Money on a Spring: Statistical Appendix and Methodology for Estimating the Growth Ceiling

Working calculation base for the analytical article. Monetary and banking data — as of September 30, 2026; macroeconomic, investment and foreign trade indicators — as published for January–September 2026. Period, coverage and type of measurement are specified separately throughout.

1. Monetary decomposition

Base: Central Bank of Uzbekistan reviews and press releases as of 30.09.2026. Breakdowns are calculated from the raw lines of the Central Bank of Uzbekistan's banking system review; the series begins in February 2013; the decomposition window is 31.12.2022 → 30.09.2026; the residual of the original breakdown is 0.0.

• September 2026. ΔM2 +24.1 trillion soums (+5.52%). Contribution of credit to the private sector +12.2; net claims on government −4.5; net foreign assets −52.1, including the gold line −55.3; change in the capital of the Central Bank and banks in the counterbalancing. The net contribution of the "gold — Central Bank capital" pair to the decomposition is about +11.2 trillion; excluding this revaluation block, M2 growth is about +12.9 trillion. These lines describe accounting counter-positions of the monetary aggregate, not the amount of money directed toward purchases.
• Instrument structure of M2 as of October 1. M2 — 461.9 trillion soums (+38.4% year on year; +22.0% since the start of the year). September growth — 24.1 trillion. National time and savings deposits, according to the money supply press release, rose by 16.8 trillion over the month, to 217.1 trillion; that is about 70% of the monthly increase in M2. Calculated on the combined balance-sheet line "other deposits" of both currencies, time and savings deposits total about 283.1 trillion (61.3% of M2). These figures have different coverage: 217.1 trillion is the national-currency component in the press-release breakdown; 283.1 trillion is the broad combined line of the underlying balance-sheet series. Do not substitute one for the other.
• Fifteen quarters. ΔM2 +272.8 trillion; credit to the private sector +327.9 (120% of the net increase); government −27.0; net claims on non-residents excluding gold −125.0 (liabilities to them +68.3); gold +451.5; capital −353.4; other +4.3. This is a balance-sheet decomposition, not an estimate of final expenditure.
• The last twelve months to 30.09.2026. ΔM2 +128.1 trillion; credit +101.6 (79%); government −41.5; external contour −72.5 (Central Bank position −53.3 trillion, roughly $4.4 billion; banks' position −19.2 trillion, about $1.6 billion); gold +176.6. The $4.4 billion equivalent is a conversion of the change in the Central Bank's balance-sheet position, not separate statistics on gross foreign exchange interventions and not the exact sum of sales over the calendar nine months. The $5.92 billion as of 01.10.2026 is the stock of the foreign currency portion of reserves on that date, so it cannot be directly subtracted from the $4.4 billion flow: the indicators differ in nature and period. In its foreign exchange market review for January–September, the Central Bank separately reported that additional liquidity from operations with monetary gold was sterilized through foreign exchange operations and other monetary policy instruments. The reported loan portfolio in the press release increased by 85.4 trillion, from 627.9 to 713.3 trillion (+13.6%); the difference with the balance-sheet credit line in the decomposition is due to line coverage and should not be concealed.
• Fiscal channel since 2013. The net change in the corresponding position is −212.2 trillion soums (48.6% of the entire increase in M2 over the calculation window). This is a long-term balance-sheet estimate, not the sum of budget expenditures that can be "returned" to circulation by a single decision.
• Denominators. Rolling M2/GDP ratios: approximately 19.1 → 22.0%; credit/GDP: 35.9 → 33.9%. A correct velocity of circulation requires nominal GDP for the same period and the average stock of M2: for example, quarterly or annual nominal GDP divided by average M2 over the corresponding interval. One cannot directly equate M2 +38.4% on a date with nominal GDP +17.4% over nine months.

Sources:

• Central Bank: monetary aggregates.
• Central Bank data on September M2 growth and deposit composition, 16.10.2026; additional analysis, 16.10.2026.

2. Investment and output: January–September 2026

State Statistics Committee:

• Fixed capital investment — 584.2 trillion soums in current prices. The official growth rate is 20.1% versus January–September 2025; the rate is calculated using the physical volume index/comparable prices methodology, not as the ratio of two nominal sums.
• By sector (value of investment in current prices): manufacturing — 160.8 trillion; housing construction — 87.1; agriculture, forestry and fishing — 52.1; construction — 51.1; electricity, gas, steam and air conditioning — 39.0; mining — 33.9; transport and storage — 27.7; trade and vehicle repair — 22.6; water supply and sewerage — 12.2; ICT — 9.3; professional, scientific and technical activities — 6.1.
• Sources of financing: foreign investment and non-guaranteed foreign loans — 371.7288 trillion (63.6%); household funds — 71.5907 (12.3%); enterprise funds — 67.8196 (11.6%); state-guaranteed foreign loans — 36.2751 (6.2%); republican budget — 25.3830 (4.3%); commercial bank loans and other borrowed funds — 7.4191 (1.3%); Reconstruction and Development Fund — 3.9379 (0.7%). Foreign investment/non-guaranteed loans together with guaranteed foreign loans give about 69.8% of the investment volume by source classification. This is not the share of foreign direct investment and not the share of external debt: the categories combine different instruments.
• Budget + guaranteed foreign loans + Reconstruction and Development Fund: 65.596 trillion, or about 11.2% of the investment volume. This is only the arithmetic sum of the three named categories; not all of them are budget expenditures.
• GDP January–September: 1,755.5 trillion soums in current prices versus 1,494.9 trillion a year earlier; nominal growth about 17.4%, real — 8.3%, GDP deflator — 8.4%. The deflator covers domestic value added and is not an alternative CPI.
• Other real growth indices for January–September: industry +8.1%; construction +12.3%; agriculture +4.6%; services +16.7%; retail trade +19.7%.

Sources:

• State Statistics Committee: fixed capital investment — 584.2 trillion, 30.10.2026.
• State Statistics Committee: sectoral composition of investment, 02.11.2026.
• SIAT: fixed capital investment growth and methodological metadata.
• State Statistics Committee: GDP for January–September 2026, 30.10.2026.
• GDP and deflator for January–September, 30.10.2026.

3. Imports, domestic demand and production

3.1. Foreign trade according to State Statistics Committee data

• January–September 2026: foreign trade turnover — $65.4 billion; exports — $25.8 billion (−5.5%); imports — $39.6 billion (+15.7%); negative balance — about $13.7 billion.
• Exports of goods excluding gold — $13.3 billion (+31.6%). Exports of services — about $9.7 billion (+31.1%). Exports of non-monetary gold — about $2.8 billion, 10.8% of total exports. The $13.3 billion figure is not all exports excluding gold: services are counted separately.
• January–August 2026: imports $34.5 billion (+16.3%); machinery and transport equipment — $11.3 billion (32.6% of imports); industrial goods — $5.1 billion; services — $4.6 billion; chemical products — $4.2 billion; food — $3.7 billion; fuel — $2.8 billion.
• January–July 2026: imports of machinery and transport equipment — $9.7 billion (+20%); industrial goods — $4.4 billion (+10.1%); services — $3.6 billion (+13.3%); chemical products — $3.6 billion (+15.6%); food and live animals — $3.3 billion (+41.9%); mineral fuel — $2.5 billion (+10.7%). The rates reflect dollar value, not the physical volume of imports.

Sources:

• State Statistics Committee: foreign trade for January–September, 28.10.2026.
• State Statistics Committee: imports by group, January–July, 24.09.2026.
• Export structure after nine months, 31.10.2026.

