by Dr Katsuyuki Shibayama, University of Kent. Discussion paper KDPE 1813, October 2018.

Non-technical summary:

After the Great Recession around 2007-8, the U.K. has experienced the slowdown of the labour productivity; known as ’productivity puzzle’. Although it is widely recognized as the stagnation of the labour productivity, not surprisingly, it coincides with the flattening of the total factor productivity (TFP). Japan also has experienced a similar but much longer and deeper phenomenon after the bubble burst at the beginning of 1990s. Interestingly, both countries have experienced the growth slowdown after some financial turmoil. This paper aims to explain these growth slowdowns quantitatively in a reasonably realistic economic growth model.

A growth slowdown is a long-lasting, significant decline of growth rate. It is not specific to high-income countries. The economic model developed in this paper also encompasses the “middle income trap”; typical examples include Latin American countries. The standard economic growth theory tells us that low-income countries should grow faster because they tend to accumulate production capital at a faster rate (Solow effect). Actually, many countries has successfully escaped from low income levels, but, out of 101 middle income countries in 1960, only 13 of them is classified as high income countries in 2008; see Larson, Loayza and Woolcock (2016, World Bank).

Extending Romer’s (1990) seminal paper, our model has an additional state variable, which we call the R&D environment, to capture social culture (scientists ’attitudes toward business, etc.), legal system (including patent laws and property rights), R&D infrastructure (such as innovators’ networks and education systems), and so on. We can regard the R&D environment as an intangible social capital, which has the following two properties; (a) society accumulates the R&D environment as an (intangible) asset by conducting R&D; and (b) the R&D activities are more productive when the R&D environment takes a higher value. For (a), an important assumption is such an accumulation of social asset is a positive externality of the R&D activities; i.e., the researchers do not intend to improve the R&D environment, when they engage in R&D. For (b), we want to capture, for example, higher education institutions are better prepared for the commercialization of academic findings when business innovations and inventions are more active.

In our model, there are two stable balanced growth paths (BGPs, long-run equilibria); one with positive R&D activities and the other without them. This is intuitively because of vicious and virtuous cycles. Around the BGP with no R&D, there is a vicious cycle; once R&D becomes inactive, it deteriorates the R&D environment, which itself discourages the R&D activities. Similarly, there is a virtuous cycle in the good BGP.

If there are no shocks (or only small shocks) in the model, then the fate of an economy depends on its initial condition. In our model, depending on the initial state, its final destiny – good BGP or bad BGP – is predetermined. Note importantly that in our model, even an economy moves toward the bad BGP, still it grows at a faster rate in early periods through the capital accumulation (Solow effect); it fails to switch from the capital accumulation as a growth engine to the R&D driven growth. We argue that countries that successfully transited from the middle-income level have had a R&D environment good enough to attain this transition; such as Japan and West Germany after the WWII. This is the explanation of the middle-income trap in our model.

If there are some shocks in the model, even if they are temporary, still they can have long-run effects. We assume that after a successful innovation, innovators can set up a firm, meaning that the firm value is the reward to innovations. Like Comin and Gertler (2003, AER), because the firm value is the present value of the current and future profits, R&D is more active in booms. Unlike Comin and Gertler’s ’medium cycle’ effects, however, business cycle fluctuations can have very persistent effects in our model; this is because a shock may push out an economy from one BGP to the other. For example, if an economy experiences a bad external shock, output, firm profit and firm value decline, which in turn discourages innovations. If such a bad shock is large enough and lasts long enough, the dormant R&D activities during the recession may deteriorate significantly, which hampers R&D even after the end of the negative external shock. In our model, this is the mechanism behind the growth slowdown in Japan and the U.K.

There are several policy implications. First, to help escape from the middle-income trap and the long-lasting growth slowdown, we need a ’big push’. For example, a large scale ODA as a positive external shock may be required. Second, however, the type of ODA matters. In our model, improving production efficiency and increasing final goods demand do not help escape from the low BGP, because there are two effects that offset each other. On the one hand, these shocks increase the firm profit and firm value, which stimulates the R&D activities. However, on the other hand, the production sector and the R&D sector compete each other in the factor market (labour market in our model). Hence, in our model, if the production efficiency improves, the production sector absorbs more labour, squeezing out the R&D sector from the labour market. There is anecdotal evidence of type of phenomenon; e.g., Wall Street and City have taken many scientists from universities and other research institutes such as NASA. In addition, this story is in parallel with the leading exposition about the natural resource curse; if the extraction of natural resources are very profitable, the natural resource sectors absorb too much production factors such as labour, squeezing out high productivity growth sectors such as manufacturing. Hence, if they only improves production efficiency, building bridges, roads and power plants, for example, has little effect in total. Third, our numerical experiments suggest that the financial sector efficiency has a strong effect. If the financial market is malfunctioning, the firm value may be discounted unduly. In our model, because the firm value is the reward to successful R&D, such a mispricing of the firm value discourages the R&D efforts. This is reminiscent of the fact that the U.K. and Japan have experienced the growth slowdown after the financial market turmoil.

