Finance: Research, Policy and Anecdotes

My last conference of the year (and even decade) was on Funding Stability and Financial Regulation and organised by ACPR and ANR in Paris. Yet, another financial stability workshop, I thought, but it turned out to include a number of very interesting and policy-relevant papers, some of whom I will mention in the following. I will not discuss my own paper, as I will dedicate a separate blog entry to it next year, when a presentable working paper version will be finally ready.


Helene Rey presented work on Machine Learning and the Financial Crisis. Using a general framework that draws on different crisis prediction models in a type of meta-analysis, it improves on standard crisis prediction models (which typically have a relatively poor out-of-sample prediction power) and is able to predict systemic financial crises 12 quarters ahead in quasi-real time with very high signal to noise ratio.  Melina Papoutsi shows the importance of lending relationship for borrowers that fall in distress. Using data from Greece, she shows that firms that experience an exogenous interruption in their loan officer relationship are less likely to renegotiate their loans and, if they manage to renegotiate, they are given relatively tougher loan terms, compared to firms whose loan officer relationships were not interrupted. Relating this to my own work, yet again, more evidence that personal lending relationships are far from dead. My former colleague Max Bruche presented fascinating work on leveraged loan syndication (a better description would be: junk loan syndication), making a strong case that supervisors better pay more attention to retention risk in this market segment. Neeltje van Horen presented a paper that is effectively an impact evaluation of Basel III on the repo market. Specifically, she shows that the adoption of the leverage ratio (which should have a dampening effect on low-margin business such as repo transactions) had transitory effects on the access to the repo markets by smaller clients. Thibault Libert shows that large borrowers can have an impact on aggregate lending, in work that uses credit registry data from France and builds on an expanding literature that looks at the importance of firm-specific shocks for macro-aggregates. Laura Blattner uses Portuguese credit registry data to show that increased capital requirement can have perverse consequences as affected banks respond by not only cutting lending but also by reallocating credit to distressed firms with underreported loan losses. Finally, Eva Schliephake presented some very interesting theory work on how informed and uninformed depositors interact in bank runs, showing that more information can actually result in a higher likelihood of panic runs.

The voters have spoken and the uncertainty is over.  Unless something unexpected happens in the next few weeks, the UK will finally leave the EU on 31 January. There must be quite a feeling of relief in Brussels and across the EU that the naughty kid has finally decided to leave. But the show will go on – far from getting Brexit done, in February the new season of the Brexit show will start, this time negotiating the future relationship between the EU and UK. It will then become clear that Boris Johnson has fooled his voters yet again – after lies about 350m extra for the NHS and about no controls in the Irish Sea. There will be a rather long drawn-out negotiation with the EU on the future trade relationship. Of course, this can be done within three months, if the UK accepts everything (EVERYTHING) the EU proposes (including in such sensitive areas as fishing and the UK following EU rules). And even that would have to imply that there are no contrasting interests among the 27 member countries of the EU in these negotiations – unlike in the withdrawal negotiations, where little conflict could be expected among EU member states on money (more is better), EU citizen rights (as water tight guarantees as possible) and avoiding a border across Ireland.  All three objectives are achieved with the withdrawal treaty, but the interests among EU member states might very well diverge when it comes to the new relationship with the UK. And let’s not forget that unlike the withdrawal agreement, a comprehensive free trade agreement between the EU and the UK has to be approved by all national parliaments as well as some regional ones.


So, here is the new Brexit trilemma – there are three objectives for the Johnson government– (i) exit from transition phase by end-2020, (ii) get as good a deal as possible for the UK economy and the Conservative Party, and (iii) avoid another no-deal cliff edge at the end of the transition period -  and at most two can be achieved. To achieve (i) and (iii), (ii) has to give, i.e., the UK has to accept everything the EU proposes. To achieve (ii) and (iii), (i) has to give, i.e., at a minimum a two-year extension of the transition period has to be accepted by the UK government. Objectives (i) and (ii) are not compatible from the start.


