I'm not sure if this is the case with these data sets, but one common difference that causes confusion are current advertised rates vs total inventory. It could

I'm not sure if this is the case with these data sets, but one common difference that causes confusion are current advertised rates vs total inventory. It could both be true that apartments on the market cost 17% more, but all rents (including those not marketed or not renewing) are only up 3.8. Another methodological difference might be mix of goods. For example, you probably can afford to pay 150/grocery-visit, but someone else might have just cut their meat intake, or switch to cheaper cuts, or have otherwise substituted such that maybe they pay $120. If you measure by the visit, or gross store receipts then the numbers will seem lower than they would comparing the same exact goods.