Value and timing of executed_productivity_investment & planned_productivity_investment not aligned

[sternluke]

The value and timing of self.ts.current("executed_productivity_investment") does not seem to be correlated with the value and timing of self.ts.current("planned_productivity_investment"). For example, I see evidence of executed_productivity_investment when none was planned and where executed_productivity_investment that exceeds the value of planned_productivity_investment.

[jose-moran]

Automated Code Analysis (LLM-assisted)

This analysis was generated using automated code review with LLM assistance.

Root Cause Identified

The misalignment between planned_productivity_investment and executed_productivity_investment is due to fundamentally different calculation methods:

Planned Investment (in plan_productivity_investment(), firms.py:542-600)

• Forward-looking, based on available cash
• Subject to hurdle rate profitability test
• Formula: budget = min(κ * cash, θ * output), then candidate = propensity * budget

Executed Investment (in compute_productivity_investment(), firms.py:2075-2107)

• Backward-looking, based on actual purchases
• Formula: executed = total_capital_bought_costs - replacement_costs
• Where replacement_costs = production * depreciation_matrix * prices

Why They Don't Correlate

Scenario	Planned	Executed	Explanation
High cash, low purchases	High	Low	Firm planned to invest but couldn't buy enough goods
Low cash, high purchases	Low	High	Firm bought capital for other reasons (not tracked as planned)
Any	X	Y	Executed subtracts depreciation replacement, planned doesn't

Proposed Fix Options

1. Track planned through execution: Store planned investment and deduct it from deposits when executed
2. Use planned consistently: Use planned_productivity_investment in TFP growth calculation instead of executed
3. Reconcile definitions: Make executed = min(planned, actual_available) to ensure correlation

Code Locations

• Planned: macromodel/agents/firms/firms.py:531-540
• Executed: macromodel/agents/firms/firms.py:2075-2119
• TFP growth uses executed: macromodel/agents/firms/firms.py:2136-2140

[sternluke]

Investments in TFP now align with planned productivity investments.

Delta planned productivity investments is correlated with the changes in slope of delta TFP multiplier. Let me know if that was your intention?

I'm not sure that it matters at all given that executed productivity investments are calculated using a different logic but they still don't align with planned productivity investments.

Recommendation: If it is accepted that these two variables have no logical connection then their names probably should be changed to reinforce this fundamental difference and reduced future confusion.

[sternluke]

Here is a more intuitive way to visualize the new correlation. Here is the tfp_multiplier and (planned productivity investment / nominal production) *100 plotted side by side.