3.2. Consumer demand and food

• Retail trade turnover for January–September was 432.2 trillion soums, up +19.7% in real terms. Wholesale trade rose +19.2% in real terms. Turnover captures the full price of goods sold, including imported goods, and is not equal to the value added of domestic trade; it can therefore outpace GDP when imports rise.
• Agricultural output for January–September was +4.6% in real terms. Within that: crop farming +6.9%; livestock +2.1%; meat (live weight) +2.1%; milk +1.7%; eggs +7.5%.
• Food imports in January–July were $3.3 billion (+41.9% in dollar terms). This cannot be compared directly with real growth in agricultural output: the period, the goods and the price base all differ. It is a signal for a detailed commodity-by-commodity comparison, not a completed import-substitution test.
• In January–July, large food-industry enterprises produced 137.7 trillion soums of output (+8.8% in value terms). This figure is also not a physical output index and does not cover small enterprises.
• Personal remittances in the first half of the year were $9.3 billion (+13%); this is a foreign-currency inflow and a source of household income, but not a sum automatically spent on consumption.
• Loans to individuals as of 01.08.2026 stood at 239.8 trillion soums (+19% year on year) according to the Central Bank's sector table. In the original analytical breakdown, a separate card line grew by roughly 38%; that is a narrow category and cannot be described as the growth rate of all consumer lending.

Sources:

• Retail trade, January–September 2026, 03.11.2026.
• State Statistics Committee: agriculture, January–September, 29.10.2026.
• State Statistics Committee: food imports, January–July, 24.09.2026.
• Central Bank: personal remittances in H1 2026, 14.08.2026.
• Central Bank: loans by sector, 01.08.2026.

3.3. The difference between foreign trade statistics and the balance of payments

• State Statistics Committee: January–September imports of $39.6 billion (+15.7%) and a trade deficit of $13.7 billion — foreign trade statistics.
• Central Bank: first-half balance-of-payments imports of $28.8 billion (+24%), with a current-account deficit of about $6.2 billion — BPM6 methodology.
• These figures cannot be subtracted from one another to estimate Q3 imports or the trade balance: the coverage of transactions, the methodology, the valuation and the period all differ. In the publication, both series should be labelled separately.

Source: Central Bank: balance of payments, international investment position and external debt in H1 2026, 29.09.2026.

3.4. Non-monetary gold exports, production and reserve accumulation

Terms. Non-monetary gold is gold recorded in the statistics as merchandise exports. Monetary gold is the Central Bank's gold asset held as part of international reserves.

The arithmetic of the change in the foreign trade balance, January–September:

The 2025 values for imports, goods excluding gold and services are reconstructed approximately from the published 2026 growth rates; the sums are rounded. This is an arithmetic reconstruction, not a separately published table. It shows that the fall in gold exports almost equals the increase in the deficit, while growth in exports of goods excluding gold and of services partly offset the rise in imports.

Price, estimated export tonnage, production and reserves:

No comparable national gold production data in tonnes for the nine months of 2025 and 2026 were found in the open-source materials used for this version. NGMK disclosed physical output; AGMK reported the percentage of its refined-gold forecast fulfilled without tonnage. Annual production estimates for Uzbekistan in 2025 diverge across public sources: about 118.4 t in an Interfax report citing official data, and about 130 t on the country's investment portal. These annual figures cannot be mechanically spread across nine months; the discrepancy further illustrates why the 9M aggregate cannot be built up from assumptions.

Testing the hypothesis of reserve sales in 2025 and gold accumulation in 2026. The Central Bank's annual balance sheet, as reported by Interfax, settles the question for 2025: 109 t of monetary gold was purchased from local producers (110 t in 2024), 100 t was exported (100 t in 2024), and the net change in the physical stock was about +0.3 million oz (+9.3 t). The arithmetic checks out: 109 − 100 = 9 ≈ 9.3 t. No sale of previously accumulated reserves was needed to support 2025 exports; the estimated 95 t for 9M-2025 is very likely the same Central Bank flow (100 t for the year, with the last shipment in September worth about $1.5 billion), valued at the nine-month average price. With annual production of 118.4 t, Central Bank purchases covered about 92% of output; the remainder is distributed among domestic sales through the exchange, direct sales and producers' inventories. Caveats: the origin of individual Central Bank shipments is not officially disclosed; the annual net figure does not rule out intra-year fluctuations in the stock; and the non-monetary gold line in trade statistics is not methodologically broken down between the Central Bank and producers.

It is not possible to compare these 95 t with nationwide production for 9M-2025 and 9M-2026 using open data. NGMK produced 72.99 t in 9M-2025 and reported rounded 2.3 million oz in 9M-2026 (central estimate 71.5 t; the possible range when rounded is roughly 70.0–73.1 t). AGMK publishes only the percentage of its refined-gold forecast fulfilled — 100.7% and 101.3% — without tonnage. The national nine-month total therefore cannot be correctly assembled from the published components. Annual estimates for 2025 also diverge: about 118.4 t in an Interfax report citing official data and about 130 t on the investment portal; these annual figures cannot be distributed across months.

In the first half of 2026, the WGC recorded about 41 t of net purchases by Uzbekistan. According to Gazeta.uz and the WGC, Central Bank shipments abroad have not resumed since September 2025, and the only month of net sales in 2026 was July (−1 t); gross sales alongside purchases in 2026 cannot be traced in the data, but are not fully ruled out either without the Central Bank's own monthly series. A definitive answer requires a monthly physical balance of NGMK and AGMK production, refining, Central Bank purchases and sales, producers' inventories, exports in tonnes and domestic use.

Domestic use of gold is not zero either. Gold sales through the Uzbek Republican Commodity and Raw Materials Exchange amounted to 3.73 t in 9M-2025 and 8.84 t in 9M-2026; exchange representatives have said that about 90% is bought by jewellery enterprises. This is a separate channel of domestic demand, but exchange turnover is not equal to all domestic consumption and does not close the country's physical balance.

Physical gold reserve series of the Central Bank, 2025–2026. The full monthly series is published by the Central Bank itself (press releases on international reserves) and by the World Gold Council (the Gold Reserves by Country table); in the open press it can be reconstructed from key points:

The September purchase (+7.2 t) coincided with a correction in the global price: the value of gold in reserves fell over the month while tonnage rose — the regulator was buying the metal into the dip. Net purchases over nine months (about 55 t) at $4,552/ounce are equivalent to roughly $8 billion. Non-monetary exports in 2026 (an estimated ~19 t), given that Central Bank shipments stopped in September 2025, point rather to direct sales by producers; there is no official breakdown of the line between the Central Bank and producers.

Reconciling the gold purchase and sterilization. 41 t of gold at a conditional price of $4,500 per troy ounce corresponds to approximately $5.9 billion. This does not match the change in the Central Bank's external balance-sheet position of −53.3 trillion soums (about $4.4 billion), because the latter figure is the change in the balance-sheet position over the twelve months to 30.09.2026, not the gross amount of foreign-currency sales in the first half. Over the same twelve-month period, the Central Bank's sterilization liabilities rose by 70.5 trillion soums (around $5.8 billion at a conditional rate of 12.1 thousand soums per dollar). These are the Central Bank's liabilities to banks in the national currency, not a foreign-currency intervention. These amounts cannot be added together or netted against each other: they differ in period, nature and balance-sheet coverage. The Central Bank has said that liquidity from monetary gold operations was sterilized through foreign-exchange operations and other instruments. Absorption therefore ran through several channels, but their exact distribution cannot be reconstructed without monthly data.

