Why Post-Soviet Financial Systems Cannot Produce Their Own Disruption
By Heorhi Tratsiak
The meeting lasted ninety minutes, which was forty-five minutes longer than it needed to be.
There were four of us in the room: myself, my co-founder, and two representatives from the retail banking division of one of the region’s largest state-aligned financial institutions. We had come to present CreditMatch — a credit decisioning product built specifically for the underserved lending market that constituted the majority of this institution’s own client base. We had data. We had a working prototype. We had six months of informal validation from potential clients who had told us, in direct language, that what we were offering solved a problem they had been trying to solve internally for years.
The meeting was excellent. The two representatives were engaged, intelligent, and genuinely interested. They asked the right questions — about model architecture, about data privacy, about implementation timelines, about the regulatory implications of algorithmic decisioning. We answered all of them. By the end of the ninety minutes, there was a warm handshake, a clear statement of intent to move forward, and a commitment to a follow-up meeting with the technology committee within the following month.
The follow-up meeting was scheduled twice and postponed twice. On the third attempt to schedule it, the contact at the institution replied that the relevant internal stakeholders were unavailable due to a strategic planning cycle and that they would be in touch when the timing was more suitable.
They were never in touch again.
CreditMatch did not die because anyone decided to kill it. It died in the space between a warm handshake and a meeting that never happened, sustained by nothing more substantial than the institutional equivalent of an unanswered phone.
I did not understand this at the time. I understood it only later, when I had watched the same pattern complete itself twice more with different products in different institutions. When a pattern repeats three times with sufficient precision, it stops being a series of unfortunate events and becomes evidence of a mechanism.
This chapter is about that mechanism. I have come to call it the institutional veto. It is, I believe, the single most important structural force that prevents post-Soviet financial markets from producing their own disruption. And it is, by design or by accident — I am genuinely uncertain which — one of the reasons that the countries most exposed to the weaponization of the global financial architecture I have described in the preceding chapters are also the countries least capable of building anything to replace it.
The Problem That Should Have Produced a Solution
The conventional theory of financial disruption runs something like this: when an incumbent financial system fails to serve a significant portion of its addressable market, a gap opens. That gap creates an incentive for innovators to enter. The innovators, unburdened by the incumbent’s cost structure and institutional inertia, can serve the gap more efficiently. The disruption propagates upmarket. The incumbent eventually responds or declines.
This theory, broadly derived from Clayton Christensen’s framework for disruptive innovation, has been the operating model for the global fintech industry since approximately 2010. It describes, with reasonable accuracy, what happened in the United Kingdom, in the United States, in Scandinavia, in parts of Southeast Asia. Monzo, Revolut, Stripe, N26 — these companies exist because the gap existed, the incentive was real, and the infrastructure on which they built was not controlled by the institutions they were disrupting.
In post-Soviet markets, the theory makes the same prediction. The gap is larger, in fact, than almost anywhere else. Credit penetration rates are low. The quality of credit decisioning at incumbent institutions is poor, relying on methods that systematically exclude creditworthy borrowers while accepting unacceptable risks elsewhere. Fee structures are opaque. Small and medium enterprises — the economic backbone of every functioning economy — are chronically underserved. The population, in many cases, carries a deep historical skepticism toward formal financial institutions that has created demand for alternatives that incumbent banks are structurally unable to provide.
The gap is there. The incentive is real. The demand is documented. By the conventional theory, the disruption should have happened.
It did not happen. Or rather: it happened partially, in limited ways, in narrow segments, before running into something that the conventional theory does not account for. Something that stops disruption not by competing with it, not by prohibiting it, but by making it permanently, invisibly, fatally slow.
The Institutional Veto: A Definition
The institutional veto is not a policy. It is not a regulation. It is not a decision made by any identifiable person on any identifiable day. This is what makes it difficult to describe and almost impossible to challenge.
Let me define it precisely, because the precision is the point.
The institutional veto is the mechanism by which financial systems dominated by state-aligned incumbents neutralize disruptive innovation through distributed procedural delay — not by prohibiting the innovation, but by deferring engagement with it indefinitely through a sequence of legitimately accountable steps, each of which is individually reasonable, collectively terminal, and attributable to no single actor.
