what determines which assets you plan to invest in?

Final Project Topic: Asset Management & Portfolio Construction
Overall goal
Come up with and test with a portfolio strategy a set of rules that determines how much you invest in a given
asset at a given point in time.
Possible assets include the risk-free asset, individual stocks, existing portfolios of stocks constructed by others
(e.g. industry portfolios from Ken French’s website), exchange-traded funds (including bond funds, which is an
easy way to add bonds to your portfolio), currencies, and cryptocurrencies. You can take negative positions in
stocks, funds, and currencies i.e. you can short them. You cannot take negative positions in cryptocurrencies —
it is quite difficult to short crypto. Altogether, your weights should add up to 1 if you propose a long-only portfolio
and sum up to 0 if you propose a long-short portfolio.
A static portfolio buys once and does not rebalance. In a static portfolio, asset weights evolve just based on the
performance of those assets. For example, if I invest 50% in Apple and 50% in Microsoft and Microsoft does
better than Apple over the next month, then the value of my Microsoft position increases more than the value of
my Apple position, and thus my portfolio weights tilt towards Microsoft.
A dynamic portfolio rebalances at some frequency. This rebalancing is a key part of an investment strategy.
Keep in mind that maintaining fixed weights requires rebalancing. In the example above, to get back to a 50/50
split between Microsoft and Apple, I would need to sell some Microsoft shares and buy Apple shares. You can
use other information to inform your rebalancing — maybe you want to go long small firms and short large firms
and rebalance when a small firm gets to be large? This is just an example.
Deliverables

1. Describe your investment thesis i.e. what views of the market, the economy, or specific asset classes or
   assets. Broadly speaking, what is the strategy? Which assets do you plan to hold long? Short? How will
   you rebalance? Explain how the strategy aligns with your investment thesis. In coming up with your
   strategy, spend some time researching data availability. Will you be able to assess historical
   performance of your strategy given the data you have access to? If not, revise your strategy.
2. Describe your strategy in detail: what determines which assets you plan to invest in? How often do you
   plan to rebalance? What will guide your rebalancing i.e. what will determine how you change portfolio
   weights?
   Collect historical data for assets in your strategy. How far back you go depends is up to you — consider
   data availability, relevance of past episodes to your expectations of future performance, etc. If your
   allocation rules (rules for determining portfolio weights) depend on some piece of information other than
   returns, collect that historical information too. Discuss the overall performance of assets in your strategy,
   perhaps by looking at some time-series plots and summary statistics. What did you learn? Now that you
   know the data better, revise your strategy. Explain your revisions.
3. Backtest your strategy: evaluate how your strategy would have performed in the past using historical
   data. Collect past returns on your assets. Make sure you don’t “cheat” i.e. form your portfolios based on
   information that was not available to investors at the time. For example, if your idea is to go long stocks
   with high profits and short stocks with low profits, make sure you don’t use end-of-year profits to choose
   portfolio weights at the start of the year. Investors don’t have a crystall ball!
   BU.232.610 – Computational Finance – Stuart Urban – Page 2 of 2
   Consider how well your portfolio does on a raw and risk-adjusted basis? Useful risk-adjusted
   performance measures are Sharpe Ratio, factor model alpha (CAPM, or some other model for example
   the Fama-French 3-factor model), and the Traynor’s ratio, which divides the alpha by the standard
   deviation of model residuals and thus captures how “risky” the alpha is.
   Discuss your results. If you want to adjust your investment strategy to improve its historical
   performance, you can do it. But, as long as your strategy is reasonable given the preliminary analysis
   and well-motivated, we won’t grade you on investment performance.
4. No matter how well your portfolio has done in the past, it’s possible that it will do poorly in the future.
   How much of a loss would you be willing to take before giving up on your strategy and cashing out?
   Minus 10%? Minus 20%? Pick your loss limit.
   Using a simulation, determine how likely you are to reach your loss limit if you keep investing into your
   portfolio strategy for the next year? Assuming that, if you do, you pull your money out of the market,
   what is your expected return over the next year? How does it depend on your expectations for the return
   on the strategy itself? Compare returns with and without the stop loss limit. Do you think it’s a valuable
   restriction to have?