Category Archives: Finance

Text-reading machines can predict share prices

The latest on machine learning applications in finance…

“Investors and researchers have suspected for decades that text could be used to predict markets, some trying and failing. But applying machine-learning techniques originated by computer scientists, Harvard’s Zheng Tracy Ke, Yale’s Bryan T. Kelly, and Chicago Booth’s Dacheng Xiu have built a model that in early tests outperformed a similar strategy based on scores from RavenPack, the leading vendor of news-sentiment scores.”

On the role of replicating portfolios in the pricing of financial derivatives in general

Replicating portfolios play a central role in terms of pricing financial derivatives. Here is what we have learned so far about replicating portfolios in Finance 4366:

  1. Replicating portfolios for long and short forward contracts. Buying forward is equivalent to buying the underlying on margin, and selling forward is equivalent to shorting the underlying and lending money. Like options, forwards and futures are priced by pricing the replicating portfolio and invoking the “no-arbitrage” condition. If the forward/futures price it too low, then one can earn positive returns with zero risk and zero net investment by buying forward, shorting the underlying and lending money. Similarly, if the forward futures price is too high, one can earn positive returns with zero risk and zero net investment by selling forward and buying the underlying with borrowed money. This is commonly referred to as “riskless arbitrage”; it’s riskless because you’re perfectly hedged, and it’s arbitrage because you are buying low and selling high.
  2. Replicating portfolio for a call option. The replicating portfolio for a call option is a margined investment in the underlying. For example, in my teaching note entitled “A Simple Model of a Financial Market”, I provide a numerical example where the interest rate is zero, there are two states of the world, a bond which pays off $1 in both states is worth $1 today, and a stock that pays off $2 in one state and $.50 in the other state is also worth one dollar. In that example, the replicating portfolio for a European call option with an exercise price of $1 consists of 2/3 of 1 share of stock (costing $0.66) and a margin balance consisting of a short position in 1/3 of a bond (which is worth -$0.33). Thus, the arbitrage-free value of the call option is $0.66 – $0.33 = $0.33.
  3. Replicating portfolio for a put option. Since the replicating portfolio for a call option is a margined investment in the underlying, it should come as no surprise that the replicating portfolio for a put option consists of a short position in the underlying combined with lending. Thus, in order to price the put, you need to determine and price the components of the replicating portfolio.  Continuing with the same numerical example described for the call option in part 2 above,  the replicating portfolio for an otherwise identical European put option with an exercise price of $1 consists of a short position in 1/3 of 1 share of stock (which is worth -$0.33), combined with a long position in 2/3 of a bond (which is worth $0.66). Thus, the arbitrage-free value of the put option is  – $0.33 + $0.66 = $0.33.
  4. Put-Call Parity. Note also that if you know the value of a call, the underlying, and the present value of the exercise price, then you can apply the put-call parity equation to figure out the price for the put option; i.e., {C_0} + PV(K) = {P_0} + {S_0} \Rightarrow {P_0} = {C_0} + PV(K) - {S_0}. Since we know the price of the call ($0.33), the present value of the exercise price ($1), and the stock price ($1), then it follows from the put-call parity equation that the value of the put is also 33 cents. More generally, if you know the values of three of the four securities that are included in the put-call parity equation, then you can infer the “no-arbitrage” value of the fourth security.

Prediction markets’ take on removal of POTUS from office

As of 2:15 p.m. central standard time today, the PredictIt.org prediction market put the odds of President Trump being removed from office at 8%.  Specifically, Predictit.org currently offers for sale a “share” which pays $1 if the answer to the question, “Will the Senate convict Donald Trump in his first term?, turns out to be “yes”.

Allow me to provide further context for this “prediction”.  PredictIt.org is a New Zealand-based prediction market that offers “shares” on political and financial events.  The idea behind  PredictIt.org shares (technically, these are binary options, but I digress) is quite simple – you can buy and sell “yes” and “no” shares which pay off $1 if the answer to the contract question ends up being “yes” or “no”.  If you buy yes (no) but no (yes) is the answer, then your share expires worthless and you have lost the full value of your original “investment”.  However, if you sell yes (no) and no (yes) is the answer, then you don’t owe your counterparty any money and you get to pocket the price received (net of transactions costs) as profit.

Since the payoffs on PredictIt.org shares feature binary payoffs (i.e., $1 if yes and $0 if no),  these shares are canonical examples of Arrow-Debreu, or “pure” securities.  Arrow-Debreu securities pay $1 if a particular state (in this case, either “yes” or “no”) occurs at a particular time in the future.  Thus, the current price for a given PredictIt.org share is the “state price”,  which corresponds to the value today of $1 received when a particular future state of the world is realized.  Breaking the state price down further, its components include 1) the probability of a particular future state of the world, 2) the rate of interest (to compensate for the time value of money), and 3) a further discount (to compensate for risk averse behavior by the bettor) or premium (to compensate for risk loving behavior by the bettor).

Prediction market prices are frequently referred to in the news media as probabilities for future state-contingent events; if prediction market participants are risk neutral and interest rates are negligible, then this is technically appropriate and roughly correct.  What’s fascinating about prediction markets is that they showcase, in very pure form, how market prices reflect the statistical odds of some future event happening.  Similarly, prices of speculative assets generally (e.g., corporate securities such as stocks and bonds and derivative securities such as options and futures) also reflect probabilistic beliefs about future states of the world, albeit in more of an opaque fashion.

