Skip to content
FinObservatory

Chinese overseas lending / BHS

Bahamas: debt owed to China

In 2021, Bahamas is estimated to have owed China $1.29bn, equal to 10.7% of its GDP, which ranks it 56 of 126 borrowers in the panel by dollars owed. The estimated stock peaked at $2.53bn in 2017.

$1.29bn
Estimated total, 2021
rank 56 of 126
10.7%
Percent of GDP
2021
$69m
Public and publicly guaranteed
2021
$1.23bn
Private non-guaranteed
94.7% of external

The estimated stock, 2000 to 2021

The largest single-year move in the estimated stock is 2013, when it rose by $370m.

Public and publicly guaranteedPrivate non-guaranteedPBoC swap drawings

Source: Horn, Reinhart and Trebesch, China’s Overseas Lending. Selection: estimated stock owed to China by Bahamas, by year and instrument, 2000 to 2021. Zeros are estimated zeros (no known loans outstanding), not missing values. Methodology

Historical loan records from the 2021 release

These records come from the 2021 HRT Mendeley replication release, not the November 2025 refreshed stock panel charted above. The refresh uses the updated Franz, Horn, Parks, Reinhart and Trebesch methodology, so this table cannot be reconciled to the chart. The earlier release marks 4 loan-level records for Bahamas as part of its debt-estimate sample, spanning 2010 to 2012. Comparable commitments are available for 4 of those loans and sum to $2.55bn. These are gross commitments, not outstanding debt stocks. The stock methodology applies disbursement and amortisation profiles, with imputation where loan terms are missing, so the commitment sum must not be compared with the stock above. See the methodology. Terms are sparse by source: interest rate appears on 2 loans, grace period on 0, maturity on 3, any of the three on 3, and all three together on 0. The earlier file also carries 4 additional project rows for Bahamas outside that sample, so they are not included in this table.

4
Loans in 2021 sample
2010 to 2012
4
Loans with CommitmentUSD
$2.55bn
3
Loans with any terms
0 with all three
4
Rows outside 2021 sample
8 raw rows total
YearProjectCommitment, USDCommitment, source currencyTermsCreditorType
2012
China EXIM loans $33 million to Bahamas for the construction of a port
ID: BHS_7 · Source: AidData, Inter-American Dialogue
$33mUSD 33,000,000rate 2% | grace n/a | maturity 20y
China Export-Import Bank (Ex-Im)
State-owned policy bank
Loan
Concessional · borrower Central government
2012
China gives a $6.5 million loan to Bahamas for bridge construction on Little Abaco
ID: BHS_8 · Source: AidData, Inter-American Dialogue
$6.50mUSD 6,500,000rate 2% | grace n/a | maturity 20y
China Export-Import Bank (Ex-Im)
State-owned policy bank
Loan
Concessional · borrower Central government
2011
China EXIM Bank loans $2.45 billion to Bahamas for the Baha Mar Resort
ID: BHS_3 · Source: AidData
$2.45bnUSD 2,450,000,000rate n/a | grace n/a | maturity n/a
China Export-Import Bank (Ex-Im)
State-owned policy bank
Loan
Commercial · borrower Private sector
2010
China loans Bahamas $58.6 million for the Airport Gateway Project
ID: BHS_2 · Source: AidData, Inter-American Dialogue
$59mUSD 58,600,000rate n/a | grace n/a | maturity 20y
China Export-Import Bank (Ex-Im)
State-owned policy bank
Loan
Concessional · borrower Central government

Source: Horn, Reinhart and Trebesch, China’s Overseas Lending (2021 Mendeley replication loan file). Selection: 2021 replication loan records mapped to Bahamas and flagged DebtEstimateSample = 1 in that release. This is an earlier source universe, not the records used for the November 2025 refreshed stock panel. CommitmentUSD is a gross commitment flow and is not comparable to the charted stock. Commitment stays in its source currency, and ProjectDescription is rendered as free-text data from the file. Methodology

Against what Bahamas reports owing to all creditors

No comparison is possible. Bahamas has no row in the World Bank International Debt Statistics table. There is no reported all-creditor stock to measure the China estimate against, so the ratio is n/a, not zero and not a hundred percent. This is a difference in the scope of the two datasets, not evidence of anything hidden. The two-sources page lists every borrower in this position.