Household Finance and Consumption - Opening remarks
Speech by José Manuel González-Páramo, Member of the Executive Board of the ECB, at the Conference organised by the ECB and the Banque Centrale du Luxembourg, Luxembourg, 25 October 2010
Ladies and Gentlemen,
It is a great pleasure to open this joint BCL/ECB conference in Luxembourg. This is the second conference organised in the context of the Eurosystem Household Finance and Consumption Network. This network is unique as it brings together experts from a wide range of fields and institutions in one single forum. Participants in today’s meeting include survey experts, statisticians and economists. Institutions represented are the ECB, the Eurosystem’s national central banks and a number of national statistical institutes and research institutes, as well as central banks from outside the Eurosystem.
Historical background of the Eurosystem Household Finance and Consumption Network
Developing, implementing and analysing a survey on the finances and consumption of households for the entire euro area does not happen overnight. The history of this network dates back to December 2006. A task force was then established to investigate the feasibility of a survey on household finance and consumption in the euro area.
On the basis of the task force’s report, the Governing Council decided to establish a network of Eurosystem experts to prepare a proposal for such a survey. For one and a half years the network worked diligently at developing such a proposal. Based on this proposal, in September 2008 the Governing Council approved the conduct of the survey and gave a three-fold mandate to the network: (1) to complete the implementation work, (2) to take care of the survey development and (3) to serve as a forum for research with the survey data.
After all this preparatory work, the fieldwork of actual data collection of the first wave of the Household Finance and Consumption Survey could finally start, with many countries in the euro area conducting their survey in the course of 2010 and 2011. At the current juncture, we are therefore looking forward to the analysis that will result of using the collected data.
Even though the data of the first wave of the euro area wide survey are not yet available today, there is already a large literature on household finance that analyses existing micro household data. The work that will be presented at this conference shows this very clearly, with contributions based on a wide array of datasets ranging from existing public household surveys to private datasets from credit card companies.
The Household Finance and Consumption Network is a good example of what is possible when the Eurosystem co-ordinates its efforts as a team. Central banks around the world have either collected household data or supported the collection of such data by other institutions for a long time. The strength of the network is that it has been able to tap into that valuable expertise from many sources.
Household surveys on wealth date back to 1962 when the US Board of Governors’ Survey of Consumer Finances started as the Survey of Financial Characteristics of Consumers. Not long thereafter, in 1965, the Bank of Italy started its Survey of Household Income and Wealth. Those two pioneers, as I may call them, were followed later by national statistical institutes and Eurosystem central banks.
In the eighties, INSEE started its Enquête Patrimoine and Statistics Finland its Survey of Household Wealth. In 1994 a survey of Portuguese households’ wealth and debt was conducted by the National Statistical Office in Portugal, encouraged by the Banco de Portugal. In 1999 the first Cyprus Survey of Consumer Finances was launched as a joint project between the University of Cyprus and the Central Bank of Cyprus. In 2002, the Bank of Spain started its Survey of Household Finances. In the same year Greece launched its survey of household debt and the Netherlands Bank began participating in what was formerly known as the CentER Savings Survey, a survey running since 1993, and currently known as the Household Survey of De Nederlandsche Bank.
Throughout the work of the network, experts from these existing surveys have given their valuable advice. Allow me to especially mention today’s keynote speaker, Arthur Kennickell from the Board of Governors of the Federal Reserve System. As the Chief of the Microeconomic Surveys Section he is “the” expert of the US Survey of Consumer Finances with a long and deep experience in this field. He has put a relentless effort in providing the most valuable advice to the network.
Through the years, the network has also drawn on the expertise of academics who have provided invaluable advice and guidance throughout. Some are present at the conference, such as Michalis Haliassos of Frankfurt University who has been involved since the beginning of the project, and Peter Tufano of Harvard Business School, who volunteered to be part of the scientific committee of this conference. Tomorrow, you will listen to an expert on financial literary: Annamaria Lusardi of Dartmouth College will give a keynote lecture on how financial literacy affects economic decisions of households.
What can we learn from a Househould Survey?
Once the data of the Eurosystem Household Finance and Consumption Survey are available, they will provide a wealth of comparable information on micro-level household decisions with regard to the holding of real and financial assets, the taking of debt, employment, income, pensions, intergenerational transfers, gifts, consumption and saving, and risk attitudes. Anonymised euro area micro data will be made available to the wider research community.
To understand better the reasons for a conference on household finance, it is worth elaborating on why central banks get involved in the development and production of micro-level data. Fundamentally, micro-level household data provide information that is not available from national accounts or flow of funds data. Central bank policymakers expand their knowledge if they complement information from macro aggregates with micro-insights. Micro-insights are a natural counterpart of the macro models used in monetary policy that are traditionally based on analyses of the aggregates such as GDP, consumption and investment. But how precisely can micro level data serve the purpose of central banking?
It seems to me that there are multiple answers to this question. I will stress here three.
First, the reaction functions of various categories of agents that are built into macro models have to be ultimately based on micro-level analysis.
Second, the micro-level data provide us with a wealth of distributional information, information that is not available from the macro aggregates and that is valuable to understand the reaction of different groups of households to shocks and the distribution of risks.
