Statistically speaking, the Office of National Statistics can't count

As a result of the ONS survey covering a tiny proportion (<20%) of consumption, it is inherently volatile and the BoE has repeatedly said it pays little attention to it, instead prefering to concentrate on their own more detailed survey (the BoE Regional Agents survey) which covers a much larger proportion of retailers and thus shows a higher correlation with the BRC measure.

Also worth noting that in late 2005-2006 the market was very bearish about UK growth prospects and the curve priced in easing, but the BoE measure of retail sales there clearly accelerated whereas the ONS measure was muted. The BoE subsequently tightened policy. In Spring 2008 as the ONS measure was showing strength, the BoE downplayed this, pointing out that both its measure and the BRC measure were showing weakness. There was no tightening.

Anyway, the point is that UK Retail Sales are a lot stronger than the ONS data suggests. The below chart clearly shows that the ONS figures broadly track the BoE survey over time but are much more volatile.

White Line - ONS Retail Sales
Yellow Line - Bank of England Regional Agents Retail survey

Real Rate Musings

It is interesting is that since mid-December there has been a stark breakdown in cross-asset correlations related to the Real Rate story as I indicated was likely to happen in my previous post on the Liquidity Trade.

But what is driving it?

The below chart shows the 5yr & 10yr Real Rates vs the S&P500. You can clearly see that while the 5yr has only sold off about 10bps since early December, the 10yr rate has sold off by over 30bps. Disassembling further, we can see that liquidity conditions (that have driven the risk asset rally since July), as measured purely by the 5yr rate, are still pretty easy.

Key:White = SPX, Green = 10yr US Real Rate, Brown = 5yr US Real Rate.

But it is arguable that some of the aggregate shift can be explained by a re-rating of growth expectations higher (see below chart of Citi's US Growth Surprise Index).

Second, the marked steepening of the real rate curve (as defined by 5yr vs 10yr) is consistent with the idea that fiscal risk premia within yield curves has been increasing.

Of course, it could all just be a function of holiday period-related illiquidity, and curves & risk assets will mean revert in the coming weeks. In that respect, March 117.5 calls on TYH0 look quite cheap at just 25tics, targetting a return of 10yr yields to around 3.4% (or TYH0 at ~119-09 - i.e. payout ratio of around 4-1).