Legal and tax guides to doing business in Kazakhstan describe the formal route into the market in considerable detail, and, on the procedures that matter at entry, the country now performs well. In the World Bank's B-READY 2025 assessment, business entry is Kazakhstan's highest-scoring topic, with a locally owned company registered in one day and a foreign-owned one in ten. For a board asking what it is like to do business in Kazakhstan, such results show how easily a company can become established, but they say much less about how reliably it will operate afterwards.
The questions that determine results arise once a company moves from setting up a presence to running an operation. Management then needs to understand how predictable execution is beyond registration and tax filing and where the decisions, capabilities and counterparties that matter are located. It also needs to know which processes depend on organisations outside its control and when the entry model will stop fitting the business built since.
The evidence reviewed here suggests that Kazakhstan has become institutionally more accessible to foreign companies without becoming a uniformly predictable operating environment. How closely the formal framework predicts execution varies with the type of process, the business model, exposure to state and quasi-state counterparties, geography and the depth of local commitment. This article, part of Tretiakov Consulting's wider coverage of Kazakhstan, examines where the framework is a reliable guide, where it is not and what the difference means for a foreign company's operating model.
Kazakhstan's Formal Business Environment Has Become More Accessible
Much of the measurable improvement lies in digital government, where Kazakhstan rose to 24th of 193 countries in the UN E-Government Development Index in 2024, up from 28th in 2022. B-READY places the country in the top fifth of measured economies for public services directed at firms, and 98% of firms now file and pay their taxes electronically. According to the authorities, the 2021–2025 privatisation plan transferred 402 state assets to the private sector, and around 500 further companies have been identified for optimisation over 2026–2030. For commercial disputes, the common-law court and arbitration centre of the Astana International Financial Centre (AIFC) give foreign investors an alternative to the national courts.
On this evidence, formal access has largely ceased to be the binding constraint for foreign companies assessing the business environment in Kazakhstan. B-READY makes the distinction between access and execution measurable, because it scores each topic separately on the quality of regulation, on the public services that deliver it and on the operational efficiency that firms experience. Kazakhstan receives 70, 72 and 66 points out of 100 on these three measures. Its topic scores, however, range from 85 for business entry to 54 for market competition, which shows how unevenly the same institutional framework performs from one process to another. Legal and tax guides describe the rules accurately but are not designed to capture this variation. Registration speed and statutory openness are central to the choice of entry route, although they reveal little about how reliably a company will execute once it is established.
The Execution Gap Is Uneven, Not Universal
How long things take in Kazakhstan depends chiefly on the type of process involved, and the data show little sign of a single national pace. Standardised, mostly digital processes with few hand-offs between organisations compare well internationally, with B-READY recording property transfers in nine days and electricity connections in 14.
Where an outcome depends on several agencies or on physical capacity, performance weakens even when the underlying systems are well designed. B-READY gives Kazakhstan full marks for customs risk management, coordinated border management and trusted-trader programmes, but no points for equipment and facilities at the border. Firms report that clearing all border agencies takes 14 days for exports and 12 for imports. On the Middle Corridor, a World Bank studyfound that in 2022 goods took an average of 50–53 days to travel from Dostyk or Khorgos on the Chinese border to the Black Sea port of Constanța. Private-sector respondents to a 2025 OECD review of the Trans-Caspian corridor attributed delays at crossings with China to outdated infrastructure, limited throughput and inadequate equipment.
The largest gaps between rules and practice appear in approvals involving substantive review and in processes that return money to firms. An environmental permit takes 112 days, compared with 21 days for a construction-related permit. In B-READY data that predate the new Tax Code, 62% of firms had not applied for a VAT refund because the process was too burdensome. Dispute resolution shows a related divergence between speed and confidence, with almost 90% of the available points awarded for the time and cost of court processes but less than half for the reliability of courts. Arbitration and mediation fare worse still, receiving under 3% of the points available for reliability.
