Fractional Reserve Banking Example

In this Example we will use 2 banks;-

Bank 1 = Minton State Bank
Bank 2 = Imperial Commercial Bank
In this Economy the Central Bank has stipulated that the “Reserve Requirement Ratio” that Banks must
implement is 10% i.e. all Banks are required to hold reserves equal to 10% of their customer deposits. When a
Banks “Excess Reserves” are at zero the bank is at full loan capacity.
Minton State Bank and Imperial Commercial Bank both have $10,000 in deposits with no Excess Reserves. So
each bank right at this moment in time has $9,000 in loans outstanding and $10,000 in deposit balances held by
customers.
A new customer now deposits $1,000 in Minton Bank. Minton will now loan the maximum amount that they
can loan out i.e. 90% of the $1,000 or $900 and hold the required 10% in Reserve. There are now $11,000 in
Deposits in Minton State Bank ($10,000+$900+$100) = $11,000 – The $900 loan is placed into a Minton State
Bank Account to debtor and $100 is held in reserve.
However the Debtor decides to take his $900 loan and deposit it in Imperial Commercial Bank. Imperial
Commercial Banks deposits now equal ($10,000 + $900) = $10,900
At the beginning of this paper trail, Total Deposits were equal to $20,000. $1,000 was added to the system and
now Total Deposits are $21,900. Even though the addition to the system was $1,000 Total Deposits have
increased by $1,900. The $900 is new money. Minton created it when they issued the $900 loan from thin air.

The chart shows the total amount of $ that can be created with the addition of $100 in reserves, using different
reserve requirements.
Mathematically the relationship between the Reserve Requirement Ratio (RR) and money creation is given by
the Deposit Multiplier (M) which is the ratio which is the maximum possible change in deposits to the change in
reserves. When Banks in the Economy have made the maximum legal amount of loans (zero Excess Reserves),
the Deposit Multiplier is equal to the reciprocal to the Reserve Requirement Ratio i.e.
M = 1/RR
In the above example the Deposit Multiplier (M) is 1/0.1 = 10 – So with a Reserve Requirement Ratio (RR) of
0.1 an increase in Reserves of $1 can increase the Money Supply by up to $10.
Distributions of Returns
Analysis - Currencies
Contents
1.0 Introduction – Distributions ..................................................................................................... 2
2.0 The Normal Distribution Example ........................................................................................... 3
2.1 Human Height .............................................................................................................................. 3
2.2 Normal Distribution Conclusion ................................................................................................ 4
3.0 Do It Yourself (DIY) Currency Analysis ......................................................................................... 5
3.1 Saving a New Excel Spreadsheet................................................................................................... 5
3.2 Finding the Data ............................................................................................................................ 6
3.3 Analysing the Data in Excel ......................................................................................................... 11
3.3.1 Initial Set-up ......................................................................................................................... 12
3.3.2 Open to Open Returns Analysis ........................................................................................ 16
3.3.3 Open to Open Returns Analysis Summary ....................................................................... 64
3.3.4 High to Low Returns Analysis ........................................................................................... 64
3.3.5 High to Low Returns Analysis Summary .......................................................................... 82
4.0 Overall Summary ........................................................................................................................... 82

1
1.0 Introduction – Distributions
In this guide we are going to use excel to examine past distributions of currency returns. Past price
behaviour may not be a perfect guide to the future; however it is our only guide. Therefore it is by
definition, useful to understand how price movements and returns have behaved in the past. With
a sufficiently large dataset, it is probable that over a long period of time the asset behaves similarly
to the way it did historically.

As traders/investors, this basic analysis can tell us whether the odds of exposing ourselves to a
currency over a given time period provides us with a suitable risk-reward profile for us to make
money. Furthermore, calculating a distribution of returns also allows us to investigate other
important parameters such as expected return. Understanding price movements can give an insight
into the volatility of a currency which can contribute towards setting stop losses, targets and other
risk management techniques. In short, the analysis can help us understand the probabilities of
making money in a currency over a period of time, while displaying data that can be intuitively
applied to establish effective risk management.

NB: In this guide we do not attempt to provide you with a rigorous lesson in theoretical statistics. For the
purposes of trading and investing it is important to understand the basics of distributions, how they apply to
currencies and how they can be used to approximate expected returns over given time periods. Most trading
educators and brokers either do not know these techniques or choose not to advise them for risk of it
conflicting with their own interest. If you would like to read further into the statistical side of distributions and
how to calculate them, there are countless websites that can help you achieve this. You can also take advanced
statistics classes in your local area to become a real expert!

2
2.0 The Normal Distribution Example
Normal or “Gaussian” Distributions are most commonly found in nature and human behaviour. To
give some practical examples, a Normal Distribution can explain the probability of being a certain
height or IQ level. It is a bell shaped curve, symmetrical around the mean (arithmetic average) of the
sample being analysed. Larger datasets will result in more accurate probability distributions and so
in cases where a normal distribution applies, more data will show empirical analysis tending towards
normality.

The Normal Distribution is the most commonly used probability distribution within the investment
industry, largely because of the simplicity in its calculation. However, traders/investors should be
aware that often it is not a perfect representation of asset returns – this will be tested and
discussed further in the analysis later in the tutorial.

2.1 Human Height

As mentioned, human height is a classic example of a naturally normally distributed data.
Remember, the larger the sample size, the higher the tendency towards normality. This distribution
was generated from a sample of 500 people.

140
120
100
Occurences

80
60
40
20
0

Height (cm)

Figure 1 - Height Distribution Histogram

The mean of the data set used was measured as 180cm, with a standard deviation of 10cm. A
standardised normal distribution can be plotted over the top of this histogram to assess its
“goodness of fit.

3
 Figure 2 - Normal Distribution fitted over Empirical Height Distribution Histogram

We can see that a normal distribution fits the data fairly well, and therefore it can be argued that
this type of distribution is a good predictor of height variation within the general population (not
just the sample used). Note that the x-axis in the second graph has been “standardised” and
displayed in terms of standard deviations. “0” standard deviations from the mean represents the
mean height value itself, in this case 180cm, whereas plus and minus 1 standard deviation
represent the mean plus (or minus) 10cm which is equal to 1 standard deviation.

In a perfect normal distribution, standard deviations from the mean account for the
following frequency of occurrences:

Standard deviations from the mean Percentage of occurrences in data set

1 68.2%
2 95.4%
3 99.8%

E.g. 95.4% of values within a normally distributed data set are predicted to lie between +2 and -2
standard deviations from the mean. In the height example mentioned above, this would predict
that 95.4% of the data
will lie between 160cm-200cm. So, by definition, 4.6% (100-95.4) of people are either smaller
than 160cm or taller than 200cm. That equates to about 1 in 20 people having heights
greater than 2 standard deviations away from the mean.

