Secular stagnation in the Netherlands

Secular stagnation in the Netherlands

Faculty of Economics and Business (FEB), University of Amsterdam (UvA)

Student: Daniel Koudijs Supervisor: Nicoleta Ciurila BSc in Economics and Finance Student Number: 10217150

Abstract

This thesis analyses whether the Netherlands are suffering from secular stagnation. The thesis is structured around two macroeconomic pillars through which secular stagnation effects can manifest itself: the long run potential growth rate and the output gap. Data on these are presented, followed by an analysis of their corresponding determinants. The main finding of this thesis is that both a declining long run potential growth rate and a persistent output gap can be causes for secular stagnation in the Netherlands. For the first, the main cause lies in the changing demography. The structural output gap is found to be the result of structural unemployment, balance sheet recession effects and a negative real interest rate gap.

Table of contents

• Chapter 1 Introduction

• Chapter 2 Defining secular stagnation

• Chapter 3 The first pillar: the long run potential growth rate

• Chapter 4 The second pillar: output gaps

• Chapter 5 Conclusion

• Chapter 6 Appendix

• Chapter 7 Bibliography

Chapter 1 Introduction

Eight years have passed since the financial crisis of 2008 and many European economies including the Netherlands have only just passed their pre-crisis GDP level (see figure 1 in the appendix). Interest rates are locked at the zero lower bound, while inflation has remained absent and deflation has been a serious risk. Meanwhile unemployment has risen, in the Netherlands from 3.7 % in 2008 to nearly 7.2 % in the beginning of 2015. While in the past we have seen economies gain momentum and soar out of the depths of depressions, recent recovery has been slow and fragile. It seems clear something has changed.

This failure of generating economic growth, lacking in both velocity and size, appears evident and spread worldwide. Yet the idea that “sick recoveries die in their infancy and depressions feed on themselves, leaving a hard and seemingly immovable core of unemployment” is not a new one, but originates with Alvin Hansen (Hansen, 1938). Hansen looked back on a period of unprecedented expansion, suddenly halted by the Great Depression, and warned about non-cyclical issues that could keep the economy from regaining its original growth rate. But as the US rose out of the Great Depression his argument faded to the background. Until recently, when Larry Summers resurrected the idea of ‘’secular stagnation’’, arguing that it was becoming more and more questionable whether economies would reach their previous growth potential ever again (Summers, 2014). Secular stagnation has returned to the

forefront of economics.

With economic theory largely structured around short-term business cycle deviations, the revived importance of structural factors will pose new challenges, which makes secular stagnation an especially interesting topic.

In my thesis I will analyse the Dutch case for secular stagnation. Specifically, I will answer the question: is there evidence that the Dutch economy is suffering from secular stagnation?

The main result of this thesis is that both decreases in the long run potential growth rate and the structural output gap have been causing secular stagnation in the Netherlands. For the first, the main cause lies in changing demographics. For the structural output gap the main causes are found to be structural unemployment, balance sheet recession effects and a negative real interest rate gap.

The thesis is structured around the two macroeconomic pillars where the effects of secular stagnation are most present: the long-term potential growth rate and the output gap. The next chapter elaborates on the definition of secular stagnation and provides further detail on the analytical framework. The third chapter focuses on the first pillar of secular stagnation, while the fourth chapter focuses on the second pillar. In the fifth chapter I summarize and conclude my findings, providing a verdict on secular stagnation in the Dutch economy. I finish my conclusions by highlighting the shortcomings of my research and possible

directions for future research. The appendix at the end contains graphs and tables referred to in the text for further clarification.

Chapter 2 Defining secular stagnation

Economists hold different interpretations about what the concept of secular stagnation exactly means, some whether it even exists. See for example a recent debate between Larry Summers and Ben Bernanke, in which Bernanke argues against Summers that secular stagnation is a ‘’temporary geographical dislocation of savings’’ (Bernanke, 2015).

In this paper I will not expand on the difference in interpretations or debate which definition of secular stagnation is more valid. Instead I will apply an existing framework to analyse secular stagnation similar to the scheme put forward by Teulings and Baldwin in the introduction to their CEPR working paper (Teulings & Baldwin, 2015). In order to structure the different views expressed in their working paper, they organize the discussion around two ‘’macro-economic pillars’’: the long-term potential growth and the output gap. In figure 2 we see both variables plotted for the Netherlands.

Figure 2. Output gap and potential growth rate

Source: European Commission, 2015

-4% -3% -2% -1% 1% 2% 3% 4% 5% 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 O ut put g ap & P ot ent ia l g ro w th r at e, in %

It should be noted that Teulings and Baldwin make use of a third pillar as well. The sclerosis pillar, defined as one-off economical setbacks without long-term growth consequences (Teulings & Baldwin, 2015). This part is left out of my thesis with the aim of being concise, as well as for a lack of appropriate data available. For the remainder of this chapter I will discuss the theory behind the two pillars I will analyse and their main supporters.

The first pillar concerns long-term potential growth factors. Secular stagnation can be the result of a decrease in the long-term potential growth rate: the economy is growing sluggishly because its potential growth has slowed down.

Potential output is defined as the maximum output that can be achieved by deploying available factors of production; the potential growth rate is the growth of the potential output level. In the European Commission’s interpretation, potential output is given by the three inputs of the production function: labour, capital and total factor productivity (TFP). In the next chapter I present data on the evolution of the potential growth rate and its

underlying elements. To further detail the findings of this data I will look at several structural forces influencing potential growth, namely what Bob Gordon has labelled as the four

structural ‘’headwinds’’: demographics, education, inequality and public debt (Gordon, 2014).

The second pillar focuses on output gaps: secular stagnation as the result of persistent gaps between real and potential growth. Larry Summer’s perspective mainly concerns this pillar. Summers argues that when the real natural interest rates drops significantly below zero and the nominal interest rate becomes stuck at the ZLB with little inflation, an economy could consistently be kept from reaching it’s potential (Summers, 2014). In my chapter on the second pillar I present data on the Dutch economy’s output gap and put forward several

explanations for its development: unemployment, income growth, the idea of a balance sheet recession and finally I discuss the role played by the real natural interest rate.

