Working Paper 2016.04

repec.nep.eec

n?u=RePEc:inf:wpaper:2016

Working Paper 2016.04

Stocks or flows? New thinking about

monetary transmission through the lending

channel

By

Javier Villar Burke

(European Commission)

www.infer-research.net

www.infer-research.eu


Stocks or flows? New thinking about

monetary transmission through the

lending channel

-- v8 --

Javier Villar Burke*

August 2016

Abstract: The lending channel is conventionally understood to transmit monetary policy through the

origination of new loans. In this paper, we postulate that the lending channel may also operate via the

stock of existing loans. Monetary shocks generate two types of income effects: 1) monthly mortgage

payments are impacted when rates are reset; 2) inflation erodes the real value of mortgage payments

and increases the disposable income of borrowers. These income effects translate into variations in

output due to the heterogeneous propensity to consume of individual economic agents. Three types of

factors determine the importance of these income effects for individual households and at macro

level: 1) borrowers’ features, such as income distribution, indebtedness and debt burden, 2) loan

features, such as the period of rate fixation and 3) price developments. Significant differences in these

factors across euro area Member States can distort a homogeneous transmission of the single

monetary policy.

JEL Codes

D33, D47, D90, E43, E51, E52, F36, F42, G21.

Keywords

Euro area, monetary policy, monetary transmission, income effects, lending channel, mortgages.

* European Commission (Contact: javier.villar-burke@ec.europa.eu). I am very grateful to Roman

Sustek and Carlos Garriga for their support and detailed comments as well as to James Daniel

Eubanks for the statistical support provided. I am also grateful to the Federal Reserve Bank of Saint

Louis and the Washington University in Saint Louis: part of this research was undertaken during a

visit to both institutions in summer 2016. Any remaining errors are the fault of the author. The

opinions and statements expressed in this paper are strictly personal and cannot be attributed in any

way to the European Commission.

1


1. INTRODUCTION

The ultimate goal of monetary policy is to support price stability and economic growth without

hampering financial stability. Central banks try to achieve this goal by influencing inflation and the

supply of money. For this, they use tools such as the auctions of liquidity to commercial banks, where

they decide how much money to allot and at what price (the policy rate). They can also inject (or

subtract) additional amount of money into (from) the economy through the outright purchase or sale

of securities. The monetary transmission is the process throughout which different agents and

processes in the economy are influenced by the central bank’s actions so that inflation and output are

steered in the direction aimed by the central bank.

Many factors affecting the supply of money remain beyond the direct control of the central bank. The

majority of the supply of money takes the form of deposits (and similar assets) generated by

commercial banks and, therefore, the base money under the direct control of the central bank

represents a very limited share of the overall supply of money (Chart 1). The so-called money

multiplier approach maintains that the central bank can determine the supply of money by controlling

the base money. However, the Bank of England refutes such an approach (McLeay, Radia and

Thomas, 2014). The authors maintain that the supply of money is determined by the amount of

lending generated by banks.

12,000

10,000

8,000

M3 (Mo ney supp ly)

Base money

M0 (Cu rrency in circulation)

Chart 1: Main monetary aggregates, euro area, € billion

6,000

4,000

2,000

0

199 0 199 5 2000 2005 2010 2015

Source: ECB.

Indeed, the vast majority of money is created by commercial banks when granting new loans as they

simultaneously generate a deposit. These deposits increase the supply of money when borrowers put

them into circulation within the economy (e.g. by purchasing a house, a car or whatever was the

purpose of the loan). Conversely, when a borrower uses her current account to reimburse a loan, she is

“destroying” the total amount of deposits available in the economy as means of payment and,

therefore, the supply of money decreases.

In this context, central banks have placed a special focus on the use of the policy rate and other central

bank tools to influence the process of money creation by commercial banks (ECB, 2011). Although

there are other channels (see, for instance, Brunner and Meltzer, 1988; Hubbard, 1995; Cecchetti,

1995; Bernanke and Gertler, 1995; Mishkin, 1996; Ireland, 2005; and ECB, 2009a), the transmission of

monetary policy through the bank lending channel is considered among the most important ones.

The literature has mainly seen the bank lending channel working through new credit. By decreasing

(or increasing) the policy rate monetary authorities would aim at influencing the cost of new loans

and, consequently, to steer the supply of money and, ultimately, output. Practitioners generally focus

on new lending as well. For instance, the ECB continues to place a particular emphasis on how it can

influence new lending through its unconventionally monetary policies. Indeed, the targeted longerterm

refinancing operations (TLTROs) implemented since mid-2014 explicitly incentivised banks to

2


increase lending. A new programme implemented in spring 2016 (the TLTROs II) reinforced even

further the need of banks to originate new lending to be able to benefit fully from the programme 1 .

Given this focus on new credit, many models of the effect of monetary policy on mortgages tend to

use single period frameworks (Rubio, 2011; Iacoviello, 2005), although sometimes models also

consider a two-period framework (Calza, Monacelli and Stracca, 2013). Given the short term of this

type of models, the flow of new loans and the stock of loans are equivalent.

However, some loans have much longer maturities. Mortgages can have a maturity of up to 20 or 30

years. These long maturities imply that the stocks become substantially larger than flows. Indeed, on

the steady state, the stock of mortgage loans becomes 20 or 30 times larger than the loans originated

in a given year. Given the large size of the stock of long term loans, it seems worth investigating the

potential of monetary policy to be transmitted through the stock of existing mortgages.

But why has the stock of loans apparently been overlooked? Somehow, it was usually assumed that

the stock of loans constituted a legacy of the past and that they could not be influenced by monetary

policy. However, in recent times, an increasing interest has emerged about the possibility of monetary

policy being able to influence current consumers' balance sheets (see, for instance, Garriga, Kydland

and Sustek, 2015; and Badarinza, Campbell and Ramadorai, 2015). The mechanism would work

through disposable income instead of through the total volume of loans and the supply of money.

Changes in the prevalent rates may affect the cost of servicing a loan and this will impact the

disposable income of the borrower. Borrowers have uneven income levels, different propensity to

consume (Carroll et al., 2015; Keys et al. 2014; Di Maggio, Kermani, and Ramcharan, 2015) and

different constrains in the liquidity of their savings (Kaplan and Violante, 2014). Given these

differences, consumption and investment – and, consequently, output – could be impacted by

monetary policy shocks, even if the supply of money remains unchanged, depending on the specific

feature of the borrowers and the loans been affected 2 .

Inflationary effects constitute an additional channel through which monetary policy can generate

income effects on borrowers. Indeed, given that mortgages are nominal contracts with long duration,

cumulative inflation erodes the burden of debt service and generates positive income effects in real

terms. A few recent studies point to these effects either through mortgage loans (Garriga, Kydland and

Sustek, 2015), more widely through the overall debt of households (Adam and Zhu, 2016) or even

through government debt (Hilscher, Raviv and Reis, 2014).

A few factors explain this renewed interest on the impact of monetary policy on existing loans.

Firstly, the total volume of mortgage debt has significantly increased over the past years. This is

explained by a continuous increase in real estate prices beyond increases in nominal wages (Hilbers et

al.). To afford the monthly instalments linked to more expensive residences, households have been

asking for longer and longer maturities and a larger indebtedness.

Secondly, a trend towards a high concentration of income and wealth is observed since the 1960s (see

for instance, Frank and Cook, 1996; OECD, 2011 and 2014; Galbraith, 2012; Stiglitz, 2012; Lakner

and Milanovic, 2013; and Piketty, 2014). Confronted with relatively declining salaries, consumption

levels of the lower and even the middle class have artificially been maintained by resorting to credit.

This has led to the sub-prime crisis and the phenomena of debt overhang, overindebtedness and the

increasing non-performing loans observed in a number of countries (see for instance Homar, Kick and

Salleo, 2015).