The word veto is deliberate. In constitutional theory, a veto is the power to prevent an action. What makes the institutional veto different from an ordinary veto is that an ordinary veto is exercised by an identifiable agent making an identifiable decision. The institutional veto is exercised by no one. It is the cumulative output of a system in which every participant has rational incentives not to make any decision at all.
Consider the incentive structure of a mid-level manager at a large state-aligned bank presented with a fintech product that genuinely solves a problem. If she champions the product internally and the implementation succeeds, the credit for the success is distributed across the organization. If she champions the product and the implementation fails, the responsibility is concentrated on the person who championed it. If she expresses interest without committing, the meeting is productive, the relationship is maintained, and nothing bad happens to her.
The rational choice, in a system without appropriate incentive structures for innovation adoption, is always the third option. Not rejection. Not approval. Expressed interest, managed ambiguity, and strategic deferral.
Multiply this individual calculation by the number of people in the decision chain — and in a large post-Soviet financial institution, the decision chain for adopting a third-party technology solution can involve seven to twelve sign-offs — and you have a system that is, in aggregate, perfectly designed to produce the appearance of openness while ensuring that nothing changes.
Nobody lied to me in that ninety-minute meeting. The interest was genuine. The questions were good. The people were competent. The mechanism that prevented the outcome had nothing to do with the individuals in the room and everything to do with the institutional environment in which those individuals operated.
CreditMatch: A Case Study in Distributed Rejection
Let me describe what the institutional veto looks like in practice through the specific history of CreditMatch, because the operational detail is more instructive than any abstraction.
CreditMatch was built to solve one of the most clearly documented problems in post-Soviet retail banking: the gap between borrowers who deserve credit and borrowers who receive it. The problem has two components. The first is technical: traditional credit scoring in the region relied heavily on formal credit history, which a large portion of the population did not have, not because they were poor credit risks, but because they had not previously had access to formal credit products. The second is structural: loan officers at incumbent institutions were evaluated on default rates, which created an incentive for conservatism that excluded creditworthy borrowers at the margin.
Our product addressed both. It used alternative data — utility payments, phone contracts, rental history, employment patterns — to build credit assessments for borrowers who were invisible to traditional scoring models. The approach was not novel in a global context; similar models were being deployed successfully in markets from Kenya to Indonesia to Brazil. What was novel was the calibration to the specific informational environment of the post-Soviet region and the integration with the data sources that were actually available and accessible in that context.
The need was real. The technology worked. The regulatory framework, at the time, did not prohibit what we were doing. We had identified at least four large financial institutions that, by their own public statements and internal strategy documents available to us through professional channels, had identified credit model modernization as a strategic priority.
The first institution: the meeting described at the opening of this chapter. The follow-up that was postponed. The silence.
The second institution: we reached the technology committee. Three months of technical due diligence produced a detailed assessment that concluded, accurately, that our model outperformed their existing approach on the metrics they had specified. The assessment was submitted to the executive committee. The executive committee requested additional analysis on regulatory compliance. We provided it. The executive committee requested a pilot proposal. We provided it. The pilot proposal was submitted to the risk committee. The risk committee noted that the pilot would require changes to their data architecture and referred it back to the technology committee for infrastructure assessment. The technology committee was engaged in a core banking migration project and could not allocate resources to infrastructure assessment for the following eight to twelve months.
We closed the engagement at month seven. The institution’s credit model modernization initiative, according to their subsequent public disclosures, remained a strategic priority.
The third institution was the most instructive, because it introduced a variation I had not previously encountered. Here, the engagement progressed further than any previous one. We reached a signed memorandum of understanding — not a contract, not a commitment to purchase, but a formal declaration of intent to explore a partnership. The institution’s project manager was engaged, competent, and genuinely invested in making progress. We had weekly calls for three months.
Then the project manager was reassigned. His replacement had no context for what had been agreed, no particular interest in the initiative, and no instructions from her superiors to treat it as a priority. We re-explained the entire rationale. She expressed interest. She noted that the memorandum of understanding would need to be reviewed by the legal department given the personnel change. Legal review took six weeks. Legal identified several clauses that needed modification. Modification required both parties’ agreement. We agreed to the modifications. The modified agreement required re-approval by the original signatories. One of the original signatories had, by this point, moved to a different role. Their replacement needed to be briefed. The briefing was scheduled. It was postponed.
