Relationship Lab

How a series has moved with others: the correlation of daily changes (% for prices, bp for rates) over 1 and 3 years, the beta, a rolling correlation and whether one tends to move first. Correlation is not causation, and it changes over time.

Retail sales YoY against…

Series	1M	3M	1Y	n 1Y	3Y	3M − 3Y ↓	Beta 1Y
LQ45 lq45	–	–	0.92	3	0.06	–	2.37	scatter
INDOGB 10YR id10y	–	–	−0.90	4	−0.25	–	−0.10	scatter
Nasdaq 100 ndx	–	–	0.85	5	0.41	–	1.16	scatter
JCI (IHSG) jci	–	–	0.77	3	0.01	–	1.36	scatter
UST 10YR ust10y	–	–	−0.74	5	−0.38	–	−0.39	scatter
PHLX Semiconductor (SOX) sox	–	–	0.71	5	0.45	–	0.31	scatter
USD/IDR usdidr	–	–	−0.65	5	0.12	–	−5.95	scatter
S&P 500 spx	–	–	0.65	5	0.18	–	2.04	scatter
UST 2YR ust2y	–	–	−0.63	5	−0.41	–	−0.31	scatter
Brent brent	–	–	−0.51	5	−0.22	–	−0.11	scatter
Bitcoin btc	–	–	0.33	6	0.35	–	0.15	scatter
DXY dxy	–	–	−0.28	5	−0.08	–	−1.79	scatter
Gold gold	–	–	0.18	5	0.38	–	0.14	scatter
Copper copper	–	–	0.15	5	−0.05	–	0.17	scatter
Newcastle coal coal	–	–	–	0	–	–	–	scatter
United States Car registrations tvm_US_car	–	–	–	0	–	–	–	scatter
United States Retail sales YoY tvm_US_rsyy	–	–	–	0	–	–	–	scatter
Retail sales (IPR) YoY id_retail	–	–	–	0	–	–	–	scatter
Brazil Car registrations tvm_BR_car	–	–	–	0	–	–	–	scatter
Brazil Retail sales YoY tvm_BR_rsyy	–	–	–	0	–	–	–	scatter
China Car registrations tvm_CN_car	–	–	–	0	–	–	–	scatter
China Retail sales YoY tvm_CN_rsyy	–	–	–	0	–	–	–	scatter
India Car registrations tvm_IN_car	–	–	–	0	–	–	–	scatter

Compared with the headline cross-asset series and the series in the same catalog category and country. Changes are matched on common dates. Beta = change in Retail sales YoY per unit change in the other series (least squares, 1 year). Amber: fewer than 20 common changes (hover a value for its n). Stability: the 3-month correlation minus the 3-year one; ⇄ marks opposite signs. A change in correlation describes the past; it does not say the relationship has broken or will return. Data ↓

Retail sales YoY · LQ45

20 60 120 252

Not enough common dates for a rolling correlation.

Rolling correlation of daily changes over the last 60 common readings, last 3 years.

Who moves first? (1 year)

Lag (readings)	-10	-9	-8	-7	-6	-5	-4	-3	-2	-1	+0	+1	+2	+3	+4	+5	+6	+7	+8	+9	+10
correlation	–	–	–	–	–	–	–	–	–	–	0.92	–	–	–	–	–	–	–	–	–	–

Lag +k: today's change in Retail sales YoY against the change in LQ45 k readings later (positive = the first one leads). A peak away from 0 is a description of the past year only.