Thirdly, once existing loans are considered to be relevant for monetary policy, the specific features of

the loan play a very important role. There is not only a diversity of borrowers, but also a great variety

of loans: from fix-rate mortgages to perpetual mortgages paying interest only; commission and fees,

1 For further details, see ECB (2016) and Villar Burke (2016).

2 There can be also additional effects working through the banks' balance sheets.

3


oth at inception of the loans (e.g. loan study fee, notary fees, life insurance requirements, etc.) as

well as throughout the life of the mortgage (e.g. penalty for early repayment or renegotiation) may

vary from one contract to the other 3 . On the wake of the financial crisis, the ECB has been confronted

with a “broken” transmission mechanism: lending rates across countries show very heterogeneous

reaction to the single monetary policy. The prevalence of certain types of mortgages in some countries

but not in other countries may explain some of this heterogeneous transmission.

Finally, on the wake of the financial crisis, interest rates declined from about 5 per cent to virtually

zero, a drop of an unprecedented magnitude. However, inflation did not increase as it would be

predicted by conventional central banking wisdom. Williamson (2016) argues that a Fisher effect

would explain the collapse of inflation to zero (Chart 2). With virtually no inflation, the income

effects linked to the erosion of value are eliminated and mortgages become in fact more expensive

even if monetary authorities aimed at easing credit conditions.

5%

Chart 2: Inflation rate and policy rate, euro area, percentage

4%

3%

ECB rate

Inflation rate

2%

1%

0%

1999 200 1 2003 200 5 2007 200 9 2011 2013 2015

-1%

Source: ECB.

The recent literature about the relevance of the stock of loans and their characteristics for monetary

policy has gone in two main directions. In some cases, theoretical models are proposed to explain how

the type of mortgage in terms of fixation period determines the strength of the transmission of

monetary policy (e.g. Garriga, Kydland and Sustek, 2015). Other authors have followed a more

empirical approach with US data (Di Maggio, Kermani, and Ramcharan, 2015; Keys et al., 2014).

In this context, the aim of this paper is twofold. First, to document how relevant the stock of loans can

be for explaining the monetary transmission through the lending channel. And second, to document

structural differences across countries which can jeopardise a homogeneous transmission of the single

monetary policy. The paper focuses on the euro area as a whole as well as the four largest euro area

countries: Germany, France, Italy and Spain. However, the relevance of heterogeneity across

countries goes beyond the euro area; this is because global interest rates have moved in a coordinated

manner in the last few decades even if not formally under a monetary union (Henriksen, Kydland and

Šustek, 2013).

Among the main takeaways of this paper, we highlight the following ones:

The stock of loans can be as relevant for monetary policy purposes as the flow of new loans. Annual

mortgage payments have a size similar to annual loan origination. This is because the ratio between

the stock of loans and annual loan origination is 5 to 1 in the case of Germany and as large as 20 to 1 in

the case of Spain.

Income effects generated by the reset of lending rates are inexistent in countries like Germany or France,

where 90 per cent of mortgages are fixed-rate mortgages (FRMs), while they are very significant in

countries like Spain, where 90 per cent of mortgages are adjustable-rate mortgages (ARMs).

3 For a review of the different features of mortgages across EU countries see, for instance, ECB (2013), Drudi et al. (2009), Hall et al. (2015)

and Nouy (2015).

4


Income effects resulting from inflation impact both ARMs and FRMs. Even under the low inflation

environment observed since the 1990s, this effect remains relevant. For instance, between 1999 and

2013, the inflation erosion meant that indebted households increased their real disposable income by 50

per cent of their mortgage payments in Spain.

Differences in inflation are a source of heterogeneity across countries as they accumulate over time. For

instance, inflation only eroded 25 per cent of the value of mortgage payments in Germany during the

same period as the previous example (with an erosion of 50 per cent in Spain).

The potential macroeconomic impact of monetary shocks through the lending channel varies widely

across countries. For instance, mortgage loans represent as little as 23 per cent of GDP in Italy and as

much as 51 per cent in Spain. On a similar fashion, low income households devote up to 40 per cent of

their income to service their debts in Spain but only 12 per cent in Germany.

The rest of the paper is organised as follows. In Section 2, we compare the stock of loans with loan

originations in terms of volumes and the cost of debt service. In Section 3, we document the

heterogeneity in mortgage contracts across countries. In Section 4, we discuss how inflation can

influence the disposable income of indebted households in real terms. In Section 5, we review

borrowers’ features in terms of income level and indebtedness. All these features are assessed against

its potential impact on the effectiveness of monetary policy. Finally, Section 6 concludes.

2. THE RELEVANCE OF THE STOCK OF LOANS FOR MONETARY POLICY

2.1. Volumes: existing vs new mortgages

The starting point to compare existing mortgages with the origination of new mortgages is to look at

their respective volumes 4 . Data show very different patterns across countries in the evolution of these

volumes (Chart 3). In Germany, both stocks and new mortgages have remained stable throughout

most of the period, with an upward trend appearing from 2013. In France, a secular increase in the

stock of loans contrasts with a cyclical behaviour of new loans. In Italy, the crisis led to a halt in the

expansion of stocks, which is driven by a certain credit crunch in loan origination observed since

2012. Since 2015, new loans have started to increase again, but this does not seem to be reflected in

the stock of loans yet 5 . Finally, in Spain, a strong expansion of both stocks and new loans is observed

at the beginning of the 2000s followed by a significant impact of the crisis. The credit crunch in loan

origination is observed since 2007 with further deepening at each new wave of the crisis (in 2009 and

in 2012). The strong impact of the crisis in Spain led not only to stagnated stocks but even to an actual

and significant decrease in the stock of loans 6 . Despites these differences observed across countries, a

common feature seems to be a recovery on both new loans and outstanding loans since 2015

(although only very timidly in the case of Spain).

To better compare the relative importance of new mortgages and the stock of mortgages, we have

calculated the ratio between both magnitudes (Chart 3, bottom-right panel). In the last 5 years, annual

loan origination represented only 20 per cent of the stock of mortgages in the upper end (Germany)

and as little as 5 per cent in the lower end (Spain). This points to a significant importance of the stock

of loans with respect to loan origination. The crisis has increased the importance of “old” loans in the

4 For new loans, we will use the amount of new loans originated throughout a year, i.e. summing up 12 consecutive months through a rolling

window.

5 This is probably explained by price effects: since they peaked in 2009, prices of residential real estate have declined by 20 per cent in Italy.

See Chart A1 in the Annex.

6 In Spain, prices of residential real estate dropped by as much as 40 per cent between peak and trough. See Chart A1 in the Annex.

5


alance sheet of banks as the proportion of new loans has, in general, declined since the mid-2000s.

The increases observed since early 2015 are explained, to a large extent, by the refinancing of existing

loans due to the prevailing very low interest rates. The renegotiations of loans are included in the

statistics of new loans but are note translated into increased on the stock of loans as they do not

represent actual net new loans.

1,200

1,000

400

350

300

250

200

150

100

800

600

400

200

50

Chart 3: Volume of loans: new loans vs total stock, credit for house purchase, € billion

Germany

France

1,000

0

2004 200 6 2008 201 0 201 2 2014 201 6

Italy

Euro area

Stock of lo ans

New lo ans

0

200 4 200 6 200 8 2010 201 2 201 4 201 6

4,500

4,000

3,500

3,000

2,500

2,000

1,500

1,000

500

Stock of lo ans

New lo ans

Stock of loans

New loans

0

2004 200 6 2008 201 0 201 2 2014 201 6

0

200 4 2006 200 8 201 0 2012 201 4 2016

Spain

New loans over stock of loans, percentage

Notes: New loans correspond to the annual moving sum of monthly data. The white gaps indicate structural breaks in the series. The percentage

of new credit is calculated as annual volume of new loans over the stocks of loans.

