topic Re: Weighted Average of a stock price in SAS Programming

Weighted Average of a stock price

san7981 — Mon, 12 Oct 2015 11:22:35 GMT

Dear All,

I am having trouble coding the average pricing of stock.

Data

FLAG QTY PRICE AVG_PRICE CALCULATION
Buy 50 100
Buy 50 101
Buy 50 102
Sell 100 101.5 101 (50*100+50*101+50*102)/150
Buy 50 99
Sell 50 99.5 100.5 (50*100+50*101+50*102+50*99)/(150+50)

I need to generate weighted average price against flag 'S' as shown above with calculation how to calculate avg_price.
In case of first sell,I need weighted average price of all buys. total quantity purchased before first sell is 150 and quantity sold is 100.
Balance quanity is 50. To generate price against second 'Sell' we nee to take average of first 3 buys and 4th buy.

Kindly help me out.

Regards
Sandeep Gupta

Re: Weighted Average of a stock price

Florent — Mon, 12 Oct 2015 16:05:21 GMT

Hello Sandeep,

The code here below should do the job:

data results (drop= SUM_:);
	set input;

	retain SUM_COST SUM_QTY;

	if _N_ = 1 then do;
		SUM_COST = 0;
		SUM_QTY = 0;
	end;

	if strip(upcase(Flag)) = 'BUY' then do;
		SUM_COST = SUM(SUM_COST, QTY * PRICE);
		SUM_QTY = SUM(SUM_QTY, QTY);
		AVG_PRICE = .;
	end; else

	if strip(upcase(Flag)) = 'SEL' then do;
		AVG_PRICE = SUM_COST / SUM_QTY;
	end;

run;

Considering that the content of your "input" dataset has the following colomns: FLAG, QTY, PRICE.

Please let me know should there be something missing in my answer.

Regards,

Florent

Re: Weighted Average of a stock price

san7981 — Tue, 13 Oct 2015 06:05:19 GMT

Hi Florent,

Thanks for a quick reply.

I am getting the desired output for a single account. But if there are multiple accounts then the code is not giving desired output.

I am attaching the sample data for multiple accounts. Kindly help me out with this problem.

Sample data is as below:

ACCOUNT_NO	TRADE_DATE	FLAG	QTY	PRICE
1	3-Mar-14	FRESH	50	6233.75
1	3-Mar-14	COVER	50	6242.1
1	4-Mar-14	FRESH	50	6250.3
1	4-Mar-14	FRESH	50	6249
1	4-Mar-14	FRESH	50	6251.2
1	4-Mar-14	COVER	50	6267.8
1	4-Mar-14	FRESH	100	6269.95
1	4-Mar-14	FRESH	50	6291.65
1	4-Mar-14	FRESH	50	6303.35
1	4-Mar-14	FRESH	50	6333
1	5-Mar-14	COVER	350	6341
2	1-Apr-14	FRESH	200	6751
2	1-Apr-14	FRESH	100	6761
2	4-Apr-14	COVER	500	6739
2	4-Apr-14	FRESH	200	6769
2	9-Apr-14	FRESH	300	6795
2	9-Apr-14	FRESH	200	6828
2	15-Apr-14	COVER	500	6787
3	28-Feb-15	FRESH	25	8858
3	28-Feb-15	COVER	25	8890
4	18-Feb-15	FRESH	50	8951.375
4	26-Feb-15	COVER	25	8750
4	28-Feb-15	COVER	25	8938
4	28-Feb-15	FRESH	25	8950
5	28-Feb-15	FRESH	25	8961.7
5	28-Feb-15	COVER	25	8949.8
6	20-Feb-15	FRESH	50	8857.2
6	20-Feb-15	COVER	25	8888
6	20-Feb-15	COVER	25	8848.3
6	23-Feb-15	FRESH	50	8859.225
6	23-Feb-15	COVER	25	8863
6	23-Feb-15	COVER	75	8740
6	24-Feb-15	COVER	25	8780.8
6	24-Feb-15	FRESH	25	8743
6	26-Feb-15	COVER	25	8684.35
6	28-Feb-15	FRESH	25	8924.35
6	28-Feb-15	COVER	25	8990
7	26-Feb-15	FRESH	50	8766.775

Regards

Sandeep

Re: Weighted Average of a stock price

Florent — Tue, 13 Oct 2015 12:51:33 GMT

Hi Sandeep,

Just to make sure, are the rows with the value "FRESH" in the variable 'FLAG' to be considered as the "Buy" you mentioned in your first message ? I would expect to have a Buy as first operation for each account (you cannot sell something you haven't bought earlier in my point of view).

