Showing posts with label portfolio management. Show all posts
Showing posts with label portfolio management. Show all posts

Sunday, September 29, 2024

Book review: beyond diversification - what every investor needs to know about asset allocation

I recently finished reading Beyond diversification from Sebastien Page.  Sebastien Page (Chief Investment Officer at T. Rowe Price) explains in Beyond diversification the different approaches to forecasting returns, risks and correlations across asset classes by combining academic research and practical hands-on examples. 

This book is most likely targeted at the sophisticated investor  and should not be the first book to pick up if you want to understand diversification but it provides great insights on how asset managers think about their portfolios. 

The book also extensively refers to a number of academic papers that Sebastien Page has written on asset allocation, risk measurement and return forecasting. It explores a number of dynamic asset allocation strategies, acknowledging that risk is time-varying and requires adaptive approaches. Sebastien Page also explains why the typical fixed weight asset allocation (60-40 portfolio) does not deliver a constant risk exposure.

The book recommends using a  heuristics to cope with changes in volatility: assume that next month's volatility for each asset class will be the same as last month's. For the longer term the opposite is true, 5 years of calm markets are more likely to be followed by 5 years of turbulence.

The stock-bond mix is the biggest decision that multi-asset investeros make, but this mix does not reliably reduce risk as correlations shift. Stock bond correlations were positive in the 1970s and 1980s when inflation and interest rates drove volality but then the correlation reversed.


PS With volatility being a proxy for risk, you might also check out the white paper "Practical Issues in Forecasting volatility" from Clive Granger and Ser Huang Poon which was published in the Financial Analyst Journal in February 2005.





Wednesday, November 16, 2022

Visualize S&P 500 data in Power BI using Azure Synapse Serverless SQL Pool

In Explore and analyze stock ticker data in Azure data lake with Azure Synapse serverless SQL Pool, I showed you can download stock ticker data from Yahoo Finance, stored it in Azure Data Lake and retrieve the data using standard T-SQL in Azure Synapse Studio. In this post, I will show how easy it is to consume the data from Synapse SQL Serverless using Power BI.


For the standard visual with the evolution of the S&P 500 closing price, I connected directly on SP500 external table in the Synapse SQL. You can connect to Synapse SQL Serverless using either the Azure SQL Database or Azure Synapse Analytics SQL connector and you will need to enter the Serverless SQL endpoint which looks something like this <yoursynapse>-ondemand.sql.azuresynapse.net


With the second reported I want to visualize the S&P 500 yearly return and the average return since December 1927. To make it easier, I created a separate view on top of the external table which calculates the yearly returns


As you see from the visual, returns can vary quite a lot both on the negative side as well as on the positive side - for the last 20 years, there was a huge drop in 2008 (-38%) and also this year is not looking great (-22%), but 2013, 2019 and 2021 all had returns above 20%. On average across the S&P 500 returned 7% (not included dividends).


For the last visual in the Power BI report, I wanted to show a histogram with the S&P 500 yearly returns. I based myself on Power BI Histogram example using DAX since  Power BI does not have a standard histogram and I did not want to use a custom visual ( I used Power BI custom visuals from Pragmatic Works in the past)

Equity returns roughly follow a normal distribution or "bell curve", meaning that most values cluster near the central peak and values farther from the average are less common.  Stock returns however have fat tails - meaning that the occurrences on the extremes are far more common than expected in a normal distribution.  The Greate Depression (1931) and the Global Financial Crisis (2008) led to two of the largest stock market losses of the S&P 500. With a loss between -20% and -30% this year, we are in the same category/bin as 1930, 1974 and 2002.