In today’s article, we are going to highlight the similarity between crypto exchanges and the credit card sphere. Do you know that both use scoring models?
What is rating and why do we need them?
Rating is a tool for evaluating the relativity of one thing to another and, it is used in many industries and activities; for example, everyone knows Sovereign Credit Rating, a country`s rating by GDP, the top-100 songs/cities/companies which are also considered to be ratings. In fact, a ranking is an integral part of our life because we always give preference to one thing amongst a set of acceptable options.
First of all, an accurate rating is based on certain criteria that are necessary and important for making financial and marketing decisions. There has always been competition which has necessitated the need for ranking within society. When crypto exchanges appeared the need to rate them did not come as a surprise. The question is: are these ratings reliable? Let’s see.
When are scorings used?
Today, the scoring model is one of the most successful examples of using mathematical and statistical methods in banking, insurance, and other industries within the economic sphere.
Scoring models play an increasingly important role in people’s everyday business lives where a lot of decisions are made and a lot of advice is given based upon these models. How do we define a scoring model and how has it become the most convenient and used data interpretation tool for evaluating crypto exchanges? Let`s scope it out.
Scoring Models in Crypto
What about crypto exchange rankings? Today more than 200 exchanges provide their services in the crypto world. People, especially traders, generally tend to follow a loss-aversion principle and for that, they need an objective crypto exchange ranking. What ratings do we know? The platforms CoinMarketCap, Coinhills, and Coingecko all come to mind. The above-mentioned services are all formulating their crypto exchange ratings which are popular with traders and other stakeholders. Each of these sites provides information in a similar form, namely, they rank crypto exchanges by their 24h trade volumes. Let us see the top 10 crypto exchanges by these 3 rating agencies (Table 1).
That looks ‘credible’ right? According to Coinhills, BitFlyer and DragonEx trade volumes are greater than Binance`s, while CoinMarketCap and Coingecko do not include these exchanges in the top10 at all! When the user takes different approaches to analyze the trade volume, they receive a different picture each time. t.s. At first, the agency can include/exclude transaction-mining fee type trading pairs (see Transaction Fee Mining: a New Way for Exchanges to Profit ). This is the main distinction between CoinMarketCap and the adjusted lists on other exchange agencies.
Secondly, the agency can also include/exclude some market categories. Compare a 5k BTC trade volume (BTC/JPY trading pair) on BitFlyer exchange published by CoinMarketCap with a 333k BTC trade volume filed by Coinhills. The explanation given for this large variance is simply that CoinMarketCap includes only the spot market, while Coinhills – spot and future markets simultaneously. Adding the permanent tendency to manipulate their total trade volume (see BigONE, BitForex cases), the conclusion is obvious: the modern crypto exchange ratings, provided by popular analytical services, are not relevant.
Traders need more sophisticated ratings, which will allow them to see the full picture in the crypto world. The real trade volume is just a drop in the ocean in crypto exchange estimation. Using scoring models, which have proved to be a reliable data displaying method, CER has reached a new stage of crypto exchange ranking. Combining liquidity, cybersecurity, public opinion, and withdrawal & limits ratios with relevant weights, CER provides users with objective and complex crypto exchange ratings to help a trader choose the right exchange for their needs(see here: Scoring Dashboard).
Summing up, all modern popular rating agencies today, use a linear, clumsy, and incorrect approach to ranking, thereby confusing users by providing inaccurate and untruthful information. Making crypto exchange ratings is not an easy task, the issue becomes complicated when there is a lack of or improper data. Thus, scoring models are the best way to compare several platforms as they help to achieve the most accurate estimate.
Addition: Historic Reference for the Use of Scoring Models in Credit Cards
In 1941, David Durand was the first person to recognize the need to differentiate between good or bad loans by measuring the applicants’ characteristics. Since then, credit analysts in financial companies and mail order firms decided whether or not to give loans or send merchandise to potential customers based on their credit scoring. The arrival of credit cards in the late 1960`s made the banks and other credit card issuers employ credit scoring. Credit scoring not only improves forecast accuracy but also decreases default rates by 50% or more. In the 1970s complete acceptance of credit scoring led to a significant increase in the number of professional credit scoring analyses. By the 1980`s, credit scoring had been applied to personal loans, home loans, small business loans, and other fields. So a credit score is a numerical expression based on the statistical analysis of a person’s credit files to represent the creditworthiness of that person.
A simple example of credit scoring
A 45-year-old customer with an income of $70.000 and without any credit history, assuming the equivalent weight of each character is assigned a total score of 120*0.33+70*0.33+10*0.33 = 66 points.
If we add weights to this simple model, for example, income is the most important characteristic with a weight of 0.5, then credit history is less important with a weight of 0.3 and age has the lowest significance with 0.2. In such a case, the total score for our customer will be 120*0.2 + 70*0.5 + 10*0.3 = 62 points. By comparing this result with another person and creating a comparative table, or score rating, this allows stakeholders to make the decision about all listed factors.
A scoring model is described as follows: a model in which different variables are weighted in divergent ways and result in a score. This final score subsequently forms the basis for a conclusion, decision or further advice. For example, a scoring model that is used by the World Economic Forum to calculate countries` Global Competitiveness Index (GCI) and finally rank them by this indicator has the next weights for each sub-indices:
We do not take methods of national distribution into account (actually, it is GDP per capita) so, with this in mind, calculating the sub-indices values through every countries’ analysis, we can estimate the dynamics of the interested region or countries’ competitiveness index. For example, Canada is the 14th country in the world by national competitiveness with a GCI of 5.3.
How is it calculated?
Source: The Global Competitiveness Report 2017-2018
On calculating the values for all pillars, they are ranked from 1-to-7 in a score table. The next step is to calculate the sub-indices scores – without weights, which means that we need their mean values. The Basic Requirements Score is calculated as follows: (5.4+5.7+5.1+6.6) / 4 = 5.7. Follows a similar pattern, Efficiency enhancers and Innovation and sophistication factors scores is equal to 5.5 and 4.8 respectively. To calculate the total GCI, the appropriate weights should be used (see Table 2). Canada is Innovation-driven economy (according to GDP per capita ratio), so the GCI for Canada is calculated as follows: 5.7*0.2 + 5.5*0.5 + 4.8*0.3 = 5.3.
Did you know that Switzerland has been the most competitive country in the world for 9 consecutive years?
Referring to rating methods in the classical finance world, we should note that scoring models are widely used in the field of microfinance and express lending, where a bank or other institution determines the customer’s creditworthiness by using credit scoring models. National creditworthiness today is associated with three well-known credit ratings – Fitch, Moody’s and S&P who also use scoring models based on a country’s economic and politics scores (see the table below).
By collecting the data from open sources, the official documents of a country’s government or Central Bank reports etc. and then applying appropriate weights, the above agencies get the table with a country’s total credit score. Unlike the World Economic Forum and its ‘Global Competitiveness Report’, the credit agencies use letters, numbers, and math symbols (e.g ‘+’ and ‘-’) to show the forecast dynamic.