Source: ECB and own calculations.

700

600

500

400

300

200

100

900

800

700

600

500

400

300

200

100

Stock of loans

New loans

Stock of lo ans

New lo ans

0

200 4 2006 200 8 201 0 201 2 2014 201 6

35%

30%

25%

20%

15%

10%

5%

FR

DE

EA

IT

ES

0%

200 4 2006 200 8 2010 201 2 201 4 2016

2.2. The cost of debt service vs loan origination

The cost of debt service

Given that past decisions cannot be changed, a potential impact of monetary policy on existing loans

would work through the cost of debt services (i.e. through the repayment instalments) or through

inflation effects (see Section 4). The former effect can be illustrated through the instalment plan for a

stylised annuity 7 mortgage of €150,000 and a maturity of 25 years when lending rates drop from 6 to

2.5 per cent (Chart 4) 8 .

7 In an annuity loan, a constant monthly instalment is distributed between decreasing interest payments and increasing amortisation

payments. While annuity loans are the most common, other redemption schemes are also possible such as constant amortisation, interestonly

mortgages, grace periods in the principal or the interest, balloon loans, etc.

8 Such a decline in lending rates from 6.0 to 2.5 per cent was indeed observed in Spain and Italy between 2008 and 2010 (see Chart 8).

6


Monthly payment (euro)

1,000

900

800

700

600

500

400

300

200

100

Chart 4: Cost of debt service for an annuity loan with 25 years of maturity and €150,000 of initial capital

Lending rate of 6.0% Lending rate of 2.5%

Interest payment

Amortisatio n p aymen t

0

0 5 10 15 20

Years

Note: An “annuity loan” is a redemption scheme with monthly fix payments.

Source: Own elaboration.

0

0 5 10 15 20

Years

Several features outstand from this illustration. First of all, the total cost of debt service can be split into

an interest payment and an amortization payment. Second, there is a non-linear relation between both

payments. Third, an increase (or decrease) in the lending rate leads to an increase (or a decrease) in total

interest payments through an increase (or decrease) in the total monthly payment and a redistribution of

the amortization payments throughout the maturity of the loan 9 . In our example, when the lending rate

decreases from 6.0 per cent to 2.5 per cent, the interest payment decreases by €88,000 (from €140,000 to

€52,000) and the monthly instalment decreases by €300 (from €970 to €670).

The traditional view of monetary policy would argue that an expansionary monetary shock that is

ultimately transmitted to lending rates would foster new lending as loans become more affordable so

that there will be more borrowers that could apply for such a loan.

Monthly payment (euro)

1,000

900

800

700

600

500

400

300

200

100

Released disposable in come

Interest payment

Amortisation payment

Reduction

Reduction

300

250

200

150

100

50

0

35%

30%

25%

20%

15%

10%

5%

Chart 5: Reduction of the cost of debt service when lending rate decrease from 6.0% to 2.5%

Monthly instalment, € Cumulative, €

0 5 10 15 20

Moment of rate change, y ears af ter the initial contract

Monthly instalment, percentage

Reduction

Reduction

90,000

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

70%

60%

50%

40%

30%

20%

10%

0

0 5 10 15 20

Moment of rate change, y ears af ter the initial contract

Cumulative, percentage

Released income (over interest payments)

Released income (over total payments)

0%

0 5 10 15 20

0%

0 5 10 15 20

Moment of rate change, y ears af ter the initial contract

Moment of rate change, y ears af ter the initial contract

The chart illustrates the effect of a reset in lending rates on the disposable income of the borrower, depending on the moment of the reset. For

instance, if the reset occurs 5 years after the loan was initially originated, the monthly instalment will decrease by €250 (top-left panel) and will

mean a cumulative increase of €60,000 in the disposable income of the borrower throughout the remaining 20 years of the life of the mortgage

(top-right panel). In relative terms, the monthly instalment will decline by 25 per cent (bottom-left panel), the cumulative mortgage payments will

decrease by 20 per cent and the overall interest payments will drop by 40 per cent (bottom-right panel). These reductions in the debt burden

imply an increase in the disposable income of the borrower.

Note: The simulation is based on an annuity loan with 25 years of maturity and €150,000 of initial capital.

Source: Own elaboration.

9 Note that total amortization is always equal to the initial capital independently of the lending rate.

7


In this paper, we argue that the reduction in loan rates stemming from a monetary expansion could also

impact existing loans via resets. This effect would work through income effects. In our example,

borrowers could have up to €300 more disposable income a month and up to €88,000 more disposable

income throughout the life of the mortgage for increasing their consumption. Given that the stock of

loans is significantly larger than loan origination (as large as 20 to 1 in the case of Spain, see Section

2.1), the potential income effects on debt service of existing loans through resets can be very significant.

This being said, the impact of a monetary shock on an existing loan depends on its age. For instance,

when the reset occurs after 15 years (instead of immediately), the increase in the disposable income is

reduced from €88,000 to less than €20,000 (Chart 5). Given the non-linearity of interest and

amortization payments, income effects are more than proportionally larger the sooner the reset occurs.

Chart 6: Volume of new loans vs interest payments on the stock of existing loans, credit for house purchase, € billion

Germany

France

300

New lo ans

250

New lo ans

Interest payments (on th e stock of loa ns)

Inte rest payments (on th e stock of loa ns)

250

Mortgage payment (2.5 % rate )

200

Mortgage payment (2.5% rate)

Mortgage payment (6% rate)

Mortgage payment (6% rate)

200

150

150

100

100

50

50

90

80

70

60

50

40

30

20

10

0

2004 2006 200 8 201 0 201 2 201 4 2016

Italy

New lo ans

Inte rest payments (on th e stock of loa ns)

Mortgage payment (2.5% rate)

Mortgage payment (6% rate)

0

2004 200 6 200 8 201 0 201 2 201 4 201 6

900

800

700

600

500

400

300

200

100

Euro area

New loans

Inte rest payments (on th e stock of loans)

Mortgage payment (2.5% rate)

Mortgage payment (6% rate)

0

2004 2006 200 8 201 0 201 2 201 4 2016

Spain

New loans

Interest payments (on the stock of loans)

Mortgage payment (2.5% rate )

Mortgage payment (6% rate)

Interest payments over the volume of new loans, percentage

40%

DE

ES

35%

EA

FR

30%

IT

25%

Besides interest payments and loan origination (new loans), the chart includes two proxies for mortgage payments for a standard loan with 25-

year maturity and €150,000 of initial capital for lending rates of 2.5 and 6 per cent. Given the evolution of actual interest rates, the true mortgage

payments are estimated to be closer to the continuous line towards the beginning of the period and to the dotted line towards the end of the

period.

Notes: New loans correspond to the annual moving sum of monthly data. The interest payments on existing loans are calculated by multiplying

the stock of loans by lending rates.

Source: ECB and own calculations.

Interest payment vs volume of new loans

0

200 4 200 6 200 8 2010 2012 2014 201 6

0

200 4 200 6 200 8 2010 2012 2014 201 6

Comparing the cost of debt service against loan origination seems a better benchmark to assess the

importance of monetary transmission through the stock of loans because it is where the monetary shock

will impact. Given data limitations, we use interest payments as a proxy for the cost of debt service. On top

200

180

160

140

120

100

80

60

40

20

20%

15%

10%

5%

0%

200 4 2006 200 8 201 0 2012 201 4 2016

8


of that, we have made an estimation of a corridor for the total mortgage payments based on a standard

mortgage as the one presented above (Chart 6).

By construction, the interest payments depend on both the stock of loans (Chart 3) and the level of

lending rates (Chart 9). Obviously, the cost of interest payments is much smaller than the stock of

loans 10 . However, the interest payments on existing loans still represent about 20 per cent of the

volume of new loans, or up to 35 per cent for a certain period in the case of Spain. While much

smaller than what we discussed in Section 2.1, this is still not negligible. Moreover, our estimate of

the corridor for total mortgage payments indicates that the latter are similar in size to loan

originations. In a similar fashion to what was observed in Section 2.1, the dynamics are very

heterogeneous across countries.

