AI Analytics

Technical writing

Treasury TIC Data: Foreign Ownership of US Securities and the Federal Capital Flow Dataset

· AI Analytics
Federal DataTreasuryCapital FlowsFinance

The Treasury International Capital system is the United States government's authoritative ledger of cross-border portfolio investment. Every month it answers the question that bond market participants, policymakers, and sovereign debt analysts most want answered: who holds US Treasury securities, how much do they hold, and how is that changing? The answer — distributed across monthly tables, annual surveys, and Federal Reserve datasets — is more complicated than it first appears, and understanding its limitations is as important as understanding its contents.

What the Treasury International Capital System Is

TIC is jointly administered by the Treasury Department and the Federal Reserve, with data collected primarily through the Federal Reserve Bank of New York. It measures cross-border portfolio investment flows and positions in US securities — bonds, equities, and short-term instruments — as well as US investors' holdings of foreign securities. The system was formalized in the 1970s as the US current account deficit grew and the question of who was financing it became increasingly consequential for dollar stability and interest rate policy.

The fundamental distinction TIC draws is between flows (monthly net purchases and sales of securities across borders) and positions (the stock of securities held at a point in time). Monthly TIC reports emphasize flows; the annual surveys provide the definitive stock positions. Both are necessary for a complete picture. A country can show low monthly flow activity while holding a large accumulated stock position built up over prior years.

TIC data is collected from a network of US custodians, broker-dealers, and financial institutions that hold or transact in US securities on behalf of foreign principals. Because most cross-border transactions clear through a small number of large custodians, coverage is high. The primary gap is securities held physically outside the US financial system — a minimal concern for marketable Treasuries but more significant for certain structured products.

The Main TIC Reports

TIC produces several distinct datasets that serve different analytical purposes. Understanding which report to use for a given question saves significant time.

Major Foreign Holders of Treasury Securities. This is the most-watched single table in the TIC system. Published monthly with roughly a 45-day lag, it shows the top foreign holders of US Treasury securities by country, expressed in billions of dollars. The table is the primary empirical basis for tracking whether Japan, China, or other large reserve holders are accumulating or reducing their Treasury positions. It is available as an Excel download from the Treasury's website and as FRED series for specific countries.

Monthly TIC Capital Flows (Forms TIC-S and TIC-B). These reports capture net purchases of long-term domestic securities by foreigners (TIC-S) and transactions in short-term domestic securities (TIC-B). TIC-S is the data used to assess whether foreign demand for US long-term bonds — Treasuries, agency securities, corporate bonds, equities — is strengthening or weakening in a given month. It provides a flow view rather than the stock view of the major holders table. A negative TIC-S print means foreigners were net sellers of US long-term securities that month, which markets interpret as a signal of reduced external demand.

Annual Survey of Foreign Portfolio Holdings of US Securities (SHCA).The most comprehensive TIC dataset. Conducted every June with reference date of December 31 of the prior year, SHCA is a census of all foreign holdings of US long-term securities — not just Treasuries but also agency securities, corporate bonds, and equities — broken down by country, security type, and maturity. SHCA attempts to attribute holdings to the beneficial owner country rather than the custodian country, making it analytically superior to the monthly data for structural questions about who actually owns what. The tradeoff is a six-month reporting lag after the reference date.

Annual Survey of US Holdings of Foreign Securities (SHLA). The mirror image of SHCA. This survey measures US investors' portfolio holdings of foreign securities as of December 31, broken down by country, equity versus debt, and maturity. The SHLA data shows the outward face of US portfolio investment — the several trillion dollars in foreign equities and bonds held by US mutual funds, pension funds, and individuals — and is the primary source for bilateral investment position statistics published by both the US and the IMF.

The Top Foreign Holders of US Treasuries

As of 2024, the five largest foreign holders of US Treasury securities are Japan (approximately $1.1 trillion), China (approximately $800 billion), the United Kingdom (approximately $700 billion), Luxembourg (approximately $400 billion), and the Cayman Islands (approximately $390 billion). The combined foreign official and private sector holds roughly $8 trillion of the approximately $27 trillion in outstanding Treasuries, making foreign holders collectively the second-largest creditor of the US federal government after the Federal Reserve.

Japan's position at the top of the table reflects decades of current account surpluses and the investment preferences of Japan Post Bank, the Government Pension Investment Fund, and Japan's commercial banks. Japanese institutional investors are structurally long-dated, dollar-denominated assets because Japanese interest rates have been near zero for much of the past three decades; US Treasuries offered yield that domestic Japanese government bonds could not.

