Investigation of Sentiment Importance on Intraday Stock Returns

Data Science master project by Michele Costa, Alessandro De Sanctis, Laurits Marschall and S. Hamed Mirsadeghi ’18

Investigation of Sentiment Importance on Intraday Stock Returns

Editor’s note: This post is part of a series showcasing Barcelona GSE master projects by students in the Class of 2018. The project is a required component of every master program.


Authors:

Michele CostaAlessandro De SanctisLaurits Marschall and S. Hamed Mirsadeghi

Master’s Program:

Data Science

Paper Abstract:

The main goal of our Master Project is to predict intraday stock market movements using two different kinds of input features: financial indicators and sentiments from news and tweets. While the former are part of the common technical analysis of financial econometric models, the extracted sentiment of news articles and tweets from Twitter are also proven to correlate with stock markets movements. Our paper aims at contributing to the existing academic and professional knowledge in two main directions. First, we evaluate three different approaches to extract the sentiment from both social and mass media based on its forecasting power. Second, we deploy a battery of engineered features based on the sentiment, together with the financial indicators, in a machine learning model for a fine-grained minute-level forecasting exercise. In the end, two different classes of models are fitted to test the forecasting power of the combined input features. We estimated a classical ARIMA-model, and an XGBoost-model as machine learning algorithm. We collected data on the companies Apple, JPMorgan Chase, Exxon Mobil, and Boeing.

Figure: Exxon Mobil
The picture shows how sentiments towards Exxon Mobil moved over time. The two lines refers to two different methodologies: Loughran-McDonald is based on a financial dictionary while SentiStrength was trained on social media such as MySpace.

More about the Data Science Program at the Barcelona Graduate School of Economics

The wisdom or stupidity of the crowd(funding)

Pedro Hinojoauthor is a student in the Barcelona GSE Master in Competition and Market Regulation. Follow him on Twitter @pedrohinojo.


Crowdfunding can be defined as the peer-to-peer provision of financial resources, from the crowd to a particular project or venture. This is usually done via online platforms that forego the need of face-to-face interactions, slashing transactions costs and allowing the fundraiser to reach a wider audience.

This phenomenon started with a non-profit orientation, as donations channelled to political or development campaigns. Reward-based crowdfunding became more relevant later on, where a product is delivered to consumers who finance the project pre-development, usually at a discount or with other ‘perks’ (such as limited editions, first releases, recognition or references in the credits). Even if reward-based crowdfunding entails (economic) advantages for the fund providers, it is tagged as non-profit because these consumers value non-economic benefits (Belleflamme et al, 2013), such as the sense of belonging to a community (a reason for the funding scheme’s popularity in creative industries like films, music and videogames). Reward-based crowdfunding is rooted in the marketing concept of crowdsourcing, whereby firms take advantage of the crowd to obtain ideas, feedback, and solutions to corporate challenges (Schwienbacher and Larralde, 2010).

image source: mindwize.nl
image source: mindwize.nl

But crowdfunding has become relevant when moving towards a profit and investment orientation, be it credit-based or (notably less frequently) equity-shaped (Wilson and Testoni, 2014). In this fashion it is bound to become an alternative source of finance to the real economy when the traditional banking channel is temporarily subdued after the crisis (if not permanently due to more stringent capital requirements). Furthermore, it should benefit primarily small, nascent and innovative firms (which are among the most credit-rationed), especially when they produce unique goods whose features can be communicated easily through the internet.

Therefore, crowdfunding platforms put in contact a crowd of investors with entrepreneurs whose projects need financing. In principle, crowdfunding allows lenders to receive a (higher, although riskier) remuneration for their investment and entrepreneurs to get (cheaper) credit for their projects. Apart from these pecuniary benefits, the entrepreneurs can also promote their brand and products and engage with potential customers through crowdfunding platforms. Therefore, these platforms become essential not only to minimize intermediation costs but also to generate network externalities that attract good projects and a huge crowd of investors. Furthermore, projects appealing to crowdfunding have less geographical constraints in finding sources of finance than with traditional vehicles (Agrawal et al, 2013).

Nonetheless, crowdfunding comes at a cost for entrepreneurs (Agrawal et al, 2013). First, they lose the contact with rather professional investors, who can provide even more valuable advice than a crowd of individual consumers. Other sources of finance for these nascent or innovative firms, like venture capital or (to a lesser extent) angel investors, do provide some technical advice (beyond the funding) to assess the project’s feasibility (and help to improve it if needed).

