needs no doubt to explain both how expectations are formed and how they affect the evolution of aggregate variables.

In Keynesian models pessimistic demand expectations can set forth a cumulative process of multiplier effects and revision of expectations, derailing the economy from full employment. Demand management, in particular fiscal policy, can and should be used to break this vicious circle and to prevent mass unemployment and deflation of the sort seen during the 1930s.

Keynesian analysis and policies gained popularity after the World War II, contributing in the 1950s and the 1960s to a sustained and, at least in the USA, non-inflationary economic growth. This success led some economists to think that the business cycle was a thing of the past, provided that appropriate and appropriately co-ordinated demand management policies were in place. This optimism was reinforced by the introduction of large econometric macro-economic models as a new powerful scientific tool of demand management. The importance of expectations in business cycle dynamics was confirmed in these models by using different hypotheses about expectations.

Nevertheless, both large macro-econometric models and demand management policies failed to fulfil all the hopes attached to them. The period of sustained growth and stable prices was followed by unexpected turbulence in the 1970s and the 1980s punctuated by the collapse of the system of fixed exchange rates, large supply shocks and inflationary surges. In spite of at least initially very expansive fiscal and monetary policies unemployment increased, and in spite of increasing unemployment, inflation accelerated.

In retrospect, it is clear that one weakness in early macro-economic models and in the early Keynesian analysis in general was inadequate treatment of expectations. One of the underlying reasons for accelerating inflation was that monetary policy lost its credibility. Without credible monetary policy price expectations lost their anchor. This was something the simple models could not account for.

A more refined analysis of expectations was provided by the rational expectations hypothesis. This became an integral element of many micro- and macro-level theories including the "efficient markets" theory of securities prices, the "permanent income" and "life-cycle" theories of consumption and the theory of "tax smoothing" or "the Ricardian equivalence" of fiscal policy.

In the mid-1970s Robert E. Lucas argued that the then conventional macro-economic models were useless or almost useless for analysing policy changes, because their behavioural equations changed whenever policy changed. He put forward also a different argument – for example in his Yrjö Jahnsson Lectures "Models of Business Cycles" here in Helsinki in 1985 - that the potential welfare gains from even successful stabilisation policies are likely to be small, if compared with policies which have permanent effects on the level or growth of economic activity. Some macroeconomic models with rational expectations - such as the real business cycle models – indicated that policies that try to manipulate the economy tend to be ineffective or even harmful.

Macro-economic literature inspired by rational expectations hypothesis has had a significant impact on the way monetary policy is conducted in practice.

The rational expectation hypothesis has found its place in the state-of-the-art macro-economic models, such as the AINO-model of the Bank of Finland, used by central banks for forecasting, risk analysis and policy simulation purposes.

At a more fundamental level the rational expectations hypothesis has helped understand the importance of credibility in economic policy, strengthening the case for an independent central bank and stability-oriented, time-consistent monetary policy. It is now generally acknowledged that monetary policy contributes best to macroeconomic stability by anchoring inflation expectations at a level consistent with price stability. This is best achieved through a forward-looking, medium-term orientated monetary policy.

Recent macro-economic literature has also been successful in encouraging central banks to increase the transparency of their policies, not only by verbal and written policy statements but also by publishing inflation reports and macro-economic forecasts, including now in some cases even forecasts for policy rates. While this may be seen as an attempt to reduce the noise in the transmission of policy intentions, it may also indicate that the central bankers assume themselves to be better informed about the state or the functioning of the economy than the markets or the public at large.

On the other hand, even though explicit policy rules such as the Taylor rule are widely used in academic analyses and often also recommended for policy purposes, they are not widely used in actual decision making. Central bankers still rely heavily on policy deliberations, even though the room for their discretion in most cases is somewhat constrained by an explicit monetary policy strategy. And even though central bankers do not believe in fine-tuning economic activity and are uncertain of the exact effects of their decisions, they seem to be confident that their decisions contribute to the overall stability of the economy.

The last two decennia have witnessed a remarkable improvement in the performance of the world economy. The growth of global output has accelerated trend-wise since the mid-1970s, at the same time as the variability of growth has declined substantially. Central banks have succeeded in squeezing inflation to a level consistent with price stability in most developed countries and to single digits in most other countries. Moreover, the world economy has shown its resilience by surmounting without much difficulty several sudden and apparently strong shocks to economic activity, including stock market crashes and financial crises.

While several factors may have contributed to economic stability, some economists, including David and Christina Romer, have attributed it to a fundamental improvement in the understanding among economists and policymakers about what macro-economic policies can accomplish. This is consistent with the view that, in the past, severe business cycle fluctuations were at least partly caused by misguided macroeconomic policies.

Can we now take the business cycle as a thing of the past or in any case as a minor nuisance of little policy relevance?

Even though recent trends in the world economy are in many respects encouraging, they have also revealed a gradual emergence of significant imbalances. Some well-informed observers – including the IMF in its last WEO – have noted that the present period shares a number of characteristics with the 1960s and early 1970s, suggesting that the low volatility of aggregate variables may not be sustainable. The recent financial market turbulence and increased credit risks have underlined this concern.

Our knowledge of the world economy - or any national economy - is bound to be limited, first because of the size and complexity of the economic system and second because incessant structural change, or industrial mutation, as Schumpeter called it. This gives rise to uncertainty more fundamental than the actuarial randomness assumed in conventional rational expectations models, in particular as the occurrence of rare or exceptional events seems to be much more common than implied by the convenient assumption of normal distribution.
Given the limitedness of our knowledge, learning seems to be relevant, including learning from our own errors. However, it is not self-evident how it should be modelled in macroeconomic models. In recent years economists have paid increasing attention to the formation of expectations and other learning behaviour, which are also one of the main topics in this conference. Some progress has been achieved already, as witnessed by the fact that learning has found its way to empirical macro-models, but no doubt there remains still much to be done. One way to study learning behaviour in the context of forming expectations is to use data on measured expectations, such as revealed for example by surveys.


On behalf of the organisers, I wish to welcome you all to this conference, which is organised annually by the Bank of Finland and the Centre for Economic Policy Research. This is the 8th in the series. I look forward for the presentations and discussions, and I am sure that they make an important contribution to our understanding of expectations and business cycle dynamics.