2.2 Normal Distribution Conclusion

It turns out that there are a lot of examples of data which can be assumed as normally distributed,
with human height being one of the most obvious. As previously mentioned, asset returns are often
estimated using the normal distribution, however due to the financial implications of these
estimates it is important to draw your own conclusions from empirical data – which is why the rest
of this tutorial is devoted to teaching how to plot and analyse your own empirical dataset of an
asset.

4
3.0 Do It Yourself (DIY) Currency Analysis

After introducing probability distributions through the normally distributed human height example,
a step-by-step method of how to conduct your own analysis on asset returns using empirical data
will be explained. You will need internet access to obtain the data and Microsoft Excel to conduct
the analysis.

In this example we are going to look at some simple analysis of day trading the AUDUSD. It is
commonly thought that around 90% of “day traders” (those that trade intra-day or over one day
horizons) lose money. If this is the case, it is worthwhile looking at the actual numbers to find out
why this is happening.

In this example, step-by-step instructions will be given on how to analyse the daily returns of the
AUDUSD currency pair.

3.1 Saving a New Excel Spreadsheet

Open a new Excel document and navigate to File > Save as. At this stage choose an appropriate
name and directory to save the Excel workbook in and click save.

5
 3.2 Finding the Data
There are many sources of currency data on the internet. www.Investing.com/currencies is highly
recommended and will be used in the following example. It has comprehensive data and is simple to
use.

a) With the spreadsheet still open, begin by navigating to www.investing.com/currencies.

b) In the US Dollar majors tab, select AUD/USD under “Pacific” region.

c) Next, click on “Historical Data” from the AUD/USD summary page.

6
The “Historical Data” page shows prices of the asset we are looking at, AUD/USD, over a default time
horizon and at daily time intervals. We want to keep the intervals as “Daily” but change the time
horizon by selecting the calendar box on the right-hand side and looking from 1st October 2008 to 1st
March 2015, in this example.

Time horizon
Intervals

Price
Data

d) Click on the time horizon box to select the two dates we want to analyse the data between.
e) At this stage, three calendar month boxes come up (the current month surrounding by the
previous and next months to each side). By clicking on the previous month title a couple of
times (e.g. “February 2015”) it changes the calendar to look in terms of months and then
years which allows us to scroll back in time more quickly. Select the 1st October 2008 using

7
this method. Note that the small arrows to the left and right of the calendar titles can also
be used to scroll through dates.

8
f) Now apply the same method to the right-hand calendar to navigate to the 1st March 2015.
This should now select all the dates including and between the 1st October 2008 and 1st
March 2015, which can be verified by checking the time horizon box before clicking “Apply”.

g) Now scroll to the bottom of the webpage where you will find the data for the first day in our
time horizon – 1st October 2008. From there, left-click and drag from the right-hand side of
the final data entry upwards a few days so that some of the price data in the table is
selected (highlighted in blue).

9
 Left-click and drag upwards to the left to select
some data in the table. Let go once a few rows
are selected as follows:

h) Now use the webpage scroll bar to scroll back to the top of the page. This time hold SHIFT on
the keyboard and left-click in front of the headings of all the columns in the table, to the left-
hand side of “Date”. This will highlight/select all the data in the table, including the column
titles.

i) With that data still selected, press CTRL+C on the keyboard to copy it. Next, open the Excel
document you started with and paste this into cell A1 by left clicking it and then pressing
CTRL+V to paste. The result should be that all the data in the AUD/USD table on the
webpage is now pasted into your excel document in the correct columns with their titles.

10
 3.3 Analysing the Data in Excel

A few Excel notes before we start:

  Columns are listed as A…Z etc running across the top of the spreadsheet. 
  Rows are numbered down the left-hand side. 
 A ‘cell’ is coded by the column-row reference grid – E.g. Cell P9 – as selected below. 

 If data presents itself as “’####”, do not worry, this just means the data text or number
has too many characters to fit in the cell/column. Similarly, characters can just appear
cut off. Adjust the column widths where these problems occur.

This can happen quite often in the “Date” column when importing data – adjust by
 clicking and dragging the intersection of column A and B to the right. 
 To zoom in and out of the spreadsheet, hold control and use the mouse wheel, or
equivalent. 

11
 Column Intersection

Selected
Cell “P9”

3.3.1 Initial Set-up

With the data downloaded, saved, and open in Excel, we can now start setting up the data ready
to analyse. Throughout this tutorial we will analyse two sets of returns – “open to open” returns
and “high to low” returns. Take the following steps to set up your data:

(i) Heading the Returns Columns

Select cell G1 and type “open to open %” to create a heading for the open to open
percentage returns column.
Select cell H1 and type “high to low %” to create a heading for the high to low
percentage returns column.

(ii) Daily Returns Formula

We are now going to calculate daily returns. For open to open returns this will show us the
percentage return if we bought at the open on one day and sold on the open of the next
day. We will calculate this for every period in the dataset. Similarly, high to low percentage

12
 returns are calculated by assuming that we invest at the low and close out at the high of
each day.

Mathematically the formula for returns over a single period can be presented like this:

𝑷𝒏 − 𝑷𝒏−𝟏
𝑹𝒏 = ( )
𝑷𝒏−𝟏

Where: Rn = Percentage Return in period between n and n-1

(open to open, high to low)
Pn = Price at time n
(current open, current high)
Pn-1 = Price at time n-1
(previous open, current low)

To apply this formula in column G, we are going to enter it into cell G2 and copy it down the
whole column to apply it to every row in our dataset.

(iii) Entering the Returns Formula

In cell G2, type =(C2-C3)/C3 and press Enter

This follows the formula outlined previously. C2 is our Opening Price for that day, C3 is
our Opening Price for the previous day. Make sure the brackets are included to avoid any
mistakes. While entering the formula into cell G2, you may also manually click the cells C2
and C3 instead of typing the references in, however you must make sure the equation’s
form remains unchanged.

Similarly, in cell H2, type =(D2-E2)/E2 and press Enter

Formula bar with equation used

to calculate highlighted cell (G2)

13
 Formula bar with equation used
to calculate highlighted cell (H2)

(iv) Copying Down the Formula

Using the open to open returns (column G) as an example, there are two methods to do this:
 EITHER 
o Highlight the cell G2
o Click and drag the small black dot on the bottom right-hand corner of the cell
all the way down the spreadsheet to 1/10/2008.
 OR 
o Navigate to the bottom of the spreadsheet using the scrollbar
o Select cell G1997 (the row with our earliest date, the last row in our
spreadsheet)
o Press CTRL+SHIFT+UP - this will select all the cells G2 to G1997.
o Press CTRL+D - copies down the formula in the top cell selected.

For very large sets of data like this one, the second method is much faster although may lead to
more mistakes if you are not confident using excel.

REMEMBER: It makes sense to check the formula has copied down correctly by highlighting a cell
in the column and checking the formula bar.

Now repeat the process using the same methods as above, this time with high to low returns in
column H.