Chart 7: Net flows of loans vs interest payments on the stock of existing loans, credit for house purchase, € billion

Germany

France

60

Net flows of loans (new loans minus amortisations) 80

Net flows of lo ans (new loans minus amortisations)

50

Interest payments (on the stock of loans)

70

Inte rest payments (on th e stock of loans)

40

30

20

10

0

2004 200 6 200 8 2010 2012 201 4 201 6

-10

50

40

30

20

10

-10

Italy

Net flows of lo ans (n ew loans minus amortisations)

Inte rest payments (on th e stock of loa ns)

0

2004 200 6 200 8 2010 2012 201 4 201 6

60

50

40

30

20

10

0

200 4 2006 200 8 201 0 2012 2014 201 6

100

80

60

40

20

-40

Spain

Net flows of loans (new loans minus amortisations)

Interest payments (on the stock of loans)

0

200 4 2006 200 8 201 0 201 2 2014 201 6

-20

350

300

250

Euro area

Net flows of lo ans (n ew loa ns minus amo rtisations)

Inte rest payments (on th e stock of loa ns)

140%

120%

100%

Interest payments over (net flows of loans + interest

payments), percentage

DE

ES

EA

FR

IT

200

80%

150

60%

100

40%

50

20%

0

200 4 200 6 200 8 2010 201 2 201 4 201 6

Notes: Net flows of loans correspond to new loans minus amortisations (annual moving sum of monthly data). Interest payments on the stock of

existing loans are calculated by multiplying the stock of loans by lending rates. Note that net flows can be positive or negative. As a consequence,

the ratio in the bottom right panel can go beyond 100%. Note that negative net flows of loans in Spain become larger than interest payments on

the stock of existing loans.

Source: ECB and own calculations.

Net flows of loans

0%

200 4 2006 2008 2010 201 2 201 4 201 6

Given the lack of direct information about amortisations (to be able to calculate total mortgage

payments), we have also compared the net flows of loans (calculate as gross new loans minus

redemptions) against interest payments, as a robustness check. Net flows of loans appear to be of a

10 Note the use of different scales in Charts 3 and 6.

9


size similar to interest payments (Chart 7), which supports the hypothesis that monetary transmission

through existing loans can be as large as through new loans. In a way, net flows can be considered as

a better indicator of the potential impact of monetary policy decisions on the supply of money than

gross loan originations. As in previous cases, data show a wide diversity in the dynamics across

countries 11 .

This Section 2 documents that the transmission of a monetary policy shock can be as important through

the stock of loans as through the origination of new loans. Moreover, within the euro area, one can

observe a significant heterogeneity across countries.

3. LOAN FEATURES AND MONETARY POLICY

The data presented in Section 2 show very diverse dynamics for the stock of loans, loan originations

and interest payments across countries. This can be explained by the wide variety of loan features and

a different prevalence of specific loan types across countries.

Loan contracts have many features which can make loans not only complex but also very different

from each other. Indeed, a mortgage contract includes features such as the initial commissions and

fees (for opening the file, for studying the borrower’s profile, for assessing the value of the collateral,

etc.), the total amount to be lent, the maturity, the amortisation scheme, the benchmark used to

calculate the lending rate, how often the lending rate will be reset, potential floors or ceilings in the

rate, potential grace periods, penalties for late payments, penalties for early repayments, etc.

Moreover, additional factors, such as the requirement to use notarial deeds or the tax system of the

country can influence the effective cost of a mortgage for a borrower. Most of these features tend to

be homogeneous within a country but can vary widely across countries 12 .

For the purpose of monetary policy, one the most relevant features of mortgage contracts is the

fixation period of lending rates, that is, how often the rate is reset. In the so called fix-rate mortgages

(FRMs), the borrower pays the same rate – and therefore, the same nominal instalment – throughout

the whole life of the mortgage. On the other hand, in adjustable-rate mortgages (ARMs), the rate is

reset regularly (at least once a year) against the evolution of a benchmark – usually the Euribor or a

similar market rate. Intermediate cases (i.e. “mix-rate mortgages” or MRMs) with rate resets after 5 or

10 years are also possible.

Table 1: Prevailing type of interest rate for newly issued mortgages

ECB (2003) EC (2009) ECB (2009) ESRB (2015) EMF (2015)

Germany FRMs & MFRMs FRMs (98%) FRMs (85%) MFRMs (84%) FRMs & MFRMs

France FRMs & MFRMs (86%) FRMs (88%) FRMs (85%) FRMs (93%) FRMs (98%)

Italy ARMs (72%) ARMs (56%) FRMs (53%) ARMs (80%) ARMs

Spain AMRs (>75%) ARMs (98%) ARMs (91%) ARMs (70%) ARMs (70%)

Notes: ARMs: adjustable-rate mortgages, with an initial period of rate fixation of up to 1 year; MFRMs: medium-term fixed-rate mortgages, with an

initial period of rate fixation over 5 years and less than 10 years; FRMs: long-term fixed-rate mortgages, with an initial period of rate fixation over

10 years.

Source: ECB (2013): Structural factors in the EU housing markets (data for 2001); EC (2009): Retail Banking Survey (non-published, data for

2008-2009); ECB (2009): Drudi et al., Housing finance in the Euro area (data for 2007); ESRB (2009): Hall et al., Report on residential real estate

and financial stability in the EU (data for 2013); EMF (2015): Hypostat 2015. A review of Europe’s mortgage and housing markets (data for 2015).

Data indicate that countries tend to have a clear preference for either FRMs (e.g. Germany and

France) or ARMs (e.g. Spain). Italy, with a mix of both FRMs and ARMs, seems to be a special case

(Table 1). This has been confirmed by Nouy (2015) and studied from a theoretical point of view by

Campbell and Cocco (2003). It is unclear what drives the preference for one or the other type of

mortgage; inertia and cultural factors seems to play an important role. With the outbreak of the crisis,

interest rates hit the zero lower bound and, therefore, they can only increase in the future. A

11 Note that net annual flows can be negative, so that the exact value of the ratio presented in Chart 7, bottom-right panel, should be

interpreted with caution.

12 Hall et al. (2015) provide a recent survey of the different features of residential real estate markets across EU countries. See also Drudi et

al. (2009) and ECB (2003).

10


significant decrease in the share of ARMs in Spain and Italy is only observed since late 2015 or early

2016 (Chart 8). The environment of very low interest rates has significantly eroded the interest rate

margin (cf. Chart 9) and the profitability of banks, particularly where ARMs are prevalent. Therefore,

banks in countries like Spain and Italy are promoting FRMs as a way to improve their margins.

Chart 8: Share of adjustable-rate mortgages on total loan originations, credit for house purchase, percentage, 3-month

moving average

100

ES

90

IT

80

EA

DE

70

FR

60

50

40

30

20

10

0

200 3 200 5 200 7 2009 2011 2013 2015 2017

Source: ECB and own calculations.

The fixation period determines whether or not a monetary shock will have an impact on existing loans

through rate resets. Indeed, a change in the monetary policy stance after a loan has been agreed will

affect ARMs but not FRMs 13 . Monetary shocks can still affect mix-rate mortgages but with a long lag

– and therefore with a smaller impact, see Section 2.2.

Large declines in the policy rate could also impact FRMs as borrowers would have an incentive to

refinance their loans, but these effects seem limited. Andersen et al. (2015) find that many borrowers fail

to refinance optimally in Denmark, where barriers for refinancing are minimal. According to the

authors, almost 90 per cent of household-quarters with positive refinancing incentives do not refinance

their mortgage. Moreover, the relationship is asymmetric as the refinancing of a mortgage can only be

triggered by the borrower and not by the bank. Therefore, monetary authorities could count on the

potential effects of refinancing when reducing the policy rate, but not when increasing it.