What will happen? What will Boris Johnson do? He threw the DUP over board when it was convenient in order to get a deal with Brussels that ensures no border in Ireland but a border in the Irish Sea. Will the ultra-Brexiteers among the Tories also just roll over when he defaults on all his promises to them?  If he gets the big majority the exit polls predict, he might not care. He can then either agree to an extension of the transition period (more likely) or give in to all of the EU’s proposals (less likely). In any case, I would argue that the new deadline is now 2024, i.e., the Conservatives will aim to get all new relationship with EU defined before the next elevations.  It does promise many more seasons of the Brexit soap opera then.


These landslide results also indicate that the opposition has failed. It has failed to successfully make the case for a second referendum. It has failed to combine forces to translate a large voting share into actual seats at parliament. It has failed to stop the populist nationalist English wave. We will see a lot of infighting in the Labour Party over the next year or so and a lot of soul-searching among the Liberal Democrats.


My comments so far have been somewhat removed, without personal touch – it does help that I am in Sydney right now, far away from the Brexit mess.  But on a personal level, there is a degree of sadness, somewhat similar to the morning of 24 June 2016, a feeling of loss and finality.  It is sad to see a country close itself off the world and embark on a long path to long-term decline. It is scary to see its governing party appeal to people’s lowest instincts of xenophobia and attack the media at every feasible point.  If this sounds familiar – yes, there are clear parallels now between the transformation of the Republican party in the US and the Tories in the UK – though there is hope of a post-Trump era for the US and the Republican party, whereas it seems harder to revert the populist trends in the UK.  As an economist and observer, the next years promise to be as interesting as the last three, as UK resident, it is dispiriting and sad.

Yesterday, I participated in an exciting panel discussion on infrastructure finance at the IFABS conference in Medellin, together with Eduardo Cavallo (IDB) and Alejandro Sanchez (Corficolombiana), thus combining practitioner, policy and academic viewpoints.  There are lots of things to be noted on this topic, so here just some highlights – linking also to some projects I have done in this area over the past years. First, on a theoretical and practical level, infrastructure finance (or project finance more broadly) shows several characteristics that increase cost and risk profile – large size (requiring syndicates), long maturities, a lack of collateralizable assets in the early stages of funding and repayment flows only feasible after the construction phase. There are also higher skill and capacity requirements in infrastructure compared to other financing modes. These challenges have been around for many decades and it is an on-going area of learning, both for practitioners and policymakers. This also implies that one segment of the financial sector might not be positioned well to take on this challenge by itself, nor the private or public sector independently. It also implies that there are many policy and regulatory challenges in this space and a close connection with the larger challenge of financial sector development.


Second, as pointed out by Eduardo, most of infrastructure financing in Latin America is undertaken by banks, rather than non-bank financial institutions, which would be much better positioned to do so. Specifically, pension and mutual funds are in a better position (especially the former due to long-term liabilities and the latter due to risk profile) to invest in infrastructure. My own case study for Colombia (undertaken a few years ago for the IDB) shows that pension funds are relatively well developed in Colombia, though mutual funds focus mostly on low-risk, low-return securities (also referred to as “AAA-itis”). So, infrastructure is still supported mainly by banks rather than by non-banks. On the upside, the Financiera de Desarrollo Nacional,  set up in 2011 by the government, with support from IFC and CAF, as well as – at a later stage – by the IDB, has evolved into a best practice example of public-private partnership taking on a critical role in structuring financing arrangements using a mix of instruments. It has successfully provided not only direct finance, but also been a catalyst in bringing in domestic and foreign private funding for infrastructure. However, broader challenges in long-term finance continue: how to bring a larger share of the active population into the pension fund system – mainly a problem of informality -, how to address the high concentration in the pension fund industry  and how to lower entry barriers. The ultimate challenge, however, is: how to increase the risk appetite of non-bank financial institutions?