To summarize, secular stagnation is associated with a decline in an economy’s potential growth rate and a persistent output gap. Where the first pillar mainly concerns the structural slowing down of the future potential growth rate, the second pillar focuses on output

structurally deviating from potential. In the following chapters I will analyse to what extent these influence the Dutch economy.

Chapter 3 The first pillar: the long run potential growth rate

In this chapter I look at data estimates on the potential growth rate and discuss several explanations for its development over time. There are some theoretical and statistical differences between short, medium and long run potential growth rates (see for example the introductory remarks on potential growth rates by Dennis, et al. (2006). This disparity however lies beyond the scope of my thesis so here I will assume potential growth estimates by the European Commission to be a valid proxy for the long run potential growth rate.

In the graph below we see the development of the potential growth rate and its

components. The potential growth rate has declined substantially since the beginning of the century, with the biggest decline present in the factors of labour and TFP. Projected

potential growth for 2015 stands at 0.5%, down from 3.4% in 2000. Contribution by labour and TFP over the last 15 years has on average been 0.6% lower compared to the 15 years before. Over the same period, the average contribution of capital has only decreased by 0.1% (see table 1 in the appendix).

Figure 3. Contribution to potential growth

Source: European Commission, 2015

0% 1% 1% 2% 2% 3% 3% 4% 4% 1985 1990 1995 2000 2005 2010 2015 Po te nti al g ro w th ra te , i n %

Contribution to potential growth

Total labour (hours) contribution Capital accumulation contribution TFP contribution

What has caused this rapid decline in potential growth, especially among the factors of TFP and labour? Starting with TFP, the basic proxies used are technology or innovation

indicators. When looking at one such indicators for the Netherlands, R&D spending as a percentage of GDP (see figure 4 in the appendix), we find little evidence supporting the decline in TFP’s contribution to growth.

For a better understanding of how TFP, labour and potential growth in general could have experienced a substantial decrease we need to look to more structural forces moving the economy, or to what Robert Gordon calls the ‘’four headwinds’’: demographics, education, inequality and government debt (Gordon, 2015). Below I will discuss these headwinds, first outlining their relation to potential growth, followed by a discussion of the relevant data from the Netherlands. I end the chapter with a short overview of the first pillar.

Demographics

On the demographic side, there are several trends relevant to the decreasing long run potential growth rate. The relation between population and economic growth is fairly straightforward: more workers means more potential growth. By definition, the labour force consist of those in the working age population that participate in the labour market. In the Netherlands, growth of the working age population has been stagnant (see figure 5 in the appendix). The effects of a stagnating working age population have for a long time been offset by increases in the participation rate. In this way the Dutch labour force kept growing and pushing economic growth. However, growth of the participation rate gradually slowed down over the past decade and has floated around 74% since 2008. As a result, the Dutch labour force is no longer growing and no more hours are being worked in the Dutch economy.

To understand how this translates into slower growth we look at output growth per capita. By definition, output per capita equals labour productivity times hours worked. With growth in hours worked coming to a halt, only labour productivity gains can lead to growth in output per capita. Figure 6 below presents the development of output growth per capita and its elements over the past 15 years.

Figure 6. Output growth per capita

Source: Eurostat, 2015 (Real labour productivity); European Commission, 2015 (Change in hours)

Labour productivity growth in the Netherlands has been weak, averaging just 0.1 % since the financial crisis . This could accelerate again in the future, though research shows there is little evidence for optimism. A 2013 study by The Conference Board finds that average long-term growth of labour productivity across the advanced economies has stalled since 2000 and is unlikely to recover (Chen et al., 2013). Unless the Netherlands are an exception, substantial growth fuelled by labour productivity gains seems doubtful.

It seems the labour force has gone from being an engine to being a break on the economy. The effects of these long-term demographic trends have only just started to unfold, but

-5% -3% -1% 1% 3% 5% 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Gr owt h, in %

Ouput growth per capita = growth labour productivity * growth in hours per capita

Real labour productivity growth Changes in hours (per empl) contribution Ouput growth per capita

what is more pressing is that they have strong roots: the direction in which the Dutch population and labour force are heading is not one swiftly turned. Demographics is likely to continue being a force of secular stagnation.

Education

The second headwind described by Gordon is education. Theory states that education can boost economic growth by increasing a worker’s productivity as well as his income. In the past century we have seen landmark educational gains with strong effects on the economy’s growth and it’s potential. Gordon states these gains have finished and that it seems unlikely they will repeat in the future. His argument is focused on the US and based on falling high school completion rates and increased dropout rates (Gordon, 2015).

To what extent does this argument applies to the Netherlands? High school dropout rates in the Netherlands have decreased since 2003 to 8.8% in 2012, below the EU average of 13% (CBS, 2014). Figure 7 in the appendix shows the enrolment rate in primary, secondary and higher education from 1950 to 2014. The enrolment rate is defined as the number of students enrolled as a percentage of total population. The graph is partly in line with Gordon’s argument, we see that enrolment rates in primary and especially secondary

education have risen in the past century, but have stabilized since 1990. This is what Gordon means when he states:‘’the education revolution is finished’’ (Gordon, 2015).

However, the graph also shows that enrolment in higher education has been rising: a potential force of economic growth not mentioned by Gordon. The findings of figure 7 are further supported by a CBS report on education which states: ‘’in 110 years (between 1901 and 2011) the number of university students has grown by a factor of 86, while the

this trend in higher education affects the labour force we find a similar upward trend. In 2003 (data is only available from this year) 26% of the labour force was higher educated, by 2013 this share had grown to 33.9% (OECD, 2015).

