The average interest rates of the 90-day CDTs (DTFs) closely track the behavior of Banco de la República’s benchmark rate, indicating a direct and stable relationship between the two over the period analyzed. However, the interest rate on home purchase loans (mortgage rate) has followed a different path: during 2023 and 2024, this rate mirrored the benchmark rate’s downward trend, but as of year-end 2024, even as the benchmark rate continued to fall, the mortgage rate stabilized and began to rise when it was surpassed by the ten-year government yield, following closely behind.
Publication date
In recent discussions about Colombia’s macroeconomic environment, the weakening relationship between Banco de la República’s (the Central Bank of Colombia) monetary policy rate (MPR, or the benchmark policy rate) and market interest rates has received comparatively little attention. Indeed, in recent months, yields on government bonds have exerted a stronger influence on long-term lending rates set by financial institutions, a development that has often gone unnoticed.
To illustrate this phenomenon, Graph 1 shows the behavior of average interest rates on 90-day term deposits (CDT for its Spanish acronym), the MPR, the ten-year government debt securities yield, and the average interest rate of home purchase loan disbursements (mortgages) from February 2023 onward. The graph displays nominal interest rates, which represent the annual financial cost per 100 pesos.
Graph 1. Nominal interest rate
Sources: Financial Superintendence of Colombia, Banco de la República, and Ministry of Finance and Public Credit.
The graph shows that the average interest rate on 90-day term deposits in the financial system (90-day CDT rate) closely tracked the MPR. However, the interest rate on housing loans (mortgages) follows a different path. During 2023 and 2024, the mortgage rate fell, mirroring the MPR’s downward trajectory and coinciding with a significant decline in inflation. Since the end of 2024, despite the continued decrease in the MPR, the mortgage rate stabilized; once it was surpassed by the interest rate on 10-year government bonds, the mortgage rate began to rise rapidly, closely tracking the 10-year government debt securities yield rather than the MPR.
It is important to underscore that, in the current scenario, the 10-year government bond yield is no longer in lockstep with the MPR. Both the interest rate on government bonds and the mortgage rate increased significantly throughout 2025, long before the BanRep raised its MPR. In fact, the steep increase in the MPR in 2026 did not appear to affect the mortgage rate, which continued to track the 10-year government bond yield closely. Note that the mortgage rate only starts to closely track the government bond yield once the latter surpasses it. This fact explains the disconnect with the MPR: when interest rates on government bonds exceed the loan interest rate, lenders will generally require returns comparable to those available on government bonds before extending long-term loans. This compensates the financial system for the opportunity cost of allocating depositors’ funds to loans rather than investing in government bonds.
The conclusions drawn from this graph are corroborated by a simple regression analysis. Table 1 presents the results of two regressions that use the mortgage rate as a dependent variable and the MPR and the 10-year government bond yield as independent variables, all in nominal terms. The coefficients in the table indicate the estimated effect of each independent variable on the mortgage rate. The asterisks next to them indicate which of these relationships are statistically significant.
Regression 1 studies the period from January 2003 to December 2024. Over this long horizon, both the MPR and the government bond rate yield are directly linked to the mortgage rate (in fact, the effect of the government bond interest rates was quantitatively greater than that of the MPR). Regression 2 focuses only on the following fifteen months: from January 2025 to March 2026. During this time, the relationship between mortgage rates and the MPR weakened considerably, while the relationship between mortgage rates and government bond rates remains strong and statistically significant (the estimated coefficient, 0.79, is nearly the same as in regression 1, 0.77).
This result indicates that the deterioration of the fiscal outlook not only increases the cost of servicing government debt but also raises the cost of mortgages and long-term investment financing across the economy.
Table 1. Results of simple mortgage rate regressions compared to the BanRep benchmark rate and the 10-year government bond rate
Note: OLS is the acronym for ordinary least squares, which is the simplest and most commonly used econometric method for estimating regressions, where n is the number of observations (sample size); the F-statistic helps determine the level of association between the independent variables and the dependent variable; and the adjusted R² indicates what percentage of the change in the mortgage rate is explained by the model.
Source: BanRep’s calculations.
These correlations are consistent with the notion that, in recent times – as exemplified in regression 2 – the mortgage interest rate is more responsive to government bond yields than to the MPR, in contrast to the historical behavior illustrated in regression 1)1. This result indicates that the deterioration of the fiscal outlook not only increases the cost of servicing government debt but also raises the cost of mortgages and long-term investment financing across the economy.
An extended version of this analysis can be found in Box 5 published in the April 2026 Monetary Policy Report, “Recent Shifts in Interest Rate Dynamics in the Colombian Economy”.
1 ↑ It is worth clarifying that these results do not reflect measures of elasticities, as the relationship between various interest rates in the economy involves a more complex dynamic structure that is beyond this descriptive scope. However, the disappearance of the correlation between the mortgage rate and the benchmark rate after 2024 is consistent with a weakening association between those rates in recent years.