In conclusion, this paper asserts that a sort of intangible social assets such as culture, legal system, etc. are important for R&D. Our model has multiple BGPs. Even though an economy has experienced a rapid growth via capital accumulation (Solow effect), it may fail to switch to the R&D driven growth (middle-income trap). In addition, even though a country successfully achieved a high-income level, it could slip down from the good BGP, particularly if it suffers from deep and long financial turmoil (like Japan and the U.K.). To regain the sustainable R&D environment, the model suggests the following policy prescriptions; (i) policy measures that directly affect the R&D productivity such as the subsidy to higher education and R&D tax credit and (ii) policies to improve the efficiency in the financial market.

You can download the complete paper here.

by Dr Anthony Savagar, University of Kent. Discussion paper KDPE 1812, October 2018.

Non-technical summary:

A standard way to measure productivity is to take output growth and subtract input growth. Typically input growth is growth in capital and growth in labor weighted by their share in production. Whatever remains after subtracting is known as total factor productivity (TFP).

Under specific circumstances, this productivity measure also reflects underlying technology growth. This is important for economists because they struggle to measure underlying technology at an aggregate level, but it is a key variable to understand the behaviour of the economy.

However when we diverge from some basic assumptions, such as perfect competition, the TFP measure no longer reflects underlying technology. Therefore using our TFP measure to represent technology could lead to incorrect conclusions.

In this paper I show that when we recognize the slow adjustment of firms to arbitrage profits and the effect that slowly entering firms have on competition, the relationship between our measure of TFP and technology becomes much more complex. In fact, when we observe changing TFP, it will compose changing technology, changing profits and changing markups. This decomposition allows us to understand how we can get a true measure of underlying technology from our calculated TFP measure. It emphasizes that the composition of profits and markups vary in importance as firm entry takes place.

You can download the complete paper here.

by Jan-Philipp Dueber, University of Kent. Discussion paper KDPE 1811, August 2018.

Non-technical summary:

Time-varying volatility plays a crucial role in understanding business cycles in emerging market economies. There is now plentiful empirical evidence that volatility as measured by the standard deviation in macroeconomic data is time-varying and strongly countercyclical. Volatility increases during an economic recession and becomes lower during an economic boom.

In addition, we observe that standard open-economy macroeconomic models widely used for business cycle analysis fail to address the specific characteristics of many emerging market economies. In emerging market economies the net export to output ratio is typically negatively correlated with output i.e. it is countercyclical. However, standard models predict a near perfect positive correlation between the two. Besides that, emerging market economies show a higher fluctuation in consumption data than in data on output. Standard models, however, predict a higher fluctuation in output than consumption. These models are therefore overemphasizing the role of consumption smoothing.

This work is motivated by the above empirical observations. We augment a standard small-open economy model and introduce time-varying volatility to the interest rate and total factor productivity. In our model the interest rate and total factor productivity automatically turn more volatile when the economy becomes more indebted or when output declines in response to a negative total factor productivity shock. Once we introduce time-varying volatility into a standard open-economy model, the model becomes able to significantly better match emerging market economy data. After the introduction of a time-varying interest rate and total factor productivity we are able to present a model where net exports are strongly negatively correlated with output and consumption shows a higher variation than output as observed in the data. By choosing different parameter values for the time-varying volatility the model is able to characterize both, emerging market and developed economies, or an economy that is in transition to a developed economy.

Although we are not the first to address the problems of macroeconomic models for emerging market economies, our approach is especially simple and does not rely on shocks to the permanent component of total factor productivity or to shocks in the level of the interest rate. Compared to other research in this area our approach only requires one source of external variation, the widely used shock to total factor productivity. From a policy point of view our model can be especially useful for economists and policy makers in emerging market economies as our model now better fits the economic cycle in those countries.

You can download the complete paper here.