Second, they may have to disclose some information in the online platforms, something which could be critical in nascent and creative/innovative activities. If some commercially sensitive information becomes publicly known to some extent, incumbents (less credit-constrained) can adapt these new ideas to their business, hammering potential competition from new entrants.

Crowdfunding also poses market challenges, given that imperfections which affect the financial sector are amplified. In financial markets information is far from perfect because it is both incomplete and asymmetric. Information is incomplete because agents cannot control results with their actions in an environment of risk and uncertainty. This problem is amplified in the context of crowdfunding where small, nascent and innovative firms are involved, with riskier projects (Llobet, 2014).

Moreover, information is asymmetric for borrowers and lenders, leading to moral hazard and adverse selection. Moral hazard arises because once borrowers have received the funds, they have the incentive to misbehave and refuse repayment, while the lenders find it difficult to monitor whether these eventual repayment problems are caused by misbehaviour or pure bad luck. Adverse selection happens because lenders cannot discriminate borrowers’ quality, so they charge a high cost to offset potential losses (or even ration credit), jeopardizing paradoxically the best borrowers.

Again, asymmetry of information may be amplified within the crowdfunding context. Given that lenders are now a crowd, it is very likely that each of them only holds a small part of the total investment, reducing incentives to carefully monitor with due diligence the borrowers’ ex ante quality or ex post conduct. In this case, the lack of geographic bonds enabled by crowdfunding is a hitch for lenders to track borrowers in the post-investment phase (Wilson and Testoni, 2014).

In addition, this crowd of lenders will be composed mostly of non-professional investors, so, even if they devoted time to assess the borrowers’ projects, they could lack the necessary skills. And, finally, a crowd of investors would have to face problems of collective action. Against this backdrop, the borrowers might also find less incentive to repay if they use crowdfunding platforms as a one-off bet to raise funds, without any discipline effect coming from repeated interactions or reputational issues.

Bearing in mind these market failures, there is some room for regulation. Considering some international cases (such as the US, the UK or Spain), the regulatory response usually adopts a paternalistic tone: restrictions to the investment by agents (especially individuals who are non-professional investors) and to the amount a project can raise. Crowdfunding platforms are subject to registry requirements similar to other financial intermediaries, although there may be exceptions for small projects. These exceptions can be sources of distortions if firms scale down their projects to fall below certain thresholds (Hornuf and Schwienbacher, 2014)

In order for public intervention to beat the market, regulation ought to be well targeted. Limits to the exposure of non-professional (low-income) investors are rational given their lack of skills, the risks of herd behaviour, path dependence (Agrawal et al, 2013), and the high risk-profile of these investments (Dorff, 2013). However, setting stringent caps on the maximum raisable amount for projects may squeeze the sector’s (and the whole economy’s) development.

Furthermore, the sector itself can provide some solutions to these market failures. For instance, crowdfunding platforms normally charge a fee for every successful project, (which raised the same or more funds than it had pledged). This strategy gives the platforms the right incentives (skin in the game) to monitor and screen projects, so that small (non-professional) investors are relieved of that assessment.

Besides, most crowdfunding platforms opt for an All-Or-Nothing (AON) or a ‘provision point mechanism’ model, whereby projects which do not raise the amount of funds they had pledged will not receive anything. Platforms opting for the Keep-It-All scheme (KIA, whereby projects receive all the funds they have raised even without having achieved their goal) would see higher funding costs charged to those projects (Cumming et al, 2014), given that underfunded projects (which would still receive the funds in the KIA scheme) have less likelihood to succeed.

To conclude, crowdfunding offers a promising venue to spur innovation, creativity and firm growth. The regulatory response must allow that development while ensuring that no substantial amounts are invested by those individuals lacking the skills and the resources needed to cope with the complex and risky investments (See Kay, 2014, whose contribution also served as an inspiration for the title of this post).

The Death of Fixed Income

alumni


Alex Hansson
’13 (International Trade, Finance and Development) is an Analyst at Tribus Capital Partners in Zurich, Switzerland. Previously he was External Asset Management Analyst at Credit Suisse.


 

cartoon
Source: The New Yorker

A few weeks ago I found myself sitting in the James Joyce Wine Cellar in Zurich listening to a talk by the Chief Investment Officer of one of the world’s largest asset managers’ fixed income division. Surrounded by ageing wine bottles and Swiss bankers in grey suits, the CIO put down his glass of Chateau Neuf du Pape and proclaimed, “You know, I’ve been investing in fixed income all my career, but I feel obliged to tell you that fixed income as an asset class is dead.” That was the last sentence of the speech.