Data seem to corroborate the fact that borrowers with FRMs do not, in general, refinance their loans

even when it could be economically rational to do so (Chart 9). Indeed, in Spain and Italy (middle

panels), the average interest rate on the stock of mortgages tracks the evolution of the policy rate with

only a short lag. This stems from the double effect of lower rates applied to new loans and the

automatic refinancing of old (ARMs) mortgages. However, in Germany and France (top panels),

despite monetary policy shocks clearly impacting new loans, interest rates on the stock of loans only

evolve smoothly. This indicates a very low, if at all, rate of refinancing of loans.

The differentiated impact of monetary policy according to the prevalence of ARMs and FRMs is even

more patent when all four countries are plotted on the same charts (Chart 9, bottom panels). It can be

clearly observed that lending rates for new loans are very similarly impacted by monetary policy in all

four countries. However, the impact on lending rates for the stock of loans is manifestly different

across countries. The prevailing fixation period seem to be playing a very significant role to explain

these dynamics.

The various constraints and rigidities such as refinancing penalties, deed fees or informational costs

contribute to explain the low refinancing shares of FRMs, to a certain extent. However, Andersen et

al. (2015) argue that a majority of household do not refinance even when there are positive incentives

to do so. In any case, one should not overestimate the impact of the interest rate on the debt burden on

borrowers, particularly in the current environment of very low interest rates. Indeed, a substantial

chunk of the monthly instalments are allocated to amortisation of the principal and only a marginal

part corresponds to interest payments (cf. Chart 4, right-hand panel). Moreover, the older the

13 The choice between FRMs and ARMs affects also the indebtedness and the debt burden on households as FRMs tend to be more

expensive than ARMs (see Greenspan, 2004).

11


mortgage, the less interesting the refinancing becomes even for significant declines in lending rates

(cf. Chart 5).

Given that the impact of monetary policy on existing loans with long fixation period can be very limited,

one could investigate a potential alternative channel for monetary policy to influence old borrowers:

inflation. This is discussed in the next section.

7

6

5

4

3

2

1

Chart 9: Lending rates for the stock of loans and for new loans, credit for house purchase, percentage

Germany (FRMs)

France (FRMs)

Rates on the stock of lo ans

Rates on new loans

ECB rate

7

6

5

4

3

2

1

Rates on the stock of lo ans

Rates on new loans

ECB rate

0

2000 200 2 200 4 2006 200 8 201 0 2012 2014 201 6

7

6

5

4

3

Italy (mix of FRMs and ARMs)

2

Rates on the stock of lo ans

1

Rates on new loans

ECB rate

0

2000 200 2 200 4 2006 200 8 201 0 2012 2014 201 6

7

6

5

4

3

2

1

Lending rates on new loans, comparative

DE (FRMs)

FR (FRMs)

IT (Mix)

ES (ARMs)

ECB rate

0

2000 200 2 200 4 2006 200 8 201 0 2012 2014 201 6

Source: ECB and own calculations.

0

200 0 2002 2004 200 6 200 8 2010 201 2 201 4 2016

7

6

5

4

3

Spain (ARMs)

2

Rates on the stock of loans

1

Rates on new loans

ECB rate

0

200 0 2002 2004 200 6 200 8 2010 201 2 201 4 2016

7

6

5

4

3

2

1

Lending rates on the stock of loans, comparative

DE (FRMs)

FR (FRMs)

IT (Mix)

ES (ARMs)

ECB rate

0

200 0 2002 2004 200 6 200 8 2010 201 2 201 4 2016

4. BEYOND INTEREST RATE: INFLATION AS A POLICY TOOL

Inflation and debt burden

Loans are denominated and repaid in nominal terms. Therefore, inflation erodes the value, in real terms,

of the debt – and of mortgage payments – and releases disposable income (Garriga, Kydland and Sustek,

2015; Adam and Zhu, 2016; Hilscher, Raviv and Reis, 2014).

Data show a wide range in inflation rates across countries (Chart 10, left-hand panel). For instance, in

2002 inflation hovered about 4 per cent in Spain while only about 1 per cent in Germany. Although

high volatility is observed in inflation over time, countries tend to stand on the upper side of the range

(e.g. Spain and Italy) or on the lower side of the range (e.g. Germany and France).

12


5%

4%

3%

2%

1%

-2%

Source: ECB and own calculations.

Annual inflation

0%

1999 2001 2003 2005 200 7 2009 2011 2013 2015

-1%

Chart 10: Inflation rate, percentage

Cumulative inflation (6-month moving average)

ECB rate 50%

DE 45%

ES

FR

IT

IT 40%

FR

ES 35%

DE

0%

1999 2001 2003 200 5 2007 200 9 2011 2013 2015

In the 1970s and the 1980s, annual inflation of 10 or even 20 per cent was not exceptional (cf. Chart

A2 in the Annex). But subsequently, inflation significantly declined and, since the late 1990s,

inflation has virtually always remained below 5 per cent. The lower the inflation rate, the longer the

time required for eroding the value of a financial asset in real terms. In any case, given the long

maturity of mortgage – up to 20 or even 30 years – the effect of inflation over time can still be

significant. Moreover, a clear divergence in the cumulative inflation across countries is clearly

observed. For instance, between 1999 and 2013, cumulative inflation was just above 25 per cent in

Germany and 50 per cent in Spain (Chart 9, right-hand panel). This leads to heterogeneous effects of

inflation across countries.

From the point of view of a borrower, inflation deflates the burden of mortgage payments in real

terms (Chart 11, left-hand panel) 14 . For instance, for our standard mortgage issued in Germany, the

monthly mortgage payment in real terms would have declined from €960 in 1999 to €660 by early

2016. On the other hand, if the same mortgage had been issued in Spain, the real monthly mortgage

payment would have declined from €960 to €490 in the same period.

30%

25%

20%

15%

10%

5%

1,000

800

600

400

200

Chart 11: Inflation effect on debt burden

Monthly mortgage payment, €

Mortgage payment as a percentage of income

DE

FR

IT

ES

20

15

10

5

ES

FR

IT

DE

0

199 9 200 3 200 7 201 1 201 5 201 9 202 3

0

199 9 2003 200 7 2011 201 5 201 9 2023

Notes: The illustration in left-hand panel is based a standard mortgage (annuity mortgage, 25 years, €150,000 of initial capital, 6 per cent of

interest) and the actual evolution of inflation through 1999-2016 for each country. The illustration in the right-hand panel is based on the median

debt service to income ratio according to the Eurosystem household finance and consumption survey and the evolution of inflation through 1999-

2016 for each country. Note that debt service to income ratio refers to all loans, not only mortgages (e.g. auto loans and other consumer loans

are also included). Also note that the survey data debt service to income ratio correspond to the year 2010; we apply these values to 1999

assuming that they have remained stable over time. In both panels, the projections of 2016-2023 are based on the average inflation in 1999-2016

for each country.

Source: ECB and own calculations.

However, households’ finances, in terms of income and mortgage size, vary enormously across

countries (cf. Section 5.2) and, therefore, also the actual debt burden and the monthly instalment. As

an illustration, we have applied the actual inflation erosion since 1999 to the average debt burden of

14 Inflation is used as a proxy for the evolution of household income by assuming that wages are indexed to inflation. In fact, the real value

of the instalments should be compared against the evolution of salaries or income rather than against consumer prices. However, given

potential career developments and promotions, even a wage index will fail to capture the real evolution of a household’s income over a

period of 25 to 30 years. Indeed, one can start as an intern in a company to become contractual staff and eventually reach a middle or senior

management position. Throughout such a career, salaries increase well beyond the standard indexation of salaries. This being said, not

everyone arrives to retirement age as a CEO of a large company with generous economic compensations; in fact, a majority of employees

have a much flatter career (cf. Section 5.1).