The economic benefits gained from an increased level of education are measured by private and public return. A 2009 study by the OECD among its member countries, including the Netherlands, finds that both private and public returns are higher for tertiary (higher) education then for secondary education (OECD, 2009). When combining evidence of this increased return with the rising education level of the Dutch labour force, we could conclude education might contribute to future economic growth, instead of limit it as suggested by Gordon. Though evidence on the economic gains from higher education is not conclusive, it seems reasonable to assume education is not a force reducing potential growth. In other words, it is reasonable to assume education is not a headwind causing secular stagnation in the Netherlands.

Inequality

Thirdly there is the issue of inequality. Increased inequality can suppress long run potential growth when income gains mainly benefit the few. When income becomes concentrated with a few, economic theory suggests the propensity to spend will decrease and effective demand will fall. Evidence on this relation is found in a recent IMF study which finds that when the income share of the top 20% increases with 1%, this lowers GDP growth over the following five years by 0.08% (Dabla-Norris et al., 2015). Another study by the OECD finds that: ‘’rising inequality by 3 Gini points, that is the average increase recorded in the OECD over the past two decades, would drag down economic growth by 0.35 percentage point per

year for 25 years: a cumulated loss in GDP at the end of the period of 8.5 per cent’’ (OECD, 2014).

Inequality is complicated to measure. In general inequality is separated over three elements: income, wealth and consumption. All are worthy of an entire thesis, but here I will only focus on income and wealth inequality in the Netherlands.

A measure for income inequality is the S90/S10 ratio: the share of all income received by the top decile divided by the share of the first (OECD, 2015). Plotting this measure for the

Netherlands we find that, though inequality has increased over the past 20 years, it seems to have stabilized around 6.6 in recent years (see figure 8 in the appendix) . As there is no economic or morally justifiable ratio for this measure, it is hard to say whether this level of inequality is sufficient to hamper growth. The only way to put the ratio in perspective is by comparing. From a European perspective the Netherland perform moderately well

compared to the European average of 7.6 and the lowest score of 5.3 by Denmark (OECD, 2015). This average EU-wide performance by the Netherland is further supported by data on the Gini coefficient of the income distribution as presented in a report by the Dutch

Scientific Council (WRR, 2014). Income inequality seems less significant when placed in a larger perspective.

Measuring wealth inequality is considerably more difficult than income inequality, if only for the great variety of definitions. Here I will assume wealth to stand for capital, or in Dutch ‘’vermogen’’, which exists out of financial capital and housing. The measure I use for capital inequality is the distribution of capital share across ten deciles. Capital inequality, as shown in the graph below, is internationally speaking relatively high in the Netherlands: the highest decile owns 61% of all capital, while the bottom 50% owns -3%.

Figure 9. Capital distribution in the Netherlands

Source: Van Bavel, 2014

Including pension benefits in this calculation would improve the distribution, but it is more reasonable to treat these pensions as deferred income and therefore leave them out of the capital distribution (Van Bavel, 2014). The negative capital share of the bottom 50% is an indicator of the indebtedness of the Dutch population, a matter to which I will return later when discussing the notion of a balance sheet recession. For now, it seems reasonable to state that wealth inequality has a stronger presence then income inequality and could qualify as a force pressing down long run potential growth.

Overall, the inequality argument is ambiguous. On the one hand, income inequality is not yet harmful, but the low disposable income growth over the past years predicts future trouble (see figure 8 in the appendix). Wealth inequality on the other hand is more severe and has possible consequences for growth in the long run. Combining the two, it seems inequality in the Netherlands is as of yet not a predominant cause for secular stagnation but has the potential to become one were current developments to continue.

-10% 0% 10% 20% 30% 40% 50% 60% 70% 1 2 3 4 5 6 7 8 9 10 Sh ar e o f t ot al c ap ita l, in % Capital distibution

Public debt

The final headwind is public debt. With public debt reaching new heights, the government’s ability to support the economy could become limited in the future. Restricted by the rules set by the Maastricht treaty (3% deficit/GDP and 60% public debt/GDP) room for fiscal expansion would be limited (ECB, 1992). In the long run this could not only deepen future recession but also structurally affect financial markets through crowding out effects. The best way to measure the public debt headwind is to use the public debt to GDP ratio.

Figure 10. The public debt to GDP ratio

Source: CBS, 2015

The graph above shows the development of the public debt to GDP ratio and its components over the past 20 years. First we see the ratio decreasing as a result of growing GDP and public debt staying relatively stable. But after 2008 we see the opposite, GDP growth stabilizing and public debt rising. In this period the ratio rose from 43% to the current level of 69%. As with inequality, there is no economic optimal value for this ratio, though the benchmark proposed by the IMF is 60% (Chowdhury, 2010). When comparing with other

20% 30% 40% 50% 60% 70% 80% 100000 200000 300000 400000 500000 600000 700000 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 De bt /G DP Ra tio , i n % GDP a nd d eb t, in m ill io ns

The public debt to GDP ratio

European countries the current ratio looks less excessive with an EU average of 86%. The current public debt to GDP ratio seems not yet detrimental to long-run potential growth.

When taking a dynamic look at the public debt issue, the structural problem emerges more clearly. In the current situation of stagnating GDP growth, there is no offsetting denominator effect; meaning every increase in debt will instantly rise the public debt to GDP ratio. These effects are summarized in the basic debt dynamic equation as described in figure 11 in the appendix. From this equation it becomes clear that if the current situation of near zero inflation and interest were to continue (see figure 12 in the appendix), changes in the public debt to GDP ratio will solely depend on the public deficit and the real GDP growth rate. A deficit will increase the ratio, whereas a rise in the GDP growth rate will lead to a decrease. Seeing as the first is more likely to happen then the second, it is reasonable to assume the public debt to GDP ratio will continue to rise in the future. Were this trend to continue for long enough, public debt could become a force of secular stagnation.

