Why growth keeps disappointing

NB: This Special is an abridged version of a study published on the Rabobank Economic Research website. The full study in Dutch can be found here.

Now that the dust around the Greek crisis has settled, it is time to take a step back and examine the eurozone’s growth prospects for the medium term. In the past few years, GDP growth per capita has taken quite a beating due to both the credit crisis and the European debt crisis. What is more worrying, however, is the apparent decline in the potential growth per capita. Potential growth refers to the increase in the level of production that is possible without the economy becoming overheated and inflation rising. While there are many uncertainties affecting the measurement of potential growth, its development does give an indication of a country’s structural capacity for growth. Figure 1 (on the following page) shows that the potential per capita growth rate was still 1.4% per annum before the crisis but declined to 0.3% over the past five years. Actual economic growth was even lower because of cyclical effects.

Although we expect growth to recover over the next two years, this is largely because of improvements in the cyclical situation due to the favourable combination of low commodity prices and a weak euro. It is not clear what kind of growth rate the Eurozone can achieve once these temporary stimuli disappear and it has to rely on structural sources of economic growth. We examine this question in detail in this special report. The report is divided into two sections. In the first section, we argue that the labour market has borne the brunt of the economic losses from the Great Recession and that this could also have a negative effect on the Eurozone’s future potential growth. In the second section, we argue that labour productivity growth has also been weak and that this is associated in part with a decline in the Eurozone’s dynamism, while the ECB’s policy has also failed to produce growth. Structural reforms are needed to resolve the structural growth problems.

The change in GDP per capita can be broken down into three components: 1) the number of people in work in an economy, 2) the number of hours worked per person in work and 3) the labour productivity per hour. In other words, the GDP per capita can increase because people are working more, because they are working for longer, or because more goods and services are being produced in the same number of hours. Changes in the three aforementioned factors therefore affect growth trends. Based on an analysis of several countries, Vijlbrief et al. (2009) show that a fall in the level of GDP is often associated with labour market effects and a one-off fall in investment whereas a fall in the GDP growth rate is mainly caused by a slowdown in productivity growth (see figure 2 as an illustration).[1]

As figure 3 clearly shows, the contraction in eurozone GDP was indeed almost entirely attributable to labour market effects. In 2009, only one fifth of the 5%-contraction came from a reduction in the labour productivity per hour. Taking the period as a whole, the negative effects of the cumulative decline in the eurozone’s GDP of 7% were entirely borne by the labour market, whether in the form of a smaller number of hours worked or in the form of a reduction in employment levels. The result has been a substantial increase since 2008 in eurozone unemployment to 11.5% in 2014 (figure 4). Long-term unemployment – people who have been unemployed for more than 12 months – has doubled from 3% in 2008 to just over 6% in 2014. In addition to the negative social effects of long-term unemployment, this could also affect future potential growth within the eurozone.

There has, for example, been a slight increase in structural unemployment during the crisis from 8.8% in 2008 to 9.6% in 2014. ‘Structural unemployment’ is the level of unemployment that is independent of cyclical fluctuations and towards which the labour market tends in the long term. If structural unemployment increases, this will also have negative consequences for potential growth in the eurozone (see figure 1 again). The question is how fast unemployment will fall in the eurozone over the next five years. In the full version of this study (which can be found here), we estimated a partial unemployment model in which we were able to adjust two parameters: 1) economic growth and 2) the level of structural unemployment (estimated as the equilibrium unemployment rate NAIRU, or the non-accelerating inflation rate of unemployment). This gave four scenarios, which are summarised along with the resulting estimates in table 1:

[1] Vijlbrief, J.A., H.P.G. Erken and W.P. Spruijt (2009), Structurele effecten van de crisis op de groei en de rol van de overheid, [structural effects of the crisis on growth and role of the government], Holland Management Review, no. 129, pp. 2-13.

Increasing productivity is crucial

It is essential for euro area Member States to reform their labour markets to prevent the cyclical problems in the labour market from becoming structural. We have shown what effect this has. It is also crucial for the Eurozone to go back to higher growth rates. The growth rate is strongly correlated with changes in labour productivity, which in turn is the result of three underlying components: 1) the contribution of the amount of physical capital per unit of labour, 2) the contribution of human capital, and 3) the contribution of total factor productivity (TFP). Although TFP is a residual effect in growth accounting, it is one of the purest measures of the level of productivity given that it measures that portion of economic growth that cannot be explained directly by an increase in labour inputs (due to demographic changes) or capital inputs (due to investments).

