Active Research
Active research: Portfolio Computational Methods
5thN’s emerging research program centers on computational portfolio construction, Expanded Portfolio Theory, and systematic backtesting with the HiFuPi asset model.
Research program
Expanded Portfolio Theory and HiFuPi
Expanded Portfolio Theory is positioned as a practical extension of classic Modern Portfolio Theory: it keeps the useful computational discipline of MPT while widening the portfolio-construction problem so the model can accommodate richer constraints, nonlinear positions, alternative asset models, and more realistic backtesting workflows.
Expanded Portfolio Theory
EPT is the allocation engine: a broader portfolio-computation framework intended to handle cases where the classic MPT assumptions become too restrictive for real advisory workflows.
HiFuPi asset model
HiFuPi is the asset-model layer used to generate the forward-looking or backtested inputs that feed the EPT allocation engine.
MPT assumptions and EPT flexibility
Modern Portfolio Theory is powerful, but it also carries an embedded set of assumptions about how the portfolio problem is framed. Those assumptions shape the optimization problem, the allowable inputs, the form of risk, and the way the final allocation is selected.
Expanded Portfolio Theory starts from a broader position. When the classic MPT assumptions are intentionally imposed on an EPT model, the EPT solution collapses back to the same result that MPT would produce. In that sense, MPT can be understood as a special case within the broader EPT framework.
The difference is that EPT does not require those assumptions to be embedded at the foundation of the model. Because the assumptions are not fixed in advance, EPT can be adapted to solve portfolio problems across a much wider range of real-world situations: alternative constraints, different asset models, turnover limits, implementation rules, nonlinear positions, tax or operational considerations, and portfolio objectives that do not fit neatly inside the classic MPT structure.
Research implication
The practical value of EPT is not that it rejects MPT. Rather, EPT preserves MPT as one possible case while opening the door to a broader computational language for portfolio design, testing, and implementation.
Current backtesting focus
The current work is best presented as a computational research pipeline:
1. Define the investable universe. Build a security universe and collect price, return, liquidity, and eligibility data.
2. Generate asset-model inputs. Use the HiFuPi model to create the return/risk estimates or signals required for allocation.
3. Apply EPT constraints. Run allocation under realistic constraints such as holdings, turnover, security bounds, and operational rules.
4. Backtest the workflow. Rebalance through time and compare allocations, returns, drawdowns, turnover, and implementation effects.
5. Produce advisor-ready output. Turn the computation into reports, graphics, audit trails, and client-facing explanations.
Message for advisors
The research story is not simply “better optimization.” The stronger message is: advisor-grade portfolio computation needs a pipeline, not a single formula. EPT and HiFuPi together create a framework where asset modeling and allocation can be separated, tested, constrained, and explained.