14
Scroll down to the bottom of the spreadsheet where you will find a DIV0 error in the final cell of our
open to open returns (column G). This error arises because the formula within that cell is trying to
reference data that is not available (i.e. before 1st October 2008). Delete this cell by selecting it and
pressing delete on the keyboard before returning back to the top of the spreadsheet for the next
section.

15
 3.3.2 Open to Open Returns Analysis

For the majority of the analysis we are going to use open to open returns. We will revisit high to
low return towards the end of the tutorial.

With the daily open to open returns in place, we are now going to examine the distribution of these
returns. We want to find out how many days the AUD/USD moves up or down 1%, 2%, or any other
percentage. To do this we first need to determine between which percentage values we want to
analyse. In this example we will analyse from -3% to +3% in 0.3% increments.

(i) Setting up Intervals

o Select cell J1 and title it “Intervals”
o In cell J2 enter -0.03 (equivalent to -3%).
o In cell J3 enter -0.027.
o Click and drag to highlight cells J2 AND J3, copy down the formula using
the click and drag method mentioned previously, stopping at cell J22.

Note: By selecting two cells and copying down, Excel uses the difference in value between
the two cells as the incremental change for copying down.

16
 (ii) Frequency Table

We can use these intervals to make a frequency table. This will allow us to see how many of our
observations (daily returns) will have values in between our intervals and beyond. From there we
can create a histogram to display our results graphically.

For this next section we are going to use an add-in called Analysis Toolpak, making our analysis
much easier. In some versions of excel this is not included automatically. To load the Analysis
Toolpak take the following steps:

For Excel 2003

1. On the Tools menu, click Add-Ins.
2. In the Add-Ins available box, select the check box next to Analysis Toolpak, and then click
OK. Tip If Analysis Toolpak is not listed, click Browse to locate it.
3. If you see a message that tells you the Analysis Toolpak is not currently installed on
your computer, click Yes to install it.
4. Click Tools on the menu bar. When you load the Analysis Toolpak, the Data
Analysis command is added to the Tools menu.

For Excel 2007, (2010)

1. Click (File)  Options, and then click the Add-Ins category.
2. Near the bottom of the Excel Options dialog box, make sure that Excel Add-ins is
selected in the Manage box, and then click Go.
3. In the Add-Ins dialog box, select the check box for Analysis ToolPak and then click OK.
4. If Excel displays a message that states it can't run this add-in and prompts you to
install it, click Yes to install the add-in.

17
This example will be carried out in Excel 2010. In Excel 2010, Data Analysis is located in the Data
ribbon. In Excel 2003, Data Analysis is located under Tools on the menu bar.

Back to the frequency table…

 Locate Data Analysis and click on it to get to the Data Analysis menu. 

 Select Histogram from the list and press OK 

 Select the Input Range for the returns we want to analyse. Highlight cell G2,
press CTRL+SHIFT+DOWN to select all the data in that column. 
 Using the same method for the Bin Range, select our intervals by highlighting cell
J2, and press CTRL+SHIFT+DOWN to highlight all the data in that column. 
 Finally, select Output Range in the Output options. Be careful here as Excel
likes to return you to the Input Range box once you have clicked this. MAKE
SURE you DO NOT change the input range, and put the cursor in the Output
Range box before selecting cell K1. Press OK. 

18
 


Your frequency table should look like the one above. From this we can plot a histogram to
display this information graphically.

AN IMPORTANT NOTE: The frequency information displayed shows the number of days the
AUD/USD realises returns between its corresponding bin value and the next lesser bin value. For
example, cell L2 tells you how many days the AUD/USD has been down by -3% or more (10). Cell
L3 tells you how many days the AUD/USD has been down between -3% and -2.7% (8).

19
 (iii) Frequency Histogram

Creating a histogram using the data is very easily. Simply take the following steps:

 Highlight the bins and histogram data, that is, cells K2 to L23.


 With that range of cells highlighted, go to Insert on the ribbon/menu bar. 

 Select Column in the ribbon. 
 From the drop-down list of graph types, select the first of the 2-D options as highlighted
below.

20




(iv) Editing the Histogram

Note earlier how we discussed the ranges that these frequencies fall into. In this section, we
are going to tidy up the graph and display the ranges so our graphical representation is more
accurate. Before continuing, click and drag the chart to move it to the right hand side of the
spreadsheet, and expand its size by dragging out the bottom right corner of the chart.

21
Changing the X-Axis
We are going to display the ranges in terms of percentages rather than decimal values.
We will also enter ranges under which the corresponding data falls in the histogram. We will
do this by creating another column solely for the purpose of displaying our X-Axis correctly.
Remember that a Bin value of 0.03 represents a frequency range of up to 3% from the Bin
value listed before (or from negative infinity if there are no previous intervals/bin values).

 Select Cells M2 to M23, right click within the highlighted area and go to ‘Format
cells’ 

 Select Category as ‘Text’ and press OK. 

22

This allows us to enter numbers into these cells in a Text format for labelling purposes,
without getting any errors.

 Select cell M1 and enter “Text Ranges”

 Select cell M2 and enter “Less than -3%”
 Select cell M3 and enter “-3% to -2.7%”
 Select cell M4 and enter “-2.7% to -2.4%”
 Select cell M5 and enter “-2.4% to -2.1%”
 Select cell M6 and enter “-2.1% to -1.8%”
 Select cell M7 and enter “-1.8% to -1.5%”
 Select cell M8 and enter “-1.5% to -1.2%”
 Select cell M9 and enter “-1.2% to -0.9%”
 Select cell M10 and enter “-0.9% to -0.6%”
 Select cell M11 and enter “-0.6% to -0.3%”
 Select cell M12 and enter “-0.3% to 0%”
 Select cell M13 and enter “0% to 0.3%”
 Select cell M14 and enter “0.3% to 0.6%”
 Select cell M15 and enter “0.6% to 0.9%”
 Select cell M16 and enter “0.9% to 1.2%”
 Select cell M17 and enter “1.2% to 1.5%”
 Select cell M18 and enter “1.5% to 1.8%”
 Select cell M19 and enter “1.8% to 2.1%”
 Select cell M20 and enter “2.1% to 2.4%”
 Select cell M21 and enter “2.4% to 2.7%”
 Select cell M22 and enter “2.7% to 3%”
 Select cell M23 and enter “More than 3%”


23
 


Now we are going to input these values as the X-Axis in our graph.

 Right-click anywhere on the chart 

 Left-click on ‘Select Data’. 

24
  Left-click on ‘Edit’ under the Horizontal Axis Labels column 

Notice how the Bins values are highlighted as the current horizontal-axis labels. We are now going to
change these labels to the ones we created in column M.

 Select Cells N2 to M23 and press OK. 

25
  Press OK again to get out of the Select Data menu. 

Now the Text Ranges will appear on the x-axis of the chart, making it easier to analyse.

Next we are going to clean up the graph a bit more and label the Horizontal and Vertical Axes.