We ran out of runway at month eleven.
What I want to draw attention to, in each of these three histories, is what did not happen. In none of them did anyone say no. In none of them did anyone identify a specific reason the product should not be adopted. In none of them did anyone make a decision that could be contested, appealed, or used as a basis for a different approach. The veto was exercised through the accumulation of legitimately accountable steps, each of which produced a small delay, and the small delays accumulated into an outcome that was functionally identical to a prohibition — without any of the clarity, accountability, or contestability that a prohibition would have provided.
The Three Phases of the Institutional Veto
Through three products and multiple engagement processes, I came to recognise that the institutional veto operates in consistent phases. Understanding these phases is practically useful for anyone attempting to innovate within these systems, but it also illuminates the mechanism at a level of detail that the abstract definition alone cannot provide.
Phase One: Performative Interest. The initial engagement is characterized by genuine-seeming enthusiasm. Meetings are substantive. Questions are good. The representatives of the institution communicate, accurately, that the problem your product addresses is real and that their institution is interested in solutions. This phase is not dishonest. The interest is, in many cases, genuinely felt at the level of the individuals in the room. What those individuals cannot communicate, because they often do not know it themselves, is that their felt interest will encounter institutional resistance the moment it attempts to translate into action.
The critical diagnostic signal in Phase One is the absence of a named decision-maker with authority and incentive to approve. Meetings with people who are interested but not empowered are, regardless of their quality, a form of productive delay.
Phase Two: The Process Labyrinth. Once the initial interest has been expressed, the engagement moves into a phase characterized by the proliferation of legitimate process. Each step is reasonable. Each request for additional documentation, additional analysis, additional committee sign-off, additional stakeholder alignment is defensible as due diligence. The cumulative effect is to move the decision horizon perpetually forward without any party needing to take responsibility for its perpetual receding.
The distinguishing feature of Phase Two is that the process requirements grow rather than diminish as the engagement progresses. In a functional procurement process, each completed step should reduce the remaining uncertainty and move the decision closer. In the institutional veto, each completed step reveals new requirements that were not previously disclosed. The process is not leading to a decision. It is substituting for one.
Phase Three: Distributed Extinction. The engagement does not end. It attenuates. Personnel changes, competing priorities, budget cycles, strategic planning periods, regulatory reviews — the reasons for delay are always legitimate and always slightly different, which prevents the accumulation of a pattern that might trigger escalation. The startup runs out of money, or time, or the belief that continued engagement will produce a different outcome. The institution’s representatives, asked about the initiative a year later, would describe it as something that was being explored but had not progressed due to timing or resource constraints.
Nobody made a decision. Nobody is accountable. The outcome — the non-adoption of a product that worked, in a market that needed it — exists without an author.
Why Silicon Valley Logic Fails Here
The dominant mental model of financial technology innovation, developed primarily in American and British contexts, contains several assumptions that are structurally invalid in the post-Soviet environment.
The first assumption is that the regulatory environment is the primary obstacle to fintech adoption. In the Western context, this has often been true. Banking regulations in the United States and Europe were designed for an earlier era of financial services and did not, initially, contemplate the compliance requirements of technology-first financial products. The fintech industry spent considerable energy navigating, and in some cases reshaping, the regulatory framework.
In post-Soviet markets, regulation is real but it is not usually the primary obstacle. The primary obstacle is the commercial relationship between the innovation and the incumbents. And the incumbents, in these markets, do not primarily compete with fintech products. They control the infrastructure on which fintech products must be built.
This is the second failed assumption: that the innovation can be built independently of the incumbents. In the American context, Stripe built a payments infrastructure that was genuinely independent of the major banks. The banks’ merchant acquiring business was disrupted because Stripe did not need their permission to operate. In the post-Soviet context, the payment rails, the banking APIs, the account infrastructure, the settlement systems are all controlled by institutions that have no incentive to provide access to companies that will use that access to reduce their revenue.
You cannot build a competitive credit product without access to credit bureau data. The credit bureaus have agreements with the banks. You cannot build a payments layer without access to the banking APIs. The APIs are provided at the discretion of the banks. You cannot scale a financial product without banking partnerships. The partnerships are controlled by the banks.