Sources:

• State Statistics Committee: external trade for January–September 2026, 28.10.2026; non-monetary gold exports $2.804 billion versus $9.8 billion a year earlier.
• World Bank Pink Sheet: commodity prices; monthly average gold quotations used to calculate the average price for 9M.
• NGMK: operating results for 9M-2025, 10.10.2025; 2.3466 million troy ounces.
• NGMK: operating results for 9M-2026, 21.10.2026; about 2.3 million ounces, figure rounded.
• AGMK: analytical note for 9M-2025, refined gold target fulfilment — 100.7%, tonnage not disclosed.
• AGMK: analytical note for 9M-2026, refined gold target fulfilment — 101.3%, tonnage not disclosed.
• Invest Uzbekistan: Metals and mining; gives a rounded estimate of gold output for 2025 of about 130 t.
• Interfax: gold exports and output in Uzbekistan, 28.08.2026; gives the official output estimate of 118.4 t for 2025 and reports gross gold exports by the Central Bank of about 100 t for the same year.
• Interfax: Uzbekistan's Central Bank bought 109 t and exported 100 t of gold in 2025, 09.07.2026; annual balance: purchases from producers 109 t (110 t in 2024), exports 100 t (100 t in 2024), net change in the stock about +0.3 million ounces; output 118.4 t; target for 2030 — 175 t.
• Gazeta.uz: Uzbekistan entered the top three largest gold buyers, 07.09.2026; July −1 t, about 40 t since the start of the year, stock about 431 t; last shipment abroad — September 2025 (~$1.5 billion).
• World Gold Council: central bank statistics, July 2026, 03.09.2026; −1 t in July, about 40 t since the start of the year, about 431 t (87% of reserves).
• World Gold Council: central bank statistics, August 2026, 06.10.2026; +8 t in August, about 50 t since the start of the year.
• UzDaily: Uzbekistan's Central Bank bought 8 tonnes of gold in August, 07.10.2026; stock about 439 t at the end of August.
• GoldSilver: Gold Reserves by Country — 2026 Rankings, 28.05.2026; +25 t in Q1, stock 416 t.
• World Gold Council: Gold Reserves by Country — monthly table; changes in gold reserves since January 2002 in tonnes and dollars.
• Kursiv: gold trading on the UZRTSB for 9M-2026, 30.10.2026; 3.73 t for 9M-2025 and 8.84 t for 9M-2026; according to an exchange representative, about 90% of purchases are by jewellery enterprises.
• World Gold Council: Central Bank Gold Statistics, June 2026, 04.08.2026; published net purchases by Uzbekistan for the first half — about 41 t.
• Spot: Uzbekistan's international reserves at the start of October 2026, 09.10.2026; physical gold volume 446.2 t, foreign-currency reserves $5.92 billion.
• Law "On Precious Metals and Precious Stones", Article 21; the state's priority right to purchase refined gold.
• IMF: Uzbekistan—2025 Article IV Consultation, Informational Annex; the mechanism for purchasing domestic gold and the Central Bank's foreign-currency supply.
• Central Bank: review of the domestic foreign-exchange market for January–September 2026, as summarized by Spot, 03.11.2026; demand $50.4 billion, supply excluding Central Bank operations $43.8 billion; liquidity from monetary gold operations was sterilized through foreign-exchange operations and other monetary policy instruments in line with the regulator's neutrality principle.

4. Idle capacity and procurement efficiency

• 24.04.2026: of 688 enterprises launched over the previous three years, 210 were not operating at full capacity; estimated lost output — 33 trillion soums; unfilled jobs — 23 thousand. The 33 trillion is an estimate of potential output, not proven direct damage or a sum of wasted funds. Relative to nominal 2025 GDP of 1,849.7 trillion it is about 1.8%, but only as an order-of-magnitude comparison.
• 05.05.2026: for a different group of enterprises, the potential for additional output and exports of up to 65.6 trillion soums was estimated if constraints were removed. This is a separate group and a separate presentation; do not add it to the 33 trillion.
• Official statements on the causes of underutilization cited shortages of working capital, raw materials, utility and engineering infrastructure, project delays, worn-out equipment and weak demand.
• The Audit Chamber reported violations of public procurement legislation detected in the first half of 2026 amounting to 9.1 trillion soums. The ratio of 9.1 to 584.2 trillion in investment is about 1.6%, but this is not the share of losses in investment: the periods and coverage of the series do not match, and the amount of violations does not equal proven damage.
• Investment from three categories — "republican budget" + "government-guaranteed foreign loans" + "Reconstruction and Development Fund" — about 65.6 trillion, 11.2% of the total. By coincidence, 65.6 trillion is also the estimate of additional output and exports for another group of enterprises in the May 5, 2026 statement. These are different indicators: the sum of investment financing sources and the estimated output potential. This is not all government spending, nor the entire volume of procurement. Foreign investment and non-guaranteed foreign loans are a separate category at 371.7 trillion, or 63.6%.
• Consequently, the data on public procurement violations justify scrutiny of projects, but by themselves they do not explain the macroeconomic growth of imports and do not prove that asset values are inflated.

Sources:

• President: enterprises not operating at full capacity, 24.04.2026.
• President: measures to support enterprises, 05.05.2026.
• Audit Chamber: results for H1 2026, 16.07.2026.
• Anti-Corruption Agency: procurement violations in 2025, 26.02.2026.

5. Banking structure and the allocation of credit

• M2 as of 01.10.2026 — about 461.9 trillion (+38.4% year on year); the monetary aggregate includes cash and deposits of varying degrees of liquidity.
• The Treasury account and accounts of budget organizations — about 304.2 trillion soums; in size this is equivalent to 65.9% of M2 as of 01.10.2026, but it is not a share of M2 itself. The net increase in the corresponding line over the year in the original decomposition is about 41.5 trillion. We do not give last year's share in this version: a correct comparison requires confirming that the composition of accounts and the denominator match; the absolute change in the balance and its share of M2 are different indicators.
• Sterilization liabilities of the Central Bank to banks — about 87.8 trillion versus 17.2 trillion a year earlier (+70.5). This is a tool for managing bank liquidity, not a monetary reserve that can be added directly to government deposits.
• Time and savings deposits: the broad combined balance-sheet line for the two currencies — about 283.1 trillion (61.3% of M2), +41.1% year on year. The press-release line for time/savings deposits in the national currency — 217.1 trillion as of 01.10.2026. Do not conflate different definitions.
• The weighted average rate on time deposits of individuals in the national currency in August 2026 — 18.4% according to the Central Bank; the September CPI — 6.4%, while September household inflation expectations — 8.7% nationwide and 11.6% in Tashkent. Real return under the Fisher formula is roughly 11.3% against the official CPI, 8.9% under nationwide expectations and 6.1% under the expectations of Tashkent residents. The GDP deflator of 8.4% is not a substitute for the CPI in calculating a depositor's purchasing power.
• The bank portfolio as of 30.09 — about 713.3 trillion (+13.6%); under the original decomposition, industry −10.9%, construction +35.5%, trade +23.5%, a separate card line about +38%, mortgages +38–48%. These lines have different economic meaning and coverage; do not treat them as a single consumer credit index.
• A separate official Central Bank sector table as of 01.08.2026: total loan portfolio 644.6 trillion (+12%); loans to individuals 239.8 trillion (+19%); loans to industry 133.4 trillion (−13% year on year), agriculture 66.4 (+14%), construction 21.6 (+32%), trade and public catering 50.6 (+25%). This is a snapshot at a different date and under a different classification than the 30.09 decomposition.