Figure 5 shows that the average annual increase in labour productivity per hour fell in all the Eurozone countries included in the figure (with the exception of Spain) over the period 2009-2014 compared with the period 2000-2008. In countries where the average level of education in the population is lagging behind the average for Europe as a whole (such as Portugal and Spain), we see a big contribution from improvements in the quality of the labour force. In the other countries, improvements in human capital have only made a limited contribution to productivity growth.

A serious recession always results in lower levels of investment. In addition, a financial crisis makes credit more difficult to obtain, which can cause investments to fall over quite a long period. Even once the financial sector has recovered completely, a financial crisis can still lead to persistently higher levels of risk aversion, pushing the required rate of return on private investment projects up to a structurally higher level. Consequently, investment projects become less attractive, investment growth declines and so does the stock of capital goods in relation to GDP. This implies a lower rate of growth of production capacity.

However, figure 5 shows that the lower level of investment over a longer period of time has had only a limited negative effect on growth in capital intensity per hour worked and consequently on productivity growth per hour. The contribution of non-IT capital (per unit of labour) to productivity growth has only increased substantially in Spain, but this is largely due to a substantial decline in employment (a denominator effect). The other countries vary considerably in the contribution to growth of the capital intensity per hour, but in all cases it is both positive and relatively stable. It is still debatable whether the current growth in investment in the Eurozone is sufficient to compensate for the losses during the crisis. At present, we do not see sufficient evidence of such ‘catch-up growth’ in Eurozone investments.

However, it is the change in TFP in the Eurozone that is most worrying. This element explains the deterioration that many member states have experienced in the growth of labour productivity per hour over the past period. The deterioration in TFP growth is the main reason for the decline in potential growth in the Eurozone. Furthermore, TFP growth in the five largest economies has weakened considerably with respect to the trend. There has even been a reduction in productivity growth in Italy and Spain. There seems to be something structurally wrong in much of the Eurozone with the efficiency with which capital and labour generate added value. Put another way, the return on each successive technological innovation or change in working methods that is implemented seems to be less, or even to be turning negative.

The deterioration in TFP growth is partly related to a reduction in business dynamism in the Eurozone. This is the result of both the restructuring process and the creation of new ventures failing to really take off. We see this reflected in the data too. The OECD2 recently presented results from a new dataset, DynEmp, which contains data on the entry, exit, employment, size category, age and sector of individual firms for 18 countries over a ten-year period. Figure 6 [2] shows that the number of new firms as a share of total firms has fallen in many of the Eurozone’s member states. The highly regulated markets for labour, services and products, for example, are impeding the reallocation of production factors. Furthermore, labour mobility within the Eurozone is hampered by cultural, institutional and language barriers.

This in turn is curbing innovative activities, which go hand in hand with business dynamism. It is therefore worrying to see a sharp drop in the share of innovative firms in Germany (figure 7), as this has a knock-on effect on productivity changes in other economies within the Eurozone. Because of the lower level of innovation in Germany, other countries are not able to benefit to the same extent from what are termed ‘knowledge spillovers’. This refers to the fact that when a new technology is applied in Germany, other countries will eventually also start using this technology and in doing so, be in a position to refresh their own products or production techniques. Technological catching up and knowledge spillovers are an extremely important source of TFP growth, in particular for small, open economies.

[2] Criscuolo, C., P.N. Gal and C. Menon (2014), The dynamics of employment growth: new evidence from 18 countries, OECD Science, Technology and Industry Policy Papers, no. 14, OECD Publishing.

Disappointing growth in the Eurozone might also be the result of insufficiently effective fiscal and/or monetary policy. We would like to consider this aspect in somewhat more detail in this second section.

