Macroeconomics has often been criticised for assuming rational expectations. Indeed, the assumption that all firms and consumers in an economy form their expectations in full awareness of all information and all underlying relationships in the economy is extremely unrealistic. Therefore, economists have started building macroeconomic models that are based on non-rational expectations (e.g. Branch and McGough 2009, Kurz et al. 2013). However, alternative methods of expectation formation have not yet been thoroughly incorporated into international macroeconomics. In a recent paper, we investigate a currency union model in which expectations are formed according to a behavioural reinforcement learning model (Bertasiute et al. 2018). The latter was previously validated on microeconomic data (Assenza et al. 2013, Hommes et al. 2017). Reinforcement learning refers to agents learning from their past mistakes; while they do not know the exact economic laws governing outcomes in an economy, they are smart enough to rely predominantly on forecasting strategies that performed well in the past.
In this macroeconomic model, economic behaviour can be quite different from a comparable model using rational expectations. In particular, one is much more likely to observe persistent deviations in output and inflation from the natural rate of output and from the inflation target in the behavioural model. This implies that economic crises can be much more severe and that they can last much longer than the rational model predicts. It also implies that different countries can have very different economic developments over a long time span.
Figure 1 Inflation in currency unions under behavioural expectations (top) and rational expectations (bottom)
Figure 1 illustrates this by showing sample paths of five three-country currency unions. The three lines in the same style and colour always represent the three countries of the same currency union. In the top panel, inflation is shown assuming that expectations are formed according to the reinforcement learning model. One can see that some countries develop differently from others with inflation above or below the inflation target of 2% for a long time. Forces drawing different countries of one currency union towards each other seem to be very limited. The bottom panel shows inflation assuming rational expectations. One can see hardly any deviations from the target, in particular no persistent deviations. These simulations are conducted with only moderate shocks to the economy, but the finding that under rational expectations inflation in all countries returns very fast to the inflation target after a shock also holds with larger shocks.
Looking at the consequences of aggregate economic behaviour, there are at least two important results from the behavioural model.
- First, economic integration is of crucial importance for the functioning of a currency union.
- Second, monetary policy has only very limited power to stabilise economic behaviour.
In what follows, we discuss these two issues.
The left panel of Figure 1 suggests that there are only very limited forces pulling the different countries of a currency union towards each other. This can be the case, but it does not necessarily have to be the case. Economic integration plays an important role here. The more economically integrated the countries are, that is, the more open they are to one another and the more they trade with one another, the stronger is this pulling force. This is intuitive. If a country in an economic downturn is economically very integrated with other countries in the currency union whose economies are doing better, the weaker economy will be able to export a lot to the stronger economies (because of gaining competitiveness in the downturn), which stabilises the economy. Vice versa, economies in a boom will start importing more from other countries, which dampens the boom. This is stabilising for the currency union as a whole.
The blue lines in Figure 2 show economic instability in a currency union (CU) as a function of the parameter γ, which is the parameter representing how economically integrated the countries are (the red line is the comparison value of a homogeneous economy, meaning that there is one single country of similar size as the whole currency union, abbreviated by EC). It can be seen that for a sufficiently high level of economic integration, economic behaviour in the currency union is stable (instability in this graph is defined as observing deviations of more than 10% from the steady state).
Figure 2 Economic instability depending on economic integration
Monetary policy has a lot of power to stabilise the economy in a country that is not part of a currency union. The central bank can, with its interest rate decisions, dampen booms and stimulate the economy in a recession. In a currency union, the central bank is still of vital importance for the currency union as a whole – poorly conducted monetary policy can bring the whole currency union into trouble. However, monetary policy faces more limitations in stabilising the economies than it does when being responsible for only one country. This is the case because of the differential developments within such a currency union. As the different countries may go through very different economic developments, the central bank will always set an interest rate that is ‘wrong’ for some of the countries (if not all). For the countries that are currently in a boom the interest rate (which is set in response to economic outcomes in the whole currency union) is too low, prolonging the boom. For the countries in a recession, the interest rate is too high, prolonging the recession. This problem can lead to economic instability in the whole currency union.
Economic integration is important for the functioning of a currency union. Reforms to strengthen this integration are thus useful. However, such reforms are probably mainly relevant in the long run. There is also a need for stabilisation policies in the short and medium run. Given that the role that monetary policy can play is somewhat limited, it seems necessary to have fiscal policy available as a stabilisation tool at the country level. This does not mean that there is a need for bigger redistribution between the countries, but countries should be able and required to use fiscal policy to stabilise their economies.
A currency union that doesn’t function well can incur huge costs. This includes the suffering of people in countries in an economic downturn, and the costs arising from populist parties being elected as a result of the economic situation. The fact that we are worried about the design of currency unions and the associated problems does not mean that entering into a currency union is necessarily a bad idea. There are potentially large benefits from it, such as the absence of risk from fluctuating exchange rates, easier comparability of prices, etc. But the currency union should be designed well, so that the benefits outweigh the costs.
Authors’ note: The views expressed in this column are those of the authors and not necessarily those of the National Audit Office of Lithuania or the Bank of Lithuania.
Assenza, T, P Heemeijer, C Hommes and D Massaro (2013), “Individual expectations and aggregate macro behaviour”, Tinbergen Institute, Discussion paper no 13-016/II.
Bertasiute, A, D Massaro and M Weber (2018), “The behavioral economics of currency unions: Economic integration and monetary policy”, Bank of Lithuania, Working paper no 49/2018.
Branch, W A and B McGough (2009), “A New Keynesian model with heterogeneous expectations”, Journal of Economic Dynamics and Control 33(5): 1036–1051.
Hommes, C, D Massaro and M Weber (2017), “Monetary policy under behavioral expectations: Theory and experiment,” Bank of Lithuania, Working paper no 42/2017.
Kurz, M, G Piccillo and H Wu (2013), “Modeling diverse expectations in an aggregated New Keynesian Model”, Journal of Economic Dynamics and Control 37(8): 1403–1433.