Sources:

• Central Bank: loans by economic sector as of 01.08.2026.
• Central Bank: monetary aggregates.
• Central Bank: inflation indicators.
• Central Bank: weighted average deposit and loan rates.

6. Structural productivity and competition

• World Bank: according to 2023 data, about 60% of employment within the services sector was in low-skilled segments; globally oriented innovative services accounted for about 4% of the sector's jobs. This is structural background, not a 2026 snapshot.
• OECD: labor productivity grew by an average of 4.9% a year in 2011–2023; most of the increase was driven by improvements within industries, not only by workers moving between industries.
• The OECD estimates that average labor productivity at companies with foreign participation is about 63.8% higher than at domestic firms. The gap is partly explained by size, capital intensity and industry structure; it cannot be attributed entirely to management or corporate culture.
• Central Bank: the Q1 2026 labor market review pointed to real wage growth of 9.5%, outpacing labor productivity growth. This is one quarter, not proof of a multi-year trend.

Sources:

• World Bank — Promise of Services-Led Growth.
• OECD — productivity and structural shifts.
• Central Bank — Q1 2026 labor market review.

7. Methodology for assessing the growth ceiling

One must start with an industry panel, not with a single national number. For each sector, one needs to know capacity utilization, actual asset commissioning, real value added, import content, productivity and resource constraints.

7.1. Potential output and the output gap

The basic production function:

Y_t^*=A_t^*K_t^αL_t^β,

where K is productive capital, L is effective labor, and A is total factor productivity. An enterprise that has been built but is not connected to the grid, not supplied with raw materials or not equipped with machinery cannot be counted as fully working capital.

The output gap:

OG_t=(Y_t-Y_t^*)/Y_t^*×100.

A positive gap shows that actual output exceeds the estimated potential level. But it does not by itself identify a specific physical constraint; that requires industry-level detail.

7.2. Conversion of investment into value added

We keep a single return-on-investment indicator, without duplicating it with the inverse ICOR:

η_s,t,k=fracΔ VA^real_s,t+kI^real_s,t,

where I^real_s,t is investment in comparable prices in sector s in period t, and Δ VA^real_s,t+k is the increase in real value added after a lag k, for example one, two and three years. The more steadily output grows relative to investment, the higher the conversion. One must account for replacement of worn-out assets, depreciation, long construction cycles and changes in utilization. This is an analytical ratio for comparing industries and project cohorts, not a ready-made universal norm.

7.3. Short-term constraint from critical resources

For industry i and resource j, one can estimate the output supported by the available volume of each critical resource:

Y_i,t+1^max=min_j(fracX_ij,t+1^availablea_ij), g_i,t+1^max=fracY_i,t+1^maxY_i,t-1,

where X_ij^available is the available volume of electricity, water, raw materials, transport, skilled labor or imported components, and a_ij is the resource requirement per unit of output. This is a short-term approximation with fixed coefficients; in reality, resource substitution and technological adaptation are possible. The indicator should therefore be presented as a range and tested against enterprise data.

7.4. Import penetration of the domestic market

For a homogeneous product group:

IP_s=M_s/(Q_s+M_s-X_s-Δ Stocks_s),

where M_s is imports, Q_s is domestic output, X_s is exports, and Δ Stocks_s is the change in inventories. The denominator is apparent domestic consumption. It should be calculated on comparable physical volumes or constant prices, excluding re-exports and aligning product classifications to a single code. For intermediate goods, the import content of domestic output is estimated separately.

7.5. Depth of localization and project cost

Depth of localization:

LD_s=fracVA^dom_sGO_s,

where VA^dom_s is value added created within the country, and GO_s is gross output. Growth in gross output with a flat or falling share of domestic value added means that production is expanding, but technological dependence on imports persists.

Unit project cost:

UC_p=fracC_p^actualK_p^commissioned,

where C_p^actual is the actually paid cost of the project, and K_p^commissioned is the capacity actually commissioned or the physical result. Only projects with similar specifications, geography, connections, safety standards and quality can be compared. A rise in unit price without improved characteristics is grounds for an audit, but not automatic proof of corruption.

7.6. External capacity

Instead of a single crude coefficient, foreign-currency receipts and needs should be calculated by the nature of the flow:

FX_headroom=FX_{recurring receipts}+FX_{stable financing}-FX_imports-FX_{debt service}-FX_{other outflows}.

Recurring receipts include exports of goods and services and remittances; direct investment and long-term project financing are shown separately; portfolio flows and short-term borrowings should not be treated as equally stable as export revenue. Imports of equipment are separated from raw materials, components and consumer goods. Double counting must be excluded, and payment timing taken into account.

7.7. Minimum panel for calculating the limit

1. Capacity utilisation by industry and region over several years.
2. Dates of actual commissioning, attainment of design capacity and output dynamics of new enterprises.
3. Real value added per employed person and productivity by sector.
4. Peak load on power grids, connection constraints, electricity and gas disruptions.
5. Availability of water, transport, warehousing and border infrastructure.
6. Vacancies, hiring times, shortages of specialties and labour costs.
7. Import content of investment and output: equipment, raw materials, components and consumer imports separately.
8. The import share of domestic consumption for comparable commodity codes.
9. Conversion of investment into output with a lag of one to three years.
10. Average M2 for the period and reconciled nominal GDP to estimate velocity.
11. Ultimate owners, prices and physical results of major projects and key procurements.

A composite index can only be built once these series are available. Until then, it is more correct to publish a map of sectoral constraints and ranges of potential growth rather than a single indicator with false precision.

8. What else must be reconciled to close the gold loop

1. Three different series cannot be added together. Net gold purchases published by the WGC, the change in the physical volume of reserves in tonnes and the Central Bank's soum-denominated gold balance-sheet line have different coverage and reflect, respectively, operations/valuation of net purchases, the stock on a given date and the value with revaluation. To get a single picture, they must be reconciled by date and methodology.
2. Flow and stock are not interchangeable. The Central Bank's negative position of −53.3 trillion soums (about $4.4 billion) in the annual balance-sheet decomposition is a change in a balance-sheet position, whereas $5.92 billion is the stock of the foreign-currency portion of reserves as of October 1. The first is not equal to gross foreign-currency sales over nine months, and the second cannot be directly subtracted from the first.
3. Monthly reconciliation of operations is missing. The WGC estimates net purchases in the first half of the year at about 41 t (around $5.9 billion at a conditional price of $4,500/oz), while the Central Bank's annual balance-sheet position on the external contour changed by about −$4.4 billion. These are not the same thing: the periods and coverage differ, and the balance-sheet change is not the gross amount of foreign-currency sales. Sterilisation liabilities rose by a further 70.5 trillion soums, but that is a separate instrument for absorbing soum liquidity, not a foreign-exchange intervention. A quantitative assessment requires monthly data on Central Bank gold purchases, foreign-currency sales, export revenue, sterilisation operations, revaluation and changes in the foreign-currency portion of reserves.

Status of verification as of 30.09.2026. Some questions have been closed by data: the Central Bank's annual balance for 2025 (purchases of 109 t, exports of 100 t, net +9.3 t — Interfax) and the 2026 data points (Q1 +25, 1H +41, July −1, August +8, September +7.2 t; stock of 446.2 t as of 01.10) have been published and tabulated in section 3.4. Net purchases for 9M-2026 (about 55 t ≈ $8 billion), against the change in the Central Bank's external position (−$4.4 billion over 12 months) and the growth of sterilisation liabilities (+70.5 trillion soums), still do not add up to a single balance: the periods and the nature of the quantities differ. Still open are: the monthly series for 2025 (apart from the annual total and the September suspension of supplies), the breakdown of 2026 non-monetary exports (~19 t) between the Central Bank and producers, and the allocation of sterilisation between foreign-currency and other instruments.

