The concept of Patent is not anything new that emerged back a century or two. The first case of intellectual property protection was for “some kind of newfangled loaf of bread” which dates back to 600 BC- giving chefs the option to keep the recipe for a unique dish to themselves and enjoy the profits alone. According to some historians, the earliest industrial patent application was made in 1421 and is credited to Filippo Brunelleschi, a Florentine architect who created a crane system for shipping and moving marble from the Carrara highlands. The Venetian Senate established the first patent law in 1474, defining the idea of intellectual property and establishing the significance of safeguarding inventors' rights.
A patent, as we know it today, is an intellectual property right to protect inventions. This protection is granted by the government of a nation grants this protection as a temporary territorial privilege (20 years). The purpose of a patent is to make it unlawful for anybody other than the owner or a person who has the owner's consent to create, use, or sell the invention. It encourages investment in R&D and increases the scope of new useful discoveries.
Not everything can be patented, an invention is patentable only if it fulfils three basic criteria:
1. Novel- for an invention to be novel, it must not be known to the public before the filing date of the patent application. It is an absolute condition for patentability
2. Useful- the invention must have some substantial utility. 3.Not Obvious- the innovation must be novel or unexpected and cannot be predicted based on the state of technology at the time.
Patents ensure complete protection against infringement, make the invention tradeable, provide monopoly, spark competition, limit competition and give the innovator advantage over their competitors by improving market presence.
Since a Patent is a territorial right, it could be a lengthy and costly process to get patents granted in different countries. It can be a humongous task to go through a separate filing process for each country. To make this process easy we have the Patent Cooperation Treaty. More than 155 countries have ratified the PCT, making it an international treaty. The PCT is designed to simplify the initial application filing procedure, making it simpler and initially less expensive to submit a patent application in a large number of nations. In what is referred to as the "national phase," the national or regional patent offices continue to have the final say over patent issuance. Instead of submitting several separate national or regional patent applications, the PCT enables simultaneous patent protection for innovation in a large number of nations by submitting a single "international" patent application.
Of all the patents that were filed in 2021 the US Patent and Trademark Office, the European Patent Office, the Japan Patent Office, the Korean Intellectual Property Office, and the National Intellectual Property Administration in China, accounted for 85.1% of the world total. There is unprecedented growth in filings in China. China’s share of the world total has almost doubled, from 24.4% in 2011 to 46.6% in 2021. By contrast, the other four top five offices experienced a decrease in their respective shares during the same period. The five largest intellectual property countries form the IP5 forum. They discuss and explore the potential for collaboration on shared issues such as patent examination workloads, backlogs, patent quality, and inefficiencies in the international patent system.
In an economy, we measure outputs - things we produce. In an information economy, however, outputs are not the only thing that matters. Just as importantly, we need to measure inputs - the resources and knowledge used in producing outputs. But measuring inputs and outputs can be challenging, especially when resources are distributed across different organizations and over long periods of time. We rely on traditional statistics to measure most economic activities, but innovation is notoriously hard to measure with traditional methods. In fact, it's not until years after innovation has been produced that we begin to see significant economic impacts. This makes it very difficult to plan effectively for investments and innovation.
On the other hand, patent data can be used to provide a useful indication of the rate at which innovation occurs within a particular country or region. By using patent indicators, which are statistical or metric measures of technology innovation, a country or region can be measured and assessed for innovation in its fields. In addition to comparing technological capacities between different countries, this type of analysis can also be used to track changes in technological innovation as they occur. These data can also offer information about opportunities for further research and economic advancement in a specific area. In recent years, more and more countries and international organizations have been using patent data to make informed decisions about investment and policy planning. For example, the Organisation for Economic Co-operation and Development (OECD) has developed a suite of indicators that measure the outcomes of research and development (R&D) investment across various countries. These indicators provide insight into trends and patterns of innovation and R&D activity in different countries over time.
Using the latest available data, nowcasting is a new forecasting model for predicting near-term economic and business developments, using the latest available evidence. Nowcasting patent indicators have become increasingly important in recent years, as the volume of available data on innovation and technological development has increased dramatically over the past decade.
