Occasionally Binding Constraints

[CF-Ssq]

Hi, I have a question about handling Occasionally Binding Constraints in RISE.

I know that the package supports several types of solution methods for occasionally binding constraints, but I’m wondering whether there is an equivalent in RISE to the equation tags + lmmcp setup used in Dynare when solving models under the perfect foresight algorithm. For equation tags I mean something like ['ir>0'] followed by, for example, a Taylor Rule.

In other words, is there a built-in or recommended approach in RISE for implementing models with occasionally binding constraints similar to how Dynare handles them with the LMMCP solver?

[jmaih]

Hi CF, In RISE, equation tags are not used to control or influence the solution method, unlike in Dynare where tags such as ['ir>0'] can guide the LMMCP solver. Tags in RISE are strictly descriptive; they are not part of the computational logic of the solver. RISE does support several approaches for modeling occasionally binding constraints (OBCs), including: - *Regime-switching representations* (constraints mapped to regimes with endogenous switching indicators) - *Penalty and relaxation techniques* - *Piecewise or indicator-based formulations* - and additional options available in the non-public version. RISE also includes a perfect foresight solver, but the implementation in the public version is not optimized for large or highly nonlinear OBC problems in the same way that specialized MCP/LMMCP algorithms are. For users who need perfect-foresight simulations with OBCs implemented exactly in the LMMCP style, there is currently *no direct analogue* in the public version of RISE. Best regards, Junior

…

On Tue, Nov 4, 2025 at 11:36 AM CF ***@***.***> wrote: *CF-Ssq* created an issue (jmaih/RISE_toolbox#222) <#222> Hi, I have a question about handling Occasionally Binding Constraints in RISE. I know that the package supports several types of solution methods for occasionally binding constraints, but I’m wondering whether there is an equivalent in RISE to the equation tags + lmmcp setup used in Dynare when solving models under the perfect foresight algorithm. For equation tags I mean something like ['ir>0'] followed by, for example, a Taylor Rule. In other words, is there a built-in or recommended approach in RISE for implementing models with occasionally binding constraints similar to how Dynare handles them with the LMMCP solver? — Reply to this email directly, view it on GitHub <#222>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AATKBT3DG26242SR7MJN2Z333B6S7AVCNFSM6AAAAACLCUYAMKVHI2DSMVQWIX3LMV43ASLTON2WKOZTGU4DKOJSGEZTQNI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[CF-Ssq]

Thank you for your sincere and direct reply. I understand it, and I just have one last question: is it possible to do this in non-public versions—for example, the most up-to-date ones you mentioned in response to one of my previous questions—or is it simply not available anywhere yet?

[jmaih]

Yes, the non-public version of RISE is the most advanced and handles those problems easily. Best, J.

…

On Fri, Nov 7, 2025 at 4:10 PM CF ***@***.***> wrote: *CF-Ssq* left a comment (jmaih/RISE_toolbox#222) <#222 (comment)> Thank you for your sincere and direct reply. I understand it, and I just have one last question: is it possible to do this in non-public versions—for example, the most up-to-date ones you mentioned in response to one of my previous questions—or is it simply not available anywhere yet? — Reply to this email directly, view it on GitHub <#222 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AATKBT2VOAUOPFI6J5QPSU333SY5RAVCNFSM6AAAAACLCUYAMKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTKMBTGA4TEMRWG4> . You are receiving this because you commented.Message ID: ***@***.***>