Third CEPR European Workshop on Household Finance

Michalis Haliassos, Vimal Balasubramaniam 01 June 2018

a

A

On 11-12 May 2018, Imperial College Business School hosted the Third CEPR European Workshop on Household Finance. It was organised by The CEPR Network on Household Finance, Imperial College Business School and London Business School, with the support of the Think Forward Initiative and the EDHEC PhD in Finance Programme.

The Network runs the European Workshop in spring of each year since 2016, and the European Conference on Household Finance in the autumn since 2015, continuing a series of conferences previously launched by founding members in 2010. Both events are designed to encourage discussion between senior and junior researchers. Even though we are now holding these events every six months, they still attract an enormous volume of submissions, about 175 per event. Each paper is reviewed independently by at least two members of the organizing committee, and the acceptance rate is of the order of 6%. While it is not possible to include all excellent papers, the chosen ones represent state-of-the-art research on household financial behaviour and on how this is influenced by other choices, government policies, and the overall economic environment.Each event includes at least three papers from PhD students, who compete for the Network’s Best Student Paper Prize, awarded by the Steering Committee and sponsored by the Think Forward Initiative.

Session one: Peer effects

Knüpfer, Rantapuska and Sarvimäki (2018) take the established result that portfolio choice correlates across generations and ask why this is the case. While previous studies have concentrated on the influence of genetics and environmental similarities, this paper uses Finnish data to investigate the influence of word-of-mouth and other communication between parents and their adult children. By investigating correlation between precise securities, rather than their classes, and controlling for fixed effects, their results suggest that these social effects explain almost all the correlation in portfolio choice.

Also discussing peer effects, Girshina, Mathä and Ziegelmeyeret (2017) use survey data to ask to what extent has the investment behaviour of immigrants in Luxembourg, where 43% of residents are immigrants, influenced natives to participate in the stock market. Again, the authors instrument and control for correlated effects to try to identify a peer effect. Their results suggest that there is a positive effect on natives, which they suggest is at least one half a social learning effect. 

Session 2: Unintended consequences

Two papers in session two suggest potentially undesirable effects of existing policies. 

Athreya, Ionescu, Neelakantan and Vidangos (2018) provocatively ask whether the college subsidy given to US students (which they calculate at around $85,000 per student) would increase wellbeing more if it was given as stocks. Across the whole population, the utility gain from access to both stocks and college is equal, but they argue that for 46% of people, access to college has almost no effect on well-being. By creating a model of human capital accumulation and portfolio choice, they conclude that 52% of high-school graduates would, all else equal, prefer a stock-index retirement fund to the subsidy currently flowing to college, in view of their limited ability to make use of the opportunity for college education.

Johnson (2018), in a paper that was awarded the Network’s first prize for a PhD student paper, suggests that rules designed to reduce risky mortgage lending in the US have had the unintended consequence of reducing entrepreneurship. Since 2014, a bank rule that restricts the debt-to-income ratio for borrowers has disproportionately rationed mortgage lending to the self-employed. By comparing areas with a large proportion of banks that were exempt or not exempt to the rule, she concludes that the policy has reduced self-employment by around 2%, and new small business employment by at least 3%.

Session 3: Effects of skewness and new asset classes

Shen (2018) investigates the skewness in consumption growth, using a model with stockholders and non-stockholders, and finds that that skewness in earnings growth, which varies over the cycle, affects the mean and generates skewness in consumption growth. Stockholders subject to more negative skewness in earnings growth, as often happens in recessions, hold a smaller share of their financial wealth in stocks.

Buss, Uppaland Vilkov (2018) investigate what happens to the dynamics of asset-allocation decisions and asset prices when financial innovation introduces new assets, such as cryptocurrencies, to households. They show that when households are less well informed about the new asset but learn about it over time, many "intuitive'' results are reversed:  financial innovation increases the volatility of investors' portfolios along with the return volatility and risk premium for the new asset, all of which decline to their pre-innovation levels only slowly.  However, despite the substantial increase in volatility, financial innovation improves the welfare of both inexperienced and experienced investors.

Session 4: Allocation of social care and credit

Bueren (2018) asks to what extent long-term care needs affect the dissaving decisions of the old, by investigating how their decisions are affected by different standards of provision of care across countries. He finds that this is a key driver of savings for high-income elderly people that is significantly more important than bequest motives and medical expenses. His micro data shows that 40% of the cross-country variation in dissaving rates can be explained by differences in provision of long-term care.

Fuster, Goldsmith-Pinkham, Ramadorai and Walther (2018) investigate whether the pace of adoption of machine learning, used to evaluate creditworthiness, risks unequal outcomes across race, age, income, or gender. Using US mortgage and applications data, they investigate how well popular machine learning techniques predict default, and they analyse their impact on exclusion and the rates offered to minority groups. Even considering that data on race, for example, cannot be used in the models, they find evidence that minorities tended to 'lose' under machine learning, compared to less sophisticated logistic regression models, especially when deciding loan rates.

Session 5: What drives investment decisions?