9. NGMK case: price, physical output, investment and the local chain

Financial comparison period: first half of 2026 versus first half of 2025. Monetary values are given in dollars or soums depending on the indicator; physical gold output is in troy ounces.

9.1. Financial result and physical output

The company states that higher world gold prices were the key driver of revenue growth. Revenue growth and value-based output growth therefore cannot be interpreted as growth in physical supply. At the same time, the 1.8% decline in production does not prove a decline in efficiency: such an assessment requires gold grade in ore, processing volume, recovery rate, stripping ratio and the mining plan by site.

9.2. Two investment metrics — not the same CAPEX

Indicative ratios to adjusted EBITDA:

• Investment Programme / EBITDA: 219.5 / 3,060 = 7.2%; 236.8 / 4,889 = 4.8%.
• Cash CAPEX / EBITDA: 412 / 3,060 = 13.5%; 543 / 4,889 = 11.1%.

The Investment Programme is an administrative programme indicator of spending; Cash CAPEX is the cash flow from investing activities in the financial statements. These amounts cannot be added together or treated as the same metric. The ratios to EBITDA show only that profit, supported by the gold price, grew faster than both investment indicators. A decline in the ratio does not in itself prove that investment is insufficient.

9.3. Costs and physical constraints

AISC rose by $483 per ounce. The financial release cites the following among the reasons:

• higher royalties (subsoil tax) at a high gold price; the royalty share of the AISC structure reached 31%;
• inflation in the cost of consumables;
• higher stripping costs at the Muruntau open pit as part of the expansion project;
• appreciation of the soum against the dollar, which NGMK estimates added $101 per ounce to dollar AISC.

Arithmetic caveat: NGMK's press release says that $101 accounts for 8.7% of the "total increase" in AISC. But the difference between $1,647 and $1,164 is $483; $101 / $483 = 20.9%. The 8.7% figure is obtained by dividing $101 by the prior-period base AISC ($1,164), not by the increase. It is safer for the article to cite the absolute estimate of $101 and not repeat the percentage from the press release without clarifying the denominator.

The operating press release names specific works: stripping at Muruntau, delivery of mining equipment, conveyor lines for the cyclic-flow technology, expansion of processing capacity at GMP-7, modernisation of grinding units at GMP-2 and development of the tailings facility. This confirms that the expansion requires a set of interconnected capacities and involves a time lag. But without ore volumes in tonnes, gold grade in g/t and the recovery rate, the precise narrow bottleneck cannot be identified.

9.4. Localisation and employment

This is not a single "local procurement" indicator: localised products are output under the relevant programme; inter-industry cooperation is procurement under a separate mechanism. They do not cover all of NGMK's procurement and do not show domestic value added in each item. The growth in the table is in value terms; without a sectoral deflator it cannot be interpreted as physical growth.

In June 2026, annual CPI stood at about 6.4%. Nominal growth of localized output (+6.8%) was close to that pace; procurement under inter-industry cooperation (+15.1%) grew faster. The comparison is approximate: the programmes have different compositions, and CPI is not a deflator for the output the combine purchases.

The number of new jobs under the programme is not a change in total headcount. NGMK reports a workforce of more than 50 thousand employees. A decline in the number of new jobs may reflect a change in the project mix, mechanization or automation, but these releases contain no supporting data on the causes.

Operating figures for the nine months of 2026 provide an update to the series: NGMK reported output of about 2.3 million ounces and fulfilment of the plan; the figure in the press release is rounded, so the exact change against 2.3466 million ounces for the nine months of 2025 cannot be calculated from this publication. Execution of the Investment Programme — $390.5 million against $342.9 million (+13.9%); localized output — 1.2258 trillion against 1.1206 trillion soums (+9.4%); inter-industry cooperation — 6.457 trillion against 5.5802 trillion (+15.7%); new jobs — 1,344 against 2,005. Value output reached 128.1 trillion soums against 95.3 trillion, but that sum reflects the value of output, not the physical volume of production.

9.5. What else is needed to explain the monetary result

The brief release does not allow net profit to be allocated across its ultimate uses. Establishing how much went to the budget through taxes and royalties, to dividends, CAPEX, debt servicing, repayment of obligations, or remained in cash balances requires the full interim financial statements and their notes, including:

1. the full statement of cash flows and a reconciliation of EBITDA to operating cash flow;
2. taxes, royalties and other mandatory payments for the period;
3. declared and actually paid dividends;
4. repayment and raising of loans, and movements in cash balances;
5. CAPEX by project and actual commissioning dates.

As of June 30, 2026, the debt portfolio included Eurobonds of $500 million and $1 billion. The ratio of net debt to adjusted EBITDA fell from 0.5x to 0.3x, but this ratio also depends on changes in EBITDA; on its own it does not prove that the principal amount of debt was repaid in the first half.

NGMK sources

• NGMK operating results for H1 2026, 14.07.2026.
• NGMK financial results under IFRS for H1 2026, 28.09.2026.
• NGMK operating results for the nine months of 2026, 21.10.2026.
• Table of NGMK production indicators for 2025–2026, updated in October 2026.
• Spot: NGMK financial results for H1, 01.10.2026.
• Spot: NGMK operating results for H1, 31.07.2026.

10. Cross-checking against independent sources

The purpose of this section is to test the official series on which the article is built against four types of independent data: physical output indicators, mirror customs statistics, external growth estimates and the alternative return on savings. Methodological framework: Henderson, Storeygard, Weil, "Measuring Economic Growth from Outer Space" (AER, 2012); Gibson (2020) survey. No ready-made satellite series (VIIRS and analogues) for Uzbekistan is publicly available — this is honestly recorded as a limitation rather than passed over in silence.

10.1. Physical indicators versus reported growth

Comparison: official real GDP growth for 9M-2026 was 8.3%; industry +8.1%, services +16.7%, construction +12.3%. The electricity, automotive and transport series are consistent with these rates. The exception is freight turnover: +43% in tonne-kilometres over the half-year would imply a sharp rise in average haul distance, which the other series do not confirm; until the origin of the figure is clarified (including a possible value rather than physical basis for the calculation), the series is not used.

10.2. Mirror customs with China (2025)

Caveats: (1) the periods do not match — 10 months against 8; the valid comparison is of directions and annual trajectories, not levels; (2) mirrors systematically diverge on valuation basis (CIF/FOB) and coverage (transit through third countries); (3) the gas line is the only one with a gap above the typical level (~14% for a comparable period), the likely explanation being different metering points and valuation bases for pipeline supplies. Conclusion: there is no asymmetric gap of the kind for which the mirror statistics of a number of countries in the region are known in certain years; the directions of change coincide (Uzbekistan's exports to China are falling, imports from China are rising).

10.3. Remittances: the series, geography and the limits of verification

In addition: Q3 2026 — inbound $5.88 billion (+2.9%), outbound $625 million (−11.6%), balance $5.25 billion (+5%). Annual values: 2024 — $14.8 billion (+30%); 2025 — $18.9 billion (+27.2%), share of total household income 16.5% → 21%. Geography: Russia's share 77.6% (Q1 2025) → 72.4% (Q1 2026); over three years, remittances from South Korea grew 2.1-fold, from Turkey 2.5-fold, from EU countries 2.4-fold. For 2024, the Central Bank of Uzbekistan disclosed the absolute size of the corridor with Russia — $11.5 billion (+29% on 2023) — and then switched to shares. There is no strict mirror check: no country breakdown of remittances from the Russian side for 2025–2026 could be found in the public domain. The available check is internal consistency: the $13.2 billion balance appears in the balance of payments, and household sales of foreign currency to banks ($18.9 billion for 9M-2026, 1.2 times more year on year) appear in the foreign exchange market review.

