As the above citation from Hyman Minsky shows, one may think of modern macroeconomic development as a sequence of boom-bust cycles. Boom-bust cycles occur not only for specific sectors, but also for the entire macroeconomy. Macroeconomic boom periods are usually characterised by overvaluation of assets, overconfidence, expectations of high returns and undervaluation of risk, and by overleveraging. Bust periods reverse confidence and expectations. The current macroeconomic developments in the USA as well as in other regions of the world have features of a typical bust period that is characteristic of boom-bust cycles. In the boom period not only do prices increase but there is often also an asset price boom and credit boom. High asset prices serve as collateral for new borrowing. When a downturn starts, often initiated by a sudden bust, and frequently entailing long-term protracted periods of low growth and low employment, prices may fall and periods of debt deflation are often experienced.

Introduction

In this chapter we continue the analysis of our applied structural model of disequilibrium growth initiated in Chiarella and Flaschel (1999b,c,d). In those earlier papers we developed a model of disequilibrium growth which is fairly complete with respect to markets, agents and sectors and consistent with respect to the various budget constraints between them. We showed in Chiarella and Flaschel (1999b,c,d) that the model could be expressed as a dynamical system of 34D intensive state variables together with a number of static relationships. We further showed how a small number of not unreasonable simplifying assumptions reduced this dynamical system to one of eighteen laws of motion solely, which we have dubbed the core model of this approach with fixed proportions in production. Our aim in this, and in the next two chapters, is to analyse in quite some detail the properties of this 18D core model. In this chapter in particular we focus on the basic partial feedback structures of the core model and their stability characteristics.

We can distinguish qualitatively at least seven important feedback chains (plus stabilising or destabilising policy reaction functions), which we will describe below in isolation from each other. These will of course interact with each other in the full 18D dynamics of the core model, so that one or another may become dominant when the parameters of the model are chosen appropriately.

The structure of the chapter is as follows: Section 7.2 describes the accounts of the various agents of our modelling framework.

The debt deflation debate

In the recent public debate on problems of the world economy, ‘deflation’, or more specifically ‘debt deflation’, has again become an important topic. The possible role of debt deflation in triggering the Great Depression of the 1930s has come back into academic studies as well as into the writings of economic and financial journalists. It has been observed that there are similarities between recent global trends and the 1930s, namely the joint occurrence of high levels of debt and falling prices, the dangerous downside of cheaper cars and TVs when debt is high. Debt deflation thus concerns the interaction of high nominal debt of firms, households and countries and shrinking economic activity due to falling output prices and increasing real debt.

There is often another mechanism accompanying the one just mentioned. That other mechanism involves how large debt may exert an impact on macroeconomic activity and works through the asset market. Asset price inflation during economic expansions normally gives rise to generous credit expansion and lending booms. Assets with inflated prices serve as collateral for borrowing by firms, households or countries. In contrast when asset prices fall the borrowing capacity of economic agents shrinks, financial failures may set in, macroeconomic activity decreases and consequently large output losses may occur.

Countries that have gone through such booms and busts are some Asian countries (in particular Japan), Russia and Brazil in 1998 and 1999.

The preceding chapters have shown that debt accumulation when combined with price dynamics may give rise to instability. A stylised fact of periods of financial fragility is that over-indebtedness leads to the insolvency of borrowers. Firms go bankrupt and default on loans. The impact of the failure of firms and non-performing loans plays a central role in the theories of financial fragility developed by Minsky and Fisher. Nonperforming loans may have a boomerang effect on the financial sector, by undermining the profitability of commercial banks. In this chapter, the preceding models are extended to take into account three aspects of debt over-indebtedness over the business cycle:

1. Bankruptcy of firms

2. Debt default

3. Non-performing loans and banking crises.

Bankruptcies may have ambiguous effects on the business cycle. On the one hand, the market sanctions bad performance by bankruptcy. It eliminates the weakest and most fragile firms and establishes favourable conditions for economic recovery. Similarly, in a Schumpeterian approach, the creative destruction argument points in the same direction. Recessions are productive as they are periods during which new technologies and new organisations are implemented. Likewise, bankruptcy also improves the average output to capital ratio, which paves the way for economic recovery. This is in fact a key element of the so-called reproductive cycle (see Gordon et al. (1983)). On the other hand, using a Keynesian line of argument bankruptcies may have a destabilising effect on consumption through unemployment and nominal wages.

The emergence of currency crises

With the end of the Bretton Woods system in the 1970s and the financial market liberalisation in the 1980s and 1990s, the international economy has experienced several financial crises in certain countries or regions entailing, in most cases, declines in economic activity and large output losses. This has occurred regardless of whether the exchange rates were pegged or flexible. There appear to be destabilising mechanisms at work from which even a flexible exchange rate regime cannot escape. In this chapter we review some of the stylised facts that appear to be common to such financial crises and develop a Mundell–Fleming–Tobin (MFT) type model based on Rødseth (2000, Ch. 6). Our approach builds on Miller and Stiglitz (1999) and takes up Krugman's (1999a, 1999b, 1999c and 2001) suggestions in order to study the real and financial crises generated by large exchange rate swings.