The infrastructure of disruption is owned by the entity being disrupted. This is a different structural situation from the one in which Western fintech developed, and it requires a different analysis. The institutional veto is not a failure of willpower or vision or product quality. It is the rational operation of a system in which the incumbents control the preconditions for competition.
The third failed assumption is that the rational behavior of individuals inside institutions will aggregate into a rational institutional outcome. A manager who is individually enthusiastic about a product that would improve their institution’s performance is not, within the incentive structure of a state-aligned post-Soviet bank, empowered to act on that enthusiasm in ways that would require significant organizational change. The gap between individual rationality and institutional rationality is, in these systems, enormous. Filling it requires something — political will at the executive level, existential competitive pressure, regulatory mandate — that is typically absent.
The Paradox at the Centre
Here is the observation that I find most troubling, and that I believe has not been adequately described in the existing literature on financial development.
The countries most exposed to the weaponization of the global financial infrastructure — the countries that, as described in the preceding chapters, are most vulnerable to the compliance cascade, most dependent on the correspondent banking relationships that can be withdrawn, most subject to the card network suspensions that can strand their citizens overnight — are also, by a mechanism that is not coincidental, the countries where the institutional veto is most powerful and most systematically applied.
Belarus. Russia. Iran. Venezuela. These are not countries that lack technically competent people. They are not countries where the domestic demand for better financial products is absent. They are not countries whose engineers are incapable of building alternatives to the global dollar-denominated infrastructure. Russia, specifically, has a fintech industry of considerable sophistication in certain domestic segments.
What they share is a financial system dominated by state-aligned incumbents whose interests are aligned with the continuation of the existing arrangement, whose institutional structures systematically veto domestic disruption, and whose relationship with the state prevents the emergence of genuinely competitive alternatives.
The causal relationship runs in a specific direction, and I want to be precise about it. The institutional veto did not create the geopolitical vulnerability. The institutional veto and the geopolitical vulnerability are both products of the same underlying structural condition: the concentration of financial infrastructure control in institutions whose primary accountability is political rather than commercial.
When financial institutions are primarily accountable to political principals, they optimize for the preservation of those relationships. Domestic fintech disruption threatens the institutions’ market position and, by extension, the political principals’ control over financial flows. The institutional veto is the mechanism through which that threat is neutralized. And the same concentration of control that produces the institutional veto also produces the vulnerability to external coercion — because a financial system controlled by a small number of state-aligned institutions is, almost by definition, susceptible to targeted pressure on those institutions.
The countries that most need alternatives to the Western financial infrastructure cannot build those alternatives, because the same political economy that makes them targets of financial weaponization also prevents the emergence of competitive financial institutions that could produce those alternatives.
This is not a policy failure. It is a structural trap. And recognizing it as a structural trap is the beginning of thinking clearly about whether and how it might be escaped.
What CreditMatch Was Actually Competing Against
I want to return, one final time, to the specific experience of building a product inside this system, because I think it clarifies something that the structural analysis can obscure.
CreditMatch was not just competing for market share. It was competing against a model of financial decision-making in which the primary function of the decision is not to identify the best credit risk but to distribute accountability in a way that protects the decision-maker. This is the operational reality of credit decisioning in many state-aligned institutions, and it is as important to understand as the technological or regulatory landscape.
In a performance-managed commercial bank, a loan officer’s career advances when her loan portfolio performs well and recedes when it performs badly. Her incentive is aligned, imperfectly but substantially, with making good credit decisions.
In a state-aligned bank with a more ambiguous performance culture, the loan officer’s primary risk is not a bad credit decision. Bad credit decisions are distributed across the portfolio and are managed at an institutional level. Her primary risk is a decision that can be attributed to her personally — a deviation from established procedure that, if it produces a bad outcome, can be named as her failure.
The rational strategy, in this environment, is not to make better decisions. It is to make more defensible decisions — decisions that conform to established practice, that can be explained as standard procedure, and that do not expose the individual decision-maker to singular accountability.