10.4. Independent growth estimates

The range of external estimates of the actual rate is 5.6–8.7%; forecasts for 2026 are 6.8–7.9%. No institutional estimate implies a multiple gap between reported and actual growth; only World Economics is systematically below the official data, and its method (scoring based on surveys and alternative data) itself carries great uncertainty. For the article's arguments this means robustness to measurement error: if actual growth is closer to 6%, the ratios of imports to GDP, the exchange rate gap and credit velocity change quantitatively but not qualitatively.

10.5. Alternative returns and the "retention price" of the deposit spring

Inflation expectations and estimated real return. In the Central Bank's September 2026 survey, households put price growth over the previous 12 months at 9.2% and expected prices to rise 8.7% over the next 12 months; in Tashkent expected inflation was 11.6%. Official annual inflation in September was 6.4%. These are different measures: perceived inflation refers to the past period, expectations to the future, and the CPI to the average basket.

The real return under the Fisher formula at a deposit rate of 18.4% is about 11.3% against official inflation of 6.4%; about 8.9% at household expectations of 8.7%; about 6.1% at Tashkent residents' expectations of 11.6%. Scenario deposit rates give the following picture:

The figures of 14%, 12% and 10% are scenarios, not forecasts of bank rates. They show that the appeal of soum deposits depends not only on the official CPI but also on the inflation depositors expect. The survey itself does not prove that the official index is unreliable.

Interpretation: while real deposit returns are positive, alternatives may lose out once risk and liquidity are taken into account; the shift into gold bars in 2025 was small relative to a deposit base of about $24 billion. A risk scenario arises if nominal rates fall faster than inflation expectations: cash foreign currency, gold bars, housing and current consumption may become more attractive. This is a scenario to monitor, not an established outflow of deposits. Indicators: net household purchases of foreign currency, bar sales, the share of foreign-currency deposits, inflows and outflows of time deposits, deposit maturities, new loan disbursements and repayments, and the gap between supply and demand in the foreign-exchange market.

Sources:

• Central Bank: deposit and lending rates, January–August 2026.
• Central Bank / Kun.uz: household and business inflation expectations for September 2026, 03.11.2026.
• Gazeta.uz: Central Bank statement on rate cuts, 04.11.2026.
• Gazeta.uz: decision on the main rate and inflation, 05.11.2026.
• Gazeta.uz: sales of gold bars to households, 07.04.2026.

11. Asset revaluation, collateral capacity and currency risk

This section separates three phenomena that must not be conflated: market revaluation of property, administrative mass appraisal and the creation of new money through bank credit.

11.1. Mass appraisal: the state's goal, the base and bank collateral

• Presidential Decree No. UP-43 of 05.03.2025 expressly provides that property valuation for taxation, purchase and sale, pledging and other transactions must be proportionate. Its stated goals also include determining the value of more than 8 million residential and non-residential properties and roughly 7 million land plots as economic assets, and creating a regularly updated market database. Collateral use is therefore a declared goal of the system, not merely an assumption about a side effect.
• Preliminary results of the mass appraisal in Tashkent: more than 600,000 apartments — 426.7 trillion soums against the previous cadastral base of 57.8 trillion; more than 213,000 private houses — 548.4 trillion against 39.1 trillion. The ratios to the previous cadastral base are about 7.4 and 14.0 times respectively. This is not a 7–14-fold rise in market prices but a revision of an outdated administrative base toward values closer to the market.
• The National Residential Property Price Index for Q3 2026: 104.8 y/y nationwide; 105.1 y/y for Tashkent apartments. Actual market indices show moderate growth, not a seven- or fourteen-fold increase in price.
• Individual valuation existed before the mass system. The Law "On Valuation Activity" provided for independent valuation organisations, qualified appraisers and a written report as the result; market value was defined as the most probable price in an open competitive market. The Law "On Mortgage" allowed the value of the mortgage object to be set by agreement between pledgor and pledgee or by valuation under the rules of valuation legislation. In the version of Central Bank Regulation No. 3269 in force until 24.07.2025 (para. 36), a bank could use a valuation organisation's report or the price in the purchase-and-sale contract to calculate LTV. The old cadastral base was therefore not the only or mandatory reference for bank collateral valuation.
• The national valuation standard requires identifying the object and collecting data on its characteristics; if the object was not inspected, the appraiser must state in the report the reasons for not inspecting it, along with assumptions and limitations. This does not mean that every bank and every borrower in all cases necessarily ordered an on-site appraisal: for mortgages, the rules also allowed the purchase-and-sale contract price.
• Mass appraisal, the previous cadastral value, individual market valuation, the transaction price and a bank's liquidation value are different figures. The ratios of 7.4 times for apartments and 14 times for houses compare the new mass appraisal with the previous cadastral base, but they do not show how much the value of collateral already recognised by banks before the reform has changed. A bank takes into account the asset's liquidity, legal title, internal discounts and the borrower's ability to pay.
• The testable question of the reform: whether banks use the mass appraisal as an additional reference, a starting base or a replacement for individual valuation; and whether the average recognised collateral value, LTV on new loans and lending against existing property have changed since its introduction. Without these data, it cannot be claimed that the banking system's collateral capacity has grown 7–14-fold.
• The current general LTV cap on mortgage loans to individuals is 80%, with limited exceptions and a separate regime for some refinanced mortgages. This limit cannot be automatically extended to all types of property-backed loans.
• The risk mechanism: if the new valuation base is widely adopted by banks, the recognised value of some existing collateral may rise; if rates fall, borrowers may find refinancing or additional credit more accessible. But actual lending is constrained by income, debt burden, tenor, collateral liquidity, legal title, capital and bank policy. A sevenfold ratio to the previous cadastral base does not mean a sevenfold increase in the credit limit.
• IMF: in March 2026 bank deposits rose 33.1% y/y; deposit dollarisation fell to 20.4%; the loan-to-deposit ratio declined and bank liquidity strengthened. Credit growth was uneven: microloans +37.9%, mortgages +17.2%, corporate credit +12.1%.
• At the end of 2025, nine state-owned banks controlled about 63% of assets and 67% of loans in the banking system and just over half of deposits. Continued directed and subsidised lending may weaken the link between the market cost of funding and the destination of new loans.

11.2. The collateral capacity formula

For a simple scenario analysis, the theoretical ceiling on a loan against the collateral value recognised by a bank can be written as:

C^max_collateral=LTV_max× V^recognized_collateral-L^prior_secured,

where V^recognized_collateral is the value the bank accepts after checks and discounts; LTV_max is the applicable limit/internal policy; L^prior_secured is existing claims secured by the same asset.

This is only the collateral ceiling. The actual loan amount is further constrained by the borrower's income, DSTI/DTI, tenor, collateral liquidity, legal title, valuation quality, risk concentration and the bank's capital. A rise in appraised value therefore does not equal an equivalent rise in actual lending.

If the recognized collateral value rises, the loan-to-value ratio (LTV) on the current loan mechanically falls. That may open the door to additional borrowing or refinancing, but only if the borrower passes an affordability assessment. If income has not grown, revaluation by itself does not create a source of servicing for new debt.