 Left-click on ‘Series 1’ in the graph and delete it by pressing delete on the keyboard. 

26
 Next, left-click on the graph to select it and go to Layout  Axis Titles  Primary Horizontal
Axis Title  Title Below Axis

27
This inserts an editable Text Box below the horizontal axis as a label.

 Select the label, called “Axis Label”, delete its content and rename it ‘Daily Return Range’. 


 Now follow the same process as the last two bullet points but this time select the 
“Vertical Axis Title” and “Rotated Title” options.



 Rename the rotated title “Frequency”

28




 Now go to Chart Tools > Layout > Chart Title > Above Chart to add a title to the chart.





29
 

(v) Data Analysis

Now that we have our probability distributions displayed both graphically and in tabular format,
we can start drawing some statistical conclusions from the data. The first thing to note is the shape
of our histogram. It looks fairly similar to the normal distribution discussed earlier, with one key
difference – the tails of the distribution (the extremities in price movements) have many more
occurrences than a normal distribution would have. Traders call these ‘fat tails’. It is important to
be aware of how likely these volatile daily movements occur.

We are going to continue with some summary statistics on our data. Excel makes these easily
obtainable. The data analysis tool will be used again but this time to obtain descriptive statistics -
simply follow these steps:

 Navigate to Data  Data Analysis as before. This time select Descriptive Statistics and press
OK.


 Select the returns column (highlight G2, press CTRL+SHIFT+DOWN), select Q1 for the output
range and check the summary statistics box. Now press OK.

30
 


Adjust column width

where necessary



Before we move on to looking at these Descriptive Statistics in more detail, first we are going to
format some of these numbers to display themselves as a percentage or number to lower decimal
places.

 Select cell R3 (value of the mean) and right click on it and select “format cells…”.

31
  Select the category as “Percentage” and show the percentage to 3 decimals places then
press OK.

 Repeat the process with percentages for cells R4-R8 and R11-R13.
 Repeat the process with numbers for cells R9-R10 and R14.

The Descriptive Statistics table should now look like this…

32

Let’s talk through some of the key statistics shown here.

Mean – This is the average daily open to open return from the sample we have analysed. It is also
known as the Expected Return as it is the amount we expect to make if we buy on the open of the
previous day and sell on the open of the current day. The value is 0.003%.
Median – This is the middle number in our sample when listed in numerical order.
Standard Deviation – If we assume that our distribution is normal we can say that 68.2% of data lies
within one standard deviation either side of the mean, 95.4% lies within two and 99.8% within
three, as discussed at the beginning of this tutorial. This is a widely used measure of volatility/risk
within the financial industry.
Kurtosis – If this value is positive using Excel’s method of calculation, then the distribution has
‘fat tails’ meaning there are more extreme values in the tails of the distribution than a normal
distribution would predict.
Skewness –Skew is another sign of a non-normal distribution. Negative skew means that
negative values are more extreme than positive ones.

Range – This tells you the difference between the highest positive and lowest negative return in the
data set. The value of 15.077% in this sample comes from a lowest negative of about – 7.6% and
highest positive of roughly 7.4%. This is confirmed by looking at the Maximum and Minimum
values. Be aware of those extremes.
Sum – The sum of all the returns we have calculated. Ignore it. Our method of calculation does
NOT give return you would expect over the entire period if you invested from 2008-2015.
Count – The total number of trading days analysed.


It would be useful to know the probability of different ranges of returns occurring given the historic
data. Next we will use the numbers in our frequency table to create some easily understandable
probability values for us to interpret. For each of the number of occurrences that falls in a given
range, we are going to divide it by the total number of trading days so we can see the percentage of
trading days that returns fall within the range specified.


33
  Select Cell N2 and type the following equation: “=L2/R15” … Before you press enter press F4
on your keyboard and it should make the formula look like this “=L2/$R$15”. If it does not,
type this formula in manually or try again. Now press Enter.




Adding dollar signs around R15 within the equation adds a ‘lock’ to that cell (R15). So when we
want to copy the formula down, R15 is continuously referenced as the denominator in the
equation. Simply, when we copy down the formula it will use the frequency on each row divided by
the total number of trading days (count).

 Next, highlight cells N2 to N23 and press CTRL+D to copy the formula down and apply
to each of the cells selected.


34
 

 Keep these cells highlighted and right-click on the highlighted area. Go to format cells.




 Click on ‘Percentage’ as the cell format and press OK. 

35
 

We now have a column of percentages showing the percentage of our past returns that fall within these
ranges. Therefore we can argue that these percentages represent our percentage probability of each of
these returns happening in the future, if we believe that historical returns provide a good prediction.

 Select Cell N1 and type ‘Probability’ to head the column. Press Enter. 

Let’s now take this a step further by adding a column for cumulative percentages, so we can start
deducing more probabilities. At this stage it is worth noting that any returns with a value exactly
equal to one of the intervals, for example -1.5, is counted as a daily return within the LOWER range.
So, a return of -1.5% is counted in the range of “-1.8% to -1.5%”, NOT “-1.5% to -1.2%”. The same is
true for -3%. An exact return of -3% is counted in the range of “Less than -3%” NOT in “-3% to -
2.7%”. This applies to the whole frequency distribution. Using this principle, let’s continue:

 In Cell O1, type ‘Cumulative Percentage’ to head the column. 

 In Cell O2, type “=N2” and press Enter. 


36
 

Here it gets a little trickier. In order to calculate cumulative percentages we are going to reference all
the percentages we have calculated so far (by using the previous column O cell) and adding the new
percentage probability for the new range. All will become clear when we go through it.

 In Cell O3, type “=O2+N3”. Press Enter.


We can see the formula used in the cell selected in the formula bar

Now you can see what is happening when we calculate the cumulative percentage. Each of column
O’s cells equal the sum of all column N’s cells up to the same row number.

 Select Cells O3 to O23 and copy down. CTRL+D. Do not include cell O2! (it’s formula is a
different construction to the one we want to copy down to the rows below)

37
 

As you can see this is another useful way to convey percentages. From this we can start concluding
things like:

 94.687% of the time, daily returns yield less than or equal to 1.2%.
 16.090% of the time, daily returns yield less than or equal to -0.6%.

Let’s now rearrange the spreadsheet to make it a little more presentable and easier to continue
working with.

 Select/highlight cells K1 to O23 and outline all the cells in the range with a border by going to
Home  Borders  All Borders.

38
 Next, rename cell Q1 from “Column1” to “Descriptive Statistics”.

39
In order to establish some more detailed information using our data we are going to use the Filter
toggle in Excel which will allow us to calculate statistics on certain filtered data from our sample.
It sounds a little complicated, but stick with it and we will have some useful numbers to look at.

To add a filter to our data we need click on “open to open %” cell G1 and apply the filter.

 Select cell G1.

 With G1 still selected, add a filter by navigating to Data  Filter on the ribbon.