13


each country 15 (Chart 11, right-hand panel). In this case, one can observe that the debt burden in Spain

would have decreased from 20 to 10 per cent of borrowers’ income between 1999 and early 2016; in

the case of Germany, debt burden would have decreased from 11 to 8 per cent of borrowers’ income

during the same period.

The role of the central bank in steering inflation

These simulations indicate that, even in an environment of relatively low inflation rate, income effects

can become substantial over time. In this context, one could ask about how central banks can steer

inflation. While, in the past, many central banks used to have money growth targets, since the 1980s,

most of them adopted an inflation rate target (Williamson, 2016). For instance, the EU treaty mandates

the ESCB to maintain price stability in the EU and the euro area 16 .

The ECB’s Governing Council has translated the price stability mandate into the following quantitative

definition: "Price stability is defined as a year-on-year increase in the Harmonised Index of Consumer

Prices (HICP) for the euro area of below, but close to, 2 per cent". The average inflation for the euro

area has indeed hovered about 2 per cent up to 2007 (Chart 2). However, since the outbreak of the crisis,

a larger range and a significant volatility is observed; for instance, inflation dropped from 4 per cent to

almost minus 1 per cent in about one year. Moreover, actual inflation has remained well below the 2 per

cent target for a long period and has been significantly different across countries.

In this context, there seem to be two contradictory views about how monetary policy decisions translate

into inflation. A traditional view based on the Taylor rule advocates that declines in the policy rate

would ease credit conditions and ultimately translate into higher inflation. Other authors argue that a

Fisher effect moves inflation in the same direction as the central bank policy rate. While the

conventional central bank wisdom had been following the first approach, there seem to be increasing

voices supporting the second one (see Williamson, 2016 and Cochrane, 2016 for further details). Getting

more clarity on how the policy interest rates translate into inflation will help monetary authorities in

understanding how they can potentially use the income effects as a policy tool.

5. BORROWERS FEATURES AND MONETARY POLICY

Sections 3 and 4 show how income effects impact current borrowers through their stock of debt.

These effects work either through the reset of interest rates in ARMs or through inflation and the

subsequent erosion in real value in both FRMs and ARMs. The actual macroeconomic impact of these

effects depends on a series of channels, including the response of borrowers to them: if borrowers are

wealthy and unconstrained, the macroeconomic impact would be limited. On the other hand, income

shocks on borrowers without savings or with illiquid assets will be translated into consumption

behaviour to a large extent (Kaplan and Violante, 2014). In other words, the microeconomic structure

of borrowers determines the potential macroeconomic impact of the above income effects. Therefore,

in this section, we review borrowers’ features such as indebtedness, income and homeownership.

5.1. Who are the borrowers?

Homeownership and mortgage debt

According to the 2013 Household Finance and Consumption Survey, 60 per cent of euro area

households owned their main residence while 40 per cent of households were renters 17 (Chart 12, top-

15 Note that the survey data for debt service to income ratio correspond to the year 2010. We apply the data of the ECB household finance

and consumption survey to the data on debt service 1999, by assuming that they have remained stable over time. This assumption does not

fundamentally affect the takeaways of the illustration.

16 Article 2 of the statute of the ESCB and the ECB, enshrined in the Protocol No. 4 of the Treaty on the Functioning of the European Union.

17 This category includes also other situations of non-owners (e.g. a household hosted by family without paying a rent).

14


left panel). A significant variation across countries can be observed: ownership rate is twice as large

in Spain (80 per cent) as in Germany (44 per cent).

Homeowners may own the residence outright or through a mortgage. About 23 per cent of euro area

households had a mortgage in 2010/2011 (either on their main residence or on other property), with

the highest proportion being in Spain (32.5 per cent) and the lowest in Italy (11 per cent) (Chart 12,

bottom-left panel). The average size of the mortgage loan was almost €70,000 for the euro area as a

whole, €80,000 in Germany and about €55,000 in the other three countries (Spain, France and Italy)

(Chart 12, bottom-right panel).

Chart 12: Households assets (residential ownership) and liabilities (debt and mortgages), 2013

Homeowners, percentage of households

Households with debt, percentage

80

80

70

70

60

60

50

50

40

40

30

30

20

20

10

10

0

80

EA DE ES FR IT

Households with a mortgage, percentage

0

80

EA DE ES FR IT

Average balance of the mortgage, € thousand

70

70

60

60

50

50

40

40

30

30

20

20

10

10

0

EA DE ES FR IT

Source: ECB (Eurosystem household finance and consumption survey).

0

EA DE ES FR IT

Household debt: from a micro to a macro perspective

According to monetary statistics and national accounts, the outstanding volumes of mortgages

represented 38 per cent of euro are GDP on average in 2015. In the four largest countries, it ranged

from 23 per cent (Italy) to 51 per cent (Spain). Therefore, monetary policy decisions through existing

loans would potentially have a larger impact in Spain than in Italy (Chart 13, left-hand panel).

50%

40%

Chart 13: Household (aggregate) indebtedness, outstanding volumes of credit for house purchase over GPD, percentage

December 2015

Evolution

70%

60%

50%

30%

40%

20%

10%

0%

ES FR EA DE IT

30%

Notes: The gaps in the series for France and Italy indicate a structural break.

Source: ECB and own calculations.

20%

DE

ES

10%

EA

FR

IT

0%

2004 2006 2008 2010 2012 2014 2016

Household indebtedness shows heterogeneous dynamics across countries (Chart 13, right-hand panel).

In France, household indebtedness has increased up to mid-2014 and then stagnated. In Italy, the

15


stagnation is observed already since late 2010. Data for Spain show a very fast expansion in

household indebtedness up to 2007; thereafter, the indebtedness has not only stagnated but even

declined, particularly after 2013. In Germany, the trend is clearly the opposite to the rest of the

countries: a secular decline in household indebtedness since the early 2000s. These dynamics are the

result of a combination of the evolution of interest rates, amortisations and housing prices.

Household income

Household income is relevant for two reasons. First, lenders decide whether or not to grant a loan

depending on the credit worthiness of the potential borrower, which in turn depends on her income

level. Second, household income ultimately defines the burden of mortgage payments on household

finance and influences some features of the loan such as the duration.

According to the ECB data, average annual household income was significantly higher in Germany

(€43,500) than in Spain (€31,000), while in France and Italy it was slightly below the euro area

average (€38,000). In terms of distribution, the highest income quintile had an average annual income

of almost 13 times the average income of the lowest quintile (almost €120,000 and less than €10,000

of annual income, respectively). The picture is similar in terms of medians and across countries;

however, income is slightly more unequally distributed in Germany (in terms of both medians and

means) and slightly less unequally distributed in Spain (in terms of medians) and in France (in terms

of means) (Chart 14).

45

40

35

30

25

20

15

10

5

0

35

30

25

20

15

10

5

0

Chart 14: Household annual income, mean and median, € thousand, 2013

Mean income

Total mean

Distribution (mean of each income segment)

140

Bottom 20% 20-4 0%

40-6 0% 60-8 0%

120

80-9 0% 90-1 00%

EA DE ES FR IT

Total median

EA DE ES FR IT

Source: ECB (Eurosystem household finance and consumption survey).

100

80

60

40

20

0

Median income

80

70

60

50

40

30

20

10

0

EA DE ES FR IT

Distribution (median of each income segment)

Bottom 20% 20-40%

40-60% 60-80%

80-90% 90-100%

All DE ES FR IT

5.2. The effort for the borrower

Data show striking differences in the level of indebtedness (measured through the debt to income

ratio) across countries and income strata. Firstly, the median indebtedness is considerably higher in

Spain than in Germany (113.5 per cent and 37 per cent, respectively). Secondly, the distribution of

indebtedness throughout income strata in Spain appears to be the opposite of the distribution in

countries like Germany or France. In Spain, the households in the lowest income quintiles are the

ones with the highest levels of indebtedness (with a debt to income ratio above 140 per cent), while

16


icher households are much less indebted (with a debt to income ratio below 60 per cent). In Germany

and France, richer households are more indebted than poorer households. Italy appears somehow in an

intermediate position (Chart 15).