11.3. How to tell revaluation of wealth from money creation

1. Revaluation of an asset already owned raises the accounting or market value of wealth, but does not create a deposit and does not increase M2 by itself.
2. Selling an asset without new credit redistributes existing money from buyer to seller.
3. A new bank loan, on issuance of which the bank credits funds to the borrower's account, creates a new deposit. This is precisely where revaluation can turn into monetary expansion: a larger collateral base allows additional financing to be raised, provided the bank and the borrower pass the remaining constraints.
4. A rise in market price may reflect future rental income, profits and productivity growth; or lower discount rates and speculative demand. To distinguish between them, one needs price-to-rent, household price-to-income, corporate price-to-earnings, actual transactions and lending against assets that already exist.

The comparison with the US is methodological, not direct: the AIERA article "What will happen to $84 trillion of American wealth" uses DFA and Z.1 data on the increase in net assets of the US household and nonprofit sector. Between Q1 2020 and Q2 2026, the increase in the Z.1 net worth change table (release of 11.09.2026) was $86.7 trillion, of which $71.9 trillion was holding gains (revaluation of assets already existing) and $15.1 trillion was accumulation through transactions and saving. The "$84 trillion" in the article's headline is the increase on the DFA perimeter (levels $101.78 → $185.58 trillion), which covers households only; Z.1 also includes nonprofit organizations (levels $109.2 → $195.9 trillion), so the totals on the two perimeters do not match. This is not equal to the increase in bank credit and does not prove that the entire rise was driven by collateral-based lending.

11.4. Currency asymmetry: what exactly needs to be measured

• Under the IMF classification, Uzbekistan's regime is de jure and de facto floating; the exchange rate is set by supply and demand on the foreign exchange market. So calling the soum a "derivative of the dollar" is literally incorrect: it is not pegged to the dollar at a fixed parity and is not a dollar monetary liability. It is more accurate to speak of a currency hierarchy and an external constraint: imports and servicing of foreign-currency debt require foreign currency, which the Central Bank cannot issue.
• The IMF notes a significant share of foreign-currency loans to corporate borrowers without natural or market hedging, as well as around 20% of companies with insufficient profit to cover interest expenses. This creates vulnerability to soum depreciation among debtors with foreign-currency liabilities and soum revenue; companies with export foreign-currency revenue are partly naturally hedged.
• Sovereign external debt should not be lumped together with corporate FX risk: according to the IMF, a significant part of sovereign external debt is owed to official creditors, has long maturities and concessional/semi-concessional terms, which reduces refinancing risk, although currency revaluation remains.
• For a scenario test, the following should be modelled separately: (a) a fall in the market value of collateral; (b) a change in the soum exchange rate; (c) the borrower's share of debt in foreign currency; (d) the currency of its revenue; (e) the import share of costs; (f) time to maturity. The same soum depreciation affects an exporter with dollar revenue, an importer with soum revenue and a household with a soum mortgage differently.

11.5. The Central Bank's official real estate stress test

The Central Bank's Financial Stability Report for the first half of 2025 contains a stress test specifically for a fall in residential real estate prices. This is a model-based assessment, not a forecast and not the current state of the banks. The cohort examined is mortgage loans issued in the first half of 2025; the price-fall scenario is added on top of an adverse macro scenario. The −30% and −50% shocks are alternative variants layered on a single macro path, not steps of one cascade; they are not additive.

The Central Bank notes that with a 50% price fall, the probability of losing the collateral rather than having the loan repaid becomes high for a significant share of the mortgages examined. The results show the sensitivity of bank capital to collateral value, but do not prove that such a shock is likely. Nor do they automatically include a separate soum devaluation scenario; a currency-amplified stress would require a joint model of collateral price and the borrower's FX position.

11.6. Sources

• Ministry of Economy and Finance: preliminary results of the mass valuation of real estate in Tashkent, 14.08.2026; the initial valuation of private houses was 587 trillion soums. Legal Information Agency: subsequent clarification of the results, October 2026; the revised valuation of 548.4 trillion soums is given.
• State Statistics Committee/SIAT: residential real estate price index, year on year, Q3 2026, updated 28.10.2026.
• Law of the Republic of Uzbekistan "On Valuation Activity", version as of 25.07.2026: independence of the valuation organization, qualification requirements for the valuer, determination of market value and formalization of results in a report.
• Law of the Republic of Uzbekistan "On Mortgage", Article 13: the value of the mortgage item is determined by agreement of the parties or by valuation in accordance with the legislation on valuation activity.
• Central Bank Regulation No. 3269 on the terms of mortgage lending, version as of 28.07.2021, paragraphs 34–37: valuation of the mortgage item under the legislation on valuation activity; for LTV purposes the bank could use the valuation organization's report or the price in the sale and purchase agreement. Paragraph 36 was amended from 24.07.2025.
• Unified National Valuation Standard No. 3487, section on real estate valuation: requirements for identifying the object, collecting information and disclosing the reasons and limitations where no inspection was carried out.
• Central Bank: LTV ratio, updated 10.03.2026.
• IMF: 2026 Article IV Consultation, Staff Report, 2026: deposit growth, loan-to-deposit ratio, the structure of credit expansion and foreign-currency lending to unhedged companies.
• IMF: Informational Annex — Exchange Rate Arrangements, 2026: classification of the exchange rate regime as floating.
• IMF: Debt Sustainability Analysis, 2026: structure and maturities of sovereign external debt.
• Central Bank: Financial Stability Report for H1 2025, November 2025, pp. 97–99 by the report's numbering: stress test of a 30% and 50% fall in housing prices.
• AIERA Frontiers: "What will happen to $84 trillion of American wealth", 01.11.2026 — the author's analysis; window Q1 2020 – Q2 2026; increase in the Z.1 net worth change table — $86.7 trillion (holding gains $71.9 trillion, transactions and saving $15.1 trillion; Z.1 release of 11.09.2026); DFA levels $101.78 → $185.58 trillion versus Z.1 $109.2 → $195.9 trillion (different accounting perimeters).
• IMF COFER: the dollar's share of global official foreign exchange reserves in Q2 2026 — 56.7%, 30.09.2026.
• The Fed: The International Role of the U.S. Dollar — 2025 Edition, 18.07.2025; the international role of the dollar in reserves, payments, bank claims/liabilities and foreign-currency debt.
• Legal Information Agency: revised results of the mass valuation of housing in Tashkent, October 2026; according to a later clarification, the valuation of private houses was 548.4 trillion soums against a cadastral value of 39.1 trillion.
• Central Bank: Financial Stability Report for H1 2025, November 2025; the gap between the market and fundamental price of housing narrowed from 17% to 4%, the market price declined by 4% y/y, while rents rose by 12%.
• The Fed: International Role of the U.S. Dollar — accessible data, 18.07.2025.

12. Institutional complexity, human capital and the theoretical framework

12.1. Institutional coordination

• In May 2026 the government reported 5,650 functions of state bodies, more than 42,000 mandatory requirements for business and 1,041 types of public services. The same statement identified redundant procedures, paperwork and duplicative requirements as a problem, and the "Eliminating Bureaucracy — 2030" programme as the reform track. These numbers show the scale of the administrative system, but they are not an independent measure of the actual delays on any individual project.
• An empirical assessment of bottlenecks requires data on the time needed to obtain permits and utility connections, the duration of procurement, repeat approvals, the number of changes to project documentation, the share of facilities commissioned late, and the time between construction readiness and the start of output.