Once you have activated the filter, a small arrow representing a drop-down menu will appear in the
top cell of each of your data columns A to H.

40
Clicking on one of these arrows will open a drop-down menu that will allow us to filter the
data in that column.

Before we continue in this section, we are going to edit our spreadsheet further by adding some
blanks rows above our data. The reason for this is so that when we conduct filter-analysis (which
collapses rows temporarily) there will be a section of the spreadsheet that is unaffected.

 Left-click and drag from the row 1 box to row 10 to select rows 1-10.
 Right-click within that selection and click insert.





41
 

The section we are building at the top of the spreadsheet is for further analysis on our open to open
daily returns. To make this clear we will label it appropriately.

 Select/highlight cell I1 to I4, and go to Home  Merge & Center

 Within these four cells now merged into one, type “open to open data” within the cell.
 With the cell still selected go to Home  Wrap Text so that the text label fits inside the cell
by using multiple lines.

42
  Finally, finish this cell by adding a border, as we did earlier in the tutorial.

We are now going to build a cell which displays the average of all the data in the “open to open %”
returns column, filtered or unfiltered. This will allow flexibility in changing the sample of returns we
are looking at and instantaneously give us an average value of those returns.

 Select cell J1 and type “Average Return”. Adjust the column width to fit in all the characters
as in previous steps. 


43
 

 Select cell K1 and type “=SUBTOTAL(“ 




 Double-click “1 – Average” from the drop-down menu that appears. Alternatively type “1”.




 Next, add a comma to the equation. Type “,”. 
 Select cells G12 to the last returns cell in our filtered open to open % returns column. Again,
use CTRL+SHIFT+DOWN. This will shift your view to the bottom of the spreadsheet. Navigate
back up to the top using the scrollbar on the right-hand side. 
 Close the bracket to the formula and press Enter.

44
 




 Right-click cell K1 and format is to show itself as a percentage to 3 decimal places, as we
have done before.

45
Now we will repeat the process in the cells below, J2 and K2, but this time to work out the “Count”
rather than average. This will count the number of data points in the filtered/unfiltered data. In this
case the “Count” will display the number of days in our dataset that match the criteria we set in the
filter.

 In cell J2, type “Count”.

 In cell K2, type “=SUBTOTAL(2,G12:G2006)” – similar to what we did above with the average
but replacing the “1” with a “2” which is the function number for count rather than average.

Notice that both the average return and our count figure that we have just calculated match those
which we found in the descriptive statistics. This is a good head check to know that our calculations
are correct. When we start filtering our open to open % returns column, cells K1 and K2 will change
values to reflect the filtered data and the descriptive statistics we calculated earlier will remain the
same since they apply to the entire dataset.

46
Now we move on to using are variable average and count cells.

The first average return we will calculate is the average positive return. We need to filter out all
returns that are negative or equal to 0% so that only positive returns are listed in the column. To do
this:

 Click the filter arrow for the open to open % returns column and navigate to Number Filters
 Greater Than. Now select the is Greater Than option to set up the filter.
 Enter “0” into the top right-hand box, opposite ‘is greater than’. Press OK.

47
 Now shows only
positive returns

Note that the entire open to open % returns column has changed to positive values. Excel has
filtered out the rows with negative or 0% daily returns and has applied it to the whole spreadsheet,
this is why other data appears lost (temporarily) which can be seen in the row numbers to the left
hand side. Do not worry, the filter can be undone (by clicking the filter drop down menu and
selecting clear filter) and reapplied at any time. Also note that the drop-down menu symbol has
changed to include a small filter symbol to show that the filter is currently active.

We can see that the ‘Average Return’ and ‘Count’ values have changed and are now calculated
based on only positive returns due to our filter.

While we have our data filtered for positive returns, let’s take permanent note of the count and
average returns values calculated.

 In cell M1, type “Av Pos Ret” for average positive return.

48
  Now select cell K1 with our average return for positive values and press CTRL+C to copy this
cell. Once copied, right-click in cell N1 and click on Paste as Values in the “Paste Options”
section. Pasting the cell as a value makes the record permanent and so this record will not
vary dependent on the filter applied to our returns.

You can see that the average positive return has now been copied across, but is displayed as a
decimal rather than a percentage. Format the cell appropriately (as a percentage) as we have done
throughout the tutorial.

 In cell P1, type “Pos Count” number of positive return days counted.
 Now select cell K2 with our count for positive values and press CTRL+C to copy this cell. Once
copied, right-click in cell Q1 and click on Paste as Values in the “Paste Options” section.

49
  In cell R1, we will calculate the number of positive return days in the dataset in terms of a
percentage. To do this, in cell R1 type “=Q1/$R$25” which divides the positive count by the
total count shown in the descriptive statistics. The $ signs are used within the R25 part of the
formula to fix this cell reference so that we can copy that formula down later.
 Display this cell as a percentage by formatting the cell – as we have done previously.

Still using our positive returns data, we can calculate a probability adjusted average positive return.
By multiplying the average positive return by the positive count frequency as a percentage of the
total count, we obtain a number that tells us our average daily return if we were to invest long in
the AUD/USD on every positive return day in the period analysed and not participate on other days.

 In cell T1, type “Adj Av Pos Ret” for probability adjusted average positive return.

50
  In cell U1 type “=R1*N1”. Format this cell as a percentage.

Now we want to repeat this process for negative returns.

 Select the filter menu from the open to open % returns column and go to Number Filters 
Less Than…

 Select “is less than” in the top-left box and 0 in the respective top-right box. Press OK.

51
All the values in the returns column have changed to negatives. The variable average return cell (K1)
now gives us the average negative return for the AUD/USD in our historical dataset. This value
should be recorded in the same way as the average positive return, as should be the count and other
computations previously recorded for the positive returns values.

 In cell M2, type “Av Neg Ret”.

 Select cell K1 and copy it using CTRL+C. Paste it as a value into the cell N2 by right-clicking on
N2 and selecting Paste as Values (the 123 clipboard icon). Format the cell as a percentage.

52
 In cell P2, type “Neg Count”.
 Select cell K2 and copy it using CTRL+C. Paste it as a value into the cell Q2 by right-clicking on
Q2 and selecting Paste as Values (the 123 clipboard icon).

53
Having established the positive and negative counts (number of trading days with positive and
negative returns respectively in our dataset), we can deduce the number of days with a 0% move.

 In cell P3, type “0 Count”.

 In cell Q3, type “=R25-Q1-Q2” which takes away the number of positive and negative days
from the total count. This will leave only the 0% returns days.

54
Now we can copy down the formula in cell R1 so that we can obtain the frequency of negative and
0% returns days as a percentage of the total count.

 Select cells R1 to R3, press CTRL+SHIFT+DOWN to copy down the formula down from R1 into
R3.