160

140

120

100

80

60

40

20

Chart 15: Indebtedness: Debt to income ratio, indebted households, percentage, 2013

Median

Distribution (median of each segment)

160

Bottom 20% 20-4 0%

140

40-6 0% 60-8 0%

80-9 0% 90-1 00%

120

100

80

60

40

20

0

EA DE ES FR IT

Source: ECB (Eurosystem household finance and consumption survey).

0

EA DE ES FR IT

The debt service burden derives from the indebtedness level. Data show that debt service burden is twice

as high in Spain (where indebted households allocate almost 20 per cent of their income to mortgage

payments) as in Germany (where they only allocate slightly above 10 per cent of their income to

mortgage payments); France and Italy appear in an intermediate position. In terms of distribution, in the

euro area as whole as well as in Spain and Italy, the debt service burden is higher the lower the income

of the household. However, in Germany and France, the debt service burden is very similar across

income segments (Chart 16).

40

35

30

25

20

15

10

5

Chart 16: Debt service burden: debt service to income ratio, indebted households, percentage, 2013

Median

Distribution (median of each segment)

40

Bottom 20 % 20-4 0%

40-60% 60-8 0%

35

80-90% 90-1 00%

30

25

20

15

10

5

0

EA DE ES FR IT

Source: ECB (Eurosystem household finance and consumption survey).

0

EA DE ES FR IT

One could argue that income effects would be neutral in aggregate terms because any increase in the

disposable income of a borrower will be matched with an equivalent decrease in the disposable

income of a lender. However, households have different propensities to consume depending on their

income, and indebtedness. Poorer and more indebted households tend to have a higher propensity to

consume (Carroll et al., 2015; Keys et al. 2014; Di Maggio, Kermani, and Ramcharan, 2015).

Moreover, income shocks may have different effects depending on the liquidity constrains of

household wealth (Kaplan and Violante, 2014). Given these differences in the propensity to consume

and in household constrains, a redistribution of disposable income between borrowers and lender is

not neutral with respect to output. The heterogeneous distribution of indebtedness and debt service

burden across countries implies that monetary policy shocks will necessarily have uneven effects in

the different countries.

To assess the significance of the debt burden from a macroeconomic perspective, we have constructed

an indicator of “implicit maturity” as the ratio between the stock of loans and annual amortisation.

This ratio indicates in how many years the stock of debt would be repaid if amortisation payments

17


were to continue at current levels 18 . Once again, data show heterogeneous dynamics across countries

(Chart 17).

Implicit maturity of mortgages in Germany seems rather stable hovering at about 6 years. In the other

three countries, a much more volatile evolution is observed. In Spain, it increased from 7 to 12 years

between 2004 and 2012. The need to increase the duration of the mortgages can be explained, to a

certain extent, by increasing real estate prices, which almost doubled between 2003 and 2008 (cf.

Chart A1 in the Annex). Increasing unemployment rates and the difficulties for many borrowers to

confront the mortgages payments may also be playing an important role since the outbreak of the

crisis 19 . Logically, these phenomena affected particularly the population at the lowest end of the

income distribution, who were already confronted with the highest relative debt burden (cf. Chart 16).

A significant raise in the implicit maturity is also observed in France between 2007 and 2010. The

drop in the maturity observed across the board since early 2015 reflects, to a large extent, the

statistical effect 20 of the increasing refinancing of loans to take advantage of the very low rates.

1,200

1,000

400

350

300

250

200

150

100

800

600

400

200

50

Chart 17: Stock of loans, annual amortisation (€ billion) and implicit maturity (years), credit for house purchase

Germany

France

Stock of lo ans

Ann ual amortisation

0

2004 200 6 2008 201 0 201 2 2014 201 6

Italy

0

200 4 200 6 200 8 2010 201 2 201 4 201 6

4,500

4,000

3,500

3,000

2,500

2,000

1,500

1,000

500

Euro area

Stock of lo ans

Ann ual amortisation

Stock of lo ans

Ann ual amortisation

0

2004 200 6 2008 201 0 201 2 2014 201 6

0

200 4 2006 200 8 201 0 2012 201 4 2016

Spain

Implicit maturity, years

Notes: Annual amortisation corresponds to the moving sum of monthly data. The white gaps indicate structural breaks in the series. The implicit

maturity is calculated as annual amortisation over the stock of loans.

Source: ECB and own calculations.

1,000

700

600

500

400

300

200

100

18

16

14

12

10

900

800

700

600

500

400

300

200

100

Stock of lo ans

Ann ual amortisation

Stock of lo ans

Ann ual amortisation

0

200 4 2006 200 8 201 0 201 2 2014 201 6

8

6

4

2

0

200 4 200 6 2008 2010 2012 2014 201 6

FR

DE

EA

IT

ES

18 Note that this implicit maturity refers to the remaining maturity of the total stock of loans in banks’ balance sheet. In other word, it

includes not only recently agreed loans but also loans agreed years ago. This explains why the implicit maturity is much shorter than the

typical maturity of 20-30 years for new loan. Moreover, early repayments and renegotiation of loans also reduce the effective maturity.

19 Indeed, non-performing loans increased significantly during the crisis (see, for instance, European Commission, 2015, p. 19). Some banks

responded to this situation through renegotiations of the loans by prolonging the maturities or by providing a grace period. In many cases, this

forbearance approach was an attempt for banks to avoid provisioning and to conceal potential losses (see Homar, Kick and Salleo, 2015).

20 The refinancing appear as amortizations even if the borrower simultaneously withdraws the same amount.

18


6. CONCLUSIONS

In this paper, we make a survey of a number of factors influencing how monetary policy decisions can

be transmitted through the stock of existing loans and how these effects can be very different across

euro area countries despite sharing the same monetary policy.

First of all, we document that the stock of loans can be as relevant for monetary policy purposes as the

flow of new loans. Annual mortgage payments have a size similar to annual loan origination. This is

because the ratio between the stock of loans and annual loan origination is 5 to 1 in the case of

Germany and as large as 20 to 1 in the case of Spain.

Secondly, we discuss the mechanism through which monetary policy is transmitted through the stock of

loans. This consist of income effects originated either on the reset of lending rates or through inflation.

The first process impacts only ARMs. Consequently, it is inexistent in countries like Germany or

France, where 90 per cent of mortgages are FRMs, while it is very significant in Spain, where 90 per

cent of mortgages are ARMs.

Income effects resulting from inflation impact both ARMs and FRMs. Even under a low inflation

environment, the erosion of value through inflation appears to be very relevant to eventually impact

aggregate output. For instance, between 1999 and 2013, the inflation erosion meant that indebted

households increased their real disposable income by 50 per cent of their mortgage payments in Spain.

The potential macroeconomic impact of monetary shocks through the lending channel varies widely

across countries within the monetary union. For instance, mortgage loans represent as little as 23 per

cent of GDP in Italy and as much as 51 per cent in Spain. On a similar fashion, low income households

devote up to 40 per cent of their income to service their debts in Spain but only 12 per cent in

Germany.

This paper documents a wide diversity in borrowers’ features, loan features and price developments

across euro area Member States. Better understanding the individual and combined implications of

these features is of the outmost importance for being able to conduct an effective monetary policy in

the euro area. Significant differences in these factors are observed across countries, what seems to be

distorting a homogeneous transmission of the single momentary policy.