Having discussed technology and the four headwinds as reasons for the decline in the long run potential growth rate, we can provide a verdict on the first pillar. Data shows the long-run potential growth rate in the Netherlands has been gradually slowing down, with the most notable decline present in labour input and total factor productivity. Of the arguments discussed in this thesis, demography is the most prominent reasons. The stagnant working age population, combined with stable participation rate in the Netherlands implies output per capita can only increase through labour productivity growth, which has been weak since the financial crisis and is not expected to grow strong in the future. The evidence on the education headwind suggests that it is not a force of secular stagnation. Though enrolment rates in primary and secondary education are no longer growing like they have the past century, the rising share of higher educated workers could continue to contribute to future

potential growth. The headwinds of inequality and public debt are present in the

Netherlands but at the moment appear not to be main causes for declining potential growth; though their prospect is alarming. Concluding, the slowing long run potential growth rate is a force of secular stagnation in the Dutch economy.

Chapter 4 The second pillar: output gaps

The second pillar focuses on the output gap: secular stagnation as the persistent

underperformance of output below potential. The key measure for this is the output gap measured as the deviation of output from potential. A negative output gap means output was below potential; a positive gap means the opposite.

The development of the output gap in the Netherlands and in the EU is shown in the graph below. Both suffer a significant setback in 2009, the EU even more then the Netherlands.

Figure 13. Output gap in the Netherlands and the EU

Source: European Commission, 2015

But after shock we see the EU quickly gaining moment and moving to positive output gaps, while the Netherlands has kept a negative output gap since 2008. Though the Dutch

economy has seen an upward trend from 2013, output gaps are still negative and below the EU benchmark. In short, Dutch output gaps have been structurally negative, as well as below overall EU performance. -5% -4% -3% -2% -1% 0% 1% 2% 3% 4% 5% 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 O uput g ap, in %

What has caused this structural and severe negative output gap in the Netherlands? In this chapter I will look at four possible reasons: unemployment, income growth, the notion of a balance sheet recession and finally the real natural interest rate gap. Below I discuss these arguments in turn, first establishing their relation to the output gap, followed by an analysis of data from the Netherlands for each subject. I end the chapter with a short summary and verdict for the second pillar.

Unemployment

This first possible reason for the output gap is unemployment: high unemployment can keep output from reaching its potential. When unemployment fails to improve in the wake of a recession, this could suggest it has become more structural. Growing structural or long-term unemployment could systematically keep output from reaching potential.

Figure 14. Unemployment by duration

Source: CBS, 2015

Figure 14 above shows that both short-term and long-term unemployment have increased steadily since 2008 to 7% of the labour force. Long-term unemployment has gone from 1.4%

0% 1% 2% 3% 4% 5% 6% 7% 8% 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 U ne mp lo yme nt , i n % o f l ab ou r f orc e

to 3.1%. Though we have seen long-term unemployment spike before 2008 as well, the current share of short-term unemployment (4.4% out of 7.4%) is exceptionally low.

The total unemployment rate has stabilized around 7% for the last two years; a level not seen since 1996 when it stood at 7.7%. When looking within the EU the Netherlands end up in the middle with overall EU unemployment standing at 10%. However, opposed to the Netherlands, most EU members have seen their unemployment rate decrease over the last two years. The current level of unemployment in the Netherlands is still rising and has a growing share of structural unemployment. This trend can be viewed as critical enough to cause consistent output gaps.

This observations is further supported when we look at data on the Dutch NAIRU, the non-accelerating inflation rate unemployment. Figure 15 below combines estimates of the NAIRU with the unemployment rate and output gap.

Figure 15. NAIRU & unemployment

Source: European Commission, 2015 (Output gap); ECB, 2015 (unemployment rate); OECD, 2015 (NAIRU) 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 -4% -3% -2% -1% 0% 1% 2% 3% 2 3 4 5 6 7 8 Ac tua l O uput G ap, in % NA IR U & U ne m pl oym ent ra te , i n %

NAIRU & unemployment

The NAIRU can be thought of as equilibrium unemployment: below this level inflation expectations will rise, above this level they will fall and cause disinflation. We find that the structural output gap of the past years largely corresponds to a period of above NAIRU unemployment rates. This suggests unemployment as a reason for economic

underperformance as well as the observed falling level of inflation. In short, unemployment can be assumed as a main cause for the output gap.

Income growth

The second reason to be discussed is income growth. Output gaps can be the result of weak income growth when some growth exists but is not translated into real income growth. The measure I will use for income growth is mean disposable income growth which the OECD defines as ‘’total income less current transfers paid’’ (OECD, 2015c).

Evidence from the Netherlands shows that household mean income has barely grown since 2008, just 0.6% on average (see figure 8 in the appendix). Furthermore, when separating over different age groups we see a strong divergence between the older and younger generations (see figure 16 in the appendix). Where the 65+ age group has seen there disposable income grow by an average of 1.2% since the crisis, those between 18 and 65 have seen theirs decrease by 0.3%.

The effect of this generational inequality on output is ambiguous. A possible way of

analysing its effect is by looking at the marginal propensity to consume (mpc). If the mpc was considerably higher for the younger, income gains solely benefiting the elder (with lower mpc) might depress economic growth. The lifecycle hypothesis by Modigliani suggests a high mpc for the young, a lower for the middle aged generations but a subsequent rise upon reaching the retirement age (the mpc is implied to move opposite to the saving rate over the

life cycle). However, a 2009 paper analysing these effects only finds conformation for the saving rate, with the impact of an ageing demographic on the marginal propensity to consume less pronounced (Dynan, Edelberg & Palumbo, 2009).

Combining this result with the Dutch data on income growth, there is little evidence generational inequality is a problem as of yet. Income growth in the Netherlands has been weak and has done little to support economic recovery. However, the problem does not seem severe enough to be a probable source for the structural output gap.