The final analysis with our negative open to open % returns data is the probability adjusted negative
return. Similar to the positive version, by multiplying the average negative return by the negative
count frequency as a percentage of the total count, we obtain a number that tells us our average
daily return if we were to invest long in the AUD/USD on every negative return day in the period
analysed and not participating on other days. Note if we reverse the sign of this value it would be
the probability adjusted return of going short the asset (investing only on negative days and not
participating on others).

 In cell T2, type “Adj Av Neg Ret”.

 In cell U2, type “=R2*N2”. Format this cell as a percentage to 3 decimal places. Remember to
adjust column widths as you go.

55
We now have the average Up/Down returns for our historical AUD/USD dataset. The approach used
here allows you to conduct your own analysis using average returns for any configurable filter to
the dataset. To get the spreadsheet back to normal, with no rows collapsed, change the filter by
doing the following:

 Select the open to open % returns filter and select “Clear Filter From open to open %”

You will see the rows are expanded back to normal (none are missing) and our variable count and
average cells display the average and count for the full dataset.

56
The final data analysis this tutorial will show is a standard deviation occurrences table. This will
compare the actual occurrences within 1,2 and 3 standard deviations from the mean against the
number of occurrences predicted by a normal distribution of the same data.

To analyse the number of trading days with returns within 1,2 and 3 standard deviations from the
mean, a simple table needs to be constructed that calculates the upper and lower bound of 1,2 and
3 standard deviations from the mean. The upper bound returns of 1 standard deviation from the
mean will be the mean value plus one standard deviation. The lower bound of 1 standard deviation
will be the mean value minus one standard deviation. Continuing this, the upper bound return of 2
standard deviations from the mean will be the mean value plus two standard deviations, and so on.

 In cell W1, type “Std Dev”.

 In cell W2, type “Upper”.
 In cell W3, type “Lower”.
 In cell X1, type “1”.
 In cell Y1, type “2”.
 In cell Z1, type “3”.
 Highlight all these cells and press CTRL+B to put the text into bold.

 In cell X2, type “=R13+(1*R17)”. This will calculate the upper bound for 1 standard deviation
from the mean.
 In cell X3, type “=R13-(1*R17)”. This will calculate the lower bound for 1 standard deviation
from the mean.

57
Now fill in the rest of the table using the following calculations:

 In cell Y2, type “=R13+(2*R17). Press Enter.

 In cell Y3, type “=R13-(2*R17). Press Enter.
 In cell Z2, type “=R13+(3*R17). Press Enter.
 In cell Z3, type “=R13-(3*R17). Press Enter.

Your table should now look like this:

Now we are ready to create our standard deviations occurrences table. This will show the
percentage of trading days where the returns lie within 1, 2 and 3 standard deviations from the
mean and compare them to those predicted by normal distribution.

 Type “Std Dev” into cell AB1. Press Enter. 

 Type “Actual” into cell AC1. Press Enter. 
 Type “Normal” into cell AD1. Press Enter. 
 Type “Actual%” into cell AE1. Press Enter. 
 Type “Normal%” into cell AF1. Press Enter. 
 Type “1” into cell AB2. Press Enter. 
 Type “2” into cell AB3. Press Enter. 
 Type “3” into cell AB4. Press Enter. 

58
 

The returns filter can be used again to find the number of occurrences between the upper and lower
bounds of the “standard deviation from the mean” measures. This will show the real/actual count.
The normal distribution percentages for 1,2 ad 3 standard deviations from the mean have been
outlined earlier in this tutorial. They can be used to find the predicted number of occurrences in the
dataset that fall within 1,2 and 3 standard deviations if the distribution was normal, by multiplying
their percentages by the overall count of all the trading days in the dataset (1995).

 Type “68.2%” into cell AF2. Press Enter.

 Type “95.4%” into cell AF3. Press Enter.
 Type “99.8%” into cell AF4. Press Enter.

 Type “=AF2*R25” into cell AD2. Press Enter.

 Type “=AF3*R25” into cell AD3. Press Enter.
 Type “=AF4*R25” into cell AD4. Press Enter




The predicted occurrences (estimated by a normal distribution) are shown. Now we will calculate
the actual values and see how they compare.

59
  Position the view of the spreadsheet so you can see both the returns column AND the
standard deviation lower and upper bounds. See below.




 Select the open to open % returns filter and choose Number Filters  Between

 Manually type in the lower bound of 1 standard deviation from the mean into the ‘Greater
than or equal to’ box and the upper bound into the ‘Less than or equal to box. Press OK. See
below. 

60

 View the count number (cell K2) and copy it with CTRL+C, then paste it as a value into cell AC2
(the actual occurrences within 1 standard deviation from the mean). Paste the value by right-
clicking and choosing paste as value under paste options. See below. 

 Repeat the process above; remember to start with a view which shows the returns column and
the upper and lower bounds so that you can easily type them in. 

 When finished with this procedure, remove the filter from the open to open % returns column by
selecting the “Clear Filter From open to open %” option in the Filter drop-down menu. 


61
 




To finish off our analysis for open to open daily returns we calculate the actual percentage of
occurrences that lie within these ranges. To do this we take the actual count for each range and
divide it by the total count for the whole dataset.

 Type “=AC2/R25” in cell AE2. Press Enter.

 Type “=AC3/R25” in cell AE3. Press Enter.
 Type “=AC4/R25” in cell AE4. Press Enter.

 Format cells AE2-AE4 to display themselves as a percentage by right-clicking them and going

62
 to Format Cells.

The simple conclusion from this table is that normal distribution predicts more occurrences than
were actually counted within 2 and 3 standard deviations. Turning that on its head; more actual
occurrences are unaccounted for by normal distribution outside 3 standard deviations. Since one
standard deviation is roughly equal to 0.87% in this dataset, it means that returns of more than 2.6%
or less than -2.6% occur more frequently than normal distribution predicts. If we look within 1
standard deviation from the mean, a normal distribution predicts fewer occurrences than the actual

63
count. All this adds up to an actual distribution that is “peakier” around the mean and has fatter tails
(more extremes) than the normal distribution would predict using the same mean and standard
deviation statistics. This empirical analysis just backs up the basic descriptive statistics we described
earlier (a higher than 0 kurtosis level).

3.3.3 Open to Open Returns Analysis Summary

The open to open daily returns statistics allow us to interpret probabilities of a range of returns
occurring. Descriptive statistics were discussed with some key interpretations for your future
analysis. Expected returns and average positive and negative returns were calculated as well as
comparisons drawn between a standard normal distribution and that of the AUD/USD. The key
conclusion about the mean (expected return) and average positive and negative returns is that they
are all very low percentage values, which results in low unleveraged opportunity when we trade the
asset over a daily timeframe. In fact, for this last point, commission was not even taken into account
which makes it even more likely that there is no money to gain by day trading in this manner.
Despite the often supressed opportunity in day trading this asset it is also important to be aware of
tail risks (extreme daily price movements) which happen more often that the standard normal
distribution would predict.

