With the new finportfolio command, you can easily create a portfolio of assets and examine this portfolio's properties. Optimize your portfolio to minimize variance or maximize the Sharpe ratio, also known as the return-to-risk ratio. This feature is a part of StataNow™.

A conventional investment strategy combines different assets to create a portfolio. Portfolio construction depends on the type of investment strategy you want to pursue. The new finportfolio command allows you to construct portfolios using a weighted average of assets. The weights may all be equal or selected optimally to minimize variance or maximize the Sharpe ratio. Weights may also be chosen ex ante and provided to finportfolio. For the minimum-variance weighted portfolio, you may even target a specific value for the returns.

Let's see it work

Below, we use data on the asset prices of 25 fictional companies. Our data consist of 780 monthly observations containing the end-of-month values for these asset prices.

. webuse finex
(Fictional stock price data)

We tell Stata that we have time-series data by using tsset on the date variable, datem:

. tsset datem

Time variable: datem, 1955m1 to 2019m12
        Delta: 1 month

Next we construct a set of simple returns based on five of the asset prices in our data by typing

. finreturns acme-tyr, simple(r_) multiply(100) nopreview
(all returns multiplied by 100)

Simple returns generated for variables acme, bat, iron, dune, and tyr.

We have created five simple returns (percentage changes in asset prices). We used the simple() option to specify that the returns have a name that starts with r_ (for example r_acme) instead of the generic name that the command provides. Also, we multiply the returns by 100 for interpretation as percentage changes instead of proportional changes.

We now construct a portfolio with returns r_acme through r_tyr. We first assign each portfolio equal weights.

. finportfolio equal r_acme-r_tyr

Equally weighted portfolio

Number of obs = 779
Sample: 1955m2 thru 2019m12


             
 
    Weight
 
 
 
        ACME 
 
        .2
         BAT 
 
        .2
        IRON 
 
        .2
        DUNE 
 
        .2
         TYR 
 
        .2

Portfolio return     =  0.8281
Portfolio std. dev.  =  4.5342
Risk-free rate       =  0.0000
Sharpe ratio         =  0.1826

We get a monthly return of 0.83% with a standard deviation of 4.53%. The Sharpe ratio is the quotient of these two numbers. It tells us the return per unit of risk. Ideally, we would like to minimize variance and maximize return. Of course, we can do one or the other or a ratio of both. Below, we explore these possibilities. Let's see what would happen if we wanted to minimize variance.

. finportfolio minvariance r_acme-r_tyr

Global minimum variance portfolio
Short selling allowed

Number of obs = 779
Sample: 1955m2 thru 2019m12


             
 
    Weight
 
 
 
        ACME 
 
  .8985511
         BAT 
 
  .0918437
        IRON 
 
 -.1387537
        DUNE 
 
  -.130712
         TYR 
 
  .2790709

Portfolio return     =  0.4050
Portfolio std. dev.  =  1.0869
Risk-free rate       =  0.0000
Sharpe ratio         =  0.3726

The return and the standard deviation have decreased, and we now have a higher Sharpe ratio. Another important consideration is that two of the weights in our portfolio are negative. This means that we are short selling these assets. In other words, we sell the assets today and agree to purchase them back later. We are betting that the asset's price is going to be lower in the future. If we do not want to pursue such a strategy, we could use the noshort option and recompute the optimal portfolio forcing all weights to be positive.

We now explore the possibility of maximizing the Sharpe ratio.

. finportfolio maxsharpe r_acme-r_tyr

Maximum Sharpe ratio portfolio
Short selling allowed

Number of obs = 779
Sample: 1955m2 thru 2019m12


             
 
    Weight
 
 
 
        ACME 
 
  .9157845
         BAT 
 
  .1024231
        IRON 
 
 -.1234373
        DUNE 
 
 -.0289488
         TYR 
 
  .1341785

Portfolio return     =  0.4439
Portfolio std. dev.  =  1.1380
Risk-free rate       =  0.0000
Sharpe ratio         =  0.3901

Returns are higher than when we minimized variance, and by construction, the Sharpe ratio is higher.

We could look for a target return level while simultaneously minimizing variance. Maybe we are willing to allow a bit more risk to attain a particular monthly return, say, 0.5%. We use the target() option and type

. finportfolio minvariance r_acme-r_tyr, target(.5)

Minimum variance portfolio with fixed return
Short selling allowed

Number of obs = 779
Sample: 1955m2 thru 2019m12


             
 
    Weight
 
 
 
        ACME 
 
  .9405779
         BAT 
 
  .1176436
        IRON 
 
 -.1014019
        DUNE 
 
  .1174566
         TYR 
 
 -.0742762

Portfolio return     =  0.5000 (fixed return)
Portfolio std. dev.  =  1.3629
Risk-free rate       =  0.0000
Sharpe ratio         =  0.3669

We increased the variance and reduced the Sharpe ratio, as expected. We did this mainly by increasing the fraction of ACME that we hold in our portfolio and shifting our short-selling strategy.

So far, we have just looked at the portfolio weights that arise from an optimization problem or from using fixed weights. We can also specify the generate() option to create variables that contain each of our portfolios.

. finportfolio equal        r_acme-r_tyr, generate(equal)
. finportfolio minvariance  r_acme-r_tyr, generate(minvar)
. finportfolio maxsharpe    r_acme-r_tyr, generate(maxsharpe)
. finportfolio minvariance  r_acme-r_tyr, target(.5) generate(minvar_noshort)

Now we can use the portfolio variables for further analyses. Perhaps we want to see how they behave relative to a benchmark market and adjust for a particular rate of return; we can use finsummarize and finregress. Maybe we want to see how risky the portfolios are in terms of extreme events; we can compute their values at risk with finvalrisk. Or we may want to model the portfolios using Stata's time-series tools. See Financial statistics for an overview of these and other commands for analyzing financial data.