With respect to exchange rate shocks due to currency runs triggering financial and real crises, there are three views, in fact three generations of models, that have been presented in the literature. The first view maintains that news on macroeconomic fundamentals (such as differences in economic growth rates, productivity differences and differences in price levels, in short-term interest rates as well as in monetary policy actions) may cause currency runs. The second view maintains that speculative forces drive exchange rates where there can be self-fulfilling expectations at work, destabilising exchange rates without deterioration of fundamentals.

Introduction

In this part of the book we lay the theoretical foundations for the construction of larger macroeconometric models where large means the approximate size of, for example, the Murphy model of the Australian economy (approximately a hundred equations). In practice this implies that about twenty laws of motion have to be considered in order to describe the evolution of such an economy. Yet, in contrast to many models that are actually applied we insist here that such models must be completely specified in terms of budget equations (identities or restrictions) and the stock-flow interactions that they imply. Moreover the models should not only be formulated on the extensive form level, but must also allow for a representation in intensive terms as well (trendless variables as far as the theoretical representation of the model is concerned). This intensive form representation should then also allow the determination of at least one steady state solution, the stability of which is to be discussed from the perspective of the partial feedback structures which are included in the general formulation of the model.

In this chapter we extend the hierarchically structured continuous-time models of Keynesian monetary growth, that have been introduced and generalised in some respects in Chiarella and Flaschel (2000, Chs. 4–7), Chiarella et al. (2000, Chs. 4–6) and Asada et al. (2003) both for closed as well as open economies, along the lines of the macroeconometric Murphy model of the Australian economy.

Mortgage and banking crises

The current financial crisis in the USA can be characterised by a fast expansion of mortgage loans to households in order to purchase real estate. A difference to the traditional Minskyan crisis and the previous type of financial crisis is that borrowers are not firms but households. The Japanese crisis was characterised by firms and banks borrowing to invest in real estate, and the East Asian crisis was characterised by firms and banks borrowing foreign-denominated debt. In the current worldwide crisis, financial fragility that has led borrowers from hedge to speculative and even Ponzi positions involves the dynamics of household income and interest payments.

There are three main reasons explaining the large increase in mortgage credit: competition between financial institutions; the interaction between real estate prices and credit constraints; financial innovations. These three elements have induced banks to relax screening and monitoring of borrowers and to increase the quantity of credit supplied to households. First, in the aftermath of the crash of the dotcom bubble in 2000 and in an environment characterised by low interest rates, banks, under the pressure of financial intermediaries such as hedge funds, found in mortgage debt a highly profitable business. Second, increasing real estate prices contributed to the relaxation of credit rationing. The increasing value of collateral has reduced default risk and has led banks to increase credit, in line with the financial accelerator model.

The current macroeconomic development of the USA as well as of most major industrial economies is characterised by boom-bust cycles. Such boom-bust cycles start with overconfidence, expectations of high returns and overleveraging. Often an asset price boom goes hand in hand with a credit boom and rising prices. When a downturn is triggered, often initiated by a sudden bankruptcy or similar event, frequently entailing long-term protracted periods of low growth and low employment, prices may fall and periods of debt deflation are experienced. Normally such boom-bust cycles are driven by specific sectors in the economy. In the most recent boom-bust cycle, the credit sector and the real estate sector were the main driving forces.

To study such phenomena, this book takes a macroeconomic perspective. It uses a dynamic framework that builds on the theoretical tradition of non-clearing markets.

Introduction

In this chapter we derive and investigate the 34D intensive (state variable) form of the applied structural model of disequilibrium growth we have introduced and discussed in its originally extensive form level in great detail in Chiarella and Flaschel (1999b) and in the preceding chapter. We will represent the resulting 34 dimensional dynamical system from various perspectives, providing compact intensive form representations of the real and the financial sector of this economy in tabular form and also in the form of a system of national accounts. We will then discuss to some extent the economic content of the resulting laws of motion from their intensive form perspective, thereby showing that the model can be understood from the outset on the intensive form level.

Presenting the system from these various perspectives serves the purpose of making the reader acquainted with the notation and the relationships that apply on the intensive form level of the model. We hope that this approach will increase the readability of the laws of motion for quantities (including rates of growth), for prices (including wages, asset prices and also expectations), financial asset accumulation and feedback fiscal and monetary policy rules to be presented and discussed in Section 6.3. Section 6.4 then calculates the (up to the determination of nominal variables) uniquely determined steady state solution of this dynamical system and briefly considers its comparative dynamic properties which are generally very simple in nature.