Get the random effects values from a `mblogit` model

Question

I want to fit a multinomial response mixed effects model. I'm using the mblogit function from the mclogit package.

From the help(mblogit) example I can fit a fixed effects model (which is structurally close enough to the real thing I want to fit, but 100 times smaller so that's good).

library(mclogit)
library(MASS) # For 'housing' data
house.mblogit <- mblogit(Sat ~ Infl + Type + Cont, weights = Freq, data=housing)

Printing this out shows the values of the fixed effects parameters:

                   Predictors
Response categories  (Intercept)  InflMedium  InflHigh  TypeApartment
         Medium/Low  -0.4192       0.4464      0.6649   -0.4357      
         High/Low    -0.1387       0.7349      1.6126   -0.7356      
                   Predictors
Response categories  TypeAtrium  TypeTerrace  ContHigh
         Medium/Low   0.1314     -0.6666       0.3609 
         High/Low    -0.4080     -1.4123       0.4818 

It's used "Low" as the baseline of the multinomial model and I've got a value for each level of the factors in the model relative to a baseline factor (ie Type has four levels and these are relative to "Tower").

If I try to use Type as a random effect parameter:

house.mblogit.mixed <- mblogit(Sat ~ Infl + Cont, random=~1|Type, weights = Freq, data=housing)

it converges and fits but I can't find the values of the random effect parameters or hyperparameters. The printed value drops the Type from the "Predictors" and adds an extra "(Co-)Variances" table:

Coefficients:
                   Predictors
Response categories  (Intercept)  InflMedium  InflHigh  ContHigh
         Medium/Low  -0.6910       0.4387      0.6609    0.3680 
         High/Low    -0.7704       0.7341      1.6089    0.4721 

(Co-)Variances:
Grouping level: Type 
          Medium~1  High~1 
Medium~1  0.08094          
High~1    0.12866   0.23883

What I'm expecting is to find three values (or four?) matched up to the levels of Type - ie Tower, Apartment, Atrium and Terrace - in the way that ranef on an nlme mixed effects model returns values for each level in a random effect factor...

> fm2 <- lme(distance ~ age + Sex, data = Orthodont, random = ~ 1|Subject)
> head(ranef(fm2))
    (Intercept)
M16  -1.7018336
M05  -1.7018336
M02  -1.3776748
...

What am I missing?

[More exploration...]

The returned object has a $random.effects component which seems undocumented. Its essentially a vector in this case but its unclear how that vector maps to combinations of response level and effect factor level.