Venture capital serves as critical infrastructure in our innovation ecosystem, directing resources toward future technologies and business models. It decides which ideas get built, which startups grow. The visions of the future held by the social VC ecosystem, and the capital they deploy has a strong say in defining what kind of future we create.
But this innovation engine has a flaw, it treats "not harming society" as optional rather than a strict requirement, and it’s often a hidden assumption that social impact can come at the cost of returns. The “move fast and break things” motto has become generally accepted beyond the realms of software code, and applies to the public plaza. Regulating startups too heavily is seen as an attack to innovation. Negative impacts on the public is deemed an acceptable cost to remain technologically innovative and competitive.
For example, Uber was promised to cut car ownership, reduce pollution and congestion. In reality, the opposite happened: More congestion, more pollution. Obviously, concerns about precarious employment practices have also been a major theme, among other ethical dilemmas.
In urban planning when a street corner sees too many accidents, fining individual drivers will not solve the problem that’s baked in the infrastructure. We need to redesign the intersection. In an ideal world, we don't accept "but faster intersections make transportation more efficient" as an excuse for dangerous design. Safety shouldn’t be optional for transport, it needs to be a critical factor of the design intention. And crucially, well-designed intersections can be both safe AND efficient.
The same should be true for startups. Not harming society should be the baseline, not a "nice to have." Having highly speculative tech-driven futures is a high variance game, designing safe infrastructure (in other words, safe guardrails) should be something we take seriously.
In practice this is often portrayed as disruptive innovators playing a cat and mouse game with regulators. My question here is: Instead of playing an adversarial game, can we not align VC interests with social good?
Mathematical misalignment
The way success is calculated for VCs is currently optimise to for a single variable.
Where:
- θ represents a startup/investment decision in the set of all possible startups Θ
- R(θ) is the expected financial return from investing in startup θ
- E[R(θ)] is the expected return across the portfolio
Omitting societal impacts could be described in mathematical terms as an incomplete optimisation function. This model encourages VCs not only to fund companies with high expected returns regardless of negative externalities, but also to actively pressure portfolio companies toward decisions that maximise financial returns even when those choices are directly at odds with societal well-being. These short-term incentives can be dangerously misaligned with both long-term societal value and the company’s own sustainable success
Systemic consequences
The VC search algorithm prioritises "fund returners" (P(Return > 20x | Features)) by evaluating market size, team quality, and growth potential, but does not account for externalities created along the way.
Despite good intentions, VCs operate within a system that structurally disincentivizes positive impact. This is a serious a misalignment between the purpose of innovation infrastructure (advancing society) and its incentives (maximising financial returns).
Why Not Just Tax Bad Companies?
Traditional Pigouvian taxes try to fix problems after they happen. This reactive approach means damage is already done before any action is taken. They also typically focus only on penalising negatives rather than rewarding positives.
Taxation affects the outputs (companies) but not the infrastructure that selects which innovations receive capital in the first place.
Can we make it easier for companies doing good things to get easier access to capital? Can we make it more difficult for companies causing problems to find it harder to get funding?
Optimisation Function Redesign
By modifying the fundamental VC optimisation function from max E[R(θ)] to max E[R(θ) + Π(θ)], where Π(θ) represents the financial value of a company's social impact, we can realign this search problem.
An impact evaluation framework could incorporate domain-specific metrics for major verticals such as climate, health, labor, and data rights. It would use weighted scoring based on relative improvement from baseline, distance from ideal state, trajectory/innovation potential, and scale of impact (users/markets affected).
The calculation method for this Net Externality Score (NES) would aggregate portfolio impact using the formula:
where wi represents a weighting factor based on stake, duration, and scale, while NES(θi) represents the net externality score of each company i in the portfolio. This accounts for both a VC's level of involvement and the magnitude of impact.
The redistribution formula operates on a conservation principle - the total financial incentives within the system remain constant, but their distribution shifts toward funds that generate the greatest social utility per dollar invested.
To implement this, VCs with negative NES would contribute to a redistribution pool with a percentage based on their impact score, while VCs with positive NES receive dividends from this pool. This maintains market-based incentives within a competitive framework, rewarding positive impact financially. The redistribution pool creates a critical feedback loop in the innovation system, where negative externalities (outflows of societal value) trigger financial outflows from responsible funds, while positive externalities (inflows of societal value) generate financial inflows to those funds.
Tackling the Measurement Problem: How the System Helps
How can we actually measure societal impact accurately and consistently? It’s hard to put a number on "doing good," and creating standards that work everywhere without being easy to fool is a major challenge.
An “impact pool” system creates clear financial winners – the VCs getting dividends because their companies have a high positive impact. This gives these winners a direct interest in making sure the impact measurement system is trustworthy and solid.
Instead of needing perfect measurement from the start, we’d want to create a process where measuring impact gets better step-by-step, driven by the VCs who gain the most from a system that truly recognises and rewards positive results. Improving impact measurement becomes part of how VCs compete and show leadership.
APPENDIX
Impact-Adjusted Optimisation
New Objective :
Where Π(θ) represents the impact-adjusted return component. This term is defined as:
Where:
- C(θ) quantifies the net social impact of investment θ (positive or negative)
- λ is a penalty parameter for negative externalities
- γ is a reward multiplier for positive externalities
- max(0, -C(θ)) isolates negative impact for penalty
- max(0, C(θ)) isolates positive impact for rewards
This formula creates a two-sided incentive structure:
1. If C(θ) < 0 (harmful impact), the VC incurs a financial penalty proportional to harm
2. If C(θ) > 0 (beneficial impact), the VC receives a financial reward proportional to benefit
Crucially, this changes the feasible investment space from Θ to a filtered subset:
VCs now rationally exclude investments with excessive harm, even if financially attractive, because the impact penalty would reduce expected returns below their threshold.
Bayesian Decision Framework
VCs operate under uncertainty, using limited information to predict future outcomes. In Bayesian terms, the current VC approach estimates:
Where:
- P(Success|D) is the posterior probability of success given data D
- P(D|Success) is the likelihood of observing data D given success
- P(Success) is the prior probability of success
Currently, the prior P(Success) is shaped primarily by financial indicators. With the new framework, this prior is expanded to include impact considerations:
Most importantly, the utility function that VCs maximise changes to an impact-inclusive version:
This means VCs must update their beliefs not just about which companies will deliver financial returns, but also about the impact profile of their investments. Two startups with similar IRR profiles but different impact profiles will have different expected utilities under this framework.