My MPC-ometer model correctly forecast the 25bp December rate cut but indicated a narrow 5-4 vote rather than the 9-0 revealed this morning. The four vote “miss” is the largest since March and compares with an average model error of just one vote since I began to use it to forecast in “real time” in October 2006.

The minutes suggest the Committee was particularly concerned by renewed widening in interbank / Bank rate spreads during November: three-month LIBOR stood at 6.6% at the time of the December meeting. According to analysis presented in the Bank’s Quarterly Bulletin, this increase reflected rising credit risk premia rather than a shortage of liquidity, which was the dominant factor during earlier spread widening in August / September.

The MPC-ometer assesses whether the prevailing level of Bank rate is appropriate given incoming economic and financial news. It implicitly assumes a normal relationship between Bank rate and interbank rates, which are the key driver of borrowing costs faced by households and companies. Under current unusual circumstances, there is a case for using the model to assess the need for a change relative to three-month LIBOR rather than Bank rate. If I rerun the December forecast using the prevailing three-month rate of 6.6% rather than the 5.75% Bank rate, the forecast changes from 5-4 for a cut to 9-0, as actually occurred.

This suggests the next cut could occur as soon as January if interbank rates fail to fall back significantly early in the New Year. Some decline is likely but three-month LIBOR probably needs to move below 6% to justify my current forecast that a move will be delayed until February.

As explained here, I think the economic outlook warrants policy being set to achieve three-month rates of about 5.5% in early 2008.