China's trajectory is more instructive for analysts tracking geopolitical financial risk. Chinese holdings peaked at approximately $1.3 trillion in late 2013, when China was accumulating foreign exchange reserves at a rapid pace as a result of its managed exchange rate policy. Since then, holdings have declined steadily — a combination of deliberate reserve diversification, currency defense operations during the 2015–2016 RMB depreciation episode, and a more recent pattern that analysts describe as quiet decoupling from dollar assets. The decline is regularly cited in discussions of US-China financial interdependence and the weaponization of financial sanctions following the freezing of Russian reserves in 2022.

The United Kingdom, Luxembourg, and the Cayman Islands appear in the top five not because British, Luxembourgish, or Caymanian entities are the ultimate investors but because those jurisdictions host major financial centers and custodians. This is the custody problem discussed below.

The Custody Problem: Why Country Labels Are Misleading

TIC monthly data is collected on a custodian country basis, not a beneficial owner basis. When a foreign entity purchases US Treasury securities, the transaction is attributed to the country where the custodian holding the securities is located. For direct central bank purchases settled in New York, the attribution is clean. For securities held through international custodians such as Euroclear in Belgium or Clearstream in Luxembourg, the attributed country is Belgium or Luxembourg regardless of who the actual investor is.

The most discussed manifestation of this problem is the “Belgium anomaly.” Between 2013 and 2015, Belgian holdings of US Treasuries surged from roughly $170 billion to $354 billion — an implausible increase for a country of Belgium's economic size. The consensus interpretation, consistent with simultaneous movements in Chinese reserves and the timing of Chinese intervention in currency markets, is that Chinese official entities were purchasing Treasuries through Euroclear rather than through direct custody in New York, causing transactions to be attributed to Belgium in the monthly TIC data. The “Belgium problem” became shorthand for the difficulty of inferring actual beneficial ownership from TIC monthly statistics.

The Annual Survey (SHCA) partially corrects this by instructing respondents to attribute holdings to the country of the beneficial owner to the extent practicable. SHCA figures for Belgium are substantially smaller than the monthly custody-based data would imply, while SHCA figures for China are correspondingly larger. For structural analyses of which countries own US securities, SHCA is more reliable than the monthly major holders table. For monitoring month-to-month changes in flows, the monthly data is the only option despite its custody-attribution limitation.

Why China's Treasury Holdings Matter

China's Treasury position is the most politically discussed number in the TIC dataset. The analytical backstory is worth understanding precisely. China accumulated large foreign exchange reserves as a structural byproduct of its exchange rate management regime. To keep the RMB from appreciating against the dollar as China ran large trade surpluses, the People's Bank of China purchased dollars in the foreign exchange market. Those accumulated dollars had to be invested somewhere; US Treasury securities — deep, liquid, dollar-denominated, and essentially default-risk-free — were the natural destination. By 2013 China held approximately $3.9 trillion in total foreign exchange reserves, with Treasuries representing the core of the dollar-denominated portion.

The “financial nuclear option” framing — that China could dump its Treasuries to punish the United States — circulates in policy discourse but does not survive serious analysis. A rapid sale of $800 billion in Treasuries would drive prices down and yields up, imposing substantial capital losses on China itself. It would simultaneously cause RMB appreciation, undermining the export competitiveness that motivates China's reserve accumulation in the first place. And the buyers for $800 billion in liquidated Treasuries would primarily be other dollar-reserve holders, limiting the long-run yield impact. The more consequential scenario — a gradual diversification away from Treasuries over years, reducing Chinese marginal demand at the margin as US deficits continue to expand — is precisely what the TIC data has been recording since 2013.

The freezing of approximately $300 billion of Russian sovereign reserves following the 2022 invasion of Ukraine accelerated reserve diversification discussions globally. Countries that hold large dollar-denominated reserves now confront the demonstrated risk that those assets can be immobilized by the United States government. TIC data monitoring became more acute among analysts tracking whether this event accelerated Chinese or other emerging market reserve shifts away from dollar assets.

TIC and the Sudden Stop Risk

The United States runs a current account deficit of roughly $800 billion to $1 trillion per year, representing the net flow of goods, services, and income to the rest of the world. Every dollar of current account deficit must be financed by an equivalent capital account surplus — foreigners acquiring US financial assets. TIC data is the primary instrument for monitoring whether this financing is stable.

A “sudden stop” — a rapid reversal of foreign capital inflows — is the canonical external financing crisis. In emerging market episodes (Mexico 1994, East Asia 1997, Russia 1998, Argentina 2001), sudden stops forced sharp currency depreciations, interest rate spikes, and deep recessions as domestic investment collapsed with the withdrawal of foreign capital. The question of whether the United States could experience an analogous sudden stop has motivated TIC surveillance since the US current account deficit widened in the 1990s.