Across these examples, predictability tends to decline as independent decision points multiply, as official discretion widens, as reliance on infrastructure operating near capacity grows and as exposure to rules set outside Kazakhstan increases. The last of these has mattered more since 2022, because payments involving Russia-linked counterparties now pass through bank compliance reviews shaped by Western sanctions, and the EU's sanctions package of October 2025extended its transaction ban to VTB's subsidiary in Kazakhstan. Within the tax system there is also an asymmetry, since filing and payment are almost fully automated while recovering input VAT has been considerably harder, at least in the data collected before the new Tax Code took effect. With the standard VAT rate raised from 12% to 16% under the new Tax Code in January 2026, slow refunds now tie up more cash for exporters and investing companies.
For a foreign operator, the practical implication is that the reliability of a plan depends on its mix of processes, and a general view of Kazakhstan as fast or slow is of little help in judging it. Management needs to know which parts of the operating model rest on standardised processes that can be scheduled with confidence and which rest on borders, refunds, permits, bank reviews or counterparties whose timing the company cannot control. The second group requires buffers in inventory, working capital and project schedules, together with people close enough to manage exceptions. Such buffers are expensive to carry when the National Bank's base rate stands at 16.25%, as it did in September 2026.
Economic Geography and Decision Geography Are Not the Same Thing
Kazakhstan is often treated as a single national market, although output, companies, population, government and infrastructure are distributed across the country in quite different ways. Economic activity is concentrated and has become more so, with Almaty accounting for 22.7% of gross regional product in 2025, Astana for 12.3% and Atyrau region for 10.4%, while the two cities' combined share rose from 32% in 2023 to 35%. Almost half of all registered legal entities were based in the two cities at the start of 2026, even though the two cities hold only about a fifth of the population. Turkistan region, the most populous with 2.15 million residents, has the lowest output per head in the country.
These layers point to different places, with Almaty the principal concentration of private commercial activity and Astana the institutional centre, home to the government, the national welfare fund Samruk-Kazyna and the AIFC. Much of the country's industrial value, and the procurement budgets attached to it, sits in the western oil-producing regions. Infrastructure follows its own pattern, with generating capacity weighted towards the north and supply to the south dependent on transmission links that are still being reinforced. The Ministry of Energy does not expect to close the electricity deficit recorded in 2025 until 2027.
The operating consequence is that where to sell, where to manage, where to maintain stakeholder coverage and where to locate physical operations may need separate answers, even though these decisions are easily taken together when the first office opens. For an oilfield equipment supplier, for example, the customers are in the western regions, the ministries and national companies that shape its market access are in Astana, and the largest concentration of private companies and service providers is in Almaty. One country office can coordinate such a footprint, but it cannot execute all of it from one city. Where to compete remains a question of demand and commercial feasibility, whereas how to organise around the resulting footprint is a question of operating model.
State Participation Changes the Stakeholder Map
Official estimates put the state's share of GDP at around 14%, a figure that understates how widely a foreign company may encounter the state in its operations. The OECD's 2024 review of state-owned enterprises in Kazakhstan found them in at least 20 of 30 sectors, held through several large holding companies of which Samruk-Kazyna is the most significant. The OECD's 2025 competition review adds that the government's list of activities carried out by state enterprises contained 419 items as of February 2025, and B-READY notes that competition law does not apply to all state-owned enterprises and legal monopolies.
The state's influence on execution is most direct in purchasing, where public procurement contracts worth about 5 trillion tenge were signed in the first five months of 2025. Samruk-Kazyna group companies bought around 2.2 trillion tenge from domestic producers over the same year. The rules favour local supply, as B-READY records restrictions on foreign firms' participation in public procurement and Samruk-Kazyna signed 363 offtake contracts in 2025, while state-owned enterprises buy under rules that sit outside the general procurement regulations. State-controlled companies also operate the railway and the national grid and hold stakes across oil and gas.
How much of this a foreign company actually encounters is determined mainly by its business model and position in the value chain. A consumer brand or private B2B service provider meets the state mainly as regulator and tax collector, although price regulation is not marginal when 26% of firms report that the government regulates their prices. An importer adds customs, certification and bank compliance, which are procedural matters, whereas a supplier to public or quasi-state buyers finds that procurement preferences begin to shape pricing, sourcing and the case for local production. An industrial investor depends in addition on state-controlled networks, land, permits and regional authorities. In the resource sector the state can act as regulator, as partner through national companies and, as recent arbitration against international oil consortia has shown, as claimant.