Balance sheet recession

The third argument considers the notion of a balance sheet recession. This concept is put forward by Richard Koo and states that following the burst of a debt-financed bubble we see a tendency towards private debt minimization. The bursting of the debt financed bubble causes asset prices to drop while liabilities remain; leaving private balance sheets

underwater (Koo, 2011). As a consequence, both companies and households prioritize paying off exisiting debts before taking on new loans or making investments. Households become more austere, paying down debt and saving more while spending less. Though reasonable on the individual level, it structuraly decreases an economy’s aggregate demand and can cause output to stagger. The long run consequences can be powerfull and lasting, as Koo shows for the case of Japan’s lost decade (Koo, 2011).

Are we suffering from a balance sheet recession in the Netherlands? Total private debt currently stands at 191% of GDP (see figure 17 in the appendix). I will not go into much detail considering the balance sheet recession effects for companies. The main reason for this is that Dutch companies are not exceptionally leveraged, with an average debt to equity ratio of 2 (OECD, 2012b). The most common type of household debt in the Netherlands are

mortgages, figure 18 in the appendix further details the composition of private debt. The most prominent reason for the mortgages majority share in private debt is the tax deductable interest on mortgages, or in Dutch ‘’hyptheekrenteaftrek’’.

The graph below shows the development of household mortgages, GDP (both drawn on the left axis) and the change in average housing prices (on the right axis) over the past 20 years.

Figure 19. Mortgages, GDP and housing prices

Source: CBS, 2015

What this graph shows is that household mortgages have seen a rapid expansion from 54% of GDP in 1995 to 100% of GDP in 2013. Over the same period the growth in value of housing has stagnated and in the wake of the recession declined; 2013 being the first year since 2007 where the change in average housing prices was positive, by 1%. The underwater mortgages have become a real problem in the Netherlands. As a 2014 report on financial stability by the Dutch Central Bank states: ‘’30% of all mortgages are underwater… which will be an ongoing concern for both banks as well as the Dutch government’’ (DNB, 2014). What the report finds as well is the high concentration of underwater mortgages with younger generations, around 60% for 20-40 year olds, as opposed to an average 15% for those age 40 to 70+ (a result

-10% -5% 0% 5% 10% 15% 20% 100000 200000 300000 400000 500000 600000 700000 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 Ho us e pr ic es , % c ha ng e to pr ev io us per io d Ho us eho ld m or tg ag es a nd G DP , i n m ill io ns

GDP, Mortgages and house prices

even more distressing when combined with the earlier findings on the generational divergence of disposable income. In short, the situation in the Netherlands prior and

following the financial crisis of 2008 seems favourable for balance sheet recession effects to take hold.

There are several indicators balance sheet effects have indeed influenced the Dutch economy following the debt crisis of 2008. First, as the previous graph shows, we see a decline in mortgages since 2012. Secondly, household austerity seems to have increased as figure 20 in the appendix shows a strong divergence between household saving and

investment after 2008. Furthermore, household consumption as a percentage of GDP has seen a drop as well, though one could argue this to be the continuing of a trend originating before 2008 (see figure 20 in the appendix). The excess household savings extents to the overall economy where we see a growing amount of excess savings as a result of both a rise in savings as well as a fall in investment (figure 21). The steady increase in excess savings can be identified as a results of balance sheet recession effects.

The population of the Netherlands are among the most leveraged.When measured by the debt to disposable income ratio, they are only surpassed within the EU by Denmark (OECD, 2012a). The rapid expansion of mortgages, encouraged by tax deductable interest payments, has led to both a significantly high debt and house price level; leaving the Netherlands especially vulnerable to balance sheet recession effects. The main consequence is continued household austerity, which can be assumed as a major reason for the persistent output gap.

Real natural interest rate gap

The final variable relevant for the output gap is the real natural interest rate gap. The real natural interest rate is a theoretical concept introduced by Wicksell (as the neutral interest

rate) and most commonly defined as ‘’ the real short-term interest rate consistent with output at its potential level and a stable rate of inflation’’ (ECB, 2004). This definition immediately shows its relevance to the topic of this thesis: if the real interest rate

structurally deviates from the real natural interest rate, this could keep output from reaching potential. The concept is also related to the previously discussed NAIRU: the real natural interest rate is the real interest rate consistent with unemployment at NAIRU level.

The real natural interest rate gap is the main reason for structural stagnation proposed by Larry Summers in his ‘’new secular stagnation hypothesis’’. He suggests that if the natural real interest rate has fallen significantly below zero, it might be the case that with nominal interest rates fixed at the zero lower bound (ZLB) and ample inflation, the real interest rate is unable to reach its natural level (Summers, 2015b).

What evidence of Summers’s ‘’new secular stagnation hypothesis’’ do we find in the Netherlands? First we need to estimate the path of the natural real interest rate in the Netherlands. Secondly, we need to compare this estimate to the real interest rate to see whether a gap between the two has existed that prevented output from reaching potential.

Since the natural real interest rate is a theoretical concept, numerical estimates are

complicated to obtain and often the result of complex statistical methods (see for example the 2007 paper by Benati and Vitale). A more intuitive understanding of the natural interest rate and its trend is found by looking at its underlying determinants, mainly productivity and population growth (ECB, 2004).

As stated earlier in this paper, both productivity and population growth in the Netherlands have stagnated in a way similar to those seen across the EU and summarized in a monthly bulletin by the ECB from 2014 (ECB, 2014). Labour productivity growth has steadily declined

while population growth has stagnated: both putting a downward pressure on the real natural interest rate. From the similarity in trends between the Netherlands and Europe in general, we can reasonable assume the natural real interest rate in the Netherlands has followed a path similar to Europe.

The current real interest rate in the Netherlands is approximated to be below zero. The nominal rate (using the annualized short-term interest rate as a proxy) has been near zero for four consecutive years, while inflation (using the basic CPI as a proxy) has steadily declined and has averaged 0.3% in 2015 (see figure 12 in the appendix). Though the real interest rate is perceived to be negative, it is likely the real natural interest rate has seen a significantly steeper decline and lies even lower; an indication of a structural real interest rate gap. The earlier observed difference between the unemployment rate and the NAIRU is consistent with this claim, in that it is impossible for the real rate to reach its potential when unemployment is substantially removed from equilibrium. We can assume the Netherlands, like the rest of Europe, is suffering from structural real interest rate gap preventing output from reaching potential.