There is broad consensus among central banks that a deep balance sheet recession requires a longer-lasting expansionary monetary policy than a ‘standard’ recession in order to kick-start consumption and investment. However, if the starting position is characterised by a low trend growth rate and low inflation, the central bank may be hampered by what is termed the zero-bound - it cannot introduce negative interest rates. This problem is particularly likely to occur when there is low inflation or deflation, as then real interest rates are too high. To circumvent this problem (to some extent), central banks have started using unconventional instruments such as large-scale injections of liquidity and quantitative easing (QE). The bond-buying programme implemented by the ECB is one example.

The question is how effective such programmes are in achieving the central bank’s objectives. If there is a liquidity trap, monetary policy will become less effective as interest rates move closer to the zero-bound; the creation of new liquidity has less and less of a dampening effect on interest rates. This is expressed primarily in the form of lower money velocity. At the end of the 20th century, Paul Krugman[3] argued that this was largely a question of ‘confidence’. If economic actors do not believe that the central bank wants, or is able to, let inflation rise again to a certain target level, inflation expectations will remain too low and real interest rates will indeed be too high for the target to be achieved. The result is that the economy continues to muddle along with excessively low rates of growth and inflation.

This is precisely why the ECB has put such an emphasis over the past few years on convincing financial markets that its policy rate will remain low for an extended period of time and why it has backed this up with its bond buying programme. Even so, some doubts still remain about the success of this policy. It has proved relatively easy in recent years to create liquidity in the financial system, but the task of transmitting this liquidity to the real economy has turned out to be much more difficult. This is evident, for example, in the limited transmission of the so-called ‘base money’ created by the central bank to the broader measurements of money held by the public in recent years (figure 8).

The three key channels whereby an increase in liquidity can be passed on to the real economy are (i) via bank loans, (ii) via a weak exchange rate and (iii) via wealth effects. However, criticisms can be made of all three of these channels in the context of the current situation in the Eurozone. For instance, the stricter rules governing banks’ capital requirements, liquidity and balance sheet positions are preventing a strong expansion in bank credit. As regards the exchange rate, the ECB’s policy in the run-up to QE did have some effect as the euro is now more than 20% down against the US dollar compared with a year ago. However the same applies to the Japanese yen against the dollar thanks to interventions by the Bank of Japan.

In a world where all central banks are struggling with the same problems and trying to devalue their currencies, there can ultimately be no winner. In this regard, quantitative easing is an example of a beggar-thy-neighbour policy. If other central banks take countermeasures, there is a significant risk of the euro not being able to maintain its position as a cheap currency relative to other currencies. This has recently become highly relevant again with the Chinese central bank’s decision to let its exchange rate become even more flexible, which resulted in a devaluation of the renminbi.

[3] Krugman, P.R., K.M. Dominquez and K. Rogoff, K. (1998), It's baaack: Japan's slump and the return of the liquidity trap, Brookings Papers on Economic Activity, vol. 2, pp. 137-205.

As regards the final channel — wealth effects — we can conclude at any rate that the aggregate savings ratio in the euro area has risen rather than fallen since 2013 (see figure 9), despite positive capital accumulation (see figure 10) due to rising equity prices, lower interest rates and stabilising house prices. The idea underlying the wealth effect is that if people ‘feel wealthier’, they will consume part of their accumulated capital. But, as of yet, there seems to be no evidence of a strong wealth effect on expenditures given the movements in the savings ratio.

The standard view is that the disappointing economic growth in the euro area in recent years is partly due to insufficient demand as a result of excessively high real interest rates (in addition to the supply effects on the total factor productivity mentioned above). Indeed, the ECB’s working hypothesis is that its policy of ultra-low nominal interest rates in combination with its bond-buying programme will ultimately bear fruit and will help bring about a recovery in demand, either via the domestic channels of investment and consumption or through the channel of exports. Our scepticism regarding the effectiveness of QE means that we expect the underlying rate of recovery to remain relatively moderate. We also expect that there will regularly be unpleasant surprises and that it could therefore take several years at least before the Eurozone economy can be considered to have truly returned to normal.

Or are interest rates actually too low?