The 2008 financial crisis provided a natural experiment. Contrary to the feared scenario in which foreigners fled US assets as the US financial system seized up, TIC data showed strong foreign demand for US Treasury securities throughout the acute phase of the crisis. The dollar appreciated, Treasury yields fell, and foreign holdings of Treasuries increased. The explanation is the dollar's reserve currency status: in a global crisis, the flight-to-safety trade goes into dollar assets, not out of them. This dynamic — the dollar smiling — is a structural feature that makes a classic sudden stop substantially less likely for the US than for any other country running a comparable current account deficit.

The scenario that TIC analysts monitor more carefully is not a sudden stop but a gradual erosion: a slow reduction in foreign central bank demand for Treasuries as reserve diversification proceeds, coinciding with expanding US fiscal deficits, that requires domestic investors to absorb an increasing share of Treasury supply at potentially higher yields. TIC monthly data provides the earliest observable signal of this shift.

Data Structure and Access

TIC data is published at home.treasury.gov/data/treasury-international-capital-tic-system. The main access formats are:

For time-series analysis, FRED (Federal Reserve Economic Data) provides convenient access to key TIC series. The series HQFLUSQ232S tracks total foreign holdings of US Treasury securities on a quarterly basis. Country-specific series include China (HQCNULQ), Japan, and other major holders. FRED series update automatically and integrate cleanly with quantitative analysis workflows.

US Holdings of Foreign Securities (SHLA)

The outward side of the TIC picture is equally striking in scale. As of recent annual surveys, US investors hold more than $16 trillion in foreign securities — roughly $12 trillion in equities and $4 trillion in debt instruments. The equity holdings are concentrated in developed markets: European equities (UK, Germany, France, Netherlands, Switzerland) and Japanese equities account for the majority, reflecting the fact that US mutual funds and pension funds hold internationally diversified portfolios dominated by MSCI World index constituents.

Emerging market equity exposure in the SHLA data is substantial in absolute terms but modest relative to market capitalization: US investors hold a smaller share of Chinese, Indian, or Brazilian equities than a market-cap-weighted benchmark would imply, reflecting both explicit investment restrictions (QFII quotas, MSCI exclusions) and investor risk preferences. The SHLA data provides the empirical basis for quantifying home bias in US international portfolio allocation.

Foreign debt holdings in SHLA are dominated by corporate bonds and sovereign debt from investment-grade issuers. The maturity profile of US holdings of foreign debt is shorter than the maturity profile of foreign holdings of US debt — the asymmetry reflects the US Treasury's role as the global reserve asset, with foreign central banks holding long-dated Treasuries while US investors take on shorter-duration foreign corporate exposure.

Python: Downloading and Analyzing the Major Foreign Holders Data

The following script downloads the Treasury's major foreign holders Excel file, parses the country-level monthly series, and produces a summary of the top ten holders over the most recent five-year window. The Treasury publishes this file at a stable URL; the structure has been consistent for over a decade, though the exact row offsets occasionally shift when Treasury reformats the header.

import requests
import pandas as pd
import io

# Treasury TIC: Major Foreign Holders of US Treasury Securities
# The Excel file is published monthly at home.treasury.gov
# We download the current file, parse the country-level monthly series,
# and produce a summary of the top 10 holders over the past 5 years.

TIC_URL = (
    "https://home.treasury.gov/system/files/276/mfh.xls"
)

resp = requests.get(TIC_URL, timeout=60)
resp.raise_for_status()

# The workbook has a single sheet; skip the header rows
# Row 0 is the title, rows 1-3 are sub-headers, data starts at row 4
raw = pd.read_excel(
    io.BytesIO(resp.content),
    sheet_name=0,
    header=None,
)

# Find the row index where country data begins (first column is "GRAND TOTAL")
start_row = None
for i, row in raw.iterrows():
    if str(row.iloc[0]).strip().upper() == "GRAND TOTAL":
        start_row = i
        break

if start_row is None:
    raise ValueError("Could not locate GRAND TOTAL row in TIC workbook")

# Column headers are in the rows just above the data block
# Row start_row - 1 typically holds month-year labels
header_row = start_row - 1
date_labels = raw.iloc[header_row, 1:].tolist()

# Build the data frame: rows are countries, columns are months
data = raw.iloc[start_row:].copy()
data.columns = ["country"] + list(range(len(date_labels)))
data = data[data["country"].notna()].copy()
data["country"] = data["country"].astype(str).str.strip()

# Keep only meaningful country rows (exclude subtotal separators)
exclude_prefixes = ("GRAND TOTAL", "Of which:", "Europe", "Caribbean", "Asia",
                    "Latin America", "Africa", "Other")
mask = ~data["country"].str.startswith(exclude_prefixes)
countries = data[mask].copy()