The Index Fund also featured as one of “50 Things That Made the Modern Economy”

Besides insurance, Tim Harford also features the index fund in his “Fifty Things That Made the Modern Economy” radio and podcast series. This 9 minute long podcast lays out the history of the development of the index fund in particular and the evolution of so-called of passive portfolio strategies in general. Much of the content of this podcast is sourced from Vanguard founder Jack Bogle’s September 2011 WSJ article entitled “How the Index Fund Was Born” (available at https://www.wsj.com/articles/SB10001424053111904583204576544681577401622). Here’s the description of this podcast:

“Warren Buffett is the world’s most successful investor. In a letter he wrote to his wife, advising her how to invest after he dies, he offers some clear advice: put almost everything into “a very low-cost S&P 500 index fund”. Index funds passively track the market as a whole by buying a little of everything, rather than trying to beat the market with clever stock picks – the kind of clever stock picks that Warren Buffett himself has been making for more than half a century. Index funds now seem completely natural. But as recently as 1976 they didn’t exist. And, as Tim Harford explains, they have become very important indeed – and not only to Mrs Buffett.”

Warren Buffett is one of the world’s great investors. His advice? Invest in an index fund

Insurance featured as one of “50 Things That Made the Modern Economy”

From November 2016 through October 2017, Financial Times writer Tim Harford presented an economic history documentary radio and podcast series called 50 Things That Made the Modern Economy. This same information is available in book under the title “Fifty Inventions That Shaped the Modern Economy“. While I recommend listening to the entire series of podcasts (as well as reading the book), I would like to call your attention to Mr. Harford’s episode on the topic of insurance, which I link below. This 9-minute long podcast lays out the history of the development of the various institutions which exist today for the sharing and trading of risk, including markets for financial derivatives as well as for insurance.

“Legally and culturally, there’s a clear distinction between gambling and insurance. Economically, the difference is not so easy to see. Both the gambler and the insurer agree that money will change hands depending on what transpires in some unknowable future. Today the biggest insurance market of all – financial derivatives – blurs the line between insuring and gambling more than ever. Tim Harford tells the story of insurance; an idea as old as gambling but one which is fundamental to the way the modern economy works.”

Khan Academy “Finance and capital markets” videos

Not only are the Khan Academy Calculus and Statistics videos that I referenced in a previous posting quite useful; I am also a big fan of the Khan Academy “Finance and capital markets” videos; these videos do a great job of effectively presenting some of the most important concepts which are typically covered in undergraduate and MBA level finance curricula. Note also that the “Options, swaps, futures, MBSs, CDOs, and other derivatives” subsection of the “Finance and capital markets” page covers many of the same topics that we will cover this semester in Finance 4366!

Plans for next week’s Finance 4366 class meetings, along with a preview of future topics

Next week in Finance 4366 will be devoted to tutorials on probability and statistics. These tools are foundational for the theory of pricing and managing risk with financial derivatives, which is what this course is all about. Next Tuesday’s class meeting will be devoted to introducing discrete and continuous probability distributions, calculating parameters such as expected value, variance, standard deviation, covariance and correlation, and applying these concepts to measuring expected returns and risks for portfolios consisting of risky assets. Next Thursday will provide a deeper dive into discrete and continuous probability distributions, in which the binomial and normal distributions are showcased.

Beginning Tuesday, January 28, we introduce financial derivatives, We will begin by defining financial derivatives and motivating their study with examples of forwards, futures, and options. Derivatives are so named because the prices of these instruments are derived from the prices of one or more underlying assets. The types of underlying assets upon which derivatives are based are often traded financial assets such as stocks, bonds, currencies, or other derivatives, but they can be pretty much anything. For example, the Chicago Mercantile Exchange (CME) offers exchange-traded weather futures contracts and options on such contracts (see “Market Futures: Introduction To Weather Derivatives“). There are also so-called “prediction” markets in which derivatives based upon the outcome of political events are actively traded (see “Prediction Markets“).

Besides introducing financial derivatives and discussing various institutional aspects of markets in which they are traded, we’ll consider various properties of forward and option contracts, since virtually all financial derivatives feature payoffs that are isomorphic to either or both of these schemes. For example, a futures contract is simply an exchange-traded version of a forward contract. Similarly, since swaps involve exchanges between counter-parties of payment streams over time, these instruments essentially represent a series of forward contracts. In the option space, besides traded stock options there are also embedded options in corporate securities; e.g., a convertible bond represents a combination of a non-convertible bond plus a call option on company stock. Similarly, when a company makes an investment, there may be embedded “real” options to expand or abandon the investment at some future date.

Perhaps the most important (pre-Midterm 1) idea that we’ll introduce is the concept of a so-called “arbitrage-free” price for a financial derivative. While details will follow, the basic idea is that one can replicate the payoffs on a forward or option by forming a portfolio consisting of the underlying asset and a riskless bond. This portfolio is called the “replicating” portfolio, since it is designed to perfectly replicate the payoffs on the forward or option. Since the forward or option and its replicating portfolio produce the same payoffs, then they must also have the same value. However, suppose the replicating portfolio (forward or option) is more expensive than the forward or option (replicating portfolio). It this occurs, then one can earn a riskless arbitrage profit by simply selling the replicating portfolio (forward or option) and buying the forward or option (replicating portfolio). However, competition will ensure that opportunities for riskless arbitrage profits vanish very quickly. Thus the forward or option will be priced such that one cannot earn arbitrage profit from playing this game.