AI-Optimized Asset Allocation for GDP Scenarios

By Big Data Federation, Inc.

Recession forecasts have been pushed out for the better part of two years. Economic growth has continually defied expectations amidst a resilient job market so far. Economic growth is one key factor in asset allocation; however, how to optimally allocate assets based on various unfolding scenarios for a spectrum of investors’ goals has been an age-old complex problem.
 

Leveraging our big data and machine learning models, we hypothesize a range of scenarios ahead of the advance estimate released by the BEA for the second calendar quarter of the year. Our sample range includes the expectations of Atlanta Fed’s GDPNow as well as other economic consensus. As displayed in the table below, our multi-asset portfolio includes bonds, S&P 500, and commodities. Each scenario produces three bespoke portfolios, optimized for either the Sharpe Ratio, return, or risk.
 

2Q23 GDP Growth*
 

# of Occurrences

(1970-01-30 and 2023-07-25)

AI-Optimized Asset Allocation

Optimized for

Bonds1

S&P 500

Commodities2

+3%

15 times

Sharpe Ratio

58%

37%

5%

Return

2%

98%

0%

Risk

66%

18%

16%

+1.5%

17 times

Sharpe Ratio

54%

29%

17%

Return

23%

77%

0%

Risk

52%

24%

24%

0%

15 times

Sharpe Ratio

98%

0%

2%

Return

100%

0%

0%

Risk

63%

13%

24%

-1.5%

14 times

Sharpe Ratio

49%

0%

51%

Return

49%

0%

51%

Risk

57%

23%

20%

-3%

14 times

Sharpe Ratio

100%

0%

0%

Return

100%

0%

0%

Risk

59%

23%

18%

*Measured as annualized QoQ change in real GDP
Source: Big Data Federation, Inc.
 

Initially, our models complete an extensive search throughout history, dating as far back as the 1920s if the data is available, to identify periods that match the defined scenario. Next, our machines capture monthly returns for all portfolio components over the identified period that match the defined scenario. Finally, our algorithms solve a nonlinear optimization problem, considering the performance of individual components and their correlations while adhering to given constraints.

1  Ibbotson® SBBI®: Long-Term Govt Bonds (Total Return)


2  Commodities as defined by Bloomberg Commodity Index


 

About OdinUltra

OdinUltra is a cutting-edge no-code specialized AI platform, designed to cater to a wide array of professionals in the asset management industry. It is a battle-tested specialized partner that empowers users including discretionary, quantitative, multi-asset, and multi-strategy managers to harness the immense potential of big data and sophisticated investment techniques. From idea exploration and construction to verification and deployment, OdinUltra is an all-encompassing tool. Asset managers can enhance existing products and rapidly develop new ones, securing a competitive edge in the dynamic world of asset management.

 

Published by Big Data Federation · All insights · Subscribe to Insights

Financial AI ®