3.3.4 High to Low Returns Analysis

Now we move our attention to the high to low % returns column created earlier in the tutorial. We
will go through similar steps in creating a returns distribution and frequency table as we did with
open to open returns. However, high to low returns cannot be any less than 0%, since the
assumption is you invest on the low and close out on the high of each day. This means that much of
the more advanced descriptive statistics (such as average positive and negative returns) we did with
open to open data are void. As well as this, the structure of high to low returns also affects the
histogram and probability distribution so that it is not really applicable to analyse the distribution
against a normal distribution either. This will become clearer as we progress.

We begin by creating intervals that define the ranges of returns that we want to analyse in a
frequency table, as we did with the open to open returns data. This time the intervals will start at
0.2% and increase in 0.2% intervals to 4%.

(ii) Setting up Intervals

o Select cell J36 and title it “Intervals”
o In cell J37 enter 0.002 (equivalent to 0.2%).
o In cell J38 enter 0.004.
o Click and drag to highlight cells J37 AND J38, copy down the formula
using the click and drag method mentioned previously, stopping at cell
J56.

Note: By selecting two cells and copying down, Excel uses the difference in value between
the two cells as the incremental change for copying down.

64
 (iii) Frequency Table

Now we will create a frequency table using these intervals, just the same as we did before with open
to open returns.

 On the ribbon, go to Data  Data Analysis.

65
 Choose Histogram as the analysis tool and click OK.
 Set the Input Range to the high to low % returns column (H12:H2007)
 Set the Bin Range to the intervals we created in cells J37 to J56.
 Set the Output Range to cell K36.
 Press OK.

66
 Now let’s create some text ranges to show the ranges that our intervals define. In cell M36
head this new column “Text Ranges”.
 Follow the method we used previously with open to open returns to names all of these
ranges, starting by formatting cells M37 to M57 as text and then writing all the ranges in
manually. The end product should look like this:

67
Now we are going to calculate some basic descriptive statistics for high to low returns using the data
analysis tool again.

 Go to Data  Data Analysis on the ribbon.

 Select “Descriptive Statistics” as the analysis tool and press OK.
 Set the Input Range to our high to low returns (H12:H2007).
 Set the Output Range to cell Q36.
 Check the summary statistics box and press OK.

68
69
 Rename Q36 from “Column1” to “Descriptive Statistics H-L”

 Format cells R38 to R43, and R46 to R48 as percentages to 3 decimal places.

70
71
  Next do the same with R44, R45 and R49 but this time as numbers to 3 decimal places.

Now we can create probability column which looks at how many occurrences lie within certain
ranges and divides them by the total number of occurrences (the count) to get a percentage
probability of returns lying between those values.

 In cell N36, type “Probability”.

 In cell N37, type “=L37/$R$50”.
 Select cells N37 to N57 and hit CTRL+D to copy down that formula in all the cells to N57.

72
 Select cells N37 to N57 and format them as a percentage by right-clicking within the
selection and going to format cells.

73
From this analysis, deductions can be made of the percentage chance of returns landing within
certain ranges based on historical data. For example based on the dataset used we could say that
the AUD/USD has a 14.58% chance of having a daily high to low return of between 0.6% and 0.8%.

We can deepen our understanding of these percentages by looking at them on a cumulative scale.

 In cell O36, type “Cumulative Percentages”.

 In cell O37, type “=N37”.
 In cell O38, type “=O37+N38”.
 Select cells O38 to O57 and press CTRL+D to copy the formula in cell O38 down all the cells to

74
 O57.

In cell O43 there is a cumulative percentage of 74.05%. This means that all the returns up to and
including that respective range (1.2%-1.4%) account for 74.05% of the total data. Turning that on its
head we can conclude than roughly 26% (100%-74.05%) of the time the daily high-low return is
higher than 1.4%.

Before we go through a summary of what can be learned from this High-Low analysis, lets plot the
probability distribution (histogram) using our high to low frequency table.

 Select cells K37 to L57.

 Navigate to Insert  Column Chart  2D Clustered Column Chart.

75
To tidy up the chart, we will edit the x and y axis and delete any unnecessary information.

 Select the “Series1” label on the right side of the chart and press delete.
 Right-click within the chart and go to “Select Data”.

76
 Click “Edit” within the Horizontal Axis Labels.

 Delete the Axis label range and replace it by selecting cells M37 to M57 to put the text
ranges onto the x-axis of the chart. Press OK, and again to confirm.

77
 To label the x-axis, with the chart selected go to Chart Tools  Layout  Axis Titles 
Primary Horizontal Axis Title  Title Below Axis.

78
 Select the text inside the “Axis Title” label below the horizontal x-axis and delete it, renaming
it to “Returns Ranges”.

79
 Now follow the same procedure but choose Vertical Axis Title  Rotated Axis Title on the
ribbon.
 Rename the axis to “Frequency”.

80
 In a similar way we will now add a Chart Title. With the chart selected go to Chart Tools 
Layout  Chart Title  Above Chart on the ribbon.

 Rename the Chart Title to high to low returns.

81
 3.3.5 High to Low Returns Analysis Summary

The key things to realise about high to low returns is firstly that structurally they cannot yield any
less than 0%, as was previously discussed, and that secondly they are essentially the high to low
range of the day as a percentage of the low for the day. This makes it a good intra-period (in this
case daily) volatility/opportunity indicator. We can clearly see from the probability distribution of
high to low returns that it is not normal distributed and heavily skewed to one side which makes it
unsuitable for comparing to normal distributions. The descriptive statistics’ kurtosis and skew levels
well above 0 also back this up in hard numbers. The whole objective with this tutorial is to get an
understanding of volatility/risk and the opportunity that comes with it for the asset that you are
looking at (in this case AUD/USD) and in the periods that you analyse (in this case daily). The most
important high to low returns statistics from that perspective are the mean high to low which is
1.153%, the maximum which is 10.892% and a general understanding of the frequency table and
probabilities respective to certain ranges of returns. This will give you a very good feel for the asset
you are trading in terms of how it behaves and what the risks and opportunities are.

4.0 Overall Summary

In this document we have discussed how to construct a distribution of returns as well as a number of
descriptive statistics for the AUD/USD currency pair in Excel. The material covered in this tutorial will
allow you to draw some key conclusions and interpret basic statistical data, while translating them
into underlying probabilities of gaining varying returns. The methods put across were intended to
provide some flexibility should you want to explore further analysis, with the data analysis and filter
tools in Excel being particularly useful. It is worth experimenting with Excel to draw more

82
personalised conclusions about your data and this tutorial should have provided a basis for you to
accomplish that. The goal with this sort of analysis is to understand where the risks and associated
opportunities lie. When we look on a short term basis as we have here with daily returns then clearly
the opportunities are suppressed and there are not many occasions when you can make good profits
by day trading - certainly when unleveraged and accounting for trading costs. The answer is then to
extend the time horizon for investment to try and take advantage or larger price movements. A very
useful way to use this volatility analysis would be by using cross-asset methods – e.g. comparing the
AUD/USD to other currency pairs. From that you can see which currencies are more volatile and
therefore offer more risk/opportunity to profit when they are traded. The key comparable statistics
to use there would be the standard deviation values for open to open returns and the average
(mean) high to low returns values. Others that are worth looking at are ranges, maximums and
minimums as well as average positive and negative returns for open to open returns. Remember
when you compare assets to use the same timeframes and periods (i.e. daily analysis between 1st
October 2008 and 1st March 2015 across all assets).