Third, micro-level household data can give insights into new trends.
Reaction functions of agents must be based on micro-data
Let me first express in more detail what we can learn about the reaction functions of households from micro data. Household survey data link information on income, assets, debt and consumption at the household level. To understand consumption and saving behaviour of households, it is crucial to combine these various elements in the analysis. Households respond in a quite heterogeneous way to shocks depending, among many other factors, on the form their wealth takes.
For instance, wealth can be very illiquid when it is concentrated in home ownership, thus affecting households’ possibilities for adapting their consumption behaviour to various shocks. Understanding how consumption is affected by changes in either stock market prices or house prices is relevant for policymakers. Especially during times of big shocks, such as a financial or economic crisis, policymakers are interested in understanding how the fluctuations in assets prices will affect aggregate demand. During these times, knowledge provided by the surveys comes at a premium.
In particular, the way household spending reacts to changes in interest rates is of continuing relevance for monetary policy, not only during crisis periods. Important elements of this reaction are the demand for housing and the different homeownership ratios across euro-area countries. Residential investment is an important channel through which interest rates changes are transmitted via households through the economy. As purchase of a house is often financed with a mortgage, understanding the access of households to mortgage credit increases our understanding of the transmission of interest rate changes to credit demand. Central banks regularly monitor the evolution of credit aggregates. Mortgages make up the bulk of household liabilities and are an important element of total credit.
Two interesting puzzles in this context are: (1) the origins of the enormous diversity of home-ownership levels, and (2) the large differences in mortgage types that we see across the countries of the euro area. In some countries, with Germany and France as examples, households tend to hold fixed rate mortgages, whereas in other countries, such as Spain and Portugal, households typically hold adjustable rate mortgages. Despite big advances in the integration of financial markets in the euro area, differences in the fixation of mortgage rates seem to persist.
Micro household data can shed light on the origins of these differences, namely whether these are caused by household characteristics, institutions, government policies or just different tastes. Understanding these differences is important because the way the assets and liabilities of households are structured has implications for their relative sensitivity to market interest rates and, ultimately, for the transmission of monetary policy. For instance, the transmission of interest rate shocks is likely to be stronger and more immediate in countries with adjustable-rate mortgages.
Importance of distributional information
Household survey data are not only useful to estimate more precisely the behaviour of households with respect to spending, saving and credit, it also provides unique information on the distribution of assets and liabilities. Distribution of assets and liabilities across households and possible mismatches between assets and liabilities of individual households affect the transmission of monetary policy to macro aggregates: different distributions of wealth will generally lead to different aggregate fluctuations in the macro aggregates in reaction to shocks.
This brings me to the second reason for the usefulness of micro data. The distribution of assets and liabilities has an effect on how risks are distributed across different household subpopulations. In that respect, policy makers cannot be content to just rely on aggregate numbers. Aggregate statistics on the indebtedness of households can be misleading if debt is concentrated among some household groups. For instance, it poses fewer risks if it is concentrated mainly among the relatively wealthy households.
In this regard, it is equally important to understand how households choose their portfolios: the risks stemming from the household sector will depend largely on how household portfolios are structured. Household portfolios are prone to all sorts of biases. The home bias in equity holdings is a well-known example. Well-functioning financial markets are a prerequisite for a well-functioning monetary transmission. Important topics for further analyses are: what influences households in their choices of assets? What role do financial intermediaries such as banks play? What determines the risk-taking behaviour of households and what are its implications for financial stability?
In the European Union we are currently in the process of setting up the European Systemic Risk Board (ESRB), a new body responsible for the macro-prudential oversight of the financial system. It is clear that macro-prudential risks may stem not only from the balance sheets of banks or other financial intermediaries, but also from the balance sheets of households. The aggregated data of the household sector are not always sufficient to understand the magnitude and the location of these risks: they need to be complemented by distributional information. I therefore foresee that the household survey on finance and consumption will also be an important source of information for the ESRB, in spite of its low frequency, especially if it can be enhanced to encompass the entire EU.
Monitoring of new trends
My third point is that survey data are also useful to monitor new important developments. The relationship between aging, pension provision and consumption is a field of research that will become more important as European populations grow older. As of this day, many countries faced with the budgetary consequences of an aging population have already embarked on policy changes in pension provision. Retirement ages are being increased and pension entitlements have changed. The micro survey data can shed light on the saving behaviour of the households when they are faced with these changes.
New developments will also come from financial innovation. As households are faced with new types of assets to invest in, or with new types of debt, these innovations will have an effect on consumption smoothing. New types of debt, for instance, may reduce the importance of borrowing constraints and could therefore have financial stability implications and change households’ reaction to shocks.
I am happy to see that all three elements I just discussed are present in this conference. For instance, in your first session today you will be discussing how stock market participation depends on financial advice. In your second session, you will discuss how consumption reacts to mortgage interest rates. Tomorrow, you will go deeper into issues of inheritance, household wealth and borrowing.
I hope I was able to convince you that what you will be discussing in these two days is not only of purely academic interest, but also of importance for us policymakers. I wish you a fruitful conference and I look forward to your presentations and discussions.