I am studying the term structure model and there is one statement in the reading that is quite confusing to me -

**"Equilibrium models require fewer parameters to be estimated relative to arbitrage-free models"**.

The reading introduces three models:

i) CIR model: a(b-r)dt + (sigma)(sqrt{r})dz

ii) Vasicek model: a(b-r)dt + (sigma)dz

iii) Ho-Lee model: theta(t) dt + (sigma)dz

For CIR and Vasicek model, we need to estimate value of a and b

For Ho-Lee model, we are using the market data to feed into the model (i.e. theta(t)) and no estimation of parameters is required(???)

Why does the reading say that the equilibrium model require fewer parameter to be estimated? I think I have some misconception here.