83
 Money Supply Definition Example

M0: Notes and coins in circulation. In some countries, M0 includes bank reserves, so M0 is referred to as the
monetary base, or narrow money.

MB: is referred to as the Monetary Base or total currency. This is the base from which other forms of money
(like checking deposits are created and is traditionally the most liquid measure of the money supply.

M1: Notes and Coins in circulation + Demand Deposits + Other Checkable Deposits (OCDs). Bank reserves
are not included in M1.

M2: Represents M1 and "close substitutes" for M1. M1 + Savings Deposits + Time / Term deposits of less
than $100,000 + Money Market Deposit Accounts. M2 is a broader classification of money than M1. M2 is the
key economic indicator used to forecast inflation.

M3: M2 plus large and long-term deposits. Since 2006, M3 is no longer published by the US central bank.
However, there are still estimates produced by various private institutions.

Demand Deposits = Funds Held in Checking (Current) Accounts

Other Checkable Deposits = Demand Deposits other than classic Checking (Current) Accounts i.e. Checking
or Current Accounts that pay interest and the depositor can write unlimited cheques on the account.
Savings Deposits = Accounts that pay interest but can not be used directly as Money in the narrow sense i.e.
writing a cheque.
Time / Term Deposits = A Savings Account in which the depositor can not withdraw / access their money fo
a specific period of time i.e. 1, 2, 3, 5 years. Interest is paid by the bank on the deposits at a higher rate than on
instant access savings deposits.
Money Market Deposit Accounts = A high interest paying account, usually requiring a large minimum
deposit ($10K, $25K, $100K, $250,000) on which cheques can be written (subject to a maximum per month). A
Money Market Account (MMA) is considered a Savings Account for most purposes. Banks are encouraged to
add high penalty fees for accessing / writing cheques on the accounts to discourage withdrawals.

The ratio of a pair of these measures, most often M2 / M0, is called an (actual, empirical) money multiplier due
to the fractional reserve system in operation.
M0

 Laura has ten US $100 bills, representing $1000 in the M0 supply for the United States.

MB = $1000
M0 = $1000
M1 = $1000
M2 = $1000

 Laura burns one of her $100 bills. The US M0 and her personal net worth, just decreased by $100.

MB = $900
M0 = $900
M1 = $900
M2 = $900

M1

 Laura takes the remaining nine bills and deposits them in her transactional account (checking account or current
account by country) at her bank.

MB = $900
M0 = 0
M1 = $900
M2 = $900

 The bank then calculates its reserve using the minimum reserve percentage given by the Fed and loans the
extra money. If the minimum reserve is 10%, this means $90 will remain in the bank's reserve. The
remaining $810 can only be used by the bank as credit, by lending money, but until that happens it will be
part of the bank's excess reserves.
 The M1 money supply increases by $810 when the loan is made. M1 money is created.

MB = $900
M0 = 0
M1 = $1710
M2 = $1710)
 Laura writes a check for $400, check number 7771. The total M1 money supply didn't change, it includes the
$400 check and the $500 left in her account.

MB = $900
M0 = 0
M1 = $1710
M2 = $1710

 Laura's check number 7771 is accidentally destroyed in the laundry. M1 and her checking account do not
change, because the check is never cashed.

MB = $900
M0 = 0
M1 = $1710
M2 = $1710

 Laura writes check number 7772 for $100 to her friend Alice, and Alice deposits it into her checking
account. MB does not change, it still has $900 in it, Alice's $100 and Laura's $800.

MB = $900
M0 = 0
M1 = $1710
M2 = $1710

 The bank lends Mandy the $810 credit that it has created. Mandy deposits the money in a checking account
at another bank. The other bank must keep $81 as a reserve and has $729 available for loans. This creates a
promise-to-pay money from a previous promise-to-pay, thus the M1 money supply is now inflated by $729.

MB = $900
M0 = 0
M1 = $2439
M2 = $2439
 Mandy's bank now lends the money to someone else who deposits it on a checking account in yet another
bank, who again stores 10% as reserve and has 90% available for loans. This process repeats itself at the
next bank and at the next bank and so on, until the money in the reserves backs up an M1 money supply of
$9000, which is 10 times the MB money.

MB = $900
M0 = 0
M1 = $9000
M2 = $9000)

M2

 Laura writes check number 7774 for $1000 and brings it to the bank to start a Money Market account (these
do not have a credit-creating charter), M1 goes down by $1000, but M2 stays the same. This is because M2
includes the Money Market account in addition to all money counted in M1.

Foreign Exchange

 Laura writes check number 7776 for $200 and brings it downtown to a foreign exchange bank teller at
Credit Suisse to convert it to British Pounds. On this particular day, the exchange rate is exactly USD 2.00 =
GBP 1.00. Credit Suisse takes her $200 check, and gives her two £50 notes (and charges her a dollar for the
service fee). Meanwhile, at the Credit Suisse branch in Hong Kong, a customer named Huang has £100 and
wants $200, and the bank does that trade (charging him an extra £.50 for the service fee). US M0 still has
the $900, although Huang now has $200 of it. The £100 notes Laura walks off with are part of Britain's M0
money supply that came from Huang.
 The next day, Credit Suisse finds they have an excess of GBP and a shortage of USD, determined by adding
up all the branches' supplies. They sell some of their GBP on the open FX market with Deutsche Bank,
which has the opposite problem. The exchange rate stays the same.
 The day after, both Credit Suisse and Deutsche Bank find they have too many GBP and not enough USD,
along with other traders. Then, to move their inventories, they have to sell GBP at USD 1.999, that is, 1/10-
cent less than $2 per pound, and the exchange rate shifts.

None of the banks has the power to increase or decrease the British M0 or the American M0 (unless they
burn bills); they are independent systems.

The power to alter the Money Supply comes from the Central Banks in each country using Classic Money
Supply levers i.e. increasing M2 (Injection) or decreasing M2 (Withdrawal), Non-Classic Money Supply
levers e.g. Quantitative Easing (QE) – Buy Bonds and Assets (Injection), Sell Bonds and Assets
(Withdrawal) and / or shifting interest rates higher or lower in order to shift global interest rate differentials
making it either more profitable to hold domestic currency for the rest of the world (interest rates up) /
(Withdrawal) – Currency Higher (Buy), or less profitable to hold domestic currency for the rest of the world
(interest rates down) (Injection) – Currency lower (Sell).