Chow Test

Calculate the Chow test for a structural break at break_date in a regression of dependent_ticker on independent_tickers.

The sample is split at break_date into a “before” and “after” period, a regression is fitted on each part separately and on the full sample, and an F-test checks whether allowing the coefficients to differ between the two periods significantly reduces the residual sum of squares.

The formula is as follows:

• The second sub-period starts at the first date on or after break_date, and k includes the constant when add_constant=True

Where RSS_pooled is the residual sum of squares of the full-sample regression, RSS_1 and RSS_2 those of the two sub-periods, k the number of parameters and n the total number of observations. Under the null hypothesis of no structural break (the same coefficients apply throughout), F follows an F(k, n - 2k) distribution. A low p-value means the relationship changed at break_date, for example an asset’s beta shifting after a merger, a regulatory change or a market regime change, so a single regression over the full sample is misspecified. The break date must be chosen in advance; testing many candidate dates and keeping the most significant overstates the evidence.

For more information about the method, see the following paper:

• Chow, G.C. (1960). “Tests of Equality Between Sets of Coefficients in Two Linear Regressions.” Econometrica, 28(3), 591-605.

Also known as: Chow breakpoint test.

Chow Test in Python or via the MCP

Install the Finance Toolkit with:

pip install financetoolkit -U

Then call get_chow_test as shown below.

from financetoolkit import Toolkit

toolkit = Toolkit(
    ["NVDA"],
    api_key="FINANCIAL_MODELING_PREP_KEY",
    start_date="2021-01-01",
    end_date="2025-12-31",
)

# NVDA is dependent and the Benchmark independent, with a break at the launch of ChatGPT
toolkit.econometrics.get_chow_test(break_date="2022-11-30", include_benchmark=True)

Or, after setting up the Finance Toolkit MCP, ask:

Which returns:

Using Your Own Data

get_chow_test runs on any data, not only prices: ratios, economic indicators, yield curves or your own dataset. Pass a DataFrame, a Series or a dictionary of them to Econometrics as data, with the dates in the rows or the columns, and set changes to how its changes are calculated: "percentage", "log", "difference" for rates and ratios, or None when the data are changes already. For example, for the monthly unemployment rates of the United States and Germany, where each country is a column:

from financetoolkit import Economics, Econometrics

economics = Economics(start_date="2000-01-01")

unemployment = economics.get_unemployment_rate(
    countries=["United States", "Germany"], period="monthly"
)

econometrics = Econometrics(data=unemployment, changes="difference")

Then call econometrics.get_chow_test with break_date.

Parameters

get_chow_test accepts the following parameters.

Related Specification Tests