Regards,
Florent

Re: Weighted Average of a stock price

san7981 — Tue, 13 Oct 2015 13:18:20 GMT

Hi Florent

Fresh is buy and cover is sell.

Regards
Sandeep

Re: Weighted Average of a stock price

Florent — Tue, 13 Oct 2015 15:58:33 GMT

Hi Sandeep,

Would the following version of the code do what you want ?

data results (drop= ACCOUNT SUM_:);
	set input;

	retain ACCOUNT SUM_COST SUM_QTY;

	if ACCOUNT ne ACCOUNT_NO then do;
		ACCOUNT = ACCOUNT_NO;
		SUM_COST = 0;
		SUM_QTY = 0;
	end;

	if strip(upcase(Flag)) = 'FRESH' then do;
		SUM_COST = SUM(SUM_COST, QTY * PRICE);
		SUM_QTY = SUM(SUM_QTY, QTY);
		AVG_PRICE = .;
	end; else

	if strip(upcase(Flag)) = 'COVER' then do;
		AVG_PRICE = SUM_COST / SUM_QTY;
	end;

run;

It's important that all the data related to the same Account are consecutive (i.e. first all lines related to the account 1 then the ones of the account 2, ...etc) in your input otherwise this will not work as expected.

Regards,

Florent

Re: Weighted Average of a stock price

san7981 — Wed, 14 Oct 2015 05:34:07 GMT

Dear Florent,

The code you shared is not giving the desired output. I am attaching excel containing desired output with formula to calculate the output. Let me know in case you require extra information.

ACCOUNT_NO	TRADE_DATE	FLAG	QTY	PRICE	Desired_AVG_PRICE
1	3-Mar-14	FRESH	50	6233.75
1	3-Mar-14	COVER	50	6242.1	6233.75
1	4-Mar-14	FRESH	50	6250.3
1	4-Mar-14	FRESH	50	6249
1	4-Mar-14	FRESH	50	6251.2
1	4-Mar-14	COVER	50	6267.8	6250.166667
1	4-Mar-14	FRESH	100	6269.95
1	4-Mar-14	FRESH	50	6291.65
1	4-Mar-14	FRESH	50	6303.35
1	4-Mar-14	FRESH	50	6333
1	5-Mar-14	COVER	350	6341	6277.3
3	28-Feb-15	FRESH	25	8858
3	28-Feb-15	COVER	25	8890	8858
4	18-Feb-15	FRESH	50	8951.375
4	26-Feb-15	COVER	25	8750	8951.375
4	28-Feb-15	COVER	25	8938	8951.375
4	28-Feb-15	FRESH	25	8950
5	28-Feb-15	FRESH	25	8961.7
5	28-Feb-15	COVER	25	8949.8	8961.7
6	20-Feb-15	FRESH	50	8857.2
6	20-Feb-15	COVER	25	8888	8857.2
6	20-Feb-15	COVER	25	8848.3	8857.2
6	23-Feb-15	FRESH	50	8859.225
6	23-Feb-15	COVER	25	8863	8859.225
6	23-Feb-15	COVER	25	8740	8859.225
6	24-Feb-15	FRESH	25	8743
6	26-Feb-15	COVER	25	8684.35	8743
6	28-Feb-15	FRESH	25	8924.35
6	28-Feb-15	COVER	25	8990	8924.35
7	26-Feb-15	FRESH	50	8766.775

Regards

Sandeep

Re: Weighted Average of a stock price

Florent — Wed, 14 Oct 2015 17:41:21 GMT

Hello Sandeep,

I don't understand why the logic to calculate the DESIRED_AVG_PRICE of the cell F7 of your Excel does not take into account the QTY and PRICE of the of row 2, whereas the calculation of the DESIRED_AVG_PRICE of the cell F12 of your Excel takes into account the QTY and PRICE values of the rows 4 to 6. It seems to me that the logic is not in aligned...