The Netherlands have experienced a significant output gap over the past years. The most relevant reasons found in this chapter are structural unemployment, balance sheet recession effects and a negative real interest rate gap. Weak income growth has not been supportive of economic growth but is here not found to be as detrimental as the other arguments. Unemployment has been above the NAIRU estimate since 2008, with especially long-term unemployment becoming more prominent. Balance sheet recession effects have been strong in the highly leveraged Dutch economy, with household austerity as a result. Finally there is evidence the real natural rate has decreased to a level below the real natural rate, which at the current level of inflation will remain unreachable and limit output.

Chapter 5 Conclusion

Secular stagnation is a cause for concern in the Netherlands. The previous chapters have shown that both a decline in the long run potential growth and a persistent output gap have been causes of secular stagnation in the Netherlands.

Long run potential growth has fallen mostly as a result of the stagnating working age population. This negative effect has in the past been offset by an increasing participation rate that kept the Dutch labour force growing. In recent years however, the participation rate has stabilized, as result of which output per capita can solely increase through labour productivity growth. Labour productivity growth has been weak since the crisis and research suggests it is unlikely to gain momentum in the future. Evidence on education as a force of secular stagnation is inconclusive. Though secondary and primary enrolment rates are no longer growing like they have in the past century, higher education enrolment has been steadily rising. As a consequence, the education level of the Dutch labour force is still increasing and has the potential to continue contributing to future economic growth. The other headwinds, inequality and public debt, are found to be present in the Netherlands but not severe enough to have limited the potential growth rate. Of these, the growing level of public debt has the most potential to become relevant to secular stagnation in the future.

The endurance of the output gap is found to be the result of structural unemployment, balance sheet recession effects and a negative real interest rate gap. First, the

unemployment rate in the Netherlands has been rising since 2008 with an increasing share being long-term. Further evidence of unemployment’s role in the output gap is found in the consistent divergence between the unemployment rate and estimates of the Dutch NAIRU. Secondly, balance sheet recession effects have been a principle cause for secular stagnation

in the Netherlands. The Dutch population is leveraged through mortgages whose value went underwater in the wake of the crisis. The subsequent household austerity has been a reason for stagnating household consumption and excess savings. Finally, a decline in the real natural interest rate is found to be a structural reason for the output gap. Assuming the real natural rate interest rate in the Netherlands has followed the same downward trend as elsewhere in the EU, the current near zero real interest rate is likely insufficiently low. An argument supported by the earlier mentioned gap between the unemployment rate and NAIRU. With the nominal interest stuck at the ZLB and little inflation it is doubtful this will improve in the future.

The most apparent deficiency of this paper is its limited gaze forward. Secular stagnation is shown to have been a gradual development over the past decades, but a stronger focus towards the future is required for making sound projections about the direction of trends like demographics and the real natural interest rate. Another limit of this thesis is the lack of precise estimates for the balance sheet recession effects. A better quantitative

understanding of these effects is useful considering the major role they play in the Dutch economy.

An interesting direction for further research is the policy consequences of secular

stagnation. As is shown in this paper secular stagnation has deep roots which will require firm policy if they are to be changed. Without an adequate revision of policy aimed at solving the structural issues raised in this thesis, economic growth in the Netherlands is likely to remain anaemic in the years to come.

Chapter 6 Appendix

Figure 1: European GDP levels

Source: OECD, 2015

Figure 2: The output gap and potential growth rate

Source: European Commission, 2015

0 500 1000 1500 2000 2500 3000 3500 4000

France Germany Italy Poland Spain Netherlands

GDP , i n b ill io ns European GDP levels 2014 2008 2002 -4% -3% -2% -1% 1% 2% 3% 4% 5% 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 O ut put g ap & P ot ent ia l g ro w th r at e, in %

The output gap and potential growth rate in the Netherlands

Table 1: Average potential growth rates

Source: European Commission, 2015

Average growth rate before 2000 Average growth rate since 2000 Labour contribution _{0.7% 0.6%} Capital contribution _{0.9% 0.3%} TFP contribution _{1.2% 0.6%}

Total potential growth rate

2.8% 1.5%

Figure 3: Contribution to potential growth

Source: European Commission, 2015

0% 1% 1% 2% 2% 3% 3% 4% 4% 1985 1990 1995 2000 2005 2010 2015 Po te nti al g ro w th ra te , i n %

Contribution to potential growth

Total labour (hours) contribution Capital accumulation contribution TFP contribution

Figure 4: R&D and potential growth

Source: OECD, 2015 (gross domestic R&D spending); European Commission

(potential growth, TFP contribution), 2015

Figure 5: Long run trend participation & population growth

Source: European Commission, 2015

0% 1% 1% 2% 2% 3% 3% 4% 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 Po te nt ia l gr ow th ra te , i n % / R& D a s % o f GD P

R&D and potential growth

Gross domestic spending on R&D Potential growth TFP contribution To Potential

0% 1% 2% 3% 4% 5% 6% 7% 8% 9% 10% 60% 62% 64% 66% 68% 70% 72% 74% 76% 78% 80% Gr ow th of w okr in g ag e pop ul at ion , i n % Tr en d pa rt ici pa tio n, % o f w or ki ng age pop ul at ion

Long Run Trend Participation & Population Growth

Figure 6: Output growth per capita

Source: Eurostat, 2015 (Real labour productivity); European Commission

(Change in hours), 2015

Figure 7: Enrolment rates

Source:CBS, 2015

Ouput growth per capita = growth labour productivity * growth in hours per capita

Real labour productivity growth Changes in hours (per empl) contribution Ouput growth per capita