# Convert holding columns to numeric (billions of dollars)
for col in range(len(date_labels)):
    countries[col] = pd.to_numeric(countries[col], errors="coerce")

# Identify columns that correspond to the most recent 60 months (~5 years)
# date_labels are strings like "Jan 2020" - parse and sort
parsed_dates = []
for idx, label in enumerate(date_labels):
    try:
        parsed_dates.append((idx, pd.to_datetime(str(label).strip())))
    except Exception:
        pass

parsed_dates.sort(key=lambda x: x[1])
recent_60 = parsed_dates[-60:] if len(parsed_dates) >= 60 else parsed_dates
recent_cols = [idx for idx, _ in recent_60]
recent_date_strs = [str(date_labels[idx]).strip() for idx, _ in recent_60]

# Compute average holdings over the past 5 years per country
countries["avg_5yr"] = countries[recent_cols].mean(axis=1)
countries_sorted = countries.nlargest(10, "avg_5yr")

# Build a clean output frame with country + monthly holdings for recent period
output_cols = ["country"] + recent_cols
output = countries_sorted[output_cols].copy()
output.columns = ["country"] + recent_date_strs

# Show latest snapshot and 5-year average for top 10 holders
latest_col = recent_date_strs[-1]
summary = pd.DataFrame({
    "country": countries_sorted["country"].values,
    "latest_holdings_bn": countries_sorted[recent_cols[-1]].values,
    "avg_5yr_holdings_bn": countries_sorted["avg_5yr"].round(1).values,
})

print("Top 10 Foreign Holders of US Treasury Securities")
print("Holdings in USD billions")
print()
print(summary.to_string(index=False))
print()
print("Earliest month in series:", recent_date_strs[0])
print("Latest month in series: ", latest_col)

A few notes on the implementation. The major holders file uses a header structure with multiple merged rows above the data table; locating the “GRAND TOTAL” row programmatically is more robust than hard-coding a row offset that can shift between releases. Holdings are reported in billions of dollars. Countries listed with asterisks or footnotes (typically indicating estimates or revisions) parse correctly as numeric values. The monthly column labels are inconsistently formatted across vintages of the file, so pd.to_datetime with error-suppression is more reliable than fixed-format parsing.

For multi-year trend analysis, looping over archived TIC files (Treasury maintains a historical archive) or using FRED API series is more efficient than parsing successive vintages of the major holders Excel file. The FRED approach is particularly clean for country-specific series where the series ID is stable.

Interpreting Monthly TIC Prints

Monthly TIC-S headlines — “foreigners bought $X billion in US long-term securities” — are regularly quoted in financial media but require context to interpret. Several factors complicate month-to-month comparison:

For these reasons, analysts focused on structural trends in foreign Treasury demand treat TIC monthly data as a noisy indicator best interpreted over three-to-six-month rolling windows, and use the annual SHCA survey data for definitive stock positions.

TIC in Macroeconomic Context

TIC data does not stand alone. Its significance depends on the macroeconomic context that drives the flows it measures. Three external data series are essential companions:

The BEA current account sets the financing requirement. If the current account deficit narrows, less capital inflow is required to fund it, reducing pressure on foreign demand for US assets. If the deficit widens — as it has in recent years driven by goods trade deficits and income outflows — TIC inflows must increase commensurately or the dollar will adjust.

The Federal Reserve balance sheet determines how much domestic demand exists for Treasuries outside the foreign sector. During quantitative easing, the Fed purchased Treasuries directly, reducing the share that had to be absorbed by foreign buyers. During quantitative tightening, that cushion is removed. TIC data must be read in light of the Fed's concurrent portfolio actions.

The dollar exchange rate both reflects and drives TIC flows. Dollar appreciation increases the local-currency value of dollar-denominated assets held by foreign investors, which can trigger rebalancing sales to maintain target allocation weights — showing up as negative TIC prints even when appetite for dollar assets is unchanged in currency-neutral terms. Dollar depreciation has the opposite mechanical effect.

TIC data captures the portfolio investment dimension of the US balance of payments. For the trade and income flows that drive the current account deficit TIC must finance, see the BEA National Income and Product Accounts. See BEA GDP Accounts: National Income, Product Accounts, and Quarterly Output Measurement.

Treasury also administers the OFAC sanctions program, which can freeze or block the foreign holdings that TIC measures — the mechanism that crystallized reserve diversification concerns globally in 2022. See Treasury OFAC: Sanctions Data and the Specially Designated Nationals List.

The domestic banking system that intermediates much of the Treasury market activity reflected in TIC flows is measured by the Federal Reserve H.8 Assets and Liabilities of Commercial Banks release. See Federal Reserve H.8: Bank Assets, Liabilities, and the Commercial Banking Balance Sheet.