Third, has regulatory tightening under Basel III resulted in banks moving even further away from infrastructure financing, given the recent regulatory focus on reducing maturity mis-matches? This FSB evaluation suggests that no,  but some caveats are due with such an assessment, including that the impact on infrastructure finance is likely to be slower moving than that on other segments because infrastructure finance involves fewer larger transactions, typically with longer maturities. In this taskforce report with Liliana Rojas Suarez, we point to several potential adverse effects that Basel III can have on infrastructure finance (through liquidity requirements, output floor, exposure limits etc.). Most important, however, seems the lack of infrastructure as asset class. If projects can be developed in a more standardized fashion, and if there is agreement on the different dimensions of risk and how they should be quantified, it may become easier to issue securities backed by infrastructure projects, potentially also resulting in lower risk weights. This could allow banks to finance infrastructure projects in the early stage before selling them off. It also makes investment by non-bank financial institutions in such projects more likely!


Finally, infrastructure finance is part of the larger long-term finance landscape. While the focus of researchers and policymakers has been on financial inclusion over the past decade, it is important to focus again more prominently on the challenges of long-term finance. I have earlier written about an effort to get data on long-term finance for Africa. The strong needs in infrastructure funding as well as long-term funding needs by firms and households calls for a comprehensive approach to strengthen the intermediation and maturity transformation capacity of banks and non-banks alike, taking into account their critical linkages and synergy effects.


I had the pleasure of discussing two excellent papers at the 1st Finance and productivity conference at the EBRD this week, on the role of finance in fostering or impeding entrepreneurship. Christoph Albert and Andrea Caggese use survey data from the Global Entrepreneurship Monitor for 21 OECD countries over the period 2002 to 2013 and show that GDP and financial sector shocks hurt the establishment of new enterprises, especially of high-growth enterprises, providing convincing evidence that financing constraints are especially binding for transformational, potentially high-growth entrepreneurs.   Nandini Gupta and Isaac Hamaco, on the other hand, document a drain brain from manufacturing into the financial sector in the US. Specifically, engineering graduates between 1998 and 2006 are more likely to work in financial sector if they start out in an area with a higher share (and thus higher growth) of financial sector employment or study in a state that deregulates interstate banking. This, in turn, results in relatively fewer start-ups founded by engineering graduates that go into finance, as well as less innovation and less VC funding by such start-ups.


Together, these papers add evidence to two strands of the macro-finance literature that have developed somewhat parallel – on the one hand, the importance of alleviating financing constraints to foster entrepreneurship and thus ultimately improve resource allocation and increase productivity growth; on the other hand, the unhealthy growth of the financial sector, extracting rents from the real economy and drawing talent away from the manufacturing sector. Both papers thus relate to the decline in start-ups (as documented for the US in this Economic Policy paper, which also shows that it is not related to higher federal regulation) and the slow-down in productivity growth. But what to make of the seeming contradiction – financing constraints vs. brain drain?  Well, these findings are consistent with an important role of financial intermediation for economic growth (which works through productivity growth and thus entrepreneurship), but also with a growth-impeding effect of an oversized financial system that does not necessarily focus on intermediation anymore. It is thus consistent with tentative results that Hans Degryse, Christiane Kneer and I documented  - financial intermediation helps growth in the long-run, while an indicator gauging the size of the financial sector (e.g., employment share) results in higher short-term growth, but higher growth volatility in the long-term. These findings are also consistent with work by Christiane Kneer who documented a brain drain, looking specifically at the US – industries with higher financing needs benefit from a larger financial sector, while industries with a higher share of R&D and skilled workers actually loose. In summary, efficient financial services are important for the real sector, while an oversized financial system is not necessarily and might even be damaging for the real sector.