0.00% 2.00% 4.00% 6.00% 8.00% 10.00% 12.00% 14.00% 16.00% 18.00% 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 St ud en ts en ro lled , a s a % o f t ot al po pu la tio n Education enrolment

Figure 8: Mean disposable income and the S90/S10 ratio

Source: OECD, 2015

Figure 9: Capital distribution in the Netherlands

Source: Van Bavel, 2014

6.00 6.20 6.40 6.60 6.80 7.00 7.20 -10% -5% 0% 5% 10% 15% 20% 25% 30% 1995 2000 2005 2006 2007 2008 2009 2010 2012 S90/ S10 r at io M ea n di spo sa bl e inc om e gr ow th, in %

Mean disposable income and S90/S10 ratio

Mean disposable income growth S90/S10 disposable income decile share

-10% 0% 10% 20% 30% 40% 50% 60% 70% 1 2 3 4 5 6 7 8 9 10 Sh ar e o f t ot al c ap ita l, in % Capital distibution

Figure 10: The public debt to GDP ratio

Source: CBS, 2015

Figure 11: The dynamics of public debt

20% 30% 40% 50% 60% 70% 80% 100000 200000 300000 400000 500000 600000 700000 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 20 10 20 11 20 12 20 13 20 14 De bt /G DP Ra tio , i n % GDP a nd d eb t, in m ill io ns

The public debt to GDP ratio

Real GDP in millions Public Debt in millions Public Debt/GDP

∆

𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷

_{𝐺𝐺𝐷𝐷𝐺𝐺}

= 𝑑𝑑𝐷𝐷𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝐷𝐷 + (𝑑𝑑 − 𝜋𝜋 − ∆𝑌𝑌) ∗

𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷

_{𝐺𝐺𝐷𝐷𝐺𝐺}

∆

𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷

_{𝐺𝐺𝐷𝐷𝐺𝐺}

= 𝑑𝑑𝐷𝐷𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝐷𝐷 − ∆𝑌𝑌 ∗

𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷

_{𝐺𝐺𝐷𝐷𝐺𝐺}

Figure 12: Interest rates & inflation

Source: OECD, 2015 (Short-term interest rate); CBS, 2015 (CPI)

Figure 13. The output gap in the Netherlands and the EU

Source: European Commission, 2015

-1% 0% 1% 2% 3% 4% 5% 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 Int er es t r at e & i nf la tio n, in %

Interest rates & inflation

Short term interest rate per annum CPI

-5% -4% -3% -2% -1% 0% 1% 2% 3% 4% 5% 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 O uput g ap, in %

Ouput gaps in the Netherlands and EU

Figure 14: Unemployment by duration

Source: CBS, 2015

Figure 15. NAIRU & unemployment

Source: European Commission, 2015 (output gap); ECB, 2015 (unemployment

rate); OECD, 2015 (NAIRU)Figure 8: Mean disposable income growth by age

group

Long-term unemployment Short-term unemployment

2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 -4% -3% -2% -1% 0% 1% 2% 3% 2 3 4 5 6 7 8 Ac tua l O uput G ap, in % NA IR U & U ne m pl oym ent ra te , i n %

NAIRU & unemployment

Figure 16: Disposable income growth by age groups

Source: European Commission, 2015

Figure 17: GDP and private debt

Source: CBS, 2015

Mean disposable income growth by age group

Average Disposable Income Growth Age 18-40 Average Disposable Income Growth Age 65+

0 200000 400000 600000 800000 1000000 1200000 1400000 GDP a nd De bt , i n m ill io ns

GDP and private debt

Figure 18: Composition of household debt

Source: CBS, 2015

Figure 19: Mortgages, GDP and house prices

Source: CBS, 2015

Household mortgages Other long term household loans Short-term loans

-10% -5% 0% 5% 10% 15% 20% 100000 200000 300000 400000 500000 600000 700000 Ho us e pr ic es , % c ha ng e to pr ev io us pe rio d Ho us eh ol d m or tga ge s an d GDP , i n m ill io

ns GDP, Mortgages and house prices

Figure 20: Household savings, investment and consumption

Source: CBS, 2015

Figure 21: Savings and investment

Source: CBS, 2015

Household savings Houshold investment Household consumption

0 20000 40000 60000 80000 100000 120000 140000 160000 180000 200000 Sa vi ngs & in ve st m en t, in m ill io ns Excess savings

Chapter 7 Bibliography

1. Benati, L., Vitale, G. (2007). Joint Estimation of the Natural Rate of Interest, the Natural Rate of unemployment, expected inflation, and potential output. ECB Working Paper Series, 797.

2. Bernanke, B. (2015). Why are interest rates so low, part 2: Secular stagnation, Ben Bernanke's Blog. Retrieved 17 June 2015 from http://www,brookings,edu/blogs/ben-bernanke/posts/2015/03/31-why-interest-rates-low-secular-stagnation.

3. CBS. (2011). Jaarboek onderwijs in cijfers 2001. CBS. Retrieved 23 June 2015 from

http://www.cbs.nl/NR/rdonlyres/FC6D3388-0F9E-4129-8F2B-53022BA3F774/0/2011f162pub.pdf

4. CBS. (2014). Number of school dropouts down in the Netherlands and the EU. Retrieved 23 June 2015 from:

http://www.cbs.nl/en-GB/menu/themas/onderwijs/publicaties/artikelen/archief/2014/2014-3940-wm.htm

5. CBS. (2015). Statline. Retrieved 19 June 2015 from: http://statline.cbs.nl/Statweb/?LA=nl.

6. Chen, V., Gubta, A., Therrien, A., Levanon, G., & Van Ark, B. (2015). Recent Productivity Developments in the World Economy: An Overview from The Conference Board Total Economy Database. The Conference Board. Retrieved 17 June 2015,

from http://www,conference-board,org/pdf_free/Productivity_TED,pdf.

