This month on Mean, Median, and Moose we look at a smorgasbord of economic indicators!
New Motor Vehicle Sales
If you drive past your local auto dealer regularly, you might notice how many cars are or not on the lot over the span of a few months. Sales of new motor vehicles is a key economic indicator since it provides a snapshot of consumer demand for a big ticket item, and it is one of the easier economic indicators for your average person to understand since many of us have bought or will buy at least one new car in our lifetime.
The New Motor Vehicle Sales table (20-10-0001-01) available on StatsCan shows the number of units sold per month, unadjusted for seasonality. In the last 5 years, from January 2018 to December 2022, a total of 8,870,067 vehicles were sold across Canada, but they were not sold in even numbers across these years or months.
In 2018 and 2019, new motor vehicle sales saw a very similar number of sales and similar seasonality to sales with 2,045,721 sold in 2018 and 1,980,150 sold in 2019. The beginning of the pandemic brought a predictable huge drop in sales, to a low of only 47,508 sold in April 2020. While you might expect new motor vehicle sales to have recovered to 2018 and 2019 levels by now, they haven’t. In 2021 and 2022, there is the usual January dip in sales to below 100,000 units, though this is notably lower than January dips in 2018, 2019, and 2020, which never went below 100,000 units sold. Similarly, while there is the usual spring spike in sales from March to June in 2021 and 2022, they remain well below the peaks of over 205,000 units sold in May 2018 and 2019, hitting only a peak of 173,881 units sold in March 2021.
Arguably, much of this effect is due to the supply chain shortage that resulted from the pandemic, though in recent months, while car lots have returned to their full capacity, car sales have not. It will be interesting to see whether a recovery occurs in the typically high sales season of March to June this year.
Measures of Productivity
The OECD hosts some measures of productivity for its member nations and well as the world. Let’s take a look at how Canada compares with some of its peers. Below is the GDP per hour worked in terms of USD. Note that these figures are not normalized to inflation:
Next up, labour compensation growth. We’ve included the means below as red. Interestingly Canada compares more favorably to the United States on this measure.
Inequality indicators
The 2021 Census saw the release for the first time of economic inequality indicators for Canada, provinces and communities. The census data includes gini-coefficients (or index) for various income types (before tax/after) as well as adjusted measures. It also includes the P90/10 ratio, which is the ratio of gap between the incomes of at the 10th and 90th decile.
Although not economic indicators in their own right, there is a range of research the connects economic growth and income inequality. One of the items that Statistics Canada did with this census is go back to 2016 data and calculate these indicators from the last Census so we have some comparison. Due to the scope of the data I focused in on Ontario CMAs. For context Canada’s Gini-Coefficient for adjusted after tax income in 2020 was 0.302, Ontario was 0.308 – internationally it can be found here (note the values are multiplied by 100 compared to Statistics Canada values).
Overall Canada and Ontario preform pretty well on income inequality. Unsurprisingly Toronto and Hamilton CMAs are near the top of the Ontario list but they are joined by Windsor which is a bit of a surprise given its affordable reputation.
The Standard Deviations of the coefficients have shrunk between the last two censuses. As inequality did drop during that period. It has to be pointed out that the 2020 Census data is skewed by COVID income supports that likely prop up the bottom end of the income spectrum.
As I was just starting to play around with this data I wanted to see how these factors related to other economic indicators. So I ran some basic correlation tests on the Gini-coefficients against census data on unemployment rates and educational attainment.
Now these correlations should be taken with a grain of salt as a more robust analysis could be done but on first pass it is certainly it does peak my interest. As a lower gini-coefficient is better the positive relationship with unemployment makes sense as low unemployment likely drives lower income inequality. The fact that the relationship got weaker between 2015 and 2020 is an interesting element that may need to be explored.
The negative correlation between educational attainment also makes sense as educational attainment rises it helps reduce the gini-coefficient across ontario. That fact that this is increasing in strength helps illustrate the growing power of educational outcomes and future economic success.
Finally there is the P90 to P10 ratio which is the income ratio between the 90th percentile income and 10th percentile income. This went down across the Ontario CMAs that were measured, despite some permanent government programs we need to think about the impact of the COVID relief programs on this data.
The Bank of Canada Valet API
Changing gears a bit and looking at data sources for baseline economic data, one of the key tools in this space is the Valet API offered by Canada’s central bank. The Bank of Canada does a significant amount of research and analysis as part of their mandate, which also includes matters like printing money, monetary policy and the financial system. One of their key data products is the Valet API which offers programmatic access to global financial data.
The Valet API is set up for ease of use and that makes it nicely beginner-friendly. It does not require authentication or special headers, and data can be accessed in CSV, JSON or XML formats. The API is organized into lists of data series and grouped data series. The API will return information about lists and series as well as the individual data points within a data series or group of data series. These data points are called “observations” within the context of the API.
The Valet API contains almost ten thousand individual data series, including highly detailed economic data feeds and survey responses on a variety of topics. Some of the highlights include exchange rate data for dozens of foreign currencies, rates of return on instruments like treasury bills and bonds, commodity price indexes by industry, interest rates, consumer price index data, and the Bank of Canada’s internal future projections for a variety of indicators.
It’s a great resource and there are a few tools out there for working with it, including an R package and a Python package. Doug decided to add his own tool to the mix. Factotum is a Python command-line utility that access the Bank of Canada API, collects data from the specified series for a specified period, and outputs an image file containing a reasonable-looking graph of the requested data. It’s a pretty simple little tool and the code should be a good starting point for someone who’d like to play with economic data.
Here are some images that we generated to show off Factotum’s capabilities:
Five Year fixed mortgage rate
Public Perception a “Big recession” is imminent
Percentage of credit cards in arrears
New cash bills printed
Crude oil and bitumen as a percentage of total exports