The future of banking in light of technological disruption has been high on the agenda of banking sector analysts, policy makers and researchers.  The Libra announcement by Facebook has raised the urgency of the topic of digital currencies in central bank corridors around the world. As most banking/corporate finance researchers, I have somewhat contributed to this literature by (i) looking at the effects of financial innovation in banking more generally (finding both a stability-reducing and growth-enhancing effect) and (ii) analysing the real sector effects of mobile money in Kenya. While both studies are (by their empirical nature) backward-looking, I am also getting involved in more conceptual discussions, both on the inclusion as on the stability side. So, in the next few months, I will publish some thoughts on the future of banking and technological disruption on my blog – all of this purely my own thoughts, but often based on other people’s research.


Financial innovation has been around for centuries and often has had disruptive effects. The introduction of the ATM in the 1970s allowed US banks facing geographical constraints to undermine them and ultimately contributed to the liberalisation of branching restrictions in the US. Financial innovation often comes in the form of new types of intermediaries. As discussed by Laeven, Levine and Michalopoulos (2015), specialized investment banks emerged to facilitate the construction of vast railroad networks in the 19th and 20th centuries across North America and Europe, screening and monitoring borrowers on behalf of disperse and distant investors. In the second half of the 20th century, venture capital funds arose to finance IT start-ups, characterized by limited if any tangible assets that could be used as collateral and thus requiring patient investment capital and close screening and monitoring as well as technical advice.  Today’s disrupters are FinTech and BigTech companies, although there is a big difference between the two. FinTech refers to technology-enabled innovation in financial services with associated new business models, applications, processes or products, all of which have a material effect on the provision of financial services.  While often undertaken by independent start-ups they do not really compete against banks – to the contrary, banks encourage experimentation in this space and often take over successful companies.   BigTech companies, on the other hand, are large existing companies whose primary activity is in the provision of digital (platform) services, rather than financial services; for these companies financial services is thus an add-on service. A critical difference between BigTech companies and other large companies that (want to) branch out into financial services (e,.g, Banco Azteca in Mexico, Walmart in the US) can be summarised with the acronym DNA, an expression coined by the BIS – data, network and artificial intelligence.  Banks have always relied on their ability to collect and process hard and soft information about borrowers; BigTech firms have such data readily available from their non-financial business with customers; artificial intelligence allows them to also convert soft information into hard information. Several papers have shown that information collected this way is better in predicting default than information shared between banks in credit registries (e.g., here and here). In addition, given their platform character, BigTechs enjoy network economies, helping to reduce transaction costs and allowing better diversification.  This provides the chance for BigTech companies to enter areas so far dominated by banks, including retail banking.


The increasing accumulation of data raises important questions on the use of such data but also the ownership of data. The Open Banking initiative in the EU allows customers to share data across different banking institutions. The question is whether this should be expanded to BigTech companies, especially when they move into financial service provision. It also raises concerns on new risk sources, such as cyber risk, and might require a stronger focus of regulators on consumer protection.


The possibility to use an increasing amount of data also raises question on how they are being used. The financing constraint view has argued that more data allows more efficient provision of financial services and allows reaching households and small enterprises that so far have been excluded due to their lack of collateral and audited accounts. Credit registries are typically seen as critical component of the institutional infrastructure underpinning financial deepening, as they allow clients to build up reputation collateral. On the other hand, there is increasing evidence that such data can also be used for client-specific targeting. More data allow price discrimination and targeted shrouding, where the latter is more consequential in finance, given intertemporal nature of contracts. I recently had the honour of discussing a paper by Antoinette Schoar showing exactly that for credit card offers in the US: Less educated consumers receive more back-loaded terms (low teaser interest rate but high late-payment fees), and more shrouded offers, exploiting behavioural biases.


In summary, the financialisation of the modern economy and society seems to give way to the digitalisation of finance. This offers great opportunities for innovation and increased competition in the financial sector, but also lots of public policy challenges. In a future blog entry, I will focus more on the stability challenges.