Now I don’t frequent too many of these events, but ending a presentation on that kind of a bombshell was not something that I’d ever experienced before. The Swiss bankers around me shifted nervously in their seats thinking something else was coming. But the CIO did not follow up the statement with any caveats. He was undoubtedly out to provoke and had we pushed him a little he would have most likely followed up his statement with the usual spiel that one should be careful with statements like ‘this time it is different.’ He would have probably added that fixed income in terms of developed markets sovereign and corporate bonds would at some point in the future again be an interesting investment.

Nonetheless, finding out fixed income was dead was pretty exciting. Moreover, the statement was in line with a trend I had been witnessing for the past couple months whereby capital was flowing into a new asset class – alternative fixed income.

But, before moving on to this, it is worth spending a few minutes on the death of fixed income, whether temporary or not. Classic portfolio management theory will tell you to put 60% of your assets in equities (e.g. at its most basic through a S&P 500 Index ETF) and 40% in bonds (e.g. at its most basic through a Bond Index) in order to achieve diversification and lower the portfolio’s volatility. Without getting too deep into the technicalities, the thinking goes that equities outperform bonds in the long run – hence the overweight. However, bonds are included because they lower the volatility and increase the diversification of the overall portfolio. They lower the volatility because they are less risky than equities. Moreover, bonds have historically moved inversely to equities and this diversification effect has served the purpose of smoothing out the return profile of a balanced portfolio in times of market turbulence. Seeing as, nearly all investors have regular liquidity needs for, for instance, a regular mortgage payment, a smoother albeit lower return profile in the long run is more attractive than high volatility with a higher return in the very long run.

So what has changed? Norman Villamin (CIO Coutts) was recently quoted as saying that bonds used to offer an asymmetric payoff relative to equities whereas now you are left with a “symmetric asset class.” Put another way, bonds still serve the purpose of protecting capital when equities are down but no longer give you any meaningful upside. In other words by having 40% bonds you lower volatility but you no longer achieve diversification. Explaining exactly why, at least from a developed markets perspective, is not straightforward. Some analysts have highlighted that rates have been kept low to encourage lending given the slower than expected recovery in the US and Europe. Other analysts have pointed to lower inflation and central banks not feeling the pressure to increase rates. In any case, with very low returns you also have very low volatility. With low volatility even if this effect moves inversely to equities it will not be ‘strong enough’ to provide the desired counterbalancing effect. Think of two grown men on a seesaw where one gets replaced by a child, no matter how much the child jumps up and down on the plank, they will not significantly propel the man into the air. Whatever the reasons for this phenomenon, most seem to agree that a low rate environment will be the norm for some time to come.

Many investors realize that if rates remain low it will become more difficult to find bonds yielding attractive returns without taking on excessive risk. In terms of fixed income this generally means looking towards emerging market debt or distressed corporate bonds. To be clear some investors have successfully gone down this route, such as, Michael Hasenstab (CIO Franklin Templeton) who has became famous for his bullish speeches on Ukrainian debt. Nonetheless this dynamic has led to many investors leaving fixed income in favor of equities. That trend is clearly reflected by the fact that the largest fund by assets under management has over the last year shifted from the PIMCO Total Return fund (a fixed income product) to the Vanguard Total Stock Market Index (an equity product). The leaving of Mr. Gross from PIMCO no doubt also contributed to this shift.

With the recent prolonged bull market most investors have shut their eyes to the increased volatility and enjoyed the attractive returns they have seen from their portfolios. However, some sophisticated investors have elected not to move their fixed income exposure to equities. Some have chosen to put assets into well-known alternatives like private equity or real estate. But these asset classes do not offer the same advantages that fixed income did.

As outlined above fixed income was liked because, first, it was relatively low risk. Second, it paid a steady regular coupon, which was liked not just by private investors but especially pension and insurance funds, which have regular liquidity needs. Third, it paid a handsome return of anywhere from 5.00% to 8.00% (historically seen). To put that into perspective, a ten year German Government Bond pays you less than 1.00% currently. A private equity or real estate investment conversely can mean having to lock up capital for ten years, with no regular coupon payment, and with no guarantee for handsome reward.