Stakeholder architecture is therefore a function of the value chain more than a national constant. It becomes clearer when the state's roles as regulator, buyer, infrastructure controller, owner or partner, and enforcer or claimant are treated as separate sources of exposure. Exposure to one of these roles can usually be handled through compliance, whereas exposure to three or four calls for stakeholder management built into the operating model. For such companies the choice of contract forum becomes a design decision in its own right, particularly as B-READY awards no points to Kazakhstan's legal framework for arbitration of commercial disputes with public bodies and state-owned enterprises.
The Local Management Question Is Really a Decision-Rights Question
Debate about how much autonomy a foreign company's Kazakhstan team should have can obscure a more precise question about decision rights. Some matters attributed to business culture in Kazakhstan, such as how quickly a local team commits to a decision or escalates it, can often be explained by where decisions may be taken and what information travels with them. The design task is to establish which decisions need local ownership, which can sit regionally or globally, and whether authority, information and accountability are located together.
Several features of Kazakhstan's economy make this allocation more consequential, beginning with the speed at which costs can change. Annual inflation was 12.3% in December 2025 and still 9.8% in August 2026, VAT rose by four percentage points in January 2026, and the National Bank has cited revised assumptions on regulated prices in raising its inflation forecast for 2027. Under these conditions a pricing process that relies on periodic approval from headquarters is likely to lag both costs and competitors. Compliance carries a different kind of risk, because Kazakhstan shares a customs union with Russia through the Eurasian Economic Union and a sanctions or export-control error has consequences for the whole group.
Local management capacity sets a practical limit on how far delegation can go, and the available evidence counsels some caution. The World Bank reports persistent gaps in foundational skills that limit firms' productivity and innovation despite low unemployment, while B-READY finds that only 12% of firms run formal training programmes and records restrictions on employing foreign personnel. None of this measures the supply of senior managers directly, but it suggests that a company planning to delegate more should not assume that the capability to use that authority can be hired quickly or brought in from abroad at will.
Because the right allocation differs between businesses, it is better derived from explicit criteria than from a standard list of local and central functions. Decisions for which delay is costly and the relevant information is mainly local are the strongest candidates for delegation. Decisions whose errors are expensive, hard to reverse or consequential for the wider group generally warrant regional or global control, informed by local judgement. Applied in Kazakhstan, these criteria tend to move routine pricing and operational exceptions closer to the country and to keep sanctions screening and long-term commitments, such as localisation undertakings to state customers, nearer the group. The exact boundary still depends on each company's business model, approval thresholds and local capability. What should be avoided is local accountability for results that depend on decisions taken elsewhere on slower cycles. Greater local autonomy is not better in itself, any more than tighter headquarters control is automatically safer. Effectiveness depends on whether authority, information and accountability sit together, which is why decision rights belong in operating model design and not in a job description.
The Operating Model Has to Change as the Commitment Deepens
An exporter selling through a distributor experiences a different Kazakhstan from a subsidiary with inventory and institutional customers, and both experience a different one again from an industrial investor with land, permits and a state-linked partner. Operating a business in Kazakhstan therefore changes character as the commitment deepens, because each step moves a category of process, and the risk attached to it, from someone else's balance sheet onto the company's own.
Under a conventional distributor-led model, many responsibilities for customs, certification, local tax, VAT refunds and receivables can stay with the distributor, depending on the contract and on who acts as importer of record. A company that imports in its own name takes on border clearance times, refund delays, bank compliance reviews and local accounting, and once it holds inventory, border variability becomes a working capital question. Selling to public or quasi-state buyers adds procurement rules, local-content requirements and pressure to localise, while industrial investment adds land and permits, grid connection in a power-constrained system, obligations attached to incentives and often a state-linked partner. Exit slows as well, with in-court liquidation taking around 36 months according to B-READY.
As a result, operating complexity tracks the number of external interfaces and irreversible commitments more closely than it tracks revenue. A company can double sales through a distributor without adding any of these interfaces, while one with flat sales that starts bidding for Samruk-Kazyna group contracts may add several at once. No universal threshold marks the point at which an entry model stops being adequate, although certain signals tend to say more than revenue does. They include a local team spending growing time on exceptions that need headquarters approval, working capital moving in ways sales do not explain, stakeholder contact shifting from regulators to state customers or partners, and capital or localisation decisions that recur. When such signals appear, the company is running a different business from the one its governance was designed for. The first year of substantive local execution is then better planned as a change of operating model than as a continuation of market entry.