12.2. Human capital and skills

• OECD PISA 2022: the share of 15-year-old students in Uzbekistan reaching at least Level 2 was 19% in mathematics, 14% in reading and 19% in science; the OECD averages are 69%, 74% and 76%. These are the results of schoolchildren tested in 2022; they do not directly describe the skills of today's workforce.
• These data can be used as an indicator of a long-term task in workforce training, but they must not be turned into a direct estimate of enterprise productivity or of the causes of low localisation.
• SWP (Andrea Schmitz, 2023) describes the role of religious online preachers and their audience in Uzbekistan; a qualitative study by Sayfitdinova (2025) examines individual spiritual-educational channels and identifies the risks of oversimplifying complex theological and social questions. These works do not measure the prevalence of videos in which religious interpretation is presented as an explanation of physical laws. This example in the article therefore flags a problem of knowledge verification, not a statistically established cause of low productivity or a characterisation of the population's religiosity.

12.3. Theoretical sources

• Irving Fisher, “The Debt-Deflation Theory of Great Depressions”, Econometrica, 1933: debt deflation and the growing real burden of obligations as prices fall.
• Hyman P. Minsky, “The Financial Instability Hypothesis”, Working Paper No. 74, Levy Economics Institute, 1992: the distinction between hedge, speculative and Ponzi finance and endogenous financial instability.
• Steve Keen, “A Monetary Minsky Model of the Great Moderation and the Great Recession”, Journal of Economic Behavior & Organization, vol. 86, 2013, pp. 221–235: a model of endogenous money, credit and financial instability based on Minsky's hypothesis.
• Steve Keen, Robert U. Ayres, Russell Standish, “A Note on the Role of Energy in Production”, Ecological Economics, vol. 157, 2019, pp. 40–46, DOI 10.1016/j.ecolecon.2018.11.002: energy as a necessary condition for labour and capital to function; exergy as the physical basis of useful work.
• Bank of England: Money creation in the modern economy, 14.03.2014: a description of how commercial banks create deposit money when they issue loans.
• The credit impulse should be calculated from the change in net lending (new disbursements minus repayments) over a comparable period, not from a single outstanding loan portfolio balance. The stock of debt, the increase in debt and the flow of credit are different measures.

12.4. Historical comparisons — not a forecast

• Japan, second half of the 1980s: a Bank of Japan study and the work of Ito and Iwasako link the initial rise in asset prices to a sharp increase in bank lending to real estate and to the interaction between land values and companies' collateral capacity. This illustrates the "credit — collateral — prices" channel, not a direct forecast for Uzbekistan.
• Thailand, 1997: the IMF noted a combination of an external deficit, short-term capital inflows, rising external debt, unhedged foreign-currency borrowing by the private sector, an inflated property market and weak banks. Differences in the exchange-rate regime and in the structure of liabilities mean this case cannot be mechanically transferred to Uzbekistan.
• Spain, 2000–2007: according to IMF estimates, private credit to GDP almost doubled, real house prices roughly doubled, and construction employment rose by 47% against 27% in the economy; the market reversal was followed by defaults and bank losses. The IMF study points to heightened risk during prolonged credit booms accompanied by rapid construction growth.

Sources:

• Leonid Krutakov: "Oil and the World. The Seven Sisters — the House That John Built", RIPOL klassik, 2025; the publisher's description of the book.
• Government of Uzbekistan: measures to cut bureaucracy, May 2026.
• OECD: PISA 2022 Country Note — Uzbekistan, 05.12.2023.
• SWP: Religious Policy in Uzbekistan — Between Liberalisation, State Ideology and Islamisation, Andrea Schmitz, 21.08.2023.
• Sayfitdinova: Spiritual Educational Leaders in Uzbekistan’s Blogosphere — A Case Study Approach, 29.11.2025.
• Bank of Japan: Explaining Asset Bubbles in Japan, 1996.
• IMF: Lessons from Southeast Asia, 13.11.1997.
• IMF: Construction Activity Can Signal When Credit Booms Go Wrong, 12.02.2020.
• Hyman P. Minsky: The Financial Instability Hypothesis, 1992.
• Irving Fisher: The Debt-Deflation Theory of Great Depressions, 1933.
• Steve Keen: A Monetary Minsky Model of the Great Moderation and the Great Recession, Journal of Economic Behavior & Organization, 2013.
• Keen, Ayres, Standish: A Note on the Role of Energy in Production, 2019.
• Bank of England: Money creation in the modern economy, 2014.

12.5. How to read the comparison table in the main article

The comparison of Japan, Thailand, Spain and Uzbekistan in the main article is a comparison of mechanisms, not a single quantitative ranking of crisis readiness. The countries differed in exchange-rate regimes, banking structures, the role of domestic and external funding sources, the timing of their credit booms and their methods of property valuation.

• Japan (1985–1991): the key channel was domestic bank credit directed into real estate and the rising value of land as collateral. This is not equivalent to Thailand's model of short-term external foreign-currency debt. The Bank of Japan and the Ito–Iwasako studies link the initial rise in asset prices to the expansion of bank lending to real estate and to the interaction between land values and credit availability.
• Thailand (1995–1997): a combination of an open capital account, short-term inflows, a fixed exchange rate, foreign-currency borrowing without sufficient hedging, a current-account deficit and weak supervision of financial institutions. The devaluation after the baht's peg was abandoned increased the real burden of foreign-currency liabilities.
• Spain (2000–2007): cheap real interest rates after euro entry, rapid expansion of private credit, mortgages and lending to developers, and external wholesale funding. Regional cajas accumulated heightened risks because of concentrated real-estate lending and weak management.
• Uzbekistan (2026): the 69.8% figure is made up of the foreign investment category and two types of foreign loans; it cannot be described as the share of external debt or of short-term foreign-currency funding. The credit/GDP ratio of around 33.9% is a level at the chosen reference date, not proof that Uzbekistan is already in the same phase of a credit boom as Spain or Thailand.
• The ratios of mass appraisal to the previous cadastral base (about 7.4 times for apartments and 14 times for houses) describe the difference between two valuation bases. They do not imply a seven- or fourteen-fold rise in market price and do not equal an automatic increase in the collateral value recognized by a bank. Testing this channel requires data on valuations actually accepted by banks, LTV at origination, and volumes of loans against existing real estate.
• The 9.1 trillion soum in procurement-law violations is the identified sum of violations, not proven damage and not direct statistics on cronyism. Assessing the effect of connections on resource allocation requires data on ultimate beneficiaries, related bidding participants, repeat winners, and contract outcomes.
• The foreign-currency portion of reserves, $5.92 billion as of October 1, 2026, is a stock at a date; the $6.6 billion gap between demand and supply in the domestic foreign-exchange market is a flow over nine months. These magnitudes cannot be subtracted from one another or read as a direct comparison of coverage.

Sources for the comparative block:

• Bank of Japan: Explaining Asset Bubbles in Japan, 1996; Yutaka Yamaguchi, Asset Price and Monetary Policy: Japan's Experience, 1999.
• IMF: Thailand Article IV Consultation, 25.06.1998; IMF: Lessons from Southeast Asia, 13.11.1997.
• IMF: Construction Activity Can Signal When Credit Booms Go Wrong, 12.02.2020; BIS: The 2008–14 banking crisis in Spain, 06.07.2023; European Commission: evaluation of Spain's financial assistance programme, 2015.
• UP-43 of 05.03.2025; State Statistics Committee/SIAT: housing price index, Q3 2026; State Statistics Committee: fixed capital investment for 9M-2026; Central Bank: domestic foreign-exchange market review for January–September 2026, as reported by Spot.

Related material

Main text: «Money on a Spring».