Nowcasting Patent Indicators in the OECD was a global initiative developed to address the question of how to "read" patterns in patent activity to provide timely information on innovation capacity in OECD member countries and their development partners.
It's an estimate of the number of patents that will be filed in a given year based on the trend in patent filings over the past few years. In addition to assessing the value of an invention in a particular country over time, the nowcast can be used to estimate the financial impact of a patent filing in a given sector. It can be used in different industries and sectors to predict demand and gauge demand factors. It can be applied to any industry or business by looking into the data that is available through the patent system. By analysing this information, organizations can anticipate trends in demand and make better strategic decisions. For example, it is possible to predict the demand for electric vehicles by analyzing data such as the number of patents filed for various electric vehicle-related technologies. Organisations can also use patent information to anticipate trends in the market that could have a significant impact on their business.
Which data to nowcast?
The OECD, EPO, USPTO and triadic patent family’s data are the core sets of patent-based indicators that would need to be nowcasted first. The forecasting method is also applied to the set of indicators in the context of developing a set of global patent indicators such as Applications filed under the PCT and filings with national patent offices- Japan Patent Office, French Institute National de la PropriÈtÈ Industrielle (INPI), the Patent Office of the United Kingdom. One OECD working paper “Nowcasting Patent Indicators” in 2007 by HÈlËne Dernis had taken only two datasets for nowcasting; the patent applications to the EPO and triadic patent families, with the aim of extending the time coverage of indicators up to 2004-2005.
Each dataset has its own specifications and no single model may fit the unique structure of the data, especially with respect to trends, Stationary, linear, exponential, etc. Different studies have already addressed the problem of nowcasting and forecasting, testing different approaches against different datasets (EPO, PCT, by country, by industry, etc.). At least three types of estimating procedure were used:
Trends analysis: For forecasting, AutoRegressive Integrated
Moving Averages model (ARIMA) is used.
Transfer models: First application transition to patent office (priority), PCT application transition to regional phase (transition coefficient).
Econometric models: The models were built on exogenous variables namely, R&D expenditures by sectors, source of funds, GDP, number of researchers, value added service, indicators of technological opportunities, indicators based on specific information from patent office (budget, number of patent examiners, patent fees), etc., probabilistic models, etc.
What is ARIMA Model?
In the field of machine learning, there is a particular collection of methods and techniques that are particularly well-suited for predicting the values of dependent variables over time. A time series can be divided into three components.
Trend: The up-and-down movement of data over a long period of time. (e.g., Air Quality Index)
Seasonality: Seasonal variations means when the data is fluctuating (e.g., increased demand for ice cream in the summer).
Noise: Random fluctuations of time series around a typical pattern.
ARIMA is a forecasting algorithm based on the concept that data on previous values of a time series can be used alone to forecast future values. This is a generalization of the simpler AutoRegressive and Moving Average and adds the concept of integration.
The descriptive explanation of ARIMA is given by:
AR (AutoRegression): A model that uses dependencies between an observation and a set of lagged observations.
I (Integrated): To make the time series stationary, we use the difference of raw observations (e.g., subtracting an observation from that of the previous time step). Stationary means mean and variance across the time period is constant. We use ADFuller test to know the stationarity.
MA(Moving Average): A model that uses the dependence between observations and residual from a moving average model applied to lagged observations.
Each of these components is explicitly specified as a parameter of the ARIMA model. These are the three terms that characterize ARIMA models. The standard notation of ARIMA is ARIMA (P, D, Q), replacing the parameters with integer values to quickly indicate the particular ARIMA model. The P,D,Q parameters can be found out by using gridsearch. ARIMA model forecast always a smooth series which will not be able to take into account the seasonal fluctuation. The parameters of the ARIMA model are defined as:
P: Number of lag observations in the model. Also called lag order or order of Auto Regression.
D: Degree of differencing, how often raw observations are differenced.
Q: The order of moving average.
Nowcasting Euro-PCT at regional phase:
A simple arithmetic method was introduced using year or average {} Euro-PCT transfer rates as an estimate of Euro-PCT transfer rates (EPCT_TR) in year t. Where stands for Euro-PCT at regional phase in year ‘t’. the number of PCT designating the EPO in year ‘t’.