I have a dataset that contains annual expenditure data for only 5 years. Here is a sample of my data:

Year   Expenditure
2019   50000
2020   52000
2021   54000
2022   56000
2023   58000

I need to forecast the expenditure on a monthly basis. However, I am facing challenges due to the limited amount of data. The current data is not enough to perform robust time series forecasting.

To overcome this limitation, I considered generating synthetic data. However, I encountered the following issues with synthetic data:

1. The synthetic data tends to have a smoother or more linear trend compared to the original data.
2. The original data exhibits its own unique trend and variability, which the synthetic data fails to capture adequately.

Given these challenges, I am seeking advice on the following points:

1. Is it advisable to use synthetic data in this scenario?
2. What are the best practices for generating synthetic data that accurately reflects the trends and variability of the original data?
3. Should I proceed with forecasting using the limited 5 years of data instead of generating synthetic data? If so, what methods would be most suitable for handling such a small dataset?

Any insights or recommendations would be greatly appreciated.