A credit scoring model that produces better outcomes but requires the loan officer to exercise judgment about its outputs is, from this perspective, not an improvement. It is a liability transfer. It moves responsibility from the institution’s established procedure to the individual officer’s judgment. Officers trained in a system where individual judgment is a risk, not an asset, are rational to resist this.
This is what I did not understand when I built CreditMatch. I had optimized for predictive accuracy. I should have optimized for accountability redistribution. I should have designed the product so that using it transferred no risk to the individual decision-maker — so that the output of the model was as formally accountable as any established procedure, so that following its recommendation was as defensible as following any other institutional standard.
This is a product design insight. But it points to a deeper truth about what innovation in these systems actually requires: not just technical superiority, but a fundamental redesign of the accountability structure around financial decisions. Technology that makes people better at their jobs is insufficient if those people are operating in an environment where being better at their jobs is not what they are primarily incentivized to do.
The Broader Implication
I have written this chapter primarily from personal experience, but the institutional veto is not unique to the specific markets where I worked. It exists, in variant forms, in any financial system where:
The incumbent institutions control the infrastructure on which competitive entry depends. The decision-making culture of those institutions is structured around accountability avoidance rather than performance optimization. The political economy of the system rewards incumbents for stability rather than penalizing them for underservice. And the regulatory environment provides no mandate for infrastructure access, interoperability, or contestable market conditions.
These conditions exist, to varying degrees, across a wide range of markets. Parts of the Middle East. Much of sub-Saharan Africa. Several Latin American markets. Even in the United States and Europe, there are segments — small business lending, agricultural finance, certain immigrant communities — where the institutional veto operates in a weaker but recognizable form.
The post-Soviet variant is particularly acute because it combines all four conditions simultaneously and because the political alignment of the incumbents is particularly deep. But the mechanism is not regionally specific. It is structurally specific. And the structural conditions that produce it can be diagnosed and, in principle, addressed.
The diagnosis requires understanding that the problem is not primarily technological. The technology for better financial services in underserved markets exists. It has existed for some time. The problem is that the deployment of that technology requires the cooperation, or at minimum the non-interference, of institutions that have rational reasons to defer indefinitely.
Addressing this requires one of three things: regulatory mandate for infrastructure access, forcing the incumbents to provide the connectivity on which competition depends; sufficiently severe competitive pressure from outside the system, which is difficult to achieve in markets where foreign entry faces its own barriers; or the emergence of an alternative infrastructure entirely — a system that does not require the incumbents’ cooperation because it does not depend on their infrastructure.
The third option is what the CBDC and digital currency experiments described in the following chapter represent. Whether they constitute a genuine alternative or simply a new form of the same incumbent control is the question that makes those developments consequential rather than merely technical.
What Three Failures Taught Me
I want to be direct about what building and losing three products inside this system cost, and what it produced.
The cost is straightforward: years of work, significant financial loss, and the specific kind of exhaustion that comes from sustained engagement with processes designed to produce no outcome. The latter is harder to quantify than the former but is, in some ways, more significant. The institutional veto does not just kill products. It consumes the energy of the people who try to bring them into existence, through the particular frustration of work that is completed correctly and still produces nothing.
What it produced is this chapter. The framework of the institutional veto is not something I derived from academic literature or theoretical reasoning. It is something I built from primary data: three specific engagement processes, described above with as much precision as the terms of the relationships permit, each of which produced the same outcome through the same mechanism. The framework is empirical. It is grounded in experience that cannot be replicated from outside the system.
This, I believe, is the only thing that transforms failure into something with value. Not the motivational narrative of failure as a teacher — that framing is too comfortable and too detached from what failure actually feels like to be fully honest. But the possibility that failure, if examined with sufficient precision, can produce a description of the mechanism that produced it. A description that did not previously exist. A description that might, in the right hands, be useful.
I do not know how useful it will be. The structural trap I have described does not have an obvious solution. The institutional veto is not a problem that can be solved by smarter founders or better products or more patient capital, though all of these help at the margin.
It is a problem that can only be solved by changing the incentive structures of the institutions that exercise it, or by rendering those institutions irrelevant through the emergence of an infrastructure they do not control.
Neither of these outcomes is close.
But naming the mechanism is the beginning of the conversation about solutions. And the conversation cannot begin until the mechanism has a name.