Although the most conventional view is that real interest rates are too high, it is also possible that the low level of growth in investments and in the economy is actually the result of policy interest rates being too low. A good example is the approach taken by the Austrian school, which argues that too low policy rates leads to overinvestment in economically unviable projects. When credit is cheap, the threshold rate of return is generally too low. While this may not fit with the recent period of both low policy rates and decreasing investment, it does provide an explanation for the boom that preceded the credit crisis. The adherents of the Austrian school claim that the effects of this are still in evidence because of the resulting misallocation of labour and capital. An excessive allocation of labour and capital to sectors with low productivity growth will eventually lead to lower productivity growth at the macro level.[4]

The excessive credit growth seen before the financial crisis is currently hampering economic growth on other fronts too. While households are trying to repair their balance sheet positions, governments have fewer options for compensating for the fall in demand through budgetary measures. They are now forced to pursue a restrictive fiscal policy. Even solvent businesses may lose the incentive to invest in new projects if the existing debt burden is high in relation to company profits. This is also known as ‘debt overhang’. Because the expected earnings from new projects will initially accrue to the existing creditors, and not the firm/shareholders itself, the firm has the undesirable incentive to keep new projects to a minimum until the debt position has improved. Research by Kalemli-Özcan et al. (2015)[5] shows that this has indeed been the case in the Eurozone in recent years. The negative effect of debt overhang on company investment actually seems to be greater than the effect of the balance sheet position of both firms and credit providers. Figure 11[6] illustrates this: the net investment ratio (as a percentage of GDP) has fallen sharply since 2008 even though the return on these investments is probably substantially higher than the market interest rate.

Following on from this, we could conclude that the recent monetary policy response by the ECB (reducing interest rates even further and promising to keep them low in the years ahead in combination with large-scale quantitative easing) may lead to even more distortion and perverse effects:

[4] See also BIS (2015), 85th Annual Report, p. 78.

[5] Kalemli-Özcan, S., L. Laeven and D. Moreno (2015), Debt overhang in Europe: Evidence from firm-bank-sovereign linkages, CEPR and ECB.

[6] A concept often used by macroeconomists is the ‘marginal efficiency of capital’ (originally introduced by John Maynard Keynes in his General Theory), which is a measure for the return on capital. The European Commission has defined this measure as ‘the change in the GDP volume per unit of invested (fixed) capital in the previous period’.

In this special report, we have shown that the Eurozone faces considerable challenges. The fall in GDP has resulted in major losses in the labour market, while deteriorating labour productivity growth has mainly affected growth prospects. Without structural reforms, unemployment will remain stuck at about 9% even in the event of a favourable growth scenario for the Eurozone. Far-reaching structural reforms of both the labour market and product markets are needed to strengthen the Eurozone’s potential growth in the medium term.

A worrying aspect for the outlook for growth is that growth of total factor productivity has deteriorated in all the major Eurozone countries. In many countries, this deterioration started long before the crisis. This means that the major economies of the Eurozone have started to function less efficiently. The weak TFP growth seems to be associated with a reduction in dynamism and innovation. This is causing lower productivity growth, both directly (through the inefficient allocation of capital and labour) and indirectly (through reduced competition in the market).

Monetary policy has been unable to counter the huge fall in demand. The effectiveness of this policy has been plagued by various difficulties. Insofar as this reduced effectiveness is the result of excessively high real interest rates (because of the zero-bound on the policy interest rate), the ECB’s new quantitative easing policy could provide a solution to the problem. However, excessively low interest rates and QE could also lead to all kinds of distortions, putting even more pressure on innovation, investment in the real economy and long-term growth. We also advocate an intensification of international policy coordination to prevent beggar-thy-neighbour policy measures aimed at devaluing currencies. However, the one-sided focus on monetary policy as the means of allaying the consequences of the crisis could have all kinds of undesirable side-effects. And that means that policymakers should perhaps concentrate more on creating the right conditions for innovation and improving the incentives for investment by firms.

The challenges posed by the labour market lie mainly in the need to increase labour mobility and to dismantle measures that encourage early retirement, make it difficult to adjust employee benefits and provide disproportionate protection from dismissal. A tour de force will be required from government leaders to accelerate the harmonisation of institutions in order to create broader support among Eurozone citizens for solidarity. It will also cost a great deal of effort to increase labour productivity by eliciting investments in innovation and encouraging new ventures. First steps in this direction have been made, for example with the new European Fund for Strategic Investments (EFSI). Whether this will be enough to bring the Eurozone’s growth potential back up to a structurally higher level, is however open to question.