19


ANNEX: ADDITIONAL CHARTS

200

180

160

140

FR

DE

IT

ES

Chart A1: Nominal house price index, 2003 Q1 = 100

120

100

80

2003 200 5 200 7 2009 201 1 201 3 201 5

Source: OECD and own calculations.

30

Chart A2: Inflation rate, annual percentage change

25

20

15

Germany

France

Italy

Spain

10

5

0

1960 1970 1980 1990 2000 2010

Source: INE.

REFERENCES

Adam, K. and Zhu, J., 2016, “Price Level Changes and the Redistribution of Nominal Wealth”, Journal of the European

Economic Association, Vol. 14, Iss. 4, pp. 871–906.

Andersen, S., J. Y. Campbell, K. M. Nielsen and T. Ramadorai, 2015, Inattention and Inertia in Household Finance:

Evidence from the Danish Mortgage Market.

Badarinza, C., J. Y. Campbell and R. Ramadorai, 2015, “What calls to ARMs? International evidence on interest rates and

the choice of adjustable-rate mortgages”.

Bernanke, B. S. and M. Gertler, 1995, “Inside the black box: the credit channel of monetary policy transmission,” Journal of

Economic Perspectives, Vol. 9(4), pp. 27-48.

Brunner, K., and A. H. Meltzer, 1988, “Money and credit in the monetary transmission process,” The American Economic

Review, Vol. 78(2), pp. 446-451.

Calza, A., Monacelli, T., Stracca, L., 2013. “Housing finance and monetary policy,” Journal of the European Economic

Association Vol. 11, Iss. s1, pp. 101–122.

Carroll, C., J. Slacalek, K. Tokuoka and M. N. White, 2015, “The Distribution of Wealth and the Marginal Propensity to

Consume”.

Campbell, J. Y. and Cocco, J. F., 2003, “Household Risk Management and Optimal Mortgage Choice”, The Quarterly

Journal of Economics (2003) Vol. 118, No. 4, pp. 1449-1494.

Cecchetti, S. G., 1995, “Distinguishing theories of the monetary transmission mechanism,” Federal Reserve Bank of St.

Louis Review, Vol. 77(3), pp. 83-97 (St. Louis: Federal Reserve Bank of St. Louis).

Cochrane, J. H., 2016, “Do Higher Interest Rates Raise or Lower Inflation?”, Working Paper.

Di Maggio, M., A. Kermani and R. Ramcharan, 2015, Monetary policy pass-through: Household consumption and voluntary

deleveraging.

Drudi, F. (Chair), P. Köhler-Ulbrich, M. Protopapa, J. Slacalek, C. K. Sorensen, G. Wolswijk (Coord.) and others, 2009,

Housing finance in the Euro area (Frankfurt: European Central Bank).

European Central Bank, 2003, Structural factors in the EU housing markets (Frankfurt).

European Central Bank, 2009a, “Monetary policy and loan supply in the Euro area,” Monthly Bulletin - October, pp. 63-80

(Frankfurt).

20


European Central Bank, 2011, “The supply of money – Bank behaviour and the implications for monetary analysis,”

Monthly Bulletin - October, pp. 63-79 (Frankfurt).

European Central Bank, 2013, Household Finance and Consumption Survey 2013 (Frankfurt).

European Central Bank, 2016, “ECB announces new series of targeted longer-term refinancing operations (TLTRO II),”

Press Release, 10 March (Frankfurt).

European Commission, 2015, European financial stability and integration report (Brussels).

European Mortgage Federation, 2015, Hypostat 2015. A review of Europe’s mortgage and housing markets (Brussels).

Frank, R. H. and P. J. Cook, 1996, The winner-take-all society: Why the few at the top get so much more than the rest of us

(Penguin).

Galbraith, J. K., 2012, Inequality and instability: A study of the world economy just before the Great Crisis (Oxford

University Press).

Garriga, C., F. E. Kydland and R. Sustek, 2015, “Mortgages and monetary policy,” Working Paper 2015-033A. (St. Louis:

Federal Reserve Bank of St. Louis).

Greenspan, A., 2004, Understanding household debt obligations, speech at the Credit Union National Association

Governmental Affairs Conference (Washington, DC).

Hall, S. (Chair), 2015, Report on residential real estate and financial stability in the EU (Frankfurt: European Systemic Risk

Board).

Henriksen, E., F. E. Kydland and R. Šustek, 2013, “Globally correlated nominal fluctuations,” Journal of Monetary

Economics, Volume 60, Issue 6, Pages 613–631.

Hilbers, P., A. W. Hoffmaister, A. Banerji and H. Shil, 2008, “House Price Developments in Europe: A Comparison,” IMF

Working Paper, WP/08/211 (Washington: IMF).

Hilscher, J., Raviv, A and Reis, R., 2014, “Inflating Away the Public Debt? An Empirical Assessment”, NBER Working

Paper No. 20339.

Homar, T., H. Kick and C. Salleo, 2015, “What drives forbearance – evidence from the ECB Comprehensive Assessment”,

Working Paper Series No. 1860 (Frankfurt: European Central Bank).

Hubbard, R. G., 1995, “Is there a ‘credit channel’ for monetary policy? Federal Reserve Bank of St. Louis Review, Vol.

77(3), pp. 67-74 (St. Louis: Federal Reserve Bank of St. Louis).

Iacoviello, M., 2005, “House prices, borrowing constraints, and monetary policy in the business cycle,” American Economic

Review Vol. 95, No. 3, pp. 739–64.

Ireland, P. N., 2005, The monetary transmission mechanism, Working Paper No. 06-1 (Boston: Federal Reserve Bank of

Boston).

Instituto National de Estadística, 2015, Notas de prensa - Estadística de Hipotecas (several months) (Madrid).

Kaplan, G. and Violante, G. L., 2014, “A Model of the Consumption Response to Fiscal Stimulus Payments”, Econometrica,

Vol. 82, Iss. 4, pp. 1199–1239.

Keys, B., T. Piskorski, A. Seru, and V. Yao, 2014, Mortgage rates, household balance sheets, and the real economy.

Lakner, C. and B. Milanovic, 2013, “Global income distribution: from the fall of the Berlin Wall to the great recession,”

Policy Research working paper No. WPS 6719 (World Bank).

McLeay, M., A. Radia, and R. Thomas, 2014, “Money creation in the modern economy”, Quarterly Bulletin 2014 Q1.

(London: Bank of England).

Mishkin, F. S., 1996, The channels of monetary transmission: lessons for monetary policy, NBER Working Paper No. 5464.

(Cambridge, MA, US: National Bureau of Economic Research).

Danièle Nouy, D., 2015, “The banking union and financial integration”, speech at the Joint conference of the European

Commission and European Central Bank on “European Financial Integration and Stability”, 27 April

OECD, 2011, Divided We Stand: Why Inequality Keeps Rising.

OECD, 2014, Does income inequality hurt economic growth? Focus on Inequality and Growth, December 2014.

Piketty, T., 2014, Capital in the Twenty-First Century (Cambridge, MA: Belknap Press).

Rubio, M., 2011, “Fixed- and variable-rate mortgages, business cycles, and monetary policy,” Journal of Money, Credit, and

Banking Vol. 43, Iss. 4, pp. 657–88.

Stiglitz, J., 2012, The price of inequality: How today's divided society endangers our future (Norton).

Treaty on the Functioning of the European Union (Consolidated version), 2012, Official Journal of the European Union, C 326,

26.10.2012.

Villar Burke, J., 2015, “The interventions of the European Central Bank throughout the crisis,” Working Paper Series No.

2555921 (Social Science Research Network – SSRN).

Williamson, S., 2016, “Neo-Fisherism. A Radical Idea, or the Most Obvious Solution to the Low-Inflation Problem?”, The

Regional Economist, Vol. 24, No. 3, The Federal Reserve Bank of St. Louis.

21

More magazines by this user
Similar magazines