I would expect to find 6246.0625 as DESIRED_AVG_PRICE in the cell F7 of your excel (calculated as with the formula =(D4*E4+D5*E5+D6*E6+D2*E2)/200).

Kr,
Florent

Re: Weighted Average of a stock price

san7981 — Fri, 23 Oct 2015 09:36:56 GMT

Hi Florent,

Please find the attached excel. I have mentioned steps how to calculate weighted avg price. Kindly let me know if have any doubts regarding data or logic.

Regards

Sandeep

Re: Weighted Average of a stock price

Florent — Fri, 08 Jan 2016 17:36:02 GMT

Hi Sandeep,

Sorry for the delay but I had a lot to do on my projects.

I don't know if you have solved your problem in the meantime but I'll still give it a last try. Please find below a new SAS program which, according to the logic I saw in your Excel, should match with your expectations.

Do not hesitate to comment the DROP statements in the code so that you can have a look at the temporary variables that are created. Also, adding PUT statements before, inside, and after the DO loops may help in visualizing what is happening during the execution of the datasteps.

Kind regards,

Florent

data tmp_Results (drop= tmp_: len_:);

	set input;

	length List_Row List_Qty List_Price $1000 tmp_Row $100;
	retain List_Row List_Qty List_Price;

	tmp_row_num = _N_;
	tmp_Total_Qty = Qty;

	/* To keep exactly the same amount of rows as in the excel example */
	if tmp_row_num <= 22 then do;

		/* In case of Fresh, add the values to the lists */
		if strip(upcase(Flag)) = 'FRESH' then do;
			List_Row = strip(List_Row) || strip(put(tmp_row_num, best12.))|| ';';
			List_Qty = strip(List_Qty) || strip(put(Qty, best12.))|| ';';
			List_Price = strip(List_Price) || strip(put(Price, best12.))|| ';';
		end; else

		/* In case of Cover, remove values from the lists */
		if strip(upcase(Flag)) = 'COVER' then do;

			do until (tmp_Total_Qty <= 0);
				tmp_Row = strip(scan(List_Row, 1, ';') || ';');
				tmp_Qty = input(scan(List_Qty, 1, ';'), best12.);
				tmp_Price = strip(scan(List_Price, 1, ';') || ';');

				len_Price = lengthn(strip(tmp_Price));
				len_Qty = lengthn(strip(scan(List_Qty, 1, ';') || ';'));

				if tmp_Qty >= tmp_Total_Qty and tmp_Total_Qty > 0 then do;
					tmp_Total_Qty = tmp_Total_Qty - tmp_Qty;

					if tmp_Total_Qty >= 0 then do;
						List_Row = strip(tranwrd(List_Row, strip(tmp_Row), ''));
						List_Qty = strip(substr(List_Qty, len_Qty+1));
						List_Price = strip(substr(List_Price, len_Price+1));
					end; else
					if tmp_Total_Qty < 0 then do;
						List_Qty = strip(put(-tmp_Total_Qty, best12.)) || ';' || strip(substr(List_Qty, len_Qty+1));
					end;
				end; else
				if tmp_Total_Qty > tmp_Qty and tmp_Total_Qty > 0 then do;
					tmp_Total_Qty = tmp_Total_Qty - tmp_Qty;

					List_Row = strip(tranwrd(List_Row, strip(tmp_Row), ''));
					List_Qty = strip(substr(List_Qty, len_Qty+1));
					List_Price = strip(substr(List_Price, len_Price+1));
				end;
			end;
		end;

		output tmp_Results;
	end;
run;


data Results (drop= List_: tmp_:);
	set tmp_Results;

	List_Row = lag1(List_Row);
	List_Qty = lag1(List_Qty);
	List_Price = lag1(List_Price);

	if strip(upcase(Flag)) = 'COVER' then do;
		tmp_nb_items = countw(List_Row);
		tmp_numerator = 0;
		tmp_denominator = 0;

		do tmp_i=1 to tmp_nb_items;
			tmp_Qty = input(scan(List_Qty, tmp_i, ';'), best12.);
			tmp_Price = input(scan(List_Price, tmp_i, ';'), best12.);

			tmp_numerator = sum(tmp_numerator, tmp_Qty * tmp_Price);
			tmp_denominator = sum(tmp_denominator, tmp_Qty);
		end;

		avg_price = divide(tmp_numerator, tmp_denominator);
	end;
run;