7. Chowdhury, A., Islam, I. (2010). Is there an optimal debt-to-GDP ratio?. VOX CEPR’s Policy Portal. Retrieved 18 June 2015,

from http://www,voxeu,org/debates/commentaries/there-optimal-debt-gdp-ratio.

8. Dabla-Norris, E., Kochhar, K., Suphaphiphat, N., Ricka, F., & Tsounta, E. (2015). Causes and Consequences of Income Inequality: A Global Perspective. IMF,org. Retrieved 17 June 2015, from https://www,imf,org/external/pubs/ft/sdn/2015/sdn1513,pdf.

9. Dennis, C., Grenouilleau, D., Mc Morrow, K., & Röger, W. (2006). Calculating potential growth rates and output gaps - A revised production function approach -. European Economy, 247, 5,6.

10. DNB. (2014). Overzicht Financiële Stabiliteit Voorjaar 2014. DNB. Retrieved 19 June 2015 from http://www,dnb,nl/binaries/OFS3_tcm46-306230,pdf.

11. Dynan, K., Edelberg, W., & Palumbo, M. (2009). The Effects of Population Aging on the Relationship between Aggregate Consumption, Saving, and Income. American Economic Review, 99(2), 380-386.

12. ECB. (1992) Treaty on European Union (p. 84). Retrieved 19 June 2015 from https://www.ecb.europa.eu/ecb/legal/pdf/maastricht_en.pdf

13. ECB. (2004). The Natural Real Interest Rate in the Euro Area (pp. 57-69). Retrieved 19 June 2015 from https://www,ecb,europa,eu/pub/pdf/mobu/mb200405en,pdf.

14. ECB. (2014). Real interest rates in the euro area: a longer-term perspective. Retrieved 19 June 2015 from https://www,ecb,europa,eu/pub/pdf/other/mb201407_focus03,en,pdf.

15. Eichengreen, B., (2015). Secular Stagnation: The Long View †. American Economic Review, 105(5), 66-70. doi:10,1257/aer,p20151104.

16. Eurostat. (2015). General government gross debt - annual data. Retrieved 17 June 2015, from http://ec,europa,eu/eurostat/tgm/table,do?tab=table&init=1&language=en&pcode =teina225&plugin=1.

17. European Commission. (2015). CIRCABC – Output gaps –. Retrieved 19 June 2015, from https://circabc.europa.eu/faces/jsp/extension/wai/navigation/container.jsp?FormP rincipal:_idcl=FormPrincipal:_idJsp32&FormPrincipal_SUBMIT=1&id=671d465b-0752- 4a2e-906c-a3effd2340ba&javax.faces.ViewState=rO0ABXVyABNbTGphdmEubGFuZy5PYmplY3Q7kM 5YnxBzKWwCAAB4cAAAAAN0AAExcHQAKy9qc3AvZXh0ZW5zaW9uL3dhaS9uYXZpZ2F0a W9uL2NvbnRhaW5lci5qc3A=

18. Gordon, R. (2015). Secular Stagnation: A Supply-Side View †. American Economic Review, 105(5), 54-59. doi:10,1257/aer,p20151102.

19. Hansen, A. (1939). Economic Progress and Declining Population Growth. American Economic Review, 29(1), 1-15.

20. IMF. (2015). IMF World Economic Outlook Database List. Retrieved 13 April 2015, from http://www,imf,org/external/ns/cs,aspx?id=28.

21. Koo, R. (2011). The world in balance sheet recession: causes, cure, and politics. Real-world economics review, 58,19-37.

22. OECD. (2009). The Economic Benefits of Education. Retrieved 19 June 2015, from http://www,oecd,org/berlin/42244848,pdf.

23. OECD. (2012a). Household Accounts – Household Debt – OECD Data. Retrieved 19 June 2015 from https://data,oecd,org/hha/household-debt,htm#indicator-chart.

24. OECD. (2012b). Corporate - Financial corporations debt to equity ratio - OECD Data. Retrieved 19 June 2015, from https://data,oecd,org/corporate/financial-corporations-debt-to-equity-ratio,htm.

25. OECD. (2015a). Focus On Inequality And Growth. Retrieved 19 June 2015, from http://www,oecd,org/els/soc/Focus-Inequality-and-Growth-2014,pdf.

26. OECD. (2015b). Education attainment - Adult education level - OECD Data. Retrieved 19 June 2015, from https://data,oecd,org/eduatt/adult-education-level,htm#indicator-chart.

27. OECD. (2015c). Income Distribution and Poverty – OECD Data. Retrieved 19 June 2015, from http://stats,oecd,org/index,aspx?queryid=46022.

28. Summers, L. (2014). U.S. Economic Prospects: Secular Stagnation, Hysteresis, and the Zero Lower Bound. National Association For Business Economics, 49(2), 65-73.

29. Summers, L. (2015a). On Secular Stagnation: A Response to Bernanke. Lawrence H. Summers. Retrieved 19 June 2015, from http://larrysummers,com/2015/04/01/on-secular-stagnation-a-response-to-bernanke/.

30. Summers, L. (2015b). Demand Side Secular Stagnation †. American Economic Review, 105(5), 60-65. doi:10,1257/aer,p20151103

31. Tuelings, C., & Baldwin, R. (2014). Secular Stagnation: Facts, Causes and Cures. VoxEU CEPR. Retrieved 19 June 2015,

from http://www,voxeu,org/sites/default/files/Vox_secular_stagnation,pdf.

32. Van Bavel, B. (2014). Vermogensongelijkheid in Nederland: De Vergeten Dimensie. WRR. Retrieved 17 June 2015, from

http://www,wrr,nl/fileadmin/nl/publicaties/PDF-verkenningen/V28_Hoe_ongelijk_is_NL_hfdst04,pdf.

33. WRR. (2015). Economische Ongelijkheid in 8 Figuren. Retrieved 17 June 2015, from http://www,wrr,nl/fileadmin/nl/publicaties/PDF-verkenningen/V28_WRR-factsheet_ec_ong-printversie_A_web4,pdf.