Where Foreign Operating Models Most Often Misalign With Kazakhstan Reality
Many of the problems described as the common mistakes foreign companies make in Kazakhstan can be traced to four recurring mismatches between an operating model and its environment. A headquarters that calibrates its expectations on fast registration and electronic tax filing and then applies them to border clearance, VAT refunds, permits and disputes confuses formal accessibility with execution predictability.
A second mismatch, between a national market and an operating footprint, arises when Kazakhstan is planned as one market run from one city while customers, institutional counterparties, labour and infrastructure constraints sit in different regions. A third separates accountability from authority, as the country manager carries the P&L while pricing, compliance clearance and investment decisions are taken abroad on slower cycles. That gap matters more in an economy with recent double-digit inflation, regulated prices and sanctions exposure.
The fourth develops gradually, as governance designed for a sales office persists after the business has acquired inventory, state customers, localisation commitments or a partner, leaving controls that no longer match the risks. None of these mismatches requires a poor initial decision, since each can arise from a design that suited an earlier stage and was not revisited as the company's position in Kazakhstan changed.
What This Means for Running a Company in Kazakhstan
Once Kazakhstan becomes a substantive part of the business, a board should be able to answer five questions that follow from this analysis. The answers will differ with business model, scale, customer structure, capital and institutional exposure, and geography, even between companies in the same sector.
Operating footprint. Where should management, commercial activity, stakeholder coverage and operational assets be located, and has each location been chosen on its merits and not by default in the city where the first office opened?
Decision rights. Which decisions sit locally, regionally and globally, and does each allocation reflect where the information is, who bears the consequences and the relative cost of delay and error?
Management capability. Does local capability in finance, compliance and stakeholder management match the business as it now is, and how will the company build that capability where it cannot be hired quickly?
Stakeholder architecture. Which external organisations materially affect execution for this business model, and how many of the state's roles as regulator, buyer, infrastructure operator, partner and enforcer does the company face?
Control architecture. Which headquarters controls, such as sanctions screening, related-party oversight and capital approval, genuinely reduce risk, and which slow routine decisions without reducing risk while weakening local accountability?
The aim is an operating model in which authority, capability, accountability and control match the Kazakhstan business the company actually runs. In practice that may mean more localisation in some functions and tighter group control in others.
Outlook: A More Formalised, but Not Homogeneous, Operating Environment
Institutional change points towards further standardisation, although what it means for operators will depend on how reforms are implemented. The new Tax Code, in force since January 2026, will be judged by how audits and VAT refunds work in practice, and the 2026–2030 optimisation plan should reduce direct state ownership further if it is carried through. Physical bottlenecks are also being addressed, as the World Bank is supporting a new rail line in Kazakhstan that is expected to help triple freight volumes and halve end-to-end transit times along the Middle Corridor by 2030.
Other sources of differentiation are structural, because distance, the location of resources and the capacity of border crossings and the Caspian leg will keep execution uneven across regions and supply chains even as individual bottlenecks ease. Support for local content is expanding, so the gap between companies that sell to the state and those that do not is more likely to widen than to close. Western sanctions policy will also continue to shape payments involving Russia-linked counterparties, whatever the pace of domestic reform.
The variables most worth monitoring are VAT refund times under the new Code, corridor reliability measured as averages instead of best cases, the southern power balance, the pace of the optimisation plan, sanctions enforcement in third countries and the unwinding of regulated prices. Progress on these fronts would make more of Kazakhstan's operating environment predictable without making it uniform across regions, processes and business models.
Conclusion
Understanding the formal rules of doing business in Kazakhstan has become considerably easier, since the guides describe those rules reliably, registration is fast and most routine compliance is digital. Designing an organisation that executes well in the actual operating environment remains a management problem, and a different one for an importer, a supplier to state buyers and an industrial investor.
That problem calls for operating footprint, stakeholder architecture, decision rights, management capability and control to be treated as connected design choices and revisited each time the commitment deepens. Entry guides remain a sound starting point, but whether the organisation built after entry fits the business the company has developed in Kazakhstan depends on those choices.