Hartzmarkand Solomon (2018) argue that what economic theory considers to be the return on an asset - the combination of the change in price, plus the cash flows from it - is often not the metric that most investors or the media consider, because indices of stock market performance, for example, show only the change in asset price. Investors also do not have the information that they would need to calculate the return on their portfolio, even if they wished to do so. The authors show that this influences asset prices: stocks covary more with market-wide price change measures than with market-wide dividend yields, even though both parts contribute equally to returns. This insight may help explain why investors tend not to reinvest dividends, and the mystery of why equities outperform bonds.

Choi and Robertson (2018) use a survey if US adults to find how well academic theories of how people invest (or why they don't invest) in stocks describe to real-world financial beliefs and decisions. They find support for many models, notably the rare disaster model, with 42% of respondents describing concern about economic disasters as a very or extremely important factor, but also for long-run risk and loss aversion. Individual investors also believe that past fund manager performance is a good signal of stock-picking skill. 

Bonaparte, Korniotisand Kumar (2018) point out that the assumption that stockholders who enter the market stay in the market is flawed: survey data shows that a quarter of stockholders enter and exit non-retirement investment accounts biennially, and that only 32.8% of households that owned stocks in 1999 consistently owned stocks until 2011. To understand this turnover in stockownership, they estimate a model showing that the income risk and time costs of investing, which they estimate are 3.7% of income, are important determinants of the stock market entry/ exit decisions. The estimation also reveals that long-term stockholders, those that almost never exit the stock market, have greater consumption risk, and their consumption growth is the most correlated with stock returns.

Session 6: Agency and trust

Afzal, d’Adda, Fafchamps and Said (2017) test two hypotheses on agency: the agency value hypothesis (AVH), that people have demand for pure agency, irrespective of the outcome, and the subordinate dependent hypothesis (SDH), that states that even the most subordinate member of a household has agency over certain consumption choices, but much less (if any) agency over more important choices. In their experiment they used a game that varied instrumental value of agency, played by lower income, adult women in Punjab, Pakistan. They find that these women have less influence on big household decisions than on small consumption choices, supporting the SDH, but no evidence that they have a pent-up demand for agency, contradicting the AVH. This suggests that women in the study population have internalised gender norms, especially outside the home.

Finally Bertsch, Hull, Qi and Zhang (2018) use U.S. consumer complaint data and survey data to investigate the increase in P2P online lending between 2008 and 2016. They find that distrust in traditional banking driven by misconduct, and negative experience of traditional finance are positively associated with the online lending share. On the contrary, higher levels of interpersonal trust means that borrowers are more likely to use in-person bank borrowing, decreasing demand for this type of online lending.

References

Afzal, U, G d’Adda, M Fafchamps, and F Said (2017), "Gender and Agency within the Household: Experimental Evidence from Pakistan", paper presented at the 2018 CEPR Third European Workshop on Household Finance​.

Athreya, K, F Ionescu, U Neelakantan and I Vidangos (2018), "Investment Opportunities and Economic Mobility: Who Benefits From College and the Stock Market?", paper presented at the 2018 CEPR Third European Workshop on Household Finance​.

Bertsch, C, I Hull, Y Qi, and X Zhang (2018), "The Role of Trust in Online Lending", paper presented at the Third CEPR European Workshop on Household Finance​.

Bonaparte, Y, G M Korniotis, A Kumar (2018), "Portfolio Choice and Asset Pricing with Investor Entry and Exit", paper presented at the Third CEPR European Workshop on Household Finance​.

Bueren, J (2018) "Long-Term Care Needs: Implications for Savings, Welfare, and Public Policy", paper presented at the Third CEPR European Workshop on Household Finance​.

Buss, A,  R Uppal, and G Vilkov (2018), "Financial Innovation and Asset Prices", paper presented at the Third CEPR European Workshop on Household Finance​.

Choi J J and A Robertson (2018), "What Matters to Individual Investors? Evidence from the Horse’s Mouth", paper presented at the Third CEPR European Workshop on Household Finance​.

Girshina, A, T Y Mathä and M Ziegelmeyer (2017), "Peer effects in stock market participation: Evidence from immigration", paper presented at the Third CEPR European Workshop on Household Finance​.

Hartzmark, S M and D H Solomon (218), "Reconsidering Returns", paper presented at the Third CEPR European Workshop on Household Finance​.

Johnson, S (2018), "Are Mortgage Regulations Affecting Entrepreneurship?", paper presented at the Third CEPR European Workshop on Household Finance​.

Knüpfer, S, E Rantapuska and M Sarvimäki (2018), "Why Does Portfolio Choice Correlate across Generations?", paper presented at the Third CEPR European Workshop on Household Finance​.

Shen, J (2018), "Countercyclical Risks and Portfolio Choice over the Life Cycle", paper presented at the the Third CEPR European Workshop on Household Finance​.

Walther A, A Fuster, P Goldsmith-Pinkham, and T Ramadorai (2018), "Predictably Unequal? The Effects of Machine Learning on Credit Markets", paper presented at the the Third CEPR European Workshop on Household Finance​.

a

A

Topics:  Frontiers of economic research

Tags:  Peer Effects, returns to education, mortgage lending, consumption, financial innovation, long-term care, machine learning, asset returns, stock-market participation

Chair of Macroeconomics and Finance at Goethe University Frankfurt; Director of CEPR Network on Household Finance

Assistant Professor of Finance, University of Warwick

Events

CEPR Policy Research