Given this, many of these sophisticated investors have turned to a different asset class that has emerged largely as a product of new lending opportunities. It is often referred to as Alternative Fixed Income. These opportunities have to a large extent emerged as a product of new banking regulations. Simply put, banks have been forced to move away from certain types of lending and new non-bank credit providers have stepped in to fill this void offering new investment opportunities. Indeed, according to Alliance Bernstein, since 1980 the nonfinancial corporate and mortgage credit outstanding has grown by ca. USD 18.9tn whereof USD 15.0tn can be attributed to non-banks.

This trend of lending by non-banks has been accelerated largely as a product of post-crisis regulation. In Europe, the most important regulation framework for this trend has been Basel III. This framework stipulated that banks now have much higher capital requirements as well as having to take on higher operational costs for certain types of lending. This has been most pronounced for investments with longer investments horizons, which therefore cannot match bank-liability structures subject to daily liquidity requirements.

There are literally hundreds of structures that have emerged to take advantage of this systemic shift. The biggest opportunities for alternative credit providers have been in corporate loans, commercial real estate loans, residential real estate loans, and infrastructure loans. Many new organizations have been set up and are raising capital from investors in order to lend to corporations and real estate developers in the same way banks used to. The lenders pay a coupon on the loan, the new firms take a cut, and the investor gets a regular coupon payment at a tolerable risk and a much higher return than they would have in traditional fixed income.

This systemic shift has created opportunities along the entire supply chain of lending. It stretches from the most logical to the most niche areas you can think of, and no doubt this is just the beginning. Below I will give two examples of the most common constellations.

Most commonly you see new organizations made up of former bankers. They leave or are made redundant by their employer. They take with them an intimate knowledge of that bank’s loan book as well as the skillset to analyze other banks’ loan books. They then raise capital in a fund structure (funded by our sophisticated investors above) and begin buying loans from banks that can no longer be kept on their balance sheets. This can take many different forms. The easiest is a pure purchase at a discount where the new organization takes over the loan. Typically this happens via a Collateralized Loan Obligation (CLO) which is a securitized asset backed by a pool of debt. Another variety is a situation where banks have debt on their balance sheet which is highly attractive from a return perspective but which they cannot keep on their balance sheet. They therefore use something called Regulatory Capital Relief Trades (CRTs) in order to temporarily transfer risky assets off their balance sheets to one of these new organizations which in turn charge interest on the assets while on their balance sheets. Finally you have situations where the bank will give an organization access to their pipeline of loan opportunities. This setup is rarer and requires a very close relationship between the bank and the new organization. Essentially, these new organization can sift through bank’s pipeline of lending opportunities and chose to take on any loans they find attractive. In return the bank receives a share of the profits.

The other more common constellation of new organizations is direct loans that banks used to be able to do. Indeed, many of these organizations are whole teams that used to do the same thing at their former employers. A great example is Renshaw Bay. Renshaw Bay was setup by the former co-CEO of J.P. Morgan’s Investment Bank – Bill Winters. The firm runs a real estate strategy that is “focused on direct, whole loan origination of commercial real estate loans… and seeks to take advantage of the lack of financing available to real estate borrowers”. They also run a structural finance strategy, which looks to “capitalize on opportunities driven by regulatory change and the retreat of capital.” There are now new organizations (or old organizations which have seen a significant uptick in demand) that are doing the same thing in corporate loans, real estate loans, infrastructure loans, or even leverage loans to private equity firms.

It is still relatively early days and no one yet completely understands the full implications of this systemic shift. Perhaps the setup is better, seeing as you have more specialized niche players, often with much of their internal capital at risk, and unhindered by bureaucracy associated with banks running these new organizations. Or perhaps this setup is riskier, since at the same time you also have small, inexperienced, and at times highly leveraged organizations with less oversight then banks used to have. Added to this increasing numbers of pension funds and insurance companies are investing in these structures. Undoubtedly there will be some financial cowboys who have cut corners in order to have first mover advantage. At the same time you will undoubtedly have a lot of smart people who will make themselves and their investors a lot of money. Clearly, it is going to be very interesting to monitor these developments to see whether this shift turns out to be a good one or a bad one.

Networks and Contagion in Financial Markets

Too well connected to fail

(This article follows on from a more general post on the study of networks in economics)

In this model, as well as those concepts Jackson discussed in the broader discussion on networks, we have the concepts of diversification and integration to separate the breadth and depth of connectivity of one organisation to others. A company/organisation/ country with many connections to others would be highly diversified; where those interests represented a higher proportion of their overall connectivity, they would be highly integrated. Continue reading “Networks and Contagion in Financial Markets”