How to change the base year of constant dollars - sql

Question

I have a table that contains the monthly values ($) of building permits, per region, per type of structure. I have them in current dollars and constant 2012 dollars. I would like to change the constant dollars to a base of the most recent month, ie 2021-05.
The Worldbank says this about changing the base year of constant dollars:
For example, you can rescale the 2010 data to 2005 by first creating an index dividing each year of the constant 2010 series by its 2005 value (thus, 2005 will equal 1). Then multiply each year's index result by the corresponding 2005 current U.S. dollar price value.
My table looks something like this (in reality, there are many cities, each having many types of structures, eg: Residential, institutional, etc.):
Period City Type of structure Value valueAdjustment
2011-01-01 New York Commercial, total 125478 Current Dollars
2011-01-01 New York Commercial, total 129276 Constant dollars
2011-02-01 New York Commercial, total 120568 Current Dollars
2011-02-01 New York Commercial, total 124110 Constant dollars
...
2021-04-01 New York Commercial, total 197296 Current Dollars
2021-04-01 New York Commercial, total 154500 Constant dollars
2021-05-01 New York Commercial, total 155043 Current Dollars
2021-05-01 New York Commercial, total 121082 Constant dollars
What I've thought of doing is to create a column, Rank, to then use some variation of ROW_NUMBER to easily compare every month to 2021-05. I populated the rank like such:
WITH cteRank AS(
SELECT t.*,
Rnk = DENSE_RANK()OVER (ORDER BY YEAR([Period]), DATEPART(MONTH,[Period]) )
- COUNT(CASE WHEN YEAR([Period]) = 2021 AND DATEPART(MONTH,[Period])=2 THEN 1 END) OVER ()
- 1 +836
FROM [buildingPermits] t)
UPDATE cteRank SET [Rank] = Rnk FROM cteRank ;
The +836 is because with the way I coded it, it counts every instance of 2021-05, so I added the count to cancel it out. Not very efficient, but it works.
The resulting Rank column looks like this:
Period Rank City Type of structure Value valueAdjustment
2011-01-01 -124 New York Commercial, total 125478 Current Dollars
2011-01-01 -124 New York Commercial, total 129276 Constant dollars
2011-02-01 -123 New York Commercial, total 120568 Current Dollars
2011-02-01 -123 New York Commercial, total 124110 Constant dollars
...
2021-04-01 -1 New York Commercial, total 197296 Current Dollars
2021-04-01 -1 New York Commercial, total 154500 Constant dollars
2021-05-01 0 New York Commercial, total 155043 Current Dollars
2021-05-01 0 New York Commercial, total 121082 Constant dollars
The last step is the Worldbank's formula adapted to my need:
For example, you can rescale the 2012 data to 2021-05 by first creating an index dividing each year of the constant 2012 series by its 2021-05 value (thus, 2021-05 will equal 1). Then multiply each year's index result by the corresponding 2021-05 current U.S. dollar price value.
So for 2011-01, it would be:
(Constant 2011-01) / (Constant 2021-05) * (Current 2021-05)
129276 / 121082 * 155043
=165535
Here's some pseudo code for the division using a subquery, but it obviously returns an error because I didn't target specific cities and Type of Structure. Partitioning returned an error as well, Incorrect syntax near the keyword 'over'
SELECT Period, [Type of structure],City,
Value/ (SELECT Value
FROM [buildingPermits]
WHERE YEAR(Period) = 2021
and DATEPART(MONTH,Period) = 5
and valueAdjustment= 'Constant dollars' and [Rank] = 0)
FROM [buildingPermits]
WHERE valueAdjustment= 'Constant dollars'
Error: Subquery returned more than 1 value.
This is not permitted when the subquery follows =, !=, <, <= , >, >=
or when the subquery is used as an expression.
I think a SELF JOIN, a temp table, subquery or somehow using MAX(Rank) to get the value at rank 0 (2021-05) could do the trick, but I don't know how to go about implementing either of those solutions.
Any help is appreciated

Answer 1

Does this get you closer to what you want?
SELECT Const201101.[Period], Const201101.City, Const201101.[Type of structure],
Const201101.[Value] as [Const 201101 Value],
Const202105.[Value] as [Const 202105 Value],
Curr202105.[Value] as [Curr 202105 Value]
FROM [buildingPermits] Const201101
JOIN [buildingPermits] Const202105 ON
Const202105.City = Const201101.City AND
Const202105.[Type of structure] = Const201101.[Type of structure] AND
Const202105.[Period] = '2021-05' AND
Const202105.valueAdjustment = 'Constant dollars'
JOIN [buildingPermits] Curr202105 ON
Curr202105.City = Const201101.City AND
Curr202105.[Type of structure] = Const201101.[Type of structure] AND
Curr202105.[Period] = '2021-05' AND
Curr202105.valueAdjustment = 'Current Dollars'
WHERE Const201101.valueAdjustment= 'Constant dollars' AND Const201101.[Period] = '2011-01-01'