Portfolio Risk Estimation
Typically, risks associated with stocks are classified as either systematic or unsystematic, and these categories are commonly measured using descriptive statistical techniques that rely on standard deviation. Standard deviation, being a second statistical moment, is utilized as an indicator of the returns' volatility.
However, at Tensor Bridge Invest, we utilize inferential statistics and sequential machine learning techniques to enhance the quantification of risks. By leveraging a broader range of information that influences changes in asset prices, we aim to provide a more accurate and comprehensive assessment of the risks associated with investment strategies.
Portfolio risk refers to the probability that the value of a portfolio will decrease over a particular time frame as a result of market volatility, economic conditions, or risks specific to the individual securities held within the portfolio.
When new information about a company becomes available, such as a negative quarterly report, investors will adjust their perception of the company's value, and the stock price will change accordingly. The stock market is said to convert information into stock prices. This conversion can be complex and difficult to model due to the diverse nature of the information and the dynamics of its effect on stock prices.
Most stocks are highly liquid and traded at high frequency, but new business information cannot arrive with the same frequency. Different stock exchanges have distinct signatures characterizing their ability to convert information into prices. This process is analytically complex when modeled with many agents.
Risks in the stock market are traditionally classified as systematic or unsystematic. Systematic risks affect the overall stock market or economy, while unsystematic risks affect individual companies. Descriptive statistics mainly based on standard deviation have been used to characterize risks. Stocks with a larger standard deviations of returns are considered to pose more risk than stocks with a smaller standard deviation.
Clearly there is no financial or business reason to assume that a stock with certain history of its standard deviation will continue to have the same history in the future – such a property for a time series is called as a stationary process. Still, under stationarity assumptions, a portfolio can be adjusted for a given unsystematic risks by an optimal diversification based on the stocks’ standard deviation. Systematic risks cannot be diversified. Due to this optimization, the concept of standard deviation as risk measure became deep-rooted in time.
Compared to basic descriptive statistics, modern inferential statistics and sequential machine learning represent a significant advancement, as they enable a more accurate characterization of risks that are closely connected to the information generating price fluctuations
At Tensor Bridge Invest, we utilize inferential statistics and sequential machine learning techniques to enhance the quantification of risks. By leveraging a broader range of information that influences changes in asset prices, we aim to provide a more accurate and